feat(security): add IP-based rate limiting on /api/v1/auth/login — fix brute force HRT-62

- saas_auth.py: in-memory sliding-window rate limiter (5 attempts/5min, 15min block)
  using collections.defaultdict + threading.Lock, stdlib only, no new deps
- portal_server.py: register rate_limit_middleware + access_log_middleware
  (was missing, leaving global 100req/min limit unApplied on portal routes)
- tests/security/test_security.py: add TestLoginRateLimit class with
  test_login_brute_force_blocked_after_5_attempts and test_login_429_has_retry_after_header

Co-Authored-By: Paperclip <noreply@paperclip.ing>

API_AUTH.md

@@ -0,0 +1,132 @@
+# API Auth JWT — Documentation
+## Sprint 2-3 (HRT-28)
+
+Base URL: `http://localhost:8792`
+
+---
+
+## Endpoints d'authentification
+
+### `POST /api/v1/auth/register`
+Inscription d'un nouvel utilisateur (plan free par défaut).
+
+**Body JSON:**
+```json
+{ "email": "user@example.com", "password": "motdepasse123" }
+```
+**Réponse 201:**
+```json
+{ "message": "Compte créé avec succès", "user_id": 1 }
+```
+**Erreurs:** `400` (email invalide / mot de passe < 8 car.), `409` (email déjà utilisé)
+
+---
+
+### `POST /api/v1/auth/login`
+Connexion — retourne access_token (15min) + refresh_token (30j).
+
+**Body JSON:**
+```json
+{ "email": "user@example.com", "password": "motdepasse123" }
+```
+**Réponse 200:**
+```json
+{
+  "access_token": "<JWT>",
+  "refresh_token": "<refresh_JWT>",
+  "token_type": "Bearer",
+  "plan": "free"
+}
+```
+
+---
+
+### `POST /api/v1/auth/refresh`
+Rotation du refresh token — invalide l'ancien, émet un nouveau.
+
+**Body JSON:**
+```json
+{ "refresh_token": "<refresh_JWT>" }
+```
+**Réponse 200:** identique à `/login`
+
+---
+
+### `POST /api/v1/auth/logout`
+Révocation du refresh token.
+
+**Body JSON:**
+```json
+{ "refresh_token": "<refresh_JWT>" }
+```
+**Réponse 200:**
+```json
+{ "message": "Déconnexion réussie" }
+```
+
+---
+
+## Routes protégées
+
+Toutes les routes protégées nécessitent le header:
+```
+Authorization: Bearer <access_token>
+```
+
+### `GET /api/v1/predictions`
+| Plan    | Accès                                      |
+|---------|---------------------------------------------|
+| free    | Top 3 uniquement, 1 course/jour             |
+| premium | Toutes les courses + alertes Telegram       |
+| pro     | API complète + lien export CSV              |
+
+### `GET /api/v1/predictions/export`
+Export CSV — **plan pro uniquement** (`403` pour free/premium).
+
+### `GET /api/v1/subscription/upgrade`
+Infos sur les plans disponibles et plan courant de l'utilisateur.
+
+### `GET /api/v1/health`
+Vérification d'état du service (pas d'auth requise).
+
+---
+
+## Sécurité
+
+- **Passwords:** hashés avec bcrypt (saltRounds=12)
+- **JWT access:** expiration 15 minutes (HS256)
+- **JWT refresh:** expiration 30 jours, stocké hashé (SHA-256) en DB, rotation à chaque usage
+- **Rate limiting:** 100 requêtes/min par IP — header `X-RateLimit-Remaining`
+- **CORS:** configuré pour `https://turf-ia.h3r7.tech` + localhost dev
+- **Logs d'accès:** horodatés ISO 8601 dans `logs/saas_api.log`
+
+---
+
+## Lancement
+
+```bash
+JWT_SECRET_KEY="votre_cle_secrete" \
+CORS_ORIGINS="https://turf-ia.h3r7.tech" \
+./venv/bin/python saas_api.py
+```
+
+---
+
+## Tests
+
+```bash
+./venv/bin/pytest tests/test_auth.py -v
+# Avec couverture:
+./venv/bin/pytest tests/test_auth.py --cov=auth --cov=auth_db --cov=middleware --cov=saas_api --cov-report=term-missing
+# Résultat: 27 tests OK, couverture globale 83%
+```
+
+---
+
+## Structure des tables DB
+
+```sql
+-- users: id, email, password_hash, plan(free/premium/pro), created_at, is_active, daily_usage, last_usage_date
+-- subscriptions: id, user_id, plan, start_date, end_date, stripe_customer_id
+-- refresh_tokens: id, user_id, token_hash, created_at, expires_at, revoked
+```

README_API_V1.md

@@ -0,0 +1,156 @@
+# Turf SaaS — API v1 Reference
+
+Sprint 3-4 · HRT-29 — Refacto API /v1/
+
+## Base URL
+
+```
+http://<host>:8792
+```
+
+## Authentication
+
+All endpoints (except `/api/v1/health` and `/api/v1/auth/*`) require a **Bearer JWT** token.
+
+```
+Authorization: Bearer <access_token>
+```
+
+### Get a token
+
+```bash
+# Register
+curl -X POST http://localhost:8792/api/v1/auth/register \
+  -H "Content-Type: application/json" \
+  -d '{"email": "user@example.com", "password": "mypassword"}'
+
+# Login → returns access_token + refresh_token
+curl -X POST http://localhost:8792/api/v1/auth/login \
+  -H "Content-Type: application/json" \
+  -d '{"email": "user@example.com", "password": "mypassword"}'
+```
+
+## Plans & Access Control
+
+| Plan      | Inclus                                             |
+|-----------|----------------------------------------------------|
+| `free`    | health, auth, courses/today, predictions/top3 (1/j)|
+| `premium` | + predictions/all, valuebets, metrics              |
+| `pro`     | + backtest, export/csv                             |
+
+## Endpoints
+
+### System
+
+| Method | Path             | Auth | Description          |
+|--------|------------------|------|----------------------|
+| GET    | `/api/v1/health` | Non  | Healthcheck public   |
+| GET    | `/api/v1/docs`   | Non  | Swagger UI           |
+
+### Auth
+
+| Method | Path                      | Description                    |
+|--------|---------------------------|--------------------------------|
+| POST   | `/api/v1/auth/register`   | Créer un compte (plan=free)    |
+| POST   | `/api/v1/auth/login`      | Login → JWT tokens             |
+| POST   | `/api/v1/auth/refresh`    | Renouveler l'access token      |
+| POST   | `/api/v1/auth/logout`     | Révoquer le refresh token      |
+
+### Courses
+
+| Method | Path                                  | Plan    | Description                        |
+|--------|---------------------------------------|---------|------------------------------------|
+| GET    | `/api/v1/courses/today`               | free+   | Courses du jour (paginé)           |
+| GET    | `/api/v1/courses/{id}/predictions`    | free+   | Prédictions ML pour une course     |
+
+Query params `courses/today`: `filter=[all|quinte|trot|plat]`, `limit`, `offset`
+
+`{id}` format: `{num_reunion}-{num_course}` ex: `1-3`
+
+### Prédictions
+
+| Method | Path                      | Plan      | Description                  |
+|--------|---------------------------|-----------|------------------------------|
+| GET    | `/api/v1/predictions/top3`| free+     | Top 3 chevaux du jour        |
+| GET    | `/api/v1/predictions/all` | premium+  | Toutes les prédictions ML    |
+
+Query params: `date=YYYY-MM-DD`, `limit`, `offset`
+
+### Value Bets
+
+| Method | Path                | Plan      | Description              |
+|--------|---------------------|-----------|--------------------------|
+| GET    | `/api/v1/valuebets` | premium+  | Value bets du jour       |
+
+Query params: `date`, `min_odds` (défaut 2.0), `limit`, `offset`
+
+### Backtest
+
+| Method | Path                | Plan | Description                      |
+|--------|---------------------|------|----------------------------------|
+| GET    | `/api/v1/backtest`  | pro  | Résultats historiques des paris  |
+
+Query params: `start`, `end` (YYYY-MM-DD), `limit`, `offset`
+
+### Export
+
+| Method | Path                    | Plan | Description          |
+|--------|-------------------------|------|----------------------|
+| GET    | `/api/v1/export/csv`    | pro  | Export CSV           |
+
+Query params: `type=[predictions|bets]`, `date`, `start`, `end`
+
+### Métriques
+
+| Method | Path                | Plan     | Description           |
+|--------|---------------------|----------|-----------------------|
+| GET    | `/api/v1/metrics`   | premium+ | Métriques ML et paris |
+
+Query params: `days` (int, défaut 30)
+
+## Réponse uniforme
+
+Toutes les erreurs retournent :
+
+```json
+{
+  "status": "error",
+  "message": "Description de l'erreur",
+  "code": 400
+}
+```
+
+Les listes paginées incluent :
+
+```json
+{
+  "pagination": {
+    "total": 150,
+    "limit": 20,
+    "offset": 0,
+    "has_more": true
+  }
+}
+```
+
+## Démarrage
+
+```bash
+cd /home/h3r7/turf_saas
+source venv/bin/activate
+python app_v1.py
+# ou
+gunicorn -w 2 -b 0.0.0.0:8792 app_v1:app
+```
+
+## Tests
+
+```bash
+cd /home/h3r7/turf_saas
+source venv/bin/activate
+python -m pytest tests/test_api_v1.py -v
+```
+
+## Documentation Swagger
+
+Accessible sur : `http://localhost:8792/api/v1/docs`

api_v1/__init__.py

@@ -0,0 +1,43 @@
+#!/usr/bin/env python3
+"""
+API v1 Blueprint package — Turf SaaS
+Sprint 3-4: HRT-29 — Refacto API /v1/
+Sprint 5-6: HRT-31 — Billing Stripe
+
+Registers sub-blueprints:
+  /api/v1/health      — public health-check
+  /api/v1/courses/    — courses du jour
+  /api/v1/predictions/— predictions ML
+  /api/v1/valuebets   — value bets (premium+)
+  /api/v1/backtest    — backtest historique (pro)
+  /api/v1/export/     — export CSV (pro)
+  /api/v1/metrics     — métriques perf ML (premium+)
+  /api/v1/billing/    — Stripe checkout, portal, webhook, status
+  /api/v1/docs        — Swagger UI (via flasgger, registered on app)
+"""
+
+from flask import Blueprint
+
+from .routes.health import health_bp
+from .routes.courses import courses_bp
+from .routes.predictions import predictions_bp
+from .routes.valuebets import valuebets_bp
+from .routes.backtest import backtest_bp
+from .routes.export import export_bp
+from .routes.metrics import metrics_bp
+from .routes.billing import billing_bp
+
+# Master blueprint that aggregates all sub-routes under /api/v1
+api_v1_bp = Blueprint("api_v1", __name__, url_prefix="/api/v1")
+
+
+def register_api_v1(app):
+    """Register all API v1 blueprints onto the Flask app."""
+    app.register_blueprint(health_bp)
+    app.register_blueprint(courses_bp)
+    app.register_blueprint(predictions_bp)
+    app.register_blueprint(valuebets_bp)
+    app.register_blueprint(backtest_bp)
+    app.register_blueprint(export_bp)
+    app.register_blueprint(metrics_bp)
+    app.register_blueprint(billing_bp)

api_v1/routes/backtest.py

@@ -0,0 +1,195 @@
+#!/usr/bin/env python3
+"""
+Backtest route for API v1.
+
+GET /api/v1/backtest  — Résultats backtest historiques (pro)
+"""
+
+from datetime import datetime, timedelta
+from flask import Blueprint, jsonify, request
+
+from api_v1.utils import (
+    get_db,
+    table_exists,
+    internal_error,
+    bad_request,
+    get_pagination_params,
+    paginate_query,
+)
+from auth import jwt_required_middleware, plan_required
+
+backtest_bp = Blueprint("v1_backtest", __name__, url_prefix="/api/v1")
+
+
+@backtest_bp.route("/backtest", methods=["GET"])
+@jwt_required_middleware
+@plan_required("pro")
+def backtest():
+    """
+    Backtest historique
+    ---
+    tags:
+      - Backtest
+    summary: Résultats backtest historiques des paris simulés — accès pro uniquement
+    security:
+      - Bearer: []
+    parameters:
+      - name: start
+        in: query
+        type: string
+        format: date
+        description: Date de début (YYYY-MM-DD), défaut = -30j
+      - name: end
+        in: query
+        type: string
+        format: date
+        description: Date de fin (YYYY-MM-DD), défaut = aujourd'hui
+      - name: limit
+        in: query
+        type: integer
+        default: 50
+      - name: offset
+        in: query
+        type: integer
+        default: 0
+    responses:
+      200:
+        description: Résultats backtest
+      401:
+        description: Token invalide
+      403:
+        description: Plan insuffisant (pro requis)
+    """
+    start = request.args.get("start")
+    end = request.args.get("end")
+
+    # Validate date formats
+    for label, val in [("start", start), ("end", end)]:
+        if val:
+            try:
+                datetime.strptime(val, "%Y-%m-%d")
+            except ValueError:
+                return bad_request(
+                    f"Paramètre '{label}' invalide, format attendu: YYYY-MM-DD"
+                )
+
+    if not start:
+        start = (datetime.now() - timedelta(days=30)).strftime("%Y-%m-%d")
+    if not end:
+        end = datetime.now().strftime("%Y-%m-%d")
+
+    limit, offset = get_pagination_params(default_limit=50, max_limit=200)
+
+    conn = get_db()
+    try:
+        if not table_exists(conn, "bet_results"):
+            return jsonify(
+                {
+                    "status": "ok",
+                    "period": {"start": start, "end": end},
+                    "summary": {
+                        "total_bets": 0,
+                        "message": "Aucune donnée bet_results",
+                    },
+                    "by_type": {},
+                    "details": [],
+                    "pagination": {
+                        "total": 0,
+                        "limit": limit,
+                        "offset": offset,
+                        "has_more": False,
+                    },
+                }
+            ), 200
+
+        # Summary
+        summary_row = conn.execute(
+            """SELECT
+                   COUNT(*)                                          AS total,
+                   SUM(CASE WHEN resultat='GAGNE' THEN 1 ELSE 0 END) AS gagne,
+                   SUM(mise)                                         AS mise,
+                   SUM(gain)                                         AS gain
+               FROM bet_results
+               WHERE date BETWEEN ? AND ?""",
+            (start, end),
+        ).fetchone()
+
+        total_bets = summary_row["total"] or 0
+        gagne = summary_row["gagne"] or 0
+        mise = float(summary_row["mise"] or 0)
+        gain = float(summary_row["gain"] or 0)
+        roi = round((gain - mise) / mise * 100, 1) if mise > 0 else 0.0
+        precision = round(gagne / total_bets * 100, 1) if total_bets > 0 else 0.0
+
+        # By type
+        by_type_rows = conn.execute(
+            """SELECT
+                   type_pari,
+                   COUNT(*)                                          AS total,
+                   SUM(CASE WHEN resultat='GAGNE' THEN 1 ELSE 0 END) AS gagne,
+                   SUM(mise)                                         AS mise,
+                   SUM(gain)                                         AS gain
+               FROM bet_results
+               WHERE date BETWEEN ? AND ?
+               GROUP BY type_pari""",
+            (start, end),
+        ).fetchall()
+
+        by_type = {}
+        for row in by_type_rows:
+            t = row["total"] or 0
+            g = row["gagne"] or 0
+            m = float(row["mise"] or 0)
+            gn = float(row["gain"] or 0)
+            by_type[row["type_pari"]] = {
+                "count": t,
+                "gagne": g,
+                "mise": round(m, 2),
+                "gain": round(gn, 2),
+                "roi": round((gn - m) / m * 100, 1) if m > 0 else 0.0,
+                "precision": round(g / t * 100, 1) if t > 0 else 0.0,
+            }
+
+        # Paginated details
+        count_row = conn.execute(
+            "SELECT COUNT(*) AS cnt FROM bet_results WHERE date BETWEEN ? AND ?",
+            (start, end),
+        ).fetchone()
+        detail_total = count_row["cnt"] if count_row else 0
+
+        detail_rows = conn.execute(
+            """SELECT date, race_name, type_pari, horse_name, horse_number,
+                      COALESCE(cote, 0) AS cote, mise, resultat, gain
+               FROM bet_results
+               WHERE date BETWEEN ? AND ?
+               ORDER BY date DESC, id DESC
+               LIMIT ? OFFSET ?""",
+            (start, end, limit, offset),
+        ).fetchall()
+
+        details = [dict(r) for r in detail_rows]
+        pagination = paginate_query(details, detail_total, limit, offset)
+
+        return jsonify(
+            {
+                "status": "ok",
+                "period": {"start": start, "end": end},
+                "summary": {
+                    "total_bets": total_bets,
+                    "gagne": gagne,
+                    "perdu": total_bets - gagne,
+                    "precision": precision,
+                    "mise_totale": round(mise, 2),
+                    "gain_total": round(gain, 2),
+                    "roi": roi,
+                },
+                "by_type": by_type,
+                "details": details,
+                **pagination,
+            }
+        ), 200
+
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()

api_v1/routes/billing.py

@@ -0,0 +1,664 @@
+#!/usr/bin/env python3
+"""
+Billing Blueprint — Stripe integration
+Sprint 5-6: HRT-31
+
+Endpoints:
+  POST /api/v1/billing/checkout   — create Stripe Checkout session (auth required)
+  POST /api/v1/billing/portal     — create Stripe Customer Portal session (auth required)
+  POST /api/v1/billing/webhook    — Stripe webhook handler (public, signature-verified)
+  GET  /api/v1/billing/status     — current subscription status (auth required)
+
+Environment variables required:
+  STRIPE_SECRET_KEY         — Stripe secret key (sk_live_... or sk_test_...)
+  STRIPE_PUBLISHABLE_KEY    — Stripe publishable key (pk_...)
+  STRIPE_WEBHOOK_SECRET     — webhook signing secret (whsec_...)
+  STRIPE_PRICE_PREMIUM      — Stripe Price ID for Premium plan (price_...)
+  STRIPE_PRICE_PRO          — Stripe Price ID for Pro plan (price_...)
+  APP_BASE_URL              — e.g. https://turf-ia.h3r7.tech (default http://localhost:8793)
+"""
+
+import json
+import logging
+import os
+from datetime import datetime, timedelta, timezone
+
+import stripe
+from flask import Blueprint, g, jsonify, request
+
+from auth import jwt_required_middleware
+from billing_db import get_db, migrate_billing_tables
+
+logger = logging.getLogger("turf_saas.billing")
+
+billing_bp = Blueprint("billing", __name__, url_prefix="/api/v1/billing")
+
+# ──────────────────────────────────────────────────────────────
+# Stripe configuration
+# ──────────────────────────────────────────────────────────────
+
+stripe.api_key = os.environ.get("STRIPE_SECRET_KEY", "")
+STRIPE_WEBHOOK_SECRET = os.environ.get("STRIPE_WEBHOOK_SECRET", "")
+STRIPE_PUBLISHABLE_KEY = os.environ.get("STRIPE_PUBLISHABLE_KEY", "")
+APP_BASE_URL = os.environ.get("APP_BASE_URL", "http://localhost:8793")
+
+# Plan → Stripe Price ID mapping
+PLAN_PRICE_IDS = {
+    "premium": os.environ.get("STRIPE_PRICE_PREMIUM", ""),
+    "pro": os.environ.get("STRIPE_PRICE_PRO", ""),
+}
+
+# Plan display names
+PLAN_NAMES = {
+    "free": "Free",
+    "premium": "Premium",
+    "pro": "Pro",
+}
+
+# ──────────────────────────────────────────────────────────────
+# DB helpers
+# ──────────────────────────────────────────────────────────────
+
+
+def _sget(obj, key, default=None):
+    """Safely get a value from a dict OR a Stripe StripeObject.
+
+    Stripe v7+ uses attribute-style access; plain dicts use [] / .get().
+    """
+    try:
+        # StripeObject supports [] but not .get(); dict supports both
+        val = obj[key]
+        return val if val is not None else default
+    except (KeyError, TypeError):
+        return default
+
+
+def _get_active_subscription(db, user_id: int):
+    """Return the most recent active subscription row for a user."""
+    return db.execute(
+        """SELECT * FROM subscriptions
+           WHERE user_id = ?
+           ORDER BY start_date DESC
+           LIMIT 1""",
+        (user_id,),
+    ).fetchone()
+
+
+def _upsert_subscription(db, user_id: int, **fields):
+    """
+    Update existing subscription or insert a new one.
+    fields: plan, stripe_customer_id, stripe_subscription_id,
+            status, current_period_end, grace_period_end, end_date
+    """
+    existing = _get_active_subscription(db, user_id)
+    if existing:
+        # Build SET clause dynamically from provided fields
+        set_parts = ", ".join(f"{k} = ?" for k in fields)
+        values = list(fields.values()) + [existing["id"]]
+        db.execute(f"UPDATE subscriptions SET {set_parts} WHERE id = ?", values)
+    else:
+        cols = ", ".join(["user_id"] + list(fields.keys()))
+        placeholders = ", ".join(["?"] * (1 + len(fields)))
+        values = [user_id] + list(fields.values())
+        db.execute(
+            f"INSERT INTO subscriptions ({cols}) VALUES ({placeholders})", values
+        )
+
+
+def _update_user_plan(db, user_id: int, plan: str):
+    """Sync users.plan field to match active subscription."""
+    db.execute("UPDATE users SET plan = ? WHERE id = ?", (plan, user_id))
+
+
+def _get_or_create_stripe_customer(user, db) -> str:
+    """Return existing stripe_customer_id or create a new Stripe Customer."""
+    sub = _get_active_subscription(db, user["id"])
+    if sub and sub["stripe_customer_id"]:
+        return sub["stripe_customer_id"]
+
+    # Create new customer in Stripe
+    customer = stripe.Customer.create(
+        email=user["email"],
+        metadata={"user_id": str(user["id"])},
+    )
+    return customer["id"]
+
+
+def _record_billing_event(
+    db, stripe_event_id: str, event_type: str, user_id=None, payload=None
+):
+    """Insert a billing_events audit row (idempotent on stripe_event_id)."""
+    try:
+        db.execute(
+            """INSERT OR IGNORE INTO billing_events
+               (stripe_event_id, event_type, user_id, payload)
+               VALUES (?, ?, ?, ?)""",
+            (
+                stripe_event_id,
+                event_type,
+                user_id,
+                json.dumps(payload) if payload else None,
+            ),
+        )
+    except Exception as e:
+        logger.warning("Could not record billing event %s: %s", stripe_event_id, e)
+
+
+# ──────────────────────────────────────────────────────────────
+# POST /api/v1/billing/checkout
+# ──────────────────────────────────────────────────────────────
+
+
+@billing_bp.route("/checkout", methods=["POST"])
+@jwt_required_middleware
+def create_checkout():
+    """
+    Create a Stripe Checkout session for upgrading to Premium or Pro.
+    ---
+    tags:
+      - Billing
+    security:
+      - Bearer: []
+    requestBody:
+      required: true
+      content:
+        application/json:
+          schema:
+            type: object
+            required: [plan]
+            properties:
+              plan:
+                type: string
+                enum: [premium, pro]
+    responses:
+      200:
+        description: Checkout session URL
+        schema:
+          type: object
+          properties:
+            checkout_url:
+              type: string
+            session_id:
+              type: string
+      400:
+        description: Invalid plan or Stripe not configured
+      503:
+        description: Stripe API error
+    """
+    if not stripe.api_key:
+        return jsonify({"error": "Stripe non configuré"}), 503
+
+    body = request.get_json(silent=True) or {}
+    plan = body.get("plan", "").lower()
+
+    if plan not in ("premium", "pro"):
+        return jsonify({"error": "Plan invalide. Choisir 'premium' ou 'pro'"}), 400
+
+    price_id = PLAN_PRICE_IDS.get(plan)
+    if not price_id:
+        return jsonify({"error": f"Prix Stripe non configuré pour le plan {plan}"}), 503
+
+    user = g.current_user
+    if user["plan"] == plan:
+        return jsonify({"error": f"Vous êtes déjà sur le plan {plan}"}), 400
+
+    db = get_db()
+    try:
+        customer_id = _get_or_create_stripe_customer(user, db)
+        # Persist customer_id early to prevent duplicates
+        _upsert_subscription(
+            db, user["id"], stripe_customer_id=customer_id, plan=user["plan"]
+        )
+        db.commit()
+
+        session = stripe.checkout.Session.create(
+            customer=customer_id,
+            payment_method_types=["card"],
+            line_items=[{"price": price_id, "quantity": 1}],
+            mode="subscription",
+            success_url=f"{APP_BASE_URL}/billing/success?session_id={{CHECKOUT_SESSION_ID}}",
+            cancel_url=f"{APP_BASE_URL}/billing/cancel",
+            metadata={"user_id": str(user["id"]), "plan": plan},
+            subscription_data={"metadata": {"user_id": str(user["id"]), "plan": plan}},
+        )
+    except stripe.StripeError as e:
+        logger.error("Stripe checkout error for user %s: %s", user["id"], e)
+        return jsonify({"error": "Erreur Stripe", "detail": str(e)}), 503
+    finally:
+        db.close()
+
+    return jsonify(
+        {
+            "checkout_url": session.url,
+            "session_id": session.id,
+            "plan": plan,
+            "publishable_key": STRIPE_PUBLISHABLE_KEY,
+        }
+    ), 200
+
+
+# ──────────────────────────────────────────────────────────────
+# POST /api/v1/billing/portal
+# ──────────────────────────────────────────────────────────────
+
+
+@billing_bp.route("/portal", methods=["POST"])
+@jwt_required_middleware
+def create_portal():
+    """
+    Create a Stripe Customer Portal session for managing subscription.
+    ---
+    tags:
+      - Billing
+    security:
+      - Bearer: []
+    responses:
+      200:
+        description: Portal session URL
+      400:
+        description: No Stripe customer found
+      503:
+        description: Stripe not configured or API error
+    """
+    if not stripe.api_key:
+        return jsonify({"error": "Stripe non configuré"}), 503
+
+    user = g.current_user
+    db = get_db()
+    try:
+        sub = _get_active_subscription(db, user["id"])
+        customer_id = sub["stripe_customer_id"] if sub else None
+
+        if not customer_id:
+            return jsonify(
+                {
+                    "error": "Aucun abonnement Stripe trouvé. "
+                    "Souscrivez d'abord à un plan payant."
+                }
+            ), 400
+
+        session = stripe.billing_portal.Session.create(
+            customer=customer_id,
+            return_url=f"{APP_BASE_URL}/account",
+        )
+    except stripe.StripeError as e:
+        logger.error("Stripe portal error for user %s: %s", user["id"], e)
+        return jsonify({"error": "Erreur Stripe", "detail": str(e)}), 503
+    finally:
+        db.close()
+
+    return jsonify({"portal_url": session.url}), 200
+
+
+# ──────────────────────────────────────────────────────────────
+# GET /api/v1/billing/status
+# ──────────────────────────────────────────────────────────────
+
+
+@billing_bp.route("/status", methods=["GET"])
+@jwt_required_middleware
+def billing_status():
+    """
+    Return current subscription status for the authenticated user.
+    ---
+    tags:
+      - Billing
+    security:
+      - Bearer: []
+    responses:
+      200:
+        description: Subscription status
+    """
+    user = g.current_user
+    db = get_db()
+    try:
+        sub = _get_active_subscription(db, user["id"])
+    finally:
+        db.close()
+
+    if not sub:
+        return jsonify(
+            {
+                "plan": "free",
+                "status": "active",
+                "stripe_customer_id": None,
+                "stripe_subscription_id": None,
+                "current_period_end": None,
+                "grace_period_end": None,
+            }
+        ), 200
+
+    return jsonify(
+        {
+            "plan": sub["plan"],
+            "status": sub["status"] or "active",
+            "stripe_customer_id": sub["stripe_customer_id"],
+            "stripe_subscription_id": sub["stripe_subscription_id"],
+            "start_date": sub["start_date"],
+            "end_date": sub["end_date"],
+            "current_period_end": sub["current_period_end"],
+            "grace_period_end": sub["grace_period_end"],
+        }
+    ), 200
+
+
+# ──────────────────────────────────────────────────────────────
+# POST /api/v1/billing/webhook
+# ──────────────────────────────────────────────────────────────
+
+
+@billing_bp.route("/webhook", methods=["POST"])
+def stripe_webhook():
+    """
+    Stripe webhook handler — no auth, signature-verified.
+
+    Handled events:
+      checkout.session.completed       → activate subscription
+      customer.subscription.updated    → sync plan/status
+      customer.subscription.deleted    → downgrade to free
+      invoice.payment_failed           → set past_due + 3-day grace period
+      invoice.payment_succeeded        → clear grace period
+    """
+    payload = request.get_data()
+    sig_header = request.headers.get("Stripe-Signature", "")
+
+    # Verify webhook signature (required in production)
+    if STRIPE_WEBHOOK_SECRET:
+        try:
+            event = stripe.Webhook.construct_event(
+                payload, sig_header, STRIPE_WEBHOOK_SECRET
+            )
+        except stripe.SignatureVerificationError as e:
+            logger.warning("Stripe webhook signature invalid: %s", e)
+            return jsonify({"error": "Signature invalide"}), 400
+        except ValueError as e:
+            logger.warning("Stripe webhook payload invalid: %s", e)
+            return jsonify({"error": "Payload invalide"}), 400
+    else:
+        # Dev/test: accept without verification (log a warning)
+        logger.warning("STRIPE_WEBHOOK_SECRET not set — skipping signature check!")
+        try:
+            event = stripe.Event.construct_from(json.loads(payload), stripe.api_key)
+        except Exception as e:
+            return jsonify({"error": "Payload invalide", "detail": str(e)}), 400
+
+    event_type = event["type"]
+    event_id = event["id"]
+    logger.info("Stripe webhook received: %s (%s)", event_type, event_id)
+
+    db = get_db()
+    try:
+        if event_type == "checkout.session.completed":
+            _handle_checkout_completed(db, event)
+
+        elif event_type in (
+            "customer.subscription.updated",
+            "customer.subscription.created",
+        ):
+            _handle_subscription_updated(db, event)
+
+        elif event_type == "customer.subscription.deleted":
+            _handle_subscription_deleted(db, event)
+
+        elif event_type == "invoice.payment_failed":
+            _handle_payment_failed(db, event)
+
+        elif event_type == "invoice.payment_succeeded":
+            _handle_payment_succeeded(db, event)
+
+        else:
+            logger.debug("Unhandled Stripe event type: %s", event_type)
+
+        db.commit()
+    except Exception as e:
+        db.rollback()
+        logger.error("Error processing Stripe webhook %s: %s", event_id, e)
+        return jsonify({"error": "Erreur interne"}), 500
+    finally:
+        db.close()
+
+    return jsonify({"status": "ok"}), 200
+
+
+# ──────────────────────────────────────────────────────────────
+# Webhook handlers
+# ──────────────────────────────────────────────────────────────
+
+
+def _resolve_user_from_customer(db, customer_id: str):
+    """Look up user_id via subscriptions.stripe_customer_id."""
+    row = db.execute(
+        "SELECT user_id FROM subscriptions WHERE stripe_customer_id = ? LIMIT 1",
+        (customer_id,),
+    ).fetchone()
+    if row:
+        return row["user_id"]
+
+    # Fallback: query Stripe for user_id metadata
+    try:
+        customer = stripe.Customer.retrieve(customer_id)
+        meta = _sget(customer, "metadata") or {}
+        uid = _sget(meta, "user_id")
+        if uid:
+            return int(uid)
+    except Exception:
+        pass
+    return None
+
+
+def _resolve_plan_from_price(price_id: str) -> str:
+    """Map Stripe price ID to internal plan name."""
+    for plan, pid in PLAN_PRICE_IDS.items():
+        if pid and pid == price_id:
+            return plan
+    # Unknown price — default to premium (safer than pro)
+    return "premium"
+
+
+def _handle_checkout_completed(db, event):
+    """checkout.session.completed → activate subscription for the user."""
+    session = event["data"]["object"]
+    customer_id = _sget(session, "customer")
+    subscription_id = _sget(session, "subscription")
+    metadata = _sget(session, "metadata") or {}
+    plan = _sget(metadata, "plan") or "premium"
+    user_id = _sget(metadata, "user_id")
+
+    if user_id:
+        user_id = int(user_id)
+    else:
+        user_id = _resolve_user_from_customer(db, customer_id)
+
+    if not user_id:
+        logger.error(
+            "checkout.session.completed: cannot resolve user for customer %s",
+            customer_id,
+        )
+        return
+
+    # Fetch subscription details from Stripe
+    current_period_end = None
+    if subscription_id:
+        try:
+            sub = stripe.Subscription.retrieve(subscription_id)
+            current_period_end = datetime.fromtimestamp(
+                sub["current_period_end"], tz=timezone.utc
+            ).isoformat()
+            # Sync plan from price if metadata plan is missing
+            if sub["items"]["data"]:
+                price_id = sub["items"]["data"][0]["price"]["id"]
+                plan = _resolve_plan_from_price(price_id)
+        except Exception as e:
+            logger.warning("Could not fetch subscription %s: %s", subscription_id, e)
+
+    _upsert_subscription(
+        db,
+        user_id,
+        plan=plan,
+        stripe_customer_id=customer_id,
+        stripe_subscription_id=subscription_id,
+        status="active",
+        current_period_end=current_period_end,
+        grace_period_end=None,
+    )
+    _update_user_plan(db, user_id, plan)
+    _record_billing_event(db, event["id"], event["type"], user_id=user_id)
+    logger.info("checkout.session.completed: user %s upgraded to %s", user_id, plan)
+
+
+def _handle_subscription_updated(db, event):
+    """customer.subscription.updated → sync status and plan."""
+    sub_obj = event["data"]["object"]
+    customer_id = _sget(sub_obj, "customer")
+    subscription_id = _sget(sub_obj, "id")
+    stripe_status = _sget(sub_obj, "status") or "active"
+    current_period_end = None
+
+    cpe = _sget(sub_obj, "current_period_end")
+    if cpe:
+        current_period_end = datetime.fromtimestamp(cpe, tz=timezone.utc).isoformat()
+
+    # Resolve plan from price
+    plan = "premium"
+    items_data = _sget(_sget(sub_obj, "items") or {}, "data")
+    if items_data:
+        price_id = items_data[0]["price"]["id"]
+        plan = _resolve_plan_from_price(price_id)
+
+    user_id = _resolve_user_from_customer(db, customer_id)
+    if not user_id:
+        # Try metadata
+        meta = _sget(sub_obj, "metadata") or {}
+        meta_uid = _sget(meta, "user_id")
+        if meta_uid:
+            user_id = int(meta_uid)
+
+    if not user_id:
+        logger.error(
+            "subscription.updated: cannot resolve user for customer %s", customer_id
+        )
+        return
+
+    _upsert_subscription(
+        db,
+        user_id,
+        plan=plan,
+        stripe_customer_id=customer_id,
+        stripe_subscription_id=subscription_id,
+        status=stripe_status,
+        current_period_end=current_period_end,
+    )
+    _update_user_plan(db, user_id, plan)
+    _record_billing_event(db, event["id"], event["type"], user_id=user_id)
+    logger.info(
+        "subscription.updated: user %s plan=%s status=%s", user_id, plan, stripe_status
+    )
+
+
+def _handle_subscription_deleted(db, event):
+    """customer.subscription.deleted → downgrade to free."""
+    sub_obj = event["data"]["object"]
+    customer_id = _sget(sub_obj, "customer")
+
+    user_id = _resolve_user_from_customer(db, customer_id)
+    if not user_id:
+        meta = _sget(sub_obj, "metadata") or {}
+        meta_uid = _sget(meta, "user_id")
+        if meta_uid:
+            user_id = int(meta_uid)
+
+    if not user_id:
+        logger.error(
+            "subscription.deleted: cannot resolve user for customer %s", customer_id
+        )
+        return
+
+    _upsert_subscription(
+        db,
+        user_id,
+        plan="free",
+        stripe_subscription_id=None,
+        status="canceled",
+        end_date=datetime.now(timezone.utc).isoformat(),
+        current_period_end=None,
+        grace_period_end=None,
+    )
+    _update_user_plan(db, user_id, "free")
+    _record_billing_event(db, event["id"], event["type"], user_id=user_id)
+    logger.info("subscription.deleted: user %s downgraded to free", user_id)
+
+
+def _handle_payment_failed(db, event):
+    """invoice.payment_failed → mark past_due + 3-day grace period."""
+    invoice = event["data"]["object"]
+    customer_id = _sget(invoice, "customer")
+    subscription_id = _sget(invoice, "subscription")
+
+    user_id = _resolve_user_from_customer(db, customer_id)
+    if not user_id:
+        logger.error(
+            "invoice.payment_failed: cannot resolve user for customer %s", customer_id
+        )
+        return
+
+    grace_end = (datetime.now(timezone.utc) + timedelta(days=3)).isoformat()
+
+    _upsert_subscription(db, user_id, status="past_due", grace_period_end=grace_end)
+    _record_billing_event(
+        db,
+        event["id"],
+        event["type"],
+        user_id=user_id,
+        payload={"subscription_id": subscription_id},
+    )
+
+    # TODO: send notification email via /api/notifications
+    logger.warning(
+        "invoice.payment_failed: user %s past_due, grace period until %s",
+        user_id,
+        grace_end,
+    )
+
+
+def _handle_payment_succeeded(db, event):
+    """invoice.payment_succeeded → clear past_due / grace period."""
+    invoice = event["data"]["object"]
+    customer_id = _sget(invoice, "customer")
+
+    user_id = _resolve_user_from_customer(db, customer_id)
+    if not user_id:
+        return
+
+    # Refresh subscription period end
+    current_period_end = None
+    lines = _sget(invoice, "lines") or {}
+    lines_data = _sget(lines, "data") or []
+    if lines_data:
+        period = lines_data[0].get("period") or {}
+        period_end = (
+            period.get("end") if isinstance(period, dict) else _sget(period, "end")
+        )
+        if period_end:
+            current_period_end = datetime.fromtimestamp(
+                period_end, tz=timezone.utc
+            ).isoformat()
+
+    _upsert_subscription(
+        db,
+        user_id,
+        status="active",
+        grace_period_end=None,
+        current_period_end=current_period_end,
+    )
+    _record_billing_event(db, event["id"], event["type"], user_id=user_id)
+    logger.info("invoice.payment_succeeded: user %s payment cleared", user_id)
+
+
+# ──────────────────────────────────────────────────────────────
+# On-import: ensure DB migration ran
+# ──────────────────────────────────────────────────────────────
+
+try:
+    migrate_billing_tables()
+except Exception as _e:
+    logger.warning("billing_db migration skipped (test env?): %s", _e)

api_v1/routes/courses.py

@@ -0,0 +1,277 @@
+#!/usr/bin/env python3
+"""
+Courses routes for API v1.
+
+GET /api/v1/courses/today          — liste des courses du jour (public, paginated)
+GET /api/v1/courses/{id}/predictions — prédictions ML pour une course (free tier, 1/day limit)
+"""
+
+import os
+from datetime import datetime, timedelta
+from flask import Blueprint, jsonify, request, g
+
+from api_v1.utils import (
+    get_db,
+    table_exists,
+    error_response,
+    bad_request,
+    not_found,
+    internal_error,
+    get_pagination_params,
+    paginate_query,
+)
+from auth import jwt_required_middleware, free_daily_limit_check
+
+courses_bp = Blueprint("v1_courses", __name__, url_prefix="/api/v1/courses")
+
+
+# ──────────────────────────────────────────────────────────────
+# GET /api/v1/courses/today
+# ──────────────────────────────────────────────────────────────
+
+
+@courses_bp.route("/today", methods=["GET"])
+@jwt_required_middleware
+def courses_today():
+    """
+    Courses du jour
+    ---
+    tags:
+      - Courses
+    summary: Liste toutes les courses du jour avec info course
+    security:
+      - Bearer: []
+    parameters:
+      - name: filter
+        in: query
+        type: string
+        enum: [all, quinte, trot, plat]
+        default: all
+        description: Filtre par type de course
+      - name: limit
+        in: query
+        type: integer
+        default: 20
+      - name: offset
+        in: query
+        type: integer
+        default: 0
+    responses:
+      200:
+        description: Liste des courses du jour
+      401:
+        description: Token manquant ou invalide
+    """
+    race_filter = request.args.get("filter", "all").lower()
+    limit, offset = get_pagination_params(default_limit=50, max_limit=200)
+    today = datetime.now().strftime("%Y-%m-%d")
+
+    # Build SQL condition
+    if race_filter == "quinte":
+        cond = "AND (c.libelle LIKE '%Quinté%' OR c.libelle LIKE '%Quinte%')"
+    elif race_filter == "trot":
+        cond = "AND c.discipline LIKE '%Trot%'"
+    elif race_filter == "plat":
+        cond = "AND c.discipline LIKE '%Plat%'"
+    else:
+        cond = ""
+
+    conn = get_db()
+    try:
+        # Graceful handling if pmu_courses table doesn't exist yet
+        if not table_exists(conn, "pmu_courses"):
+            return jsonify(
+                {
+                    "status": "ok",
+                    "date": today,
+                    "filter": race_filter,
+                    "courses": [],
+                    "pagination": {
+                        "total": 0,
+                        "limit": limit,
+                        "offset": offset,
+                        "has_more": False,
+                    },
+                }
+            ), 200
+
+        # Count total
+        count_row = conn.execute(
+            f"""SELECT COUNT(*) as cnt
+                FROM pmu_courses c
+                WHERE c.date_programme = ? {cond}""",
+            (today,),
+        ).fetchone()
+        total = count_row["cnt"] if count_row else 0
+
+        rows = conn.execute(
+            f"""SELECT
+                    c.date_programme,
+                    c.num_reunion,
+                    c.num_course,
+                    c.libelle,
+                    c.discipline,
+                    c.distance,
+                    c.hippodrome,
+                    c.px_type,
+                    COUNT(p.id_cheval) as nb_partants
+                FROM pmu_courses c
+                LEFT JOIN pmu_partants p
+                    ON p.date_programme = c.date_programme
+                   AND p.num_reunion    = c.num_reunion
+                   AND p.num_course     = c.num_course
+                WHERE c.date_programme = ? {cond}
+                GROUP BY c.date_programme, c.num_reunion, c.num_course
+                ORDER BY c.num_reunion ASC, c.num_course ASC
+                LIMIT ? OFFSET ?""",
+            (today, limit, offset),
+        ).fetchall()
+
+        courses = []
+        for r in rows:
+            course_id = f"{r['num_reunion']}-{r['num_course']}"
+            courses.append(
+                {
+                    "id": course_id,
+                    "date": r["date_programme"],
+                    "num_reunion": r["num_reunion"],
+                    "num_course": r["num_course"],
+                    "libelle": r["libelle"],
+                    "discipline": r["discipline"],
+                    "distance": r["distance"],
+                    "hippodrome": r["hippodrome"],
+                    "type_pari": r["px_type"],
+                    "nb_partants": r["nb_partants"],
+                }
+            )
+
+        pagination = paginate_query(courses, total, limit, offset)
+
+        return jsonify(
+            {
+                "status": "ok",
+                "date": today,
+                "filter": race_filter,
+                "courses": courses,
+                **pagination,
+            }
+        ), 200
+
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()
+
+
+# ──────────────────────────────────────────────────────────────
+# GET /api/v1/courses/<course_id>/predictions
+# course_id format: "{num_reunion}-{num_course}"  e.g. "1-3"
+# ──────────────────────────────────────────────────────────────
+
+
+@courses_bp.route("/<course_id>/predictions", methods=["GET"])
+@jwt_required_middleware
+@free_daily_limit_check
+def course_predictions(course_id):
+    """
+    Prédictions pour une course
+    ---
+    tags:
+      - Courses
+    summary: Prédictions ML pour une course identifiée par {num_reunion}-{num_course}
+    security:
+      - Bearer: []
+    parameters:
+      - name: course_id
+        in: path
+        type: string
+        required: true
+        description: Identifiant de la course (format num_reunion-num_course, ex "1-3")
+      - name: date
+        in: query
+        type: string
+        format: date
+        description: Date de la course (YYYY-MM-DD), défaut = aujourd'hui
+    responses:
+      200:
+        description: Prédictions ML pour la course
+      400:
+        description: Paramètres invalides
+      404:
+        description: Course introuvable
+      429:
+        description: Limite quotidienne free tier atteinte
+    """
+    # Parse course_id
+    parts = course_id.split("-")
+    if len(parts) != 2:
+        return bad_request(
+            "course_id doit être au format {num_reunion}-{num_course}, ex: 1-3"
+        )
+
+    try:
+        num_reunion = int(parts[0])
+        num_course = int(parts[1])
+    except ValueError:
+        return bad_request("num_reunion et num_course doivent être des entiers")
+
+    date_param = request.args.get("date", datetime.now().strftime("%Y-%m-%d"))
+
+    conn = get_db()
+    try:
+        # Fetch course info
+        course_row = conn.execute(
+            """SELECT libelle, discipline, distance, hippodrome, px_type
+               FROM pmu_courses
+               WHERE date_programme = ? AND num_reunion = ? AND num_course = ?""",
+            (date_param, num_reunion, num_course),
+        ).fetchone()
+
+        if not course_row:
+            return not_found(
+                f"Course R{num_reunion}C{num_course} introuvable pour le {date_param}"
+            )
+
+        # Fetch ML predictions from cache
+        preds = []
+        if table_exists(conn, "ml_predictions_cache"):
+            preds = conn.execute(
+                """SELECT horse_name, horse_number, odds, prob_top1, prob_top3,
+                          ml_score, recommendation, is_value_bet, risque_label, risque_score
+                   FROM ml_predictions_cache
+                   WHERE date = ? AND num_reunion = ? AND num_course = ?
+                   ORDER BY ml_score DESC""",
+                (date_param, num_reunion, num_course),
+            ).fetchall()
+
+        # Fetch partants
+        partants = conn.execute(
+            """SELECT nom, num_pmu, cote_direct, cote_reference, tendance_cote, favoris,
+                      tx_victoire, tx_place, forme_recente, driver, entraineur, musique
+               FROM pmu_partants
+               WHERE date_programme = ? AND num_reunion = ? AND num_course = ?
+               ORDER BY num_pmu ASC""",
+            (date_param, num_reunion, num_course),
+        ).fetchall()
+
+        return jsonify(
+            {
+                "status": "ok",
+                "date": date_param,
+                "course": {
+                    "id": course_id,
+                    "libelle": course_row["libelle"],
+                    "discipline": course_row["discipline"],
+                    "distance": course_row["distance"],
+                    "hippodrome": course_row["hippodrome"],
+                    "type_pari": course_row["px_type"],
+                },
+                "predictions": [dict(p) for p in preds],
+                "partants": [dict(p) for p in partants],
+            }
+        ), 200
+
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()

api_v1/routes/export.py

@@ -0,0 +1,185 @@
+#!/usr/bin/env python3
+"""
+Export route for API v1.
+
+GET /api/v1/export/csv  — Export CSV des prédictions ou paris (pro)
+"""
+
+import csv
+import io
+from datetime import datetime, timedelta
+from flask import Blueprint, jsonify, request, Response
+
+from api_v1.utils import (
+    get_db,
+    table_exists,
+    internal_error,
+    bad_request,
+    forbidden,
+)
+from auth import jwt_required_middleware, plan_required
+
+export_bp = Blueprint("v1_export", __name__, url_prefix="/api/v1/export")
+
+# Maximum rows exportable in one request
+EXPORT_MAX_ROWS = 5000
+
+
+@export_bp.route("/csv", methods=["GET"])
+@jwt_required_middleware
+@plan_required("pro")
+def export_csv():
+    """
+    Export CSV
+    ---
+    tags:
+      - Export
+    summary: Export CSV des prédictions ML ou des paris historiques — accès pro uniquement
+    security:
+      - Bearer: []
+    parameters:
+      - name: type
+        in: query
+        type: string
+        enum: [predictions, bets]
+        default: predictions
+        description: Type de données à exporter
+      - name: start
+        in: query
+        type: string
+        format: date
+        description: Date de début (YYYY-MM-DD)
+      - name: end
+        in: query
+        type: string
+        format: date
+        description: Date de fin (YYYY-MM-DD)
+      - name: date
+        in: query
+        type: string
+        format: date
+        description: Date unique (YYYY-MM-DD), ignoré si start/end fournis
+    responses:
+      200:
+        description: Fichier CSV
+        content:
+          text/csv:
+            schema:
+              type: string
+      400:
+        description: Paramètre invalide
+      401:
+        description: Token invalide
+      403:
+        description: Plan insuffisant (pro requis)
+    """
+    export_type = request.args.get("type", "predictions").lower()
+    if export_type not in ("predictions", "bets"):
+        return bad_request(
+            "Paramètre 'type' invalide. Valeurs acceptées: predictions, bets"
+        )
+
+    start = request.args.get("start")
+    end = request.args.get("end")
+    date = request.args.get("date", datetime.now().strftime("%Y-%m-%d"))
+
+    for label, val in [("start", start), ("end", end), ("date", date)]:
+        if val:
+            try:
+                datetime.strptime(val, "%Y-%m-%d")
+            except ValueError:
+                return bad_request(
+                    f"Paramètre '{label}' invalide, format attendu: YYYY-MM-DD"
+                )
+
+    # Build date range
+    if start and end:
+        date_cond = "date BETWEEN ? AND ?"
+        date_params = [start, end]
+    elif start:
+        date_cond = "date >= ?"
+        date_params = [start]
+    else:
+        date_cond = "date = ?"
+        date_params = [date]
+
+    conn = get_db()
+    try:
+        output = io.StringIO()
+
+        if export_type == "predictions":
+            if not table_exists(conn, "ml_predictions_cache"):
+                return bad_request("Table ml_predictions_cache introuvable")
+
+            rows = conn.execute(
+                f"""SELECT date, race_label, hippodrome, discipline, distance, heure,
+                           horse_name, horse_number, odds, prob_top1, prob_top3,
+                           ml_score, recommendation, is_value_bet, risque_label
+                    FROM ml_predictions_cache
+                    WHERE {date_cond}
+                    ORDER BY date DESC, ml_score DESC
+                    LIMIT {EXPORT_MAX_ROWS}""",
+                date_params,
+            ).fetchall()
+
+            fieldnames = [
+                "date",
+                "race_label",
+                "hippodrome",
+                "discipline",
+                "distance",
+                "heure",
+                "horse_name",
+                "horse_number",
+                "odds",
+                "prob_top1",
+                "prob_top3",
+                "ml_score",
+                "recommendation",
+                "is_value_bet",
+                "risque_label",
+            ]
+
+        else:  # bets
+            if not table_exists(conn, "bet_results"):
+                return bad_request("Table bet_results introuvable")
+
+            rows = conn.execute(
+                f"""SELECT date, race_name, type_pari, horse_name, horse_number,
+                           COALESCE(cote, 0) AS cote, mise, resultat, gain
+                    FROM bet_results
+                    WHERE {date_cond}
+                    ORDER BY date DESC
+                    LIMIT {EXPORT_MAX_ROWS}""",
+                date_params,
+            ).fetchall()
+
+            fieldnames = [
+                "date",
+                "race_name",
+                "type_pari",
+                "horse_name",
+                "horse_number",
+                "cote",
+                "mise",
+                "resultat",
+                "gain",
+            ]
+
+        writer = csv.DictWriter(output, fieldnames=fieldnames, extrasaction="ignore")
+        writer.writeheader()
+        for row in rows:
+            writer.writerow(dict(row))
+
+        filename = f"turf_{export_type}_{date_params[0]}.csv"
+        return Response(
+            output.getvalue(),
+            status=200,
+            mimetype="text/csv",
+            headers={"Content-Disposition": f'attachment; filename="{filename}"'},
+        )
+
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()

api_v1/routes/health.py

@@ -0,0 +1,44 @@
+#!/usr/bin/env python3
+"""
+GET /api/v1/health — public healthcheck endpoint.
+No authentication required.
+"""
+
+from flask import Blueprint, jsonify
+from datetime import datetime, timezone
+
+health_bp = Blueprint("v1_health", __name__, url_prefix="/api/v1")
+
+
+@health_bp.route("/health", methods=["GET"])
+def health():
+    """
+    Health check
+    ---
+    tags:
+      - System
+    summary: Public healthcheck — returns API status and timestamp
+    responses:
+      200:
+        description: API is healthy
+        schema:
+          type: object
+          properties:
+            status:
+              type: string
+              example: ok
+            version:
+              type: string
+              example: "1.0"
+            timestamp:
+              type: string
+              format: date-time
+    """
+    return jsonify(
+        {
+            "status": "ok",
+            "version": "1.0",
+            "api": "Turf SaaS API v1",
+            "timestamp": datetime.now(timezone.utc).isoformat(),
+        }
+    ), 200

api_v1/routes/metrics.py

@@ -0,0 +1,144 @@
+#!/usr/bin/env python3
+"""
+Metrics route for API v1.
+
+GET /api/v1/metrics  — Métriques performances ML (premium+)
+"""
+
+from datetime import datetime, timedelta
+from flask import Blueprint, jsonify, request
+
+from api_v1.utils import (
+    get_db,
+    table_exists,
+    internal_error,
+    bad_request,
+)
+from auth import jwt_required_middleware, plan_required
+
+metrics_bp = Blueprint("v1_metrics", __name__, url_prefix="/api/v1")
+
+
+@metrics_bp.route("/metrics", methods=["GET"])
+@jwt_required_middleware
+@plan_required("premium", "pro")
+def metrics():
+    """
+    Métriques ML
+    ---
+    tags:
+      - Métriques
+    summary: Métriques de performance du modèle ML (precision, ROI, top-3 rate) — premium+
+    security:
+      - Bearer: []
+    parameters:
+      - name: days
+        in: query
+        type: integer
+        default: 30
+        description: Nombre de jours à analyser (max 365)
+    responses:
+      200:
+        description: Métriques de performance ML
+      401:
+        description: Token invalide
+      403:
+        description: Plan insuffisant (premium ou pro requis)
+    """
+    try:
+        days = int(request.args.get("days", 30))
+    except (ValueError, TypeError):
+        return bad_request("Paramètre 'days' doit être un entier")
+
+    days = max(1, min(days, 365))
+    end_date = datetime.now().strftime("%Y-%m-%d")
+    start_date = (datetime.now() - timedelta(days=days)).strftime("%Y-%m-%d")
+
+    conn = get_db()
+    try:
+        # ── Bet-level metrics from bet_results ──
+        bet_metrics = {
+            "available": False,
+            "period": {"start": start_date, "end": end_date, "days": days},
+        }
+        ml_metrics = {"available": False}
+        daily_stats = []
+
+        if table_exists(conn, "bet_results"):
+            row = conn.execute(
+                """SELECT
+                       COUNT(*)                                           AS total,
+                       SUM(CASE WHEN resultat='GAGNE' THEN 1 ELSE 0 END)  AS gagne,
+                       SUM(mise)                                           AS mise,
+                       SUM(gain)                                           AS gain
+                   FROM bet_results
+                   WHERE date BETWEEN ? AND ?""",
+                (start_date, end_date),
+            ).fetchone()
+
+            total = row["total"] or 0
+            gagne = row["gagne"] or 0
+            mise = float(row["mise"] or 0)
+            gain = float(row["gain"] or 0)
+
+            bet_metrics = {
+                "available": True,
+                "period": {"start": start_date, "end": end_date, "days": days},
+                "total_bets": total,
+                "precision_pct": round(gagne / total * 100, 2) if total > 0 else 0.0,
+                "roi_pct": round((gain - mise) / mise * 100, 2) if mise > 0 else 0.0,
+                "mise_totale": round(mise, 2),
+                "gain_total": round(gain, 2),
+            }
+
+        # ── ML predictions cache metrics ──
+        if table_exists(conn, "ml_predictions_cache"):
+            cache_row = conn.execute(
+                """SELECT
+                       COUNT(*)                            AS total,
+                       SUM(is_value_bet)                   AS value_bets,
+                       AVG(prob_top1)                      AS avg_prob_top1,
+                       AVG(prob_top3)                      AS avg_prob_top3,
+                       AVG(ml_score)                       AS avg_ml_score
+                   FROM ml_predictions_cache
+                   WHERE date BETWEEN ? AND ?""",
+                (start_date, end_date),
+            ).fetchone()
+
+            if cache_row and cache_row["total"]:
+                ml_metrics = {
+                    "available": True,
+                    "total_predictions": cache_row["total"],
+                    "value_bets": cache_row["value_bets"] or 0,
+                    "avg_prob_top1": round(float(cache_row["avg_prob_top1"] or 0), 4),
+                    "avg_prob_top3": round(float(cache_row["avg_prob_top3"] or 0), 4),
+                    "avg_ml_score": round(float(cache_row["avg_ml_score"] or 0), 4),
+                }
+
+        # ── Daily breakdown ──
+        if table_exists(conn, "daily_stats"):
+            daily_rows = conn.execute(
+                """SELECT date, total_bets, bets_gagne, precision_pct, roi_pct,
+                          mise_totale, gain_total
+                   FROM daily_stats
+                   WHERE date BETWEEN ? AND ?
+                   ORDER BY date DESC
+                   LIMIT 60""",
+                (start_date, end_date),
+            ).fetchall()
+            daily_stats = [dict(r) for r in daily_rows]
+
+        return jsonify(
+            {
+                "status": "ok",
+                "period": {"start": start_date, "end": end_date, "days": days},
+                "bet_metrics": bet_metrics,
+                "ml_metrics": ml_metrics,
+                "daily": daily_stats,
+            }
+        ), 200
+
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()

api_v1/routes/predictions.py

@@ -0,0 +1,163 @@
+#!/usr/bin/env python3
+"""
+Predictions routes for API v1.
+
+GET /api/v1/predictions/top3   — Top 3 global du jour (free tier, 1/day limit)
+GET /api/v1/predictions/all    — Toutes prédictions (premium+)
+"""
+
+from datetime import datetime, timedelta
+from flask import Blueprint, jsonify, request
+
+from api_v1.utils import (
+    get_db,
+    table_exists,
+    internal_error,
+    not_found,
+    get_pagination_params,
+    paginate_query,
+)
+from auth import jwt_required_middleware, plan_required, free_daily_limit_check
+
+predictions_bp = Blueprint("v1_predictions", __name__, url_prefix="/api/v1/predictions")
+
+
+def _fetch_ml_predictions(conn, date: str, limit: int = None, offset: int = 0):
+    """Shared helper — returns rows from ml_predictions_cache."""
+    if not table_exists(conn, "ml_predictions_cache"):
+        return [], 0
+
+    count_row = conn.execute(
+        "SELECT COUNT(*) as cnt FROM ml_predictions_cache WHERE date = ?",
+        (date,),
+    ).fetchone()
+    total = count_row["cnt"] if count_row else 0
+
+    sql = """SELECT
+                 race_label, hippodrome, discipline, distance, heure,
+                 horse_name, horse_number, odds, prob_top1, prob_top3,
+                 ml_score, recommendation, is_value_bet, risque_label, risque_score
+             FROM ml_predictions_cache
+             WHERE date = ?
+             ORDER BY ml_score DESC"""
+    params = [date]
+
+    if limit is not None:
+        sql += " LIMIT ? OFFSET ?"
+        params += [limit, offset]
+
+    rows = conn.execute(sql, params).fetchall()
+    return [dict(r) for r in rows], total
+
+
+# ──────────────────────────────────────────────────────────────
+# GET /api/v1/predictions/top3
+# ──────────────────────────────────────────────────────────────
+
+
+@predictions_bp.route("/top3", methods=["GET"])
+@jwt_required_middleware
+@free_daily_limit_check
+def predictions_top3():
+    """
+    Top 3 prédictions du jour
+    ---
+    tags:
+      - Prédictions
+    summary: Top 3 chevaux avec le meilleur score ML du jour (free tier inclus)
+    security:
+      - Bearer: []
+    parameters:
+      - name: date
+        in: query
+        type: string
+        format: date
+        description: Date au format YYYY-MM-DD (défaut aujourd'hui)
+    responses:
+      200:
+        description: Top 3 prédictions ML du jour
+      401:
+        description: Token invalide
+      429:
+        description: Limite quotidienne free tier atteinte
+    """
+    date_param = request.args.get("date", datetime.now().strftime("%Y-%m-%d"))
+
+    conn = get_db()
+    try:
+        predictions, _ = _fetch_ml_predictions(conn, date_param, limit=3, offset=0)
+
+        return jsonify(
+            {
+                "status": "ok",
+                "date": date_param,
+                "top3": predictions,
+            }
+        ), 200
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()
+
+
+# ──────────────────────────────────────────────────────────────
+# GET /api/v1/predictions/all
+# ──────────────────────────────────────────────────────────────
+
+
+@predictions_bp.route("/all", methods=["GET"])
+@jwt_required_middleware
+@plan_required("premium", "pro")
+def predictions_all():
+    """
+    Toutes les prédictions du jour
+    ---
+    tags:
+      - Prédictions
+    summary: Toutes les prédictions ML du jour — accès premium et pro uniquement
+    security:
+      - Bearer: []
+    parameters:
+      - name: date
+        in: query
+        type: string
+        format: date
+        description: Date au format YYYY-MM-DD (défaut aujourd'hui)
+      - name: limit
+        in: query
+        type: integer
+        default: 20
+      - name: offset
+        in: query
+        type: integer
+        default: 0
+    responses:
+      200:
+        description: Toutes les prédictions ML
+      401:
+        description: Token invalide
+      403:
+        description: Plan insuffisant (premium ou pro requis)
+    """
+    date_param = request.args.get("date", datetime.now().strftime("%Y-%m-%d"))
+    limit, offset = get_pagination_params(default_limit=50, max_limit=500)
+
+    conn = get_db()
+    try:
+        predictions, total = _fetch_ml_predictions(
+            conn, date_param, limit=limit, offset=offset
+        )
+        pagination = paginate_query(predictions, total, limit, offset)
+
+        return jsonify(
+            {
+                "status": "ok",
+                "date": date_param,
+                "predictions": predictions,
+                **pagination,
+            }
+        ), 200
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()

api_v1/routes/valuebets.py

@@ -0,0 +1,111 @@
+#!/usr/bin/env python3
+"""
+Value bets route for API v1.
+
+GET /api/v1/valuebets  — Value bets du jour (premium+)
+"""
+
+from datetime import datetime
+from flask import Blueprint, jsonify, request
+
+from api_v1.utils import (
+    get_db,
+    table_exists,
+    internal_error,
+    get_pagination_params,
+    paginate_query,
+)
+from auth import jwt_required_middleware, plan_required
+
+valuebets_bp = Blueprint("v1_valuebets", __name__, url_prefix="/api/v1")
+
+
+@valuebets_bp.route("/valuebets", methods=["GET"])
+@jwt_required_middleware
+@plan_required("premium", "pro")
+def valuebets():
+    """
+    Value bets du jour
+    ---
+    tags:
+      - Value Bets
+    summary: Value bets du jour — chevaux à cote surévaluée par le marché (premium+)
+    security:
+      - Bearer: []
+    parameters:
+      - name: date
+        in: query
+        type: string
+        format: date
+        description: Date YYYY-MM-DD (défaut aujourd'hui)
+      - name: min_odds
+        in: query
+        type: number
+        default: 2.0
+        description: Cote minimale pour filtrer les value bets
+      - name: limit
+        in: query
+        type: integer
+        default: 20
+      - name: offset
+        in: query
+        type: integer
+        default: 0
+    responses:
+      200:
+        description: Value bets du jour
+      401:
+        description: Token invalide
+      403:
+        description: Plan insuffisant (premium ou pro requis)
+    """
+    date_param = request.args.get("date", datetime.now().strftime("%Y-%m-%d"))
+    limit, offset = get_pagination_params(default_limit=20, max_limit=100)
+
+    try:
+        min_odds = float(request.args.get("min_odds", 2.0))
+    except (ValueError, TypeError):
+        min_odds = 2.0
+
+    conn = get_db()
+    try:
+        rows = []
+        total = 0
+
+        if table_exists(conn, "ml_predictions_cache"):
+            count_row = conn.execute(
+                """SELECT COUNT(*) as cnt
+                   FROM ml_predictions_cache
+                   WHERE date = ? AND is_value_bet = 1 AND odds >= ?""",
+                (date_param, min_odds),
+            ).fetchone()
+            total = count_row["cnt"] if count_row else 0
+
+            rows = conn.execute(
+                """SELECT race_label, hippodrome, discipline, distance, heure,
+                          horse_name, horse_number, odds, prob_top1, prob_top3,
+                          ml_score, recommendation, risque_label, risque_score
+                   FROM ml_predictions_cache
+                   WHERE date = ? AND is_value_bet = 1 AND odds >= ?
+                   ORDER BY ml_score DESC
+                   LIMIT ? OFFSET ?""",
+                (date_param, min_odds, limit, offset),
+            ).fetchall()
+
+        valuebets_list = [dict(r) for r in rows]
+        pagination = paginate_query(valuebets_list, total, limit, offset)
+
+        return jsonify(
+            {
+                "status": "ok",
+                "date": date_param,
+                "min_odds": min_odds,
+                "valuebets": valuebets_list,
+                **pagination,
+            }
+        ), 200
+
+    except Exception as e:
+        return internal_error(str(e))
+    finally:
+        conn.close()

api_v1/utils.py

@@ -0,0 +1,98 @@
+#!/usr/bin/env python3
+"""
+Shared utilities for API v1 — error helpers, pagination, DB access.
+"""
+
+import sqlite3
+import os
+from flask import jsonify, request
+
+DB_PATH = os.environ.get("TURF_SAAS_DB", "/home/h3r7/turf_saas/turf_saas.db")
+
+
+# ──────────────────────────────────────────────────────────────
+# Database
+# ──────────────────────────────────────────────────────────────
+
+
+def get_db():
+    """Return a SQLite connection with Row factory."""
+    conn = sqlite3.connect(DB_PATH)
+    conn.row_factory = sqlite3.Row
+    return conn
+
+
+def table_exists(conn, table_name: str) -> bool:
+    row = conn.execute(
+        "SELECT 1 FROM sqlite_master WHERE type='table' AND name=?", (table_name,)
+    ).fetchone()
+    return row is not None
+
+
+# ──────────────────────────────────────────────────────────────
+# Uniform error responses
+# ──────────────────────────────────────────────────────────────
+
+
+def error_response(message: str, code: int, status: str = "error"):
+    """Return a JSON error envelope consistent with the API contract.
+
+    Shape: {"status": "error", "message": "...", "code": 400}
+    """
+    return jsonify({"status": status, "message": message, "code": code}), code
+
+
+def not_found(message: str = "Resource not found"):
+    return error_response(message, 404)
+
+
+def bad_request(message: str = "Bad request"):
+    return error_response(message, 400)
+
+
+def forbidden(message: str = "Forbidden", required_plans=None, current_plan=None):
+    payload = {"status": "error", "message": message, "code": 403}
+    if required_plans:
+        payload["required_plans"] = required_plans
+    if current_plan:
+        payload["current_plan"] = current_plan
+        payload["upgrade_url"] = "/api/v1/subscription/upgrade"
+    return jsonify(payload), 403
+
+
+def internal_error(message: str = "Internal server error"):
+    return error_response(message, 500)
+
+
+# ──────────────────────────────────────────────────────────────
+# Pagination helpers
+# ──────────────────────────────────────────────────────────────
+
+
+def get_pagination_params(default_limit: int = 20, max_limit: int = 100):
+    """Extract and validate limit/offset from query-string."""
+    try:
+        limit = int(request.args.get("limit", default_limit))
+    except (ValueError, TypeError):
+        limit = default_limit
+
+    try:
+        offset = int(request.args.get("offset", 0))
+    except (ValueError, TypeError):
+        offset = 0
+
+    limit = max(1, min(limit, max_limit))
+    offset = max(0, offset)
+    return limit, offset
+
+
+def paginate_query(rows, total: int, limit: int, offset: int):
+    """Wrap a list of rows in a pagination envelope."""
+    return {
+        "pagination": {
+            "total": total,
+            "limit": limit,
+            "offset": offset,
+            "has_more": (offset + limit) < total,
+        }
+    }

app_v1.py

@@ -0,0 +1,138 @@
+#!/usr/bin/env python3
+"""
+app_v1.py — Turf SaaS Flask application with versioned API /v1/
+
+This module creates the Flask app, registers:
+  - Auth JWT (from Sprint 2-3)
+  - API v1 blueprints
+  - Swagger/OpenAPI documentation at /api/v1/docs
+
+Usage:
+    python app_v1.py
+    # or via gunicorn:
+    gunicorn -w 2 -b 0.0.0.0:8792 app_v1:app
+
+Sprint 3-4: HRT-29 — Refacto API /v1/
+"""
+
+import os
+import logging
+from datetime import timedelta
+
+from flask import Flask, jsonify
+from flask_cors import CORS
+from flask_jwt_extended import JWTManager
+from flasgger import Swagger
+
+from auth_db import init_auth_tables
+from auth import auth_bp
+from api_v1 import register_api_v1
+
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
+)
+logger = logging.getLogger("turf_saas.app_v1")
+
+
+def create_app() -> Flask:
+    """Application factory."""
+    app = Flask(__name__)
+
+    # ── CORS ──
+    CORS(app, origins=["*"], methods=["GET", "POST", "PUT", "DELETE", "OPTIONS"])
+
+    # ── JWT config ──
+    app.config["JWT_SECRET_KEY"] = os.environ.get(
+        "JWT_SECRET_KEY", "change-me-in-production-use-strong-random-secret"
+    )
+    app.config["JWT_ACCESS_TOKEN_EXPIRES"] = timedelta(minutes=15)
+    app.config["JWT_REFRESH_TOKEN_EXPIRES"] = timedelta(days=30)
+    JWTManager(app)
+
+    # ── Swagger / OpenAPI ──
+    swagger_config = {
+        "headers": [],
+        "specs": [
+            {
+                "endpoint": "apispec_v1",
+                "route": "/api/v1/apispec.json",
+                "rule_filter": lambda rule: str(rule).startswith("/api/v1"),
+                "model_filter": lambda tag: True,
+            }
+        ],
+        "static_url_path": "/flasgger_static",
+        "swagger_ui": True,
+        "specs_route": "/api/v1/docs",
+    }
+
+    swagger_template = {
+        "swagger": "2.0",
+        "info": {
+            "title": "Turf SaaS API",
+            "description": (
+                "API v1 — Prédictions turf IA, value bets, backtest & métriques.\n\n"
+                "**Plans:** `free` | `premium` | `pro`\n\n"
+                "**Auth:** Bearer JWT — obtenir un token via `POST /api/v1/auth/login`"
+            ),
+            "version": "1.0.0",
+            "contact": {"name": "H3R7 Tech"},
+        },
+        "basePath": "/",
+        "schemes": ["http", "https"],
+        "securityDefinitions": {
+            "Bearer": {
+                "type": "apiKey",
+                "name": "Authorization",
+                "in": "header",
+                "description": "Entrer: **Bearer <token>**",
+            }
+        },
+        "consumes": ["application/json"],
+        "produces": ["application/json"],
+    }
+
+    Swagger(app, config=swagger_config, template=swagger_template)
+
+    # ── Auth DB init ──
+    with app.app_context():
+        try:
+            init_auth_tables()
+        except Exception as e:
+            logger.warning("init_auth_tables warning: %s", e)
+
+    # ── Register auth blueprint ──
+    app.register_blueprint(auth_bp)
+
+    # ── Register API v1 blueprints ──
+    register_api_v1(app)
+
+    # ── Global error handlers ──
+    @app.errorhandler(404)
+    def not_found_handler(e):
+        return jsonify(
+            {"status": "error", "message": "Route introuvable", "code": 404}
+        ), 404
+
+    @app.errorhandler(405)
+    def method_not_allowed_handler(e):
+        return jsonify(
+            {"status": "error", "message": "Méthode non autorisée", "code": 405}
+        ), 405
+
+    @app.errorhandler(500)
+    def internal_error_handler(e):
+        logger.exception("Unhandled 500 error")
+        return jsonify(
+            {"status": "error", "message": "Erreur serveur interne", "code": 500}
+        ), 500
+
+    logger.info("Turf SaaS API v1 ready — docs at /api/v1/docs")
+    return app
+
+
+app = create_app()
+
+if __name__ == "__main__":
+    port = int(os.environ.get("PORT", 8792))
+    app.run(host="0.0.0.0", port=port, debug=False)

auth.py

@@ -0,0 +1,362 @@
+#!/usr/bin/env python3
+"""
+Auth Blueprint — JWT authentication + multi-tenant plan enforcement
+Sprint 2-3: HRT-28
+
+Endpoints:
+  POST /api/v1/auth/register  — email/password registration
+  POST /api/v1/auth/login     — returns access_token (15min) + refresh_token (30d)
+  POST /api/v1/auth/refresh   — rotate refresh token, issue new access_token
+  POST /api/v1/auth/logout    — revoke refresh token
+
+Middleware exposed:
+  jwt_required_middleware()   — decorator: valid access JWT required
+  plan_required(plans)        — decorator: user plan must be in given list
+"""
+
+import os
+import hashlib
+import secrets
+import logging
+from datetime import datetime, timedelta, timezone
+from functools import wraps
+
+import bcrypt
+from flask import Blueprint, request, jsonify, g, current_app
+from flask_jwt_extended import (
+    JWTManager,
+    create_access_token,
+    create_refresh_token,
+    decode_token,
+    get_jwt_identity,
+    verify_jwt_in_request,
+)
+from flask_jwt_extended.exceptions import JWTExtendedException
+from jwt.exceptions import PyJWTError
+
+from auth_db import get_db
+
+logger = logging.getLogger("turf_saas.auth")
+
+auth_bp = Blueprint("auth", __name__, url_prefix="/api/v1/auth")
+
+# ──────────────────────────────────────────────────────────────
+# Helpers
+# ──────────────────────────────────────────────────────────────
+
+
+def _hash_token(raw_token: str) -> str:
+    """SHA-256 hash of a token string for secure DB storage."""
+    return hashlib.sha256(raw_token.encode()).hexdigest()
+
+
+def _get_user_by_email(email: str):
+    db = get_db()
+    user = db.execute(
+        "SELECT * FROM users WHERE email = ? AND is_active = 1", (email.lower(),)
+    ).fetchone()
+    db.close()
+    return user
+
+
+def _get_user_by_id(user_id: int):
+    db = get_db()
+    user = db.execute(
+        "SELECT * FROM users WHERE id = ? AND is_active = 1", (user_id,)
+    ).fetchone()
+    db.close()
+    return user
+
+
+def _store_refresh_token(user_id: int, raw_token: str, expires_at: datetime):
+    token_hash = _hash_token(raw_token)
+    db = get_db()
+    db.execute(
+        "INSERT INTO refresh_tokens (user_id, token_hash, expires_at) VALUES (?,?,?)",
+        (user_id, token_hash, expires_at.isoformat()),
+    )
+    db.commit()
+    db.close()
+
+
+def _revoke_refresh_token(raw_token: str):
+    token_hash = _hash_token(raw_token)
+    db = get_db()
+    db.execute(
+        "UPDATE refresh_tokens SET revoked = 1 WHERE token_hash = ?", (token_hash,)
+    )
+    db.commit()
+    db.close()
+
+
+def _is_refresh_token_valid(raw_token: str, user_id: int) -> bool:
+    token_hash = _hash_token(raw_token)
+    db = get_db()
+    row = db.execute(
+        """SELECT id FROM refresh_tokens
+           WHERE token_hash = ? AND user_id = ? AND revoked = 0
+             AND expires_at > datetime('now')""",
+        (token_hash, user_id),
+    ).fetchone()
+    db.close()
+    return row is not None
+
+
+# ──────────────────────────────────────────────────────────────
+# Auth endpoints
+# ──────────────────────────────────────────────────────────────
+
+
+@auth_bp.route("/register", methods=["POST"])
+def register():
+    """POST /api/v1/auth/register — create a new user account (plan=free)."""
+    data = request.get_json(silent=True) or {}
+    email = (data.get("email") or "").strip().lower()
+    password = data.get("password") or ""
+
+    if not email or "@" not in email:
+        return jsonify({"error": "Email invalide"}), 400
+    if len(password) < 8:
+        return jsonify({"error": "Mot de passe trop court (min 8 caractères)"}), 400
+
+    # Check uniqueness
+    existing = _get_user_by_email(email)
+    if existing:
+        return jsonify({"error": "Email déjà enregistré"}), 409
+
+    password_hash = bcrypt.hashpw(password.encode(), bcrypt.gensalt()).decode()
+
+    db = get_db()
+    try:
+        cursor = db.execute(
+            "INSERT INTO users (email, password_hash, plan) VALUES (?,?,?)",
+            (email, password_hash, "free"),
+        )
+        user_id = cursor.lastrowid
+        # Create initial subscription record
+        db.execute(
+            "INSERT INTO subscriptions (user_id, plan) VALUES (?,?)",
+            (user_id, "free"),
+        )
+        db.commit()
+    except Exception as e:
+        db.rollback()
+        logger.error("register error: %s", e)
+        return jsonify({"error": "Erreur interne"}), 500
+    finally:
+        db.close()
+
+    logger.info("New user registered: %s (id=%s)", email, user_id)
+    return jsonify({"message": "Compte créé avec succès", "user_id": user_id}), 201
+
+
+@auth_bp.route("/login", methods=["POST"])
+def login():
+    """POST /api/v1/auth/login — returns JWT access_token + refresh_token."""
+    data = request.get_json(silent=True) or {}
+    email = (data.get("email") or "").strip().lower()
+    password = data.get("password") or ""
+
+    if not email or not password:
+        return jsonify({"error": "Email et mot de passe requis"}), 400
+
+    user = _get_user_by_email(email)
+    if not user:
+        return jsonify({"error": "Identifiants invalides"}), 401
+
+    if not bcrypt.checkpw(password.encode(), user["password_hash"].encode()):
+        logger.warning("Failed login attempt for %s", email)
+        return jsonify({"error": "Identifiants invalides"}), 401
+
+    # Create tokens
+    identity = str(user["id"])
+    additional_claims = {"plan": user["plan"], "email": user["email"]}
+
+    access_token = create_access_token(
+        identity=identity,
+        additional_claims=additional_claims,
+    )
+    raw_refresh = create_refresh_token(identity=identity)
+
+    refresh_expires = datetime.now(timezone.utc) + timedelta(days=30)
+    _store_refresh_token(user["id"], raw_refresh, refresh_expires)
+
+    logger.info("User %s logged in (plan=%s)", email, user["plan"])
+    return jsonify(
+        {
+            "access_token": access_token,
+            "refresh_token": raw_refresh,
+            "token_type": "Bearer",
+            "plan": user["plan"],
+        }
+    ), 200
+
+
+@auth_bp.route("/refresh", methods=["POST"])
+def refresh():
+    """POST /api/v1/auth/refresh — rotate refresh token, issue new access_token."""
+    data = request.get_json(silent=True) or {}
+    raw_refresh = (data.get("refresh_token") or "").strip()
+
+    if not raw_refresh:
+        return jsonify({"error": "refresh_token manquant"}), 400
+
+    # Decode without verifying in DB first (to get user_id)
+    try:
+        decoded = decode_token(raw_refresh)
+    except Exception:
+        return jsonify({"error": "Refresh token invalide ou expiré"}), 401
+
+    user_id = int(decoded.get("sub", 0))
+
+    if not _is_refresh_token_valid(raw_refresh, user_id):
+        return jsonify({"error": "Refresh token invalide, révoqué ou expiré"}), 401
+
+    user = _get_user_by_id(user_id)
+    if not user:
+        return jsonify({"error": "Utilisateur introuvable"}), 401
+
+    # Revoke old refresh token (rotation)
+    _revoke_refresh_token(raw_refresh)
+
+    # Issue new tokens
+    identity = str(user["id"])
+    additional_claims = {"plan": user["plan"], "email": user["email"]}
+    new_access = create_access_token(
+        identity=identity, additional_claims=additional_claims
+    )
+    new_refresh = create_refresh_token(identity=identity)
+
+    refresh_expires = datetime.now(timezone.utc) + timedelta(days=30)
+    _store_refresh_token(user["id"], new_refresh, refresh_expires)
+
+    logger.info("Token refreshed for user_id=%s", user_id)
+    return jsonify(
+        {
+            "access_token": new_access,
+            "refresh_token": new_refresh,
+            "token_type": "Bearer",
+            "plan": user["plan"],
+        }
+    ), 200
+
+
+@auth_bp.route("/logout", methods=["POST"])
+def logout():
+    """POST /api/v1/auth/logout — revoke refresh token."""
+    data = request.get_json(silent=True) or {}
+    raw_refresh = (data.get("refresh_token") or "").strip()
+
+    if raw_refresh:
+        _revoke_refresh_token(raw_refresh)
+
+    return jsonify({"message": "Déconnexion réussie"}), 200
+
+
+# ──────────────────────────────────────────────────────────────
+# JWT-protected middleware
+# ──────────────────────────────────────────────────────────────
+
+
+def jwt_required_middleware(fn):
+    """Decorator: require a valid Bearer JWT access token."""
+
+    @wraps(fn)
+    def wrapper(*args, **kwargs):
+        try:
+            verify_jwt_in_request()
+            user_id = int(get_jwt_identity())
+            user = _get_user_by_id(user_id)
+            if not user:
+                return jsonify({"error": "Utilisateur introuvable"}), 401
+            g.current_user = dict(user)
+            g.current_user_id = user_id
+        except (JWTExtendedException, PyJWTError) as e:
+            logger.debug("JWT auth failed: %s", e)
+            return jsonify({"error": "Token invalide ou expiré", "detail": str(e)}), 401
+        return fn(*args, **kwargs)
+
+    return wrapper
+
+
+def plan_required(*allowed_plans):
+    """
+    Decorator factory: user's plan must be in allowed_plans.
+    Must be applied AFTER @jwt_required_middleware.
+
+    Example:
+        @app.route("/api/v1/predictions")
+        @jwt_required_middleware
+        @plan_required("premium", "pro")
+        def premium_predictions():
+            ...
+    """
+
+    def decorator(fn):
+        @wraps(fn)
+        def wrapper(*args, **kwargs):
+            user = getattr(g, "current_user", None)
+            if not user:
+                return jsonify({"error": "Non authentifié"}), 401
+            if user["plan"] not in allowed_plans:
+                return jsonify(
+                    {
+                        "error": "Plan insuffisant",
+                        "required": list(allowed_plans),
+                        "current_plan": user["plan"],
+                        "upgrade_url": "/api/v1/subscription/upgrade",
+                    }
+                ), 403
+            return fn(*args, **kwargs)
+
+        return wrapper
+
+    return decorator
+
+
+def free_daily_limit_check(fn):
+    """
+    Decorator: enforce free plan daily limit (1 course/jour).
+    Must be applied AFTER @jwt_required_middleware.
+    """
+
+    @wraps(fn)
+    def wrapper(*args, **kwargs):
+        user = getattr(g, "current_user", None)
+        if not user or user["plan"] != "free":
+            return fn(*args, **kwargs)
+
+        today = datetime.now(timezone.utc).date().isoformat()
+        db = get_db()
+        row = db.execute(
+            "SELECT daily_usage, last_usage_date FROM users WHERE id = ?",
+            (user["id"],),
+        ).fetchone()
+        db.close()
+
+        if row and row["last_usage_date"] == today and row["daily_usage"] >= 1:
+            return jsonify(
+                {
+                    "error": "Limite quotidienne atteinte (plan free: 1 course/jour)",
+                    "upgrade_url": "/api/v1/subscription/upgrade",
+                }
+            ), 429
+
+        # Increment usage
+        db = get_db()
+        if row and row["last_usage_date"] == today:
+            db.execute(
+                "UPDATE users SET daily_usage = daily_usage + 1 WHERE id = ?",
+                (user["id"],),
+            )
+        else:
+            db.execute(
+                "UPDATE users SET daily_usage = 1, last_usage_date = ? WHERE id = ?",
+                (today, user["id"]),
+            )
+        db.commit()
+        db.close()
+
+        return fn(*args, **kwargs)
+
+    return wrapper

auth_db.py

@@ -0,0 +1,68 @@
+#!/usr/bin/env python3
+"""
+Auth DB — users and subscriptions schema for turf_saas.db
+Sprint 2-3: Auth JWT + Multi-tenant (HRT-28)
+"""
+
+import sqlite3
+import os
+
+DB_PATH = os.environ.get("TURF_SAAS_DB", "/home/h3r7/turf_saas/turf_saas.db")
+
+
+def get_db():
+    conn = sqlite3.connect(DB_PATH)
+    conn.row_factory = sqlite3.Row
+    return conn
+
+
+def init_auth_tables():
+    """Create users and subscriptions tables if they don't exist."""
+    conn = get_db()
+    c = conn.cursor()
+
+    c.executescript("""
+        CREATE TABLE IF NOT EXISTS users (
+            id              INTEGER PRIMARY KEY AUTOINCREMENT,
+            email           TEXT    NOT NULL UNIQUE,
+            password_hash   TEXT    NOT NULL,
+            plan            TEXT    NOT NULL DEFAULT 'free'
+                            CHECK(plan IN ('free','premium','pro')),
+            created_at      DATETIME NOT NULL DEFAULT (datetime('now')),
+            is_active       INTEGER  NOT NULL DEFAULT 1,
+            daily_usage     INTEGER  NOT NULL DEFAULT 0,
+            last_usage_date TEXT     DEFAULT NULL
+        );
+
+        CREATE TABLE IF NOT EXISTS subscriptions (
+            id                  INTEGER PRIMARY KEY AUTOINCREMENT,
+            user_id             INTEGER NOT NULL REFERENCES users(id),
+            plan                TEXT    NOT NULL CHECK(plan IN ('free','premium','pro')),
+            start_date          DATETIME NOT NULL DEFAULT (datetime('now')),
+            end_date            DATETIME,
+            stripe_customer_id  TEXT,
+            FOREIGN KEY (user_id) REFERENCES users(id)
+        );
+
+        CREATE TABLE IF NOT EXISTS refresh_tokens (
+            id          INTEGER PRIMARY KEY AUTOINCREMENT,
+            user_id     INTEGER NOT NULL REFERENCES users(id),
+            token_hash  TEXT    NOT NULL UNIQUE,
+            created_at  DATETIME NOT NULL DEFAULT (datetime('now')),
+            expires_at  DATETIME NOT NULL,
+            revoked     INTEGER  NOT NULL DEFAULT 0
+        );
+
+        CREATE INDEX IF NOT EXISTS idx_users_email          ON users(email);
+        CREATE INDEX IF NOT EXISTS idx_subscriptions_user   ON subscriptions(user_id);
+        CREATE INDEX IF NOT EXISTS idx_refresh_tokens_user  ON refresh_tokens(user_id);
+        CREATE INDEX IF NOT EXISTS idx_refresh_tokens_hash  ON refresh_tokens(token_hash);
+    """)
+
+    conn.commit()
+    conn.close()
+    print("[auth_db] Tables users, subscriptions, refresh_tokens created/verified.")
+
+
+if __name__ == "__main__":
+    init_auth_tables()

billing_db.py

@@ -0,0 +1,127 @@
+#!/usr/bin/env python3
+"""
+DB Migration — Billing Stripe
+Sprint 5-6: HRT-31
+
+Adds stripe_subscription_id and status columns to subscriptions table,
+and an invoices / grace-period tracking table.
+
+Run once:
+    ./venv/bin/python billing_db.py
+"""
+
+import sqlite3
+import os
+import logging
+
+DB_PATH = os.environ.get("TURF_SAAS_DB", "/home/h3r7/turf_saas/turf_saas.db")
+logger = logging.getLogger("turf_saas.billing_db")
+
+
+def get_db():
+    conn = sqlite3.connect(DB_PATH)
+    conn.row_factory = sqlite3.Row
+    return conn
+
+
+def migrate_billing_tables():
+    """Idempotent migration: add billing columns and billing_events table.
+
+    Requires auth tables (users, subscriptions) to exist first.
+    Calls init_auth_tables() automatically if subscriptions is absent.
+    """
+    from auth_db import init_auth_tables as _init_auth
+
+    conn = get_db()
+    c = conn.cursor()
+
+    # Ensure base auth tables exist
+    tables = {
+        row[0] for row in c.execute("SELECT name FROM sqlite_master WHERE type='table'")
+    }
+    conn.close()
+
+    if "subscriptions" not in tables:
+        _init_auth()
+
+    conn = get_db()
+    c = conn.cursor()
+
+    # Add stripe_subscription_id if missing
+    columns = {row[1] for row in c.execute("PRAGMA table_info(subscriptions)")}
+
+    if "stripe_subscription_id" not in columns:
+        c.execute("ALTER TABLE subscriptions ADD COLUMN stripe_subscription_id TEXT")
+        logger.info("[billing_db] Added stripe_subscription_id column to subscriptions")
+
+    if "status" not in columns:
+        c.execute(
+            "ALTER TABLE subscriptions ADD COLUMN "
+            "status TEXT NOT NULL DEFAULT 'active' "
+            "CHECK(status IN ('active','past_due','canceled','trialing','incomplete'))"
+        )
+        logger.info("[billing_db] Added status column to subscriptions")
+
+    if "grace_period_end" not in columns:
+        c.execute("ALTER TABLE subscriptions ADD COLUMN grace_period_end DATETIME")
+        logger.info("[billing_db] Added grace_period_end column to subscriptions")
+
+    if "current_period_end" not in columns:
+        c.execute("ALTER TABLE subscriptions ADD COLUMN current_period_end DATETIME")
+        logger.info("[billing_db] Added current_period_end column to subscriptions")
+
+    # billing_events table — audit trail for all webhook events
+    c.executescript("""
+        CREATE TABLE IF NOT EXISTS billing_events (
+            id              INTEGER PRIMARY KEY AUTOINCREMENT,
+            stripe_event_id TEXT    NOT NULL UNIQUE,
+            event_type      TEXT    NOT NULL,
+            user_id         INTEGER REFERENCES users(id),
+            payload         TEXT,
+            processed_at    DATETIME NOT NULL DEFAULT (datetime('now'))
+        );
+
+        CREATE INDEX IF NOT EXISTS idx_billing_events_user  ON billing_events(user_id);
+        CREATE INDEX IF NOT EXISTS idx_billing_events_type  ON billing_events(event_type);
+        CREATE INDEX IF NOT EXISTS idx_subscriptions_stripe ON subscriptions(stripe_subscription_id);
+        CREATE INDEX IF NOT EXISTS idx_subscriptions_customer ON subscriptions(stripe_customer_id);
+    """)
+
+    conn.commit()
+    conn.close()
+    print(
+        "[billing_db] Migration complete: subscriptions + billing_events tables ready."
+    )
+
+
+if __name__ == "__main__":
+    logging.basicConfig(level=logging.INFO)
+    migrate_billing_tables()
+
+
+# ──────────────────────────────────────────────────────────────
+# Re-exported helpers for test usage
+# (primary implementations live in api_v1/routes/billing.py)
+# ──────────────────────────────────────────────────────────────
+
+
+def _upsert_subscription(db, user_id: int, **fields):
+    """
+    Update existing subscription row or insert a new one.
+    Convenience re-export for test helpers.
+    """
+    existing = db.execute(
+        "SELECT id FROM subscriptions WHERE user_id = ? ORDER BY start_date DESC LIMIT 1",
+        (user_id,),
+    ).fetchone()
+    if existing:
+        set_parts = ", ".join(f"{k} = ?" for k in fields)
+        values = list(fields.values()) + [existing["id"]]
+        db.execute(f"UPDATE subscriptions SET {set_parts} WHERE id = ?", values)
+    else:
+        cols = ", ".join(["user_id"] + list(fields.keys()))
+        placeholders = ", ".join(["?"] * (1 + len(fields)))
+        values = [user_id] + list(fields.values())
+        db.execute(
+            f"INSERT INTO subscriptions ({cols}) VALUES ({placeholders})", values
+        )

combined_api.py

@@ -3614,5 +3614,219 @@ def api_predictions_analysis():
     return jsonify({"stats": stats, "period": {"start": start_date, "end": end_date}})
 
 
+# ─────────────────────────────────────────────────────────────────────────────
+# /api/v1/predictions  — Ensemble model endpoint (Sprint 6-7 ML Upgrade)
+# ─────────────────────────────────────────────────────────────────────────────
+_predict_v2 = None
+
+
+def _load_predict_v2():
+    """Lazy import of predict_v2 module (ensemble model)."""
+    global _predict_v2
+    if _predict_v2 is None:
+        try:
+            import importlib.util, sys
+
+            spec = importlib.util.spec_from_file_location(
+                "predict_v2", "/home/h3r7/turf_saas/predict_v2.py"
+            )
+            mod = importlib.util.module_from_spec(spec)
+            spec.loader.exec_module(mod)
+            _predict_v2 = mod
+        except Exception as e:
+            import logging
+
+            logging.error(f"[v1/predictions] predict_v2 import failed: {e}")
+    return _predict_v2
+
+
+@app.route("/api/v1/predictions", methods=["GET"])
+@app.route("/turf/api/v1/predictions", methods=["GET"])
+def api_v1_predictions():
+    """
+    Ensemble ML predictions using XGBoost + LightGBM + MLP (Optuna-tuned).
+    Query params:
+      - date: YYYY-MM-DD (default: today / latest available)
+      - reunion: int (default: all)
+      - course: int (default: all)
+    """
+    import time as _time
+
+    t0 = _time.perf_counter()
+
+    mod = _load_predict_v2()
+    if mod is None:
+        # Graceful fallback: redirect to legacy ml_predictions
+        return jsonify(
+            {
+                "error": "Ensemble model not available yet",
+                "fallback": "/api/ml_predictions",
+                "message": "Model is still training. Use /api/ml_predictions for legacy XGBoost predictions.",
+            }
+        ), 503
+
+    ensemble = mod.load_ensemble()
+    if ensemble is None:
+        return jsonify(
+            {
+                "error": "Ensemble model file not found",
+                "model_path": str(mod.ENSEMBLE_PATH),
+                "message": "Run train_ensemble.py to generate the model.",
+                "fallback": "/api/ml_predictions",
+            }
+        ), 503
+
+    date_param = request.args.get("date", None)
+    reunion_param = request.args.get("reunion", None)
+    course_param = request.args.get("course", None)
+
+    conn = sqlite3.connect("/home/h3r7/turf_saas/turf.db")
+    conn.row_factory = sqlite3.Row
+
+    # Determine date to use
+    if date_param:
+        date_used = date_param
+    else:
+        row = conn.execute(
+            "SELECT MAX(date_programme) as d FROM pmu_partants"
+        ).fetchone()
+        date_used = (
+            row["d"] if row and row["d"] else datetime.now().strftime("%Y-%m-%d")
+        )
+
+    # Build query
+    where_clauses = ["p.date_programme = ?"]
+    params = [date_used]
+    if reunion_param:
+        where_clauses.append("p.num_reunion = ?")
+        params.append(int(reunion_param))
+    if course_param:
+        where_clauses.append("p.num_course = ?")
+        params.append(int(course_param))
+
+    query = f"""
+        SELECT p.*, c.distance, c.discipline, c.specialite,
+               c.nb_declares_partants, c.montant_prix, c.penetrometre_intitule,
+               c.libelle as course_libelle, c.libelle_court as hippodrome,
+               c.heure_depart_str, c.parcours
+        FROM pmu_partants p
+        LEFT JOIN pmu_courses c ON p.date_programme = c.date_programme
+            AND p.num_reunion = c.num_reunion AND p.num_course = c.num_course
+        WHERE {" AND ".join(where_clauses)}
+        ORDER BY p.num_reunion, p.num_course, p.num_pmu
+    """
+    rows = conn.execute(query, params).fetchall()
+    conn.close()
+
+    if not rows:
+        return jsonify(
+            {
+                "date": date_used,
+                "model_version": mod.get_model_version(),
+                "predictions": [],
+                "message": f"No partants found for date {date_used}",
+            }
+        )
+
+    # Convert to list of dicts
+    partants = [dict(r) for r in rows]
+
+    # Run ensemble prediction
+    preds = mod.predict_top3(partants, model=ensemble)
+
+    # Group by race
+    races = {}
+    for pred in preds:
+        key = f"R{pred.get('num_reunion', 0)}C{pred.get('num_course', 0)}"
+        if key not in races:
+            # Find race metadata from partants
+            for p in partants:
+                if p.get("num_reunion") == pred.get("num_reunion") and p.get(
+                    "num_course"
+                ) == pred.get("num_course"):
+                    races[key] = {
+                        "reunion": pred.get("num_reunion"),
+                        "course": pred.get("num_course"),
+                        "label": key,
+                        "race_name": p.get("course_libelle", ""),
+                        "hippodrome": p.get("hippodrome", ""),
+                        "heure": p.get("heure_depart_str", ""),
+                        "discipline": p.get("discipline", ""),
+                        "distance": p.get("distance", 0),
+                        "horses": [],
+                    }
+                    break
+        if key in races:
+            races[key]["horses"].append(pred)
+
+    latency_ms = (_time.perf_counter() - t0) * 1000
+
+    return jsonify(
+        {
+            "date": date_used,
+            "model_version": mod.get_model_version(),
+            "latency_ms": round(latency_ms, 1),
+            "total_horses": len(preds),
+            "races": list(races.values()),
+        }
+    )
+
+
+@app.route("/api/v1/model/invalidate-cache", methods=["POST"])
+@app.route("/turf/api/v1/model/invalidate-cache", methods=["POST"])
+def api_v1_invalidate_cache():
+    """Force reload of ensemble model on next prediction call."""
+    mod = _load_predict_v2()
+    if mod:
+        mod.invalidate_model_cache()
+        return jsonify({"status": "ok", "message": "Model cache invalidated"})
+    return jsonify({"status": "error", "message": "predict_v2 module not loaded"}), 500
+
+
+@app.route("/api/v1/model/status", methods=["GET"])
+@app.route("/turf/api/v1/model/status", methods=["GET"])
+def api_v1_model_status():
+    """Return ensemble model status and version."""
+    import os as _os
+    from pathlib import Path as _Path
+
+    ensemble_path = _Path("/home/h3r7/turf_saas/models/ensemble_top3.pkl")
+    benchmark_path = _Path("/home/h3r7/turf_saas/models/benchmark_report.json")
+
+    status = {
+        "ensemble_available": ensemble_path.exists(),
+        "ensemble_path": str(ensemble_path),
+    }
+    if ensemble_path.exists():
+        mtime = _os.path.getmtime(str(ensemble_path))
+        status["last_trained"] = datetime.fromtimestamp(mtime).isoformat()
+
+    if benchmark_path.exists():
+        try:
+            with open(benchmark_path) as f:
+                import json as _json
+
+                report = _json.load(f)
+            status["benchmark"] = {
+                "baseline_precision_at3": report.get("baseline", {}).get(
+                    "precision_at3"
+                ),
+                "ensemble_precision_at3": report.get("ensemble", {}).get(
+                    "precision_at3"
+                ),
+                "delta": report.get("delta_precision_at3"),
+                "deployed": report.get("deploy"),
+                "run_date": report.get("run_date"),
+            }
+        except Exception:
+            pass
+
+    mod = _load_predict_v2()
+    if mod and ensemble_path.exists():
+        status["model_version"] = mod.get_model_version()
+
+    return jsonify(status)
+
+
 if __name__ == "__main__":
     app.run(host="0.0.0.0", port=8790, debug=False)

middleware.py

@@ -0,0 +1,90 @@
+#!/usr/bin/env python3
+"""
+Middleware — rate limiting, CORS, and access logging
+Sprint 2-3: HRT-28
+"""
+
+import logging
+import time
+from collections import defaultdict
+from datetime import datetime, timezone
+from functools import wraps
+from threading import Lock
+
+from flask import request, jsonify, g
+
+logger = logging.getLogger("turf_saas.middleware")
+
+# ──────────────────────────────────────────────────────────────
+# In-memory rate limiter (100 req/min per IP)
+# For production: replace with Redis-backed counter
+# ──────────────────────────────────────────────────────────────
+
+_rate_store: dict = defaultdict(lambda: {"count": 0, "window_start": 0.0})
+_rate_lock = Lock()
+
+RATE_LIMIT = 100  # max requests
+RATE_WINDOW = 60  # seconds
+
+
+def rate_limit_middleware(app):
+    """Register before_request rate limiting on the Flask app."""
+
+    @app.before_request
+    def check_rate_limit():
+        ip = request.remote_addr or "unknown"
+        now = time.time()
+
+        with _rate_lock:
+            bucket = _rate_store[ip]
+            if now - bucket["window_start"] >= RATE_WINDOW:
+                bucket["count"] = 0
+                bucket["window_start"] = now
+            bucket["count"] += 1
+            count = bucket["count"]
+            remaining = max(0, RATE_LIMIT - count)
+
+        if count > RATE_LIMIT:
+            logger.warning("Rate limit exceeded for IP %s", ip)
+            resp = jsonify({"error": "Trop de requêtes. Limite: 100/min par IP."})
+            resp.status_code = 429
+            resp.headers["X-RateLimit-Limit"] = str(RATE_LIMIT)
+            resp.headers["X-RateLimit-Remaining"] = "0"
+            resp.headers["Retry-After"] = str(RATE_WINDOW)
+            return resp
+
+        # Attach headers on all responses via after_request
+        g.rl_remaining = remaining
+
+
+# ──────────────────────────────────────────────────────────────
+# Access logs (timestamped)
+# ──────────────────────────────────────────────────────────────
+
+access_log = logging.getLogger("turf_saas.access")
+
+
+def access_log_middleware(app):
+    """Register after_request access logging on the Flask app."""
+
+    @app.after_request
+    def log_access(response):
+        ip = request.remote_addr or "unknown"
+        ts = datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
+        user_id = getattr(g, "current_user_id", "-")
+        access_log.info(
+            '%s %s %s "%s %s" %s %s',
+            ts,
+            ip,
+            user_id,
+            request.method,
+            request.path,
+            response.status_code,
+            response.content_length or 0,
+        )
+        # Attach rate-limit headers
+        remaining = getattr(g, "rl_remaining", None)
+        if remaining is not None:
+            response.headers["X-RateLimit-Limit"] = str(RATE_LIMIT)
+            response.headers["X-RateLimit-Remaining"] = str(remaining)
+        return response

models/benchmark_report.json

@@ -0,0 +1,174 @@
+{
+  "run_date": "2026-04-25T19:09:46.629142",
+  "dataset": {
+    "db_path": "/home/h3r7/turf_saas/turf.db",
+    "total_rows": 10899,
+    "train_rows": 8719,
+    "holdout_rows": 2180,
+    "train_date_range": [
+      "2026-03-31",
+      "2026-04-19"
+    ],
+    "holdout_date_range": [
+      "2026-04-19",
+      "2026-04-24"
+    ]
+  },
+  "baseline": {
+    "model": "XGBoost (baseline)",
+    "precision_at3": 0.5286821705426358,
+    "auc": 0.7254057665061495
+  },
+  "individual_models": {
+    "xgboost": {
+      "model": "xgboost",
+      "auc": 0.7856,
+      "accuracy": 0.6917,
+      "precision": 0.4865,
+      "recall": 0.7229,
+      "precision_at3": 0.5783,
+      "latency_ms_per_row": 0.0112
+    },
+    "lightgbm": {
+      "model": "lightgbm",
+      "auc": 0.7833,
+      "accuracy": 0.6995,
+      "precision": 0.4951,
+      "recall": 0.709,
+      "precision_at3": 0.5736,
+      "latency_ms_per_row": 0.0041
+    },
+    "mlp": {
+      "model": "mlp",
+      "auc": 0.7743,
+      "accuracy": 0.7445,
+      "precision": 0.5743,
+      "recall": 0.5325,
+      "precision_at3": 0.5643,
+      "latency_ms_per_row": 0.0052
+    }
+  },
+  "ensemble": {
+    "model": "ensemble",
+    "auc": 0.784,
+    "accuracy": 0.7147,
+    "precision": 0.5142,
+    "recall": 0.6718,
+    "precision_at3": 0.5814,
+    "latency_ms_per_row": 0.0208
+  },
+  "delta_precision_at3": 0.0527,
+  "deploy": true,
+  "optuna": {
+    "n_trials": 100,
+    "xgboost_best_params": {
+      "n_estimators": 141,
+      "max_depth": 5,
+      "learning_rate": 0.016298172447266404,
+      "subsample": 0.7660470794373848,
+      "colsample_bytree": 0.471124415020467,
+      "min_child_weight": 14,
+      "reg_alpha": 1.9364166463791586,
+      "reg_lambda": 6.018030083488602,
+      "gamma": 4.614943551368141
+    },
+    "lightgbm_best_params": {
+      "n_estimators": 186,
+      "max_depth": 4,
+      "learning_rate": 0.012915117465216954,
+      "num_leaves": 141,
+      "subsample": 0.6193119116922561,
+      "colsample_bytree": 0.539310022549326,
+      "min_child_samples": 9,
+      "reg_alpha": 0.6864583098112754,
+      "reg_lambda": 0.0549259590914184
+    }
+  },
+  "features": {
+    "total": 43,
+    "selected_by_shap": 31,
+    "feature_list": [
+      "age",
+      "sexe_enc",
+      "nombre_courses",
+      "nombre_victoires",
+      "nombre_places",
+      "tx_victoire",
+      "tx_place",
+      "forme_recente",
+      "tendance_num",
+      "gains_annee_en_cours",
+      "cote_direct",
+      "cote_reference",
+      "distance",
+      "nb_partants",
+      "discipline_enc",
+      "specialite_enc",
+      "oeilleres_enc",
+      "tendance_cote_enc",
+      "penetrometre_intitule_enc",
+      "form_1",
+      "form_2",
+      "form_3",
+      "form_4",
+      "form_5",
+      "form_weighted",
+      "form_avg",
+      "form_best",
+      "form_worst",
+      "win_ratio",
+      "place_ratio",
+      "implied_prob",
+      "win_rate_adj",
+      "place_rate_adj",
+      "earnings_per_race",
+      "cote_diff",
+      "cote_ratio",
+      "rang_cote",
+      "ratio_cote_field",
+      "distance_cat",
+      "age_win_interact",
+      "is_favorite",
+      "poids",
+      "prize_norm"
+    ],
+    "shap_selected": [
+      "rang_cote",
+      "implied_prob",
+      "cote_direct",
+      "ratio_cote_field",
+      "nb_partants",
+      "cote_diff",
+      "cote_ratio",
+      "specialite_enc",
+      "earnings_per_race",
+      "nombre_courses",
+      "cote_reference",
+      "distance",
+      "discipline_enc",
+      "is_favorite",
+      "prize_norm",
+      "win_ratio",
+      "place_rate_adj",
+      "gains_annee_en_cours",
+      "poids",
+      "tx_place",
+      "penetrometre_intitule_enc",
+      "age_win_interact",
+      "nombre_places",
+      "tendance_num",
+      "age",
+      "form_avg",
+      "form_weighted",
+      "place_ratio",
+      "form_3",
+      "oeilleres_enc",
+      "form_5"
+    ]
+  },
+  "ensemble_weights": {
+    "xgboost": 0.23161801824035544,
+    "lightgbm": 0.23415467282905,
+    "mlp": 0.21290370528252356
+  }
+}

models/benchmark_report.md

@@ -0,0 +1,68 @@
+# Benchmark ML Ensemble — Turf Prédictions
+
+**Date:** 2026-04-25  
+**Dataset:** 10,899 partants  
+**Holdout:** 2,180 lignes (2026-04-19 → 2026-04-24)
+
+## Résultats
+
+| Modèle | Precision@3 | AUC | Latence/prédiction |
+|--------|-------------|-----|-------------------|
+| XGBoost (baseline) | 0.5287 | 0.7254 | — |
+| xgboost | 0.5783 | 0.7856 | 0.01 ms |
+| lightgbm | 0.5736 | 0.7833 | 0.00 ms |
+| mlp | 0.5643 | 0.7743 | 0.01 ms |
+| **Ensemble** | **0.5814** | **0.7840** | **0.02 ms** |
+
+## Décision de déploiement
+
+- Delta Precision@3 : **+0.0527** (+5.3%)
+- Seuil requis : **+5%**
+- Résultat : **✅ DEPLOIEMENT RECOMMANDE**
+
+## Optimisation Optuna
+
+- Trials XGBoost : 100
+- Trials LightGBM : 100
+- Pruning : MedianPruner
+
+### Meilleurs hyperparamètres XGBoost
+```json
+{
+  "n_estimators": 141,
+  "max_depth": 5,
+  "learning_rate": 0.016298172447266404,
+  "subsample": 0.7660470794373848,
+  "colsample_bytree": 0.471124415020467,
+  "min_child_weight": 14,
+  "reg_alpha": 1.9364166463791586,
+  "reg_lambda": 6.018030083488602,
+  "gamma": 4.614943551368141
+}
+```
+
+### Meilleurs hyperparamètres LightGBM
+```json
+{
+  "n_estimators": 186,
+  "max_depth": 4,
+  "learning_rate": 0.012915117465216954,
+  "num_leaves": 141,
+  "subsample": 0.6193119116922561,
+  "colsample_bytree": 0.539310022549326,
+  "min_child_samples": 9,
+  "reg_alpha": 0.6864583098112754,
+  "reg_lambda": 0.0549259590914184
+}
+```
+
+## Features
+
+- Total features : 43
+- Retenues par SHAP : 31
+
+## Poids de l'ensemble
+
+- xgboost : 0.2316
+- lightgbm : 0.2342
+- mlp : 0.2129

portal_server.py

@@ -5,8 +5,11 @@ import json
 import requests
 import subprocess
 import db
+from middleware import rate_limit_middleware, access_log_middleware
 
 app = Flask(__name__)
+rate_limit_middleware(app)
+access_log_middleware(app)
 
 DASHBOARD_API_URL = "http://localhost:8791"
 COMBINED_API_URL = "http://localhost:8790"

predict_v2.py

@@ -0,0 +1,387 @@
+#!/usr/bin/env python3
+"""
+Ensemble prediction module for /api/v1/predictions.
+
+Loads the trained ensemble model and provides a high-level predict_top3()
+function compatible with the existing combined_api.py interface.
+
+Cache: model is loaded once at import time (or on first call).
+       Invalidation: reload if models/ensemble_top3.pkl mtime changes.
+"""
+
+import logging
+import os
+import pickle
+import re
+import threading
+import time
+from datetime import datetime
+from pathlib import Path
+from typing import Optional
+
+import numpy as np
+import pandas as pd
+from sklearn.preprocessing import LabelEncoder
+
+logger = logging.getLogger(__name__)
+
+MODELS_DIR = Path("/home/h3r7/turf_saas/models")
+ENSEMBLE_PATH = MODELS_DIR / "ensemble_top3.pkl"
+
+# ── Cache ─────────────────────────────────────────────────────────────────────
+_model_cache = {
+    "ensemble": None,
+    "mtime": None,
+    "lock": threading.Lock(),
+}
+
+# ── Feature list (must match train_ensemble.py FEATURE_COLS) ─────────────────
+FEATURE_COLS = [
+    "age",
+    "sexe_enc",
+    "nombre_courses",
+    "nombre_victoires",
+    "nombre_places",
+    "tx_victoire",
+    "tx_place",
+    "forme_recente",
+    "tendance_num",
+    "gains_annee_en_cours",
+    "cote_direct",
+    "cote_reference",
+    "distance",
+    "nb_partants",
+    "discipline_enc",
+    "specialite_enc",
+    "oeilleres_enc",
+    "tendance_cote_enc",
+    "penetrometre_intitule_enc",
+    "form_1",
+    "form_2",
+    "form_3",
+    "form_4",
+    "form_5",
+    "form_weighted",
+    "form_avg",
+    "form_best",
+    "form_worst",
+    "win_ratio",
+    "place_ratio",
+    "implied_prob",
+    "win_rate_adj",
+    "place_rate_adj",
+    "earnings_per_race",
+    "cote_diff",
+    "cote_ratio",
+    "rang_cote",
+    "ratio_cote_field",
+    "distance_cat",
+    "age_win_interact",
+    "is_favorite",
+    "poids",
+    "prize_norm",
+]
+
+
+# ── Encoders (built per-prediction batch for live data) ──────────────────────
+def _fit_encoder(values, default):
+    le = LabelEncoder()
+    unique = list(set(str(v) if v else default for v in values)) + [default]
+    le.fit(unique)
+    return le
+
+
+def _safe_transform(le: LabelEncoder, value, default: str):
+    v = str(value) if value else default
+    if v not in le.classes_:
+        v = default
+    return int(le.transform([v])[0])
+
+
+# ── Model loading with auto-invalidation ─────────────────────────────────────
+def load_ensemble(force: bool = False) -> Optional[object]:
+    """Load ensemble model, reload if file changed."""
+    with _model_cache["lock"]:
+        if not ENSEMBLE_PATH.exists():
+            return None
+        mtime = ENSEMBLE_PATH.stat().st_mtime
+        if force or _model_cache["ensemble"] is None or mtime != _model_cache["mtime"]:
+            try:
+                with open(ENSEMBLE_PATH, "rb") as f:
+                    _model_cache["ensemble"] = pickle.load(f)
+                _model_cache["mtime"] = mtime
+                logger.info(f"[predict_v2] Loaded ensemble model from {ENSEMBLE_PATH}")
+            except Exception as e:
+                logger.error(f"[predict_v2] Failed to load ensemble: {e}")
+                return None
+        return _model_cache["ensemble"]
+
+
+def invalidate_model_cache():
+    """Force reload on next prediction call."""
+    with _model_cache["lock"]:
+        _model_cache["mtime"] = None
+
+
+# ── Feature engineering for live pmu_partants rows ───────────────────────────
+def _parse_musique(musique) -> list:
+    if not musique or pd.isna(str(musique)):
+        return [0, 0, 0, 0, 0]
+    try:
+        clean = re.sub(r"\(\d+\)", "", str(musique))
+        numbers = re.findall(r"\d+", clean)
+        result = [int(n) for n in numbers[:5]]
+        result += [0] * (5 - len(result))
+        return result[:5]
+    except Exception:
+        return [0, 0, 0, 0, 0]
+
+
+def build_feature_df(partants: list) -> pd.DataFrame:
+    """
+    Convert a list of pmu_partants dicts to a feature DataFrame.
+
+    Expected keys (same as pmu_partants columns):
+        date_programme, num_reunion, num_course, num_pmu,
+        age, sexe, musique, nombre_courses, nombre_victoires, nombre_places,
+        gains_annee_en_cours, handicap_poids, oeilleres, cote_direct,
+        cote_reference, tendance_cote, favoris, tx_victoire, tx_place,
+        forme_recente, tendance_forme, indicateur_inedit,
+        distance, discipline, specialite, nb_declares_partants,
+        montant_prix, penetrometre_intitule
+    """
+    if not partants:
+        return pd.DataFrame()
+
+    df = pd.DataFrame(partants)
+
+    # ── Categorical encoders fitted on this batch ─────────────────────────────
+    le_sexe = _fit_encoder(df.get("sexe", ["U"]), "U")
+    le_oeilleres = _fit_encoder(df.get("oeilleres", ["SANS"]), "SANS")
+    le_discipline = _fit_encoder(df.get("discipline", ["UNKNOWN"]), "UNKNOWN")
+    le_specialite = _fit_encoder(df.get("specialite", ["UNKNOWN"]), "UNKNOWN")
+    le_tendance = _fit_encoder(df.get("tendance_cote", ["STABLE"]), "STABLE")
+    le_penet = _fit_encoder(df.get("penetrometre_intitule", ["BON"]), "BON")
+
+    df["sexe_enc"] = df["sexe"].apply(lambda v: _safe_transform(le_sexe, v, "U"))
+    df["oeilleres_enc"] = df["oeilleres"].apply(
+        lambda v: _safe_transform(le_oeilleres, v, "SANS")
+    )
+    df["discipline_enc"] = df.get("discipline", pd.Series(["UNKNOWN"] * len(df))).apply(
+        lambda v: _safe_transform(le_discipline, v, "UNKNOWN")
+    )
+    df["specialite_enc"] = df.get("specialite", pd.Series(["UNKNOWN"] * len(df))).apply(
+        lambda v: _safe_transform(le_specialite, v, "UNKNOWN")
+    )
+    df["tendance_cote_enc"] = df.get(
+        "tendance_cote", pd.Series(["STABLE"] * len(df))
+    ).apply(lambda v: _safe_transform(le_tendance, v, "STABLE"))
+    df["penetrometre_intitule_enc"] = df.get(
+        "penetrometre_intitule", pd.Series(["BON"] * len(df))
+    ).apply(lambda v: _safe_transform(le_penet, v, "BON"))
+
+    # ── Musique ────────────────────────────────────────────────────────────────
+    music_parsed = df["musique"].apply(_parse_musique)
+    for i in range(5):
+        df[f"form_{i + 1}"] = music_parsed.apply(lambda x: x[i])
+    weights = np.array([0.4, 0.25, 0.15, 0.12, 0.08])
+    df["form_weighted"] = music_parsed.apply(
+        lambda x: sum(w * v for w, v in zip(weights, x))
+    )
+    df["form_avg"] = music_parsed.apply(np.mean)
+    df["form_best"] = music_parsed.apply(min)
+    df["form_worst"] = music_parsed.apply(max)
+
+    # ── Numeric features ───────────────────────────────────────────────────────
+    for col in [
+        "nombre_courses",
+        "nombre_victoires",
+        "nombre_places",
+        "tx_victoire",
+        "tx_place",
+        "forme_recente",
+        "tendance_forme",
+        "gains_annee_en_cours",
+        "cote_direct",
+        "cote_reference",
+        "distance",
+        "handicap_poids",
+        "age",
+        "montant_prix",
+        "nb_declares_partants",
+    ]:
+        if col not in df.columns:
+            df[col] = 0.0
+        df[col] = pd.to_numeric(df[col], errors="coerce").fillna(0)
+
+    df["tendance_num"] = df["tendance_forme"].fillna(0)
+    df["win_ratio"] = df["nombre_victoires"] / df["nombre_courses"].replace(0, 1)
+    df["place_ratio"] = df["nombre_places"] / df["nombre_courses"].replace(0, 1)
+    df["implied_prob"] = 1.0 / df["cote_direct"].replace(0, np.nan)
+    df["win_rate_adj"] = df["tx_victoire"] * np.log1p(df["nombre_courses"])
+    df["place_rate_adj"] = df["tx_place"] * np.log1p(df["nombre_courses"])
+    df["earnings_per_race"] = df["gains_annee_en_cours"] / df["nombre_courses"].replace(
+        0, 1
+    )
+    df["cote_diff"] = (df["cote_direct"] - df["cote_reference"]).fillna(0)
+    df["cote_ratio"] = (
+        df["cote_direct"] / df["cote_reference"].replace(0, np.nan)
+    ).fillna(1)
+
+    # ── Per-race rank features ─────────────────────────────────────────────────
+    if "num_reunion" in df.columns and "num_course" in df.columns:
+        grp = ["date_programme", "num_reunion", "num_course"]
+        # Some fields may be missing
+        for g in grp:
+            if g not in df.columns:
+                df[g] = 0
+        df["rang_cote"] = df.groupby(grp)["cote_direct"].rank(
+            method="min", na_option="bottom"
+        )
+        race_mean = df.groupby(grp)["cote_direct"].transform("mean")
+        df["ratio_cote_field"] = df["cote_direct"] / race_mean.replace(0, np.nan)
+        df["nb_partants"] = df.groupby(grp)["cote_direct"].transform("count")
+    else:
+        df["rang_cote"] = 1.0
+        df["ratio_cote_field"] = 1.0
+        df["nb_partants"] = df.get("nb_declares_partants", pd.Series([10] * len(df)))
+
+    df["distance_cat"] = pd.cut(
+        df["distance"].fillna(1600),
+        bins=[0, 1400, 1800, 2200, 2600, 10000],
+        labels=[1, 2, 3, 4, 5],
+    ).astype(float)
+    df["age_win_interact"] = df["age"] * df["tx_victoire"]
+    df["is_favorite"] = (
+        df.get("favoris", pd.Series([0] * len(df))).fillna(0).astype(int)
+    )
+    df["poids"] = df["handicap_poids"].fillna(60)
+    df["prize_norm"] = np.log1p(df["montant_prix"].fillna(0))
+
+    return df
+
+
+# ── Main prediction function ───────────────────────────────────────────────────
+def predict_top3(partants: list, model=None) -> list:
+    """
+    Given a list of partant dicts (from pmu_partants), return predictions.
+
+    Returns list of {horse_name, num_pmu, prob_top3, prob_top1_approx, ...}
+    sorted by prob_top3 descending.
+
+    Falls back to empty list if model not available.
+    """
+    t_start = time.perf_counter()
+
+    if model is None:
+        model = load_ensemble()
+    if model is None:
+        logger.warning("[predict_v2] Ensemble model not available — no predictions")
+        return []
+
+    df = build_feature_df(partants)
+    if df.empty:
+        return []
+
+    available = [c for c in FEATURE_COLS if c in df.columns]
+    X = df[available].fillna(0)
+
+    try:
+        proba = model.predict_proba(X)[:, 1]
+    except Exception as e:
+        logger.error(f"[predict_v2] predict_proba failed: {e}")
+        return []
+
+    latency_ms = (time.perf_counter() - t_start) * 1000
+
+    results = []
+    for i, (p, row) in enumerate(zip(proba, partants)):
+        results.append(
+            {
+                "horse_name": row.get("nom", row.get("horse_name", f"H{i}")),
+                "num_pmu": row.get("num_pmu", i + 1),
+                "num_reunion": row.get("num_reunion"),
+                "num_course": row.get("num_course"),
+                "prob_top3": round(float(p) * 100, 1),
+                # approx top1 from top3 score (divide by ~2.5 empirically)
+                "prob_top1": round(float(p) / 2.5 * 100, 1),
+                "ml_score": round(float(p) * 100, 1),
+                "recommendation": "top3"
+                if p >= 0.40
+                else ("watch" if p >= 0.28 else "pass"),
+                "is_value_bet": int(
+                    p >= 0.35 and float(row.get("cote_direct", 0) or 0) > 10
+                ),
+                "model_version": getattr(model, "version", "ensemble_v1"),
+            }
+        )
+
+    results.sort(key=lambda x: x["prob_top3"], reverse=True)
+
+    # Mark top-3 predicted
+    for i, r in enumerate(results[:3]):
+        r["predicted_rank"] = i + 1
+
+    if results:
+        logger.info(
+            f"[predict_v2] {len(results)} horses predicted in {latency_ms:.1f} ms "
+            f"({latency_ms / len(results):.2f} ms/horse)"
+        )
+
+    return results
+
+
+# ── API-compatible wrapper keeping model_version & structure ──────────────────
+def get_model_version() -> str:
+    m = load_ensemble()
+    if m is None:
+        return "ensemble_v1_not_loaded"
+    return getattr(m, "version", "ensemble_v1")
+
+
+if __name__ == "__main__":
+    # Quick self-test
+    import sqlite3
+
+    conn = sqlite3.connect("/home/h3r7/turf_saas/turf.db")
+    rows = conn.execute(
+        """SELECT p.*, c.distance, c.discipline, c.specialite,
+                  c.nb_declares_partants, c.montant_prix, c.penetrometre_intitule
+           FROM pmu_partants p
+           LEFT JOIN pmu_courses c ON p.date_programme=c.date_programme
+                AND p.num_reunion=c.num_reunion AND p.num_course=c.num_course
+           WHERE p.date_programme=(SELECT MAX(date_programme) FROM pmu_partants)
+             AND p.num_reunion=1 AND p.num_course=1
+           LIMIT 20"""
+    ).fetchall()
+    conn.close()
+
+    if not rows:
+        print("No data found for self-test")
+    else:
+        cols = [d[0] for d in conn.description] if hasattr(conn, "description") else []
+        # Fallback column list
+        import sqlite3 as sq3
+
+        conn2 = sq3.connect("/home/h3r7/turf_saas/turf.db")
+        cur = conn2.execute(
+            """SELECT p.*, c.distance, c.discipline, c.specialite,
+                      c.nb_declares_partants, c.montant_prix, c.penetrometre_intitule
+               FROM pmu_partants p
+               LEFT JOIN pmu_courses c ON p.date_programme=c.date_programme
+                    AND p.num_reunion=c.num_reunion AND p.num_course=c.num_course
+               WHERE p.date_programme=(SELECT MAX(date_programme) FROM pmu_partants)
+                 AND p.num_reunion=1 AND p.num_course=1
+               LIMIT 20"""
+        )
+        cols = [d[0] for d in cur.description]
+        rows2 = cur.fetchall()
+        conn2.close()
+
+        partants = [dict(zip(cols, row)) for row in rows2]
+        preds = predict_top3(partants)
+        print(f"Self-test: {len(preds)} predictions")
+        for p in preds[:5]:
+            print(
+                f"  {p['horse_name']:20s}  prob_top3={p['prob_top3']}%  rec={p['recommendation']}"
+            )

pytest.ini

@@ -0,0 +1,13 @@
+[pytest]
+asyncio_mode = auto
+testpaths = tests
+python_files = test_*.py
+python_classes = Test*
+python_functions = test_*
+addopts = --tb=short -v
+markers =
+    e2e: Tests End-to-End Playwright
+    load: Tests de charge Locust
+    security: Tests de sécurité
+    smoke: Tests rapides de smoke
+    integration: Tests d'intégration DB et pipeline ML

rebuild_ensemble.py

@@ -0,0 +1,182 @@
+#!/usr/bin/env python3
+"""
+Rebuild ensemble using known best Optuna params (from completed study).
+Skips the 100-trial Optuna search and goes straight to training + pickling.
+"""
+import sys
+sys.path.insert(0, '/home/h3r7/turf_saas')
+
+from train_ensemble import (
+    load_data, engineer_features, temporal_split, get_features_and_target,
+    evaluate_baseline, train_xgboost, train_lightgbm, train_mlp,
+    shap_feature_selection, compute_ensemble_weights,
+    evaluate_model, compute_precision_at3, TurfEnsemble,
+    MODELS_DIR, DEPLOY_THRESHOLD, _write_markdown_report
+)
+import json, pickle, numpy as np
+from datetime import datetime
+from pathlib import Path
+
+DB_PATH = '/home/h3r7/turf_saas/turf.db'
+
+# Best params from the 100-trial Optuna run
+XGB_BEST = {
+    'n_estimators': 141, 'max_depth': 5,
+    'learning_rate': 0.016298172447266404,
+    'subsample': 0.7660470794373848,
+    'colsample_bytree': 0.471124415020467,
+    'min_child_weight': 14,
+    'reg_alpha': 1.9364166463791586,
+    'reg_lambda': 6.018030083488602,
+    'gamma': 4.614943551368141,
+}
+LGB_BEST = {
+    'n_estimators': 186, 'max_depth': 4,
+    'learning_rate': 0.012915117465216954,
+    'num_leaves': 141,
+    'subsample': 0.6193119116922561,
+    'colsample_bytree': 0.539310022549326,
+    'min_child_samples': 9,
+    'reg_alpha': 0.6864583098112754,
+    'reg_lambda': 0.0549259590914184,
+}
+
+print("=" * 65)
+print("TURF ENSEMBLE REBUILD (using pre-computed Optuna params)")
+print("=" * 65)
+
+print("\n[1/7] Loading data...")
+df = load_data(DB_PATH)
+df = engineer_features(df)
+
+print("\n[2/7] Temporal split...")
+train_df, holdout_df = temporal_split(df)
+X_train, y_train, feat_cols = get_features_and_target(train_df)
+X_holdout, y_holdout, _ = get_features_and_target(holdout_df)
+
+n = len(X_train); n_val = int(n * 0.15)
+X_tr = X_train.iloc[:n-n_val]; y_tr = y_train.iloc[:n-n_val]
+X_val = X_train.iloc[n-n_val:]; y_val = y_train.iloc[n-n_val:]
+
+print("\n[3/7] Evaluating baseline XGBoost...")
+baseline = evaluate_baseline(holdout_df, '/home/h3r7/turf_saas/xgboost_models.pkl')
+print(f"  Baseline P@3={baseline['precision_at3']:.4f}  AUC={baseline['auc']:.4f}")
+
+print("\n[4/7] Training models with best params...")
+print("  XGBoost...")
+xgb_model = train_xgboost(X_tr, y_tr, XGB_BEST)
+print("  LightGBM...")
+lgb_model = train_lightgbm(X_tr, y_tr, LGB_BEST)
+print("  MLP...")
+mlp_model = train_mlp(X_tr.values, y_tr)
+
+print("\n[5/7] SHAP analysis...")
+selected_features, shap_df = shap_feature_selection(xgb_model, X_tr)
+
+print("\n[6/7] Computing ensemble weights...")
+class WrappedMLP:
+    def __init__(self, pipeline, cols):
+        self.pipeline = pipeline
+        self.feature_cols = cols
+    def predict_proba(self, X):
+        import pandas as pd
+        available = [c for c in self.feature_cols if c in X.columns]
+        return self.pipeline.predict_proba(X[available].values)
+
+class WrappedTree:
+    def __init__(self, model, cols):
+        self.model = model
+        self.feature_cols = cols
+    def predict_proba(self, X):
+        available = [c for c in self.feature_cols if c in X.columns]
+        return self.model.predict_proba(X[available])
+
+wrapped_xgb = WrappedTree(xgb_model, feat_cols)
+wrapped_lgb = WrappedTree(lgb_model, feat_cols)
+wrapped_mlp = WrappedMLP(mlp_model, feat_cols)
+model_dict = {'xgboost': wrapped_xgb, 'lightgbm': wrapped_lgb, 'mlp': wrapped_mlp}
+
+weights = compute_ensemble_weights(model_dict, X_val, y_val, feat_cols)
+print("  Weights:", weights)
+
+print("\n[7/7] Evaluating + saving ensemble...")
+ensemble = TurfEnsemble(xgb_model, lgb_model, mlp_model, weights, feat_cols)
+
+results = {}
+for name, wrapped in model_dict.items():
+    res = evaluate_model(wrapped, X_holdout, y_holdout, holdout_df, name)
+    results[name] = res
+    print(f"  {name:12s} P@3={res['precision_at3']:.4f}  AUC={res['auc']:.4f}")
+
+ens_res = evaluate_model(ensemble, X_holdout, y_holdout, holdout_df, "ensemble")
+results["ensemble"] = ens_res
+print(f"  {'ensemble':12s} P@3={ens_res['precision_at3']:.4f}  AUC={ens_res['auc']:.4f}")
+
+delta = ens_res['precision_at3'] - baseline['precision_at3']
+deploy = delta >= DEPLOY_THRESHOLD
+print(f"\n  Delta: {delta:+.4f} ({delta*100:+.1f}%)  Deploy={'YES' if deploy else 'NO'}")
+
+# Save ensemble
+ensemble_path = MODELS_DIR / "ensemble_top3.pkl"
+with open(ensemble_path, "wb") as f:
+    pickle.dump(ensemble, f)
+print(f"\n  ✅ ensemble_top3.pkl saved ({ensemble_path.stat().st_size//1024} KB)")
+
+# Save individual models
+for name, model in [("xgboost_optimized", xgb_model), ("lightgbm", lgb_model)]:
+    path = MODELS_DIR / f"{name}_top3.pkl"
+    with open(path, "wb") as f:
+        pickle.dump({"model": model, "feature_cols": feat_cols}, f)
+    print(f"  ✅ {name}_top3.pkl saved")
+
+mlp_path = MODELS_DIR / "mlp_top3.pkl"
+with open(mlp_path, "wb") as f:
+    pickle.dump({"pipeline": mlp_model, "feature_cols": feat_cols}, f)
+print(f"  ✅ mlp_top3.pkl saved")
+
+# Benchmark report
+report = {
+    "run_date": datetime.now().isoformat(),
+    "dataset": {
+        "db_path": DB_PATH,
+        "total_rows": len(df),
+        "train_rows": len(X_train),
+        "holdout_rows": len(X_holdout),
+        "train_date_range": [str(train_df["date_programme"].min()), str(train_df["date_programme"].max())],
+        "holdout_date_range": [str(holdout_df["date_programme"].min()), str(holdout_df["date_programme"].max())],
+    },
+    "baseline": baseline,
+    "individual_models": {k: v for k, v in results.items() if k != "ensemble"},
+    "ensemble": ens_res,
+    "delta_precision_at3": round(delta, 4),
+    "deploy": deploy,
+    "optuna": {
+        "n_trials": 100,
+        "xgboost_best_params": XGB_BEST,
+        "lightgbm_best_params": LGB_BEST,
+    },
+    "features": {
+        "total": len(feat_cols),
+        "selected_by_shap": len(selected_features),
+        "feature_list": feat_cols,
+        "shap_selected": selected_features,
+    },
+    "ensemble_weights": weights,
+}
+
+report_path = MODELS_DIR / "benchmark_report.json"
+with open(report_path, "w") as f:
+    json.dump(report, f, indent=2)
+print(f"  ✅ benchmark_report.json saved")
+
+md_path = MODELS_DIR / "benchmark_report.md"
+_write_markdown_report(report, md_path)
+print(f"  ✅ benchmark_report.md saved")
+
+print("\n" + "=" * 65)
+print("DONE")
+print(f"  Baseline P@3:  {baseline['precision_at3']:.4f}")
+print(f"  Ensemble P@3:  {ens_res['precision_at3']:.4f}")
+print(f"  Delta:         {delta:+.4f} ({delta*100:+.1f}%)")
+print(f"  Deploy:        {'✅ YES' if deploy else '❌ NO'}")
+print("=" * 65)

saas_api.py

@@ -0,0 +1,247 @@
+#!/usr/bin/env python3
+"""
+Turf SaaS API v1 — Auth JWT + Multi-tenant
+Sprint 2-3: HRT-28
+
+Run:
+    FLASK_ENV=development ./venv/bin/python saas_api.py
+
+Ports (isolated from production):
+  Portal:       8793
+  SaaS API:     8792   ← this file
+  Dashboard:    8791
+  Combined API: 8790
+"""
+
+import os
+import logging
+import logging.handlers
+import sys
+
+from flask import Flask, jsonify, g, request
+from flask_cors import CORS
+from flask_jwt_extended import JWTManager, get_jwt
+
+from auth_db import init_auth_tables
+from auth import (
+    auth_bp,
+    jwt_required_middleware,
+    plan_required,
+    free_daily_limit_check,
+    _get_user_by_id,
+)
+from middleware import rate_limit_middleware, access_log_middleware
+
+# ──────────────────────────────────────────────────────────────
+# Logging setup
+# ──────────────────────────────────────────────────────────────
+
+LOG_DIR = os.path.join(os.path.dirname(__file__), "logs")
+os.makedirs(LOG_DIR, exist_ok=True)
+
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
+    handlers=[
+        logging.StreamHandler(sys.stdout),
+        logging.handlers.RotatingFileHandler(
+            os.path.join(LOG_DIR, "saas_api.log"),
+            maxBytes=5 * 1024 * 1024,
+            backupCount=3,
+        ),
+    ],
+)
+
+# ──────────────────────────────────────────────────────────────
+# App factory
+# ──────────────────────────────────────────────────────────────
+
+
+def create_app(test_config=None):
+    app = Flask(__name__)
+
+    # JWT config
+    app.config["JWT_SECRET_KEY"] = os.environ.get(
+        "JWT_SECRET_KEY", "CHANGE_ME_IN_PRODUCTION_" + os.urandom(24).hex()
+    )
+    app.config["JWT_ACCESS_TOKEN_EXPIRES"] = 900  # 15 minutes
+    app.config["JWT_REFRESH_TOKEN_EXPIRES"] = 2592000  # 30 days
+
+    if test_config:
+        app.config.update(test_config)
+
+    # CORS — SaaS domain + localhost for dev
+    CORS(
+        app,
+        origins=os.environ.get(
+            "CORS_ORIGINS",
+            "http://localhost:8793,http://127.0.0.1:8793,https://turf-ia.h3r7.tech",
+        ).split(","),
+        methods=["GET", "POST", "PUT", "DELETE", "OPTIONS"],
+        allow_headers=["Content-Type", "Authorization"],
+        supports_credentials=True,
+    )
+
+    # JWT
+    jwt = JWTManager(app)
+
+    # ── JWT error handlers ────────────────────────────────────
+    @jwt.expired_token_loader
+    def expired_token(_jwt_header, _jwt_payload):
+        return jsonify({"error": "Token expiré"}), 401
+
+    @jwt.invalid_token_loader
+    def invalid_token(reason):
+        return jsonify({"error": "Token invalide", "detail": reason}), 422
+
+    @jwt.unauthorized_loader
+    def unauthorized(reason):
+        return jsonify({"error": "Token manquant ou invalide", "detail": reason}), 401
+
+    # ── Register middleware ───────────────────────────────────
+    rate_limit_middleware(app)
+    access_log_middleware(app)
+
+    # ── Blueprints ────────────────────────────────────────────
+    app.register_blueprint(auth_bp)
+
+    # ── Predictions routes (multi-tenant plan check) ──────────
+
+    @app.route("/api/v1/predictions", methods=["GET"])
+    @jwt_required_middleware
+    @free_daily_limit_check
+    def predictions():
+        """
+        GET /api/v1/predictions
+        - free:    Top 3 uniquement (déjà filtrées par le moteur ML)
+        - premium: toutes courses + alertes Telegram
+        - pro:     API complète + export CSV disponible
+        """
+        user = g.current_user
+        plan = user["plan"]
+
+        # Forward to combined_api for actual predictions
+        import requests as req
+
+        try:
+            params = dict(request.args)
+            resp = req.get(
+                "http://localhost:8790/api/predictions",
+                params=params,
+                timeout=10,
+            )
+            data = resp.json()
+        except Exception as e:
+            return jsonify(
+                {"error": "Service prédictions indisponible", "detail": str(e)}
+            ), 503
+
+        # Plan filtering
+        if plan == "free":
+            # Top 3 only
+            if isinstance(data, list):
+                data = [
+                    {k: v for k, v in p.items() if k not in ("score_detaille",)}
+                    for p in data[:3]
+                ]
+            return jsonify({"plan": plan, "predictions": data, "limit": "Top 3"}), 200
+
+        elif plan == "premium":
+            # All courses, but no CSV export
+            return jsonify(
+                {"plan": plan, "predictions": data, "telegram_alerts": True}
+            ), 200
+
+        else:  # pro
+            return jsonify(
+                {
+                    "plan": plan,
+                    "predictions": data,
+                    "telegram_alerts": True,
+                    "csv_export_url": "/api/v1/predictions/export",
+                }
+            ), 200
+
+    @app.route("/api/v1/predictions/export", methods=["GET"])
+    @jwt_required_middleware
+    @plan_required("pro")
+    def predictions_export():
+        """CSV export — pro plan only."""
+        import requests as req
+        import io
+
+        try:
+            resp = req.get(
+                "http://localhost:8790/api/predictions/export",
+                params=dict(request.args),
+                timeout=15,
+            )
+            from flask import Response
+
+            return Response(
+                resp.content,
+                mimetype="text/csv",
+                headers={"Content-Disposition": "attachment; filename=predictions.csv"},
+            )
+        except Exception as e:
+            return jsonify({"error": "Export indisponible", "detail": str(e)}), 503
+
+    @app.route("/api/v1/subscription/upgrade", methods=["GET"])
+    @jwt_required_middleware
+    def subscription_info():
+        """Return available plans and current user plan."""
+        user = g.current_user
+        return jsonify(
+            {
+                "current_plan": user["plan"],
+                "plans": {
+                    "free": {
+                        "price": "0€/mois",
+                        "features": ["Top 3 prédictions", "1 course/jour"],
+                    },
+                    "premium": {
+                        "price": "9.99€/mois",
+                        "features": [
+                            "Toutes les courses",
+                            "Alertes Telegram",
+                            "Historique 30j",
+                        ],
+                    },
+                    "pro": {
+                        "price": "29.99€/mois",
+                        "features": [
+                            "API complète",
+                            "Export CSV",
+                            "Alertes Telegram",
+                            "Historique illimité",
+                            "Support prioritaire",
+                        ],
+                    },
+                },
+                "upgrade_contact": "contact@h3r7.tech",
+            }
+        ), 200
+
+    # ── Health check ──────────────────────────────────────────
+
+    @app.route("/api/v1/health", methods=["GET"])
+    def health():
+        return jsonify(
+            {"status": "ok", "service": "turf-saas-api", "version": "2.3.0"}
+        ), 200
+
+    # Init DB tables on startup
+    with app.app_context():
+        init_auth_tables()
+
+    return app
+
+
+# ──────────────────────────────────────────────────────────────
+# Entrypoint
+# ──────────────────────────────────────────────────────────────
+
+if __name__ == "__main__":
+    app = create_app()
+    port = int(os.environ.get("SAAS_API_PORT", 8792))
+    app.run(host="0.0.0.0", port=port, debug=False)

saas_auth.py

@@ -14,6 +14,18 @@ import time
 import json
 from functools import wraps
 from datetime import datetime
+from collections import defaultdict
+from threading import Lock
+
+# ─── Rate limiting login ───────────────────────────────────────────────────────
+_login_attempts: dict = defaultdict(
+    lambda: {"count": 0, "window_start": 0.0, "blocked_until": 0.0}
+)
+_login_lock = Lock()
+
+LOGIN_RATE_MAX = 5  # max tentatives par fenêtre
+LOGIN_RATE_WINDOW = 300  # 5 minutes (en secondes)
+LOGIN_BLOCK_DURATION = 900  # 15 min de blocage après dépassement
 
 # ─── Config ───────────────────────────────────────────────────────────────────
 DB_PATH = os.environ.get("TURF_SAAS_DB", "/home/h3r7/turf_saas/turf_saas.db")
@@ -184,6 +196,37 @@ def login():
     if not email or not password:
         return jsonify({"error": "Email et mot de passe requis."}), 400
 
+    # ── Rate limit par IP ────────────────────────────────────────
+    ip = request.remote_addr or "unknown"
+    now = time.time()
+
+    with _login_lock:
+        bucket = _login_attempts[ip]
+        # Lever le blocage si la durée est écoulée
+        if now >= bucket["blocked_until"]:
+            if now - bucket["window_start"] >= LOGIN_RATE_WINDOW:
+                bucket["count"] = 0
+                bucket["window_start"] = now
+            bucket["count"] += 1
+            count = bucket["count"]
+            if count > LOGIN_RATE_MAX:
+                bucket["blocked_until"] = now + LOGIN_BLOCK_DURATION
+                retry_after = LOGIN_BLOCK_DURATION
+                blocked = True
+            else:
+                retry_after = int(LOGIN_RATE_WINDOW - (now - bucket["window_start"]))
+                blocked = False
+        else:
+            blocked = True
+            retry_after = int(bucket["blocked_until"] - now)
+
+    if blocked:
+        resp = jsonify({"error": "Trop de tentatives. Réessayez plus tard."})
+        resp.status_code = 429
+        resp.headers["Retry-After"] = str(retry_after)
+        return resp
+    # ─────────────────────────────────────────────────────────────
+
     pw_hash = hash_password(password)
     conn = get_db()
     user = conn.execute(

tests/beta_monitor.py

@@ -0,0 +1,448 @@
+"""
+Beta Monitoring — SaaS Turf Prédictions IA
+Sprint 8 — QA, Beta Fermee, Go/No-Go
+Ticket: HRT-34
+
+Ce module :
+- Collecte les feedbacks beta via l'API in-app
+- Envoie des alertes Telegram en cas d'erreur détectée pendant la beta
+- Génère le rapport beta final (bugs, UX, NPS)
+
+Usage :
+  # Démarrer le monitoring beta
+  python tests/beta_monitor.py --watch --interval 60
+
+  # Générer le rapport beta final
+  python tests/beta_monitor.py --report
+
+  # Test d'envoi Telegram
+  python tests/beta_monitor.py --test-telegram
+"""
+
+import os
+import sys
+import json
+import time
+import sqlite3
+import requests
+import argparse
+from datetime import datetime, timedelta
+from pathlib import Path
+
+# ============================================================
+# Configuration
+# ============================================================
+
+BASE_URL = os.environ.get("APP_URL", "http://localhost:8792")
+TELEGRAM_TOKEN = os.environ.get(
+    "TELEGRAM_TOKEN", "8649773134:AAFqzZVtSHfPPFDadcte1B-1h23nZ8DmdYE"
+)
+TELEGRAM_CHAT_ID = os.environ.get("TELEGRAM_CHAT_ID", "")  # À configurer
+
+BETA_DB_PATH = os.environ.get("BETA_DB_PATH", "/home/h3r7/turf_saas/turf_saas.db")
+REPORTS_DIR = Path("tests/reports")
+REPORTS_DIR.mkdir(parents=True, exist_ok=True)
+
+# Seuils d'alerte
+ERROR_RATE_THRESHOLD = 0.01  # 1% d'erreurs → alerte
+LATENCY_P95_THRESHOLD_MS = 500  # p95 > 500ms → alerte
+BETA_MIN_USERS = 10  # Minimum d'utilisateurs beta requis
+NPS_TARGET = 7.0  # NPS cible (sur 10)
+
+
+# ============================================================
+# Alertes Telegram
+# ============================================================
+
+
+def send_telegram(message: str, parse_mode: str = "Markdown") -> bool:
+    """Envoie un message Telegram d'alerte."""
+    if not TELEGRAM_TOKEN or not TELEGRAM_CHAT_ID:
+        print(f"⚠️  Telegram non configuré. Message: {message[:100]}")
+        return False
+
+    try:
+        resp = requests.post(
+            f"https://api.telegram.org/bot{TELEGRAM_TOKEN}/sendMessage",
+            json={
+                "chat_id": TELEGRAM_CHAT_ID,
+                "text": message,
+                "parse_mode": parse_mode,
+            },
+            timeout=10,
+        )
+        if resp.status_code == 200:
+            print(f"✅ Alerte Telegram envoyée")
+            return True
+        else:
+            print(f"❌ Telegram erreur: {resp.status_code} — {resp.text}")
+            return False
+    except Exception as e:
+        print(f"❌ Telegram exception: {e}")
+        return False
+
+
+def alert_error(endpoint: str, status_code: int, message: str):
+    """Alerte Telegram sur erreur critique."""
+    text = (
+        f"🚨 *ALERTE BETA — SaaS Turf IA*\n\n"
+        f"Erreur détectée sur `{endpoint}`\n"
+        f"Status: `{status_code}`\n"
+        f"Message: {message[:200]}\n"
+        f"Heure: {datetime.now().strftime('%H:%M:%S')}\n\n"
+        f"_Ticket: HRT-34_"
+    )
+    send_telegram(text)
+
+
+def alert_performance(p95_ms: float, error_rate: float):
+    """Alerte Telegram sur dégradation de performance."""
+    text = (
+        f"⚠️ *ALERTE PERFORMANCE — SaaS Turf IA*\n\n"
+        f"p95 latence: `{p95_ms:.0f}ms` (seuil: {LATENCY_P95_THRESHOLD_MS}ms)\n"
+        f"Error rate: `{error_rate * 100:.2f}%` (seuil: {ERROR_RATE_THRESHOLD * 100:.1f}%)\n"
+        f"Heure: {datetime.now().strftime('%H:%M:%S')}\n\n"
+        f"_Ticket: HRT-34_"
+    )
+    send_telegram(text)
+
+
+# ============================================================
+# Collecte de métriques
+# ============================================================
+
+
+class BetaMonitor:
+    """Moniteur actif pendant la beta fermée."""
+
+    ENDPOINTS_TO_CHECK = [
+        "/api",
+        "/api/races",
+        "/api/scoring",
+        "/dashboard",
+        "/",
+    ]
+
+    def __init__(self, base_url: str = BASE_URL):
+        self.base_url = base_url.rstrip("/")
+        self.errors: list[dict] = []
+        self.latencies: list[float] = []
+        self.check_count = 0
+
+    def check_endpoint(self, path: str) -> dict:
+        """Vérifie un endpoint et retourne le résultat."""
+        start = time.time()
+        try:
+            resp = requests.get(f"{self.base_url}{path}", timeout=10)
+            latency_ms = (time.time() - start) * 1000
+            return {
+                "path": path,
+                "status": resp.status_code,
+                "latency_ms": latency_ms,
+                "ok": resp.status_code < 500,
+                "timestamp": datetime.now().isoformat(),
+            }
+        except requests.exceptions.ConnectionError as e:
+            return {
+                "path": path,
+                "status": 0,
+                "latency_ms": 0,
+                "ok": False,
+                "error": str(e),
+                "timestamp": datetime.now().isoformat(),
+            }
+        except Exception as e:
+            return {
+                "path": path,
+                "status": 0,
+                "latency_ms": 0,
+                "ok": False,
+                "error": str(e),
+                "timestamp": datetime.now().isoformat(),
+            }
+
+    def run_checks(self) -> dict:
+        """Exécute tous les checks et retourne un résumé."""
+        results = [self.check_endpoint(p) for p in self.ENDPOINTS_TO_CHECK]
+        self.check_count += 1
+
+        failures = [r for r in results if not r["ok"]]
+        latencies = [r["latency_ms"] for r in results if r["latency_ms"] > 0]
+
+        p95 = (
+            sorted(latencies)[int(len(latencies) * 0.95)]
+            if len(latencies) >= 2
+            else (latencies[0] if latencies else 0)
+        )
+        error_rate = len(failures) / len(results) if results else 0
+
+        # Stocker pour rapport
+        self.latencies.extend(latencies)
+        self.errors.extend(failures)
+
+        return {
+            "check_number": self.check_count,
+            "timestamp": datetime.now().isoformat(),
+            "total_checks": len(results),
+            "failures": len(failures),
+            "error_rate": error_rate,
+            "p95_ms": p95,
+            "results": results,
+        }
+
+    def watch(self, interval_seconds: int = 60):
+        """Surveillance continue avec alertes Telegram."""
+        print(f"🔍 Beta monitoring démarré — {self.base_url}")
+        print(f"   Intervalle: {interval_seconds}s")
+        print(f"   Endpoints: {len(self.ENDPOINTS_TO_CHECK)}")
+        print(f"   Ctrl+C pour arrêter\n")
+
+        consecutive_errors = 0
+
+        try:
+            while True:
+                summary = self.run_checks()
+                timestamp = datetime.now().strftime("%H:%M:%S")
+
+                status_icon = "✅" if summary["error_rate"] == 0 else "❌"
+                print(
+                    f"[{timestamp}] {status_icon} "
+                    f"Check #{summary['check_number']} — "
+                    f"p95={summary['p95_ms']:.0f}ms, "
+                    f"errors={summary['failures']}/{summary['total_checks']}"
+                )
+
+                # Alertes
+                if summary["error_rate"] > ERROR_RATE_THRESHOLD:
+                    consecutive_errors += 1
+                    if consecutive_errors >= 2:  # 2 checks consécutifs en erreur
+                        for failure in summary["results"]:
+                            if not failure["ok"]:
+                                alert_error(
+                                    failure["path"],
+                                    failure.get("status", 0),
+                                    failure.get("error", "Non-2xx response"),
+                                )
+                else:
+                    consecutive_errors = 0
+
+                if summary["p95_ms"] > LATENCY_P95_THRESHOLD_MS:
+                    print(f"⚠️  Latence p95 élevée: {summary['p95_ms']:.0f}ms")
+                    if summary["p95_ms"] > LATENCY_P95_THRESHOLD_MS * 2:
+                        alert_performance(summary["p95_ms"], summary["error_rate"])
+
+                # Sauvegarder les résultats
+                log_file = REPORTS_DIR / "beta_monitor_log.jsonl"
+                with open(log_file, "a") as f:
+                    f.write(json.dumps(summary) + "\n")
+
+                time.sleep(interval_seconds)
+
+        except KeyboardInterrupt:
+            print(f"\n⏹️  Monitoring arrêté après {self.check_count} checks")
+            self.generate_report()
+
+
+# ============================================================
+# Rapport beta final
+# ============================================================
+
+
+class BetaReport:
+    """Générateur de rapport beta fermée."""
+
+    def __init__(self, base_url: str = BASE_URL):
+        self.base_url = base_url
+        self.timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+
+    def collect_feedback_from_db(self) -> list[dict]:
+        """Collecte les feedbacks depuis la BDD (table beta_feedback si elle existe)."""
+        try:
+            conn = sqlite3.connect(BETA_DB_PATH)
+            c = conn.cursor()
+            c.execute(
+                "SELECT name FROM sqlite_master WHERE type='table' AND name='beta_feedback'"
+            )
+            if not c.fetchone():
+                conn.close()
+                return []
+            c.execute("SELECT * FROM beta_feedback ORDER BY created_at DESC")
+            rows = c.fetchall()
+            conn.close()
+            return [dict(zip([col[0] for col in c.description], row)) for row in rows]
+        except Exception as e:
+            print(f"⚠️  Impossible de lire beta_feedback: {e}")
+            return []
+
+    def collect_monitor_logs(self) -> list[dict]:
+        """Lit les logs du monitoring beta."""
+        log_file = REPORTS_DIR / "beta_monitor_log.jsonl"
+        if not log_file.exists():
+            return []
+        entries = []
+        with open(log_file) as f:
+            for line in f:
+                try:
+                    entries.append(json.loads(line))
+                except Exception:
+                    pass
+        return entries
+
+    def generate(self) -> str:
+        """Génère le rapport complet et le sauvegarde."""
+        feedbacks = self.collect_feedback_from_db()
+        monitor_logs = self.collect_monitor_logs()
+
+        # Calculer NPS depuis les feedbacks
+        nps_scores = [
+            f.get("nps_score") for f in feedbacks if f.get("nps_score") is not None
+        ]
+        avg_nps = sum(nps_scores) / len(nps_scores) if nps_scores else None
+
+        # Statistiques monitoring
+        if monitor_logs:
+            all_latencies = []
+            total_errors = 0
+            total_checks = 0
+            for entry in monitor_logs:
+                all_latencies.extend(
+                    [
+                        r["latency_ms"]
+                        for r in entry.get("results", [])
+                        if r.get("latency_ms", 0) > 0
+                    ]
+                )
+                total_errors += entry.get("failures", 0)
+                total_checks += entry.get("total_checks", 0)
+            avg_latency = (
+                sum(all_latencies) / len(all_latencies) if all_latencies else 0
+            )
+            overall_error_rate = total_errors / total_checks if total_checks > 0 else 0
+        else:
+            avg_latency = 0
+            overall_error_rate = 0
+            total_checks = 0
+
+        # Construire le rapport
+        report = []
+        report.append("=" * 60)
+        report.append("RAPPORT BETA FERMÉE — SaaS Turf Prédictions IA")
+        report.append(f"Généré le : {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
+        report.append(f"Ticket    : HRT-34")
+        report.append("=" * 60)
+        report.append("")
+        report.append("## 1. PARTICIPANTS BETA")
+        report.append(f"   Feedbacks reçus : {len(feedbacks)}")
+        report.append(
+            f"   NPS moyen       : {avg_nps:.1f}/10"
+            if avg_nps
+            else "   NPS moyen       : (en attente feedbacks)"
+        )
+        report.append(f"   Cible NPS       : ≥ {NPS_TARGET}/10")
+        nps_ok = avg_nps is not None and avg_nps >= NPS_TARGET
+        report.append(
+            f"   Statut NPS      : {'✅ OBJECTIF ATTEINT' if nps_ok else '⏳ En attente' if avg_nps is None else '❌ OBJECTIF NON ATTEINT'}"
+        )
+        report.append("")
+        report.append("## 2. BUGS SIGNALÉS")
+        bugs = [f for f in feedbacks if f.get("type") == "bug"]
+        critical_bugs = [b for b in bugs if b.get("severity") in ("critical", "high")]
+        report.append(f"   Total bugs      : {len(bugs)}")
+        report.append(f"   Critiques/High  : {len(critical_bugs)}")
+        report.append(
+            f"   Statut          : {'✅ 0 bug critique' if len(critical_bugs) == 0 else f'❌ {len(critical_bugs)} bug(s) critique(s)'}"
+        )
+        report.append("")
+        report.append("## 3. PERFORMANCE RÉELLE (monitoring)")
+        report.append(f"   Checks effectués: {total_checks}")
+        report.append(f"   Latence moyenne : {avg_latency:.1f}ms")
+        report.append(f"   Error rate      : {overall_error_rate * 100:.2f}%")
+        report.append(f"   Seuil latence   : {LATENCY_P95_THRESHOLD_MS}ms")
+        perf_ok = (
+            avg_latency < LATENCY_P95_THRESHOLD_MS
+            and overall_error_rate < ERROR_RATE_THRESHOLD
+        )
+        report.append(
+            f"   Statut          : {'✅ OBJECTIF ATTEINT' if perf_ok else '⏳ Données insuffisantes' if total_checks == 0 else '❌ OBJECTIF NON ATTEINT'}"
+        )
+        report.append("")
+        report.append("## 4. FEEDBACKS UX")
+        ux_feedbacks = [f for f in feedbacks if f.get("type") == "ux"]
+        report.append(f"   Retours UX      : {len(ux_feedbacks)}")
+        if ux_feedbacks:
+            for fb in ux_feedbacks[:5]:  # Top 5
+                report.append(f"   - {fb.get('comment', '')[:100]}")
+        report.append("")
+        report.append("## 5. VERDICT BETA FERMÉE")
+        users_ok = len(feedbacks) >= 5  # Au moins 5 feedbacks = 5 users satisfaits
+        verdict = all([users_ok, nps_ok, len(critical_bugs) == 0])
+        report.append(
+            f"   Participants suffisants (≥5)  : {'✅' if users_ok else '❌'}"
+        )
+        report.append(f"   NPS ≥ 7/10                   : {'✅' if nps_ok else '❌'}")
+        report.append(
+            f"   0 bug critique               : {'✅' if len(critical_bugs) == 0 else '❌'}"
+        )
+        report.append("")
+        report.append(
+            f"   VERDICT GLOBAL : {'✅ GO — Beta réussie' if verdict else '❌ NO-GO — Conditions non remplies'}"
+        )
+        report.append("=" * 60)
+
+        report_text = "\n".join(report)
+
+        # Sauvegarder
+        report_file = REPORTS_DIR / f"beta_report_{self.timestamp}.txt"
+        with open(report_file, "w") as f:
+            f.write(report_text)
+
+        print(report_text)
+        print(f"\nRapport sauvegardé : {report_file}")
+
+        return report_text
+
+
+# ============================================================
+# CLI
+# ============================================================
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Beta Monitor — SaaS Turf IA")
+    parser.add_argument("--watch", action="store_true", help="Surveillance continue")
+    parser.add_argument(
+        "--interval", type=int, default=60, help="Intervalle en secondes (défaut: 60)"
+    )
+    parser.add_argument(
+        "--report", action="store_true", help="Générer le rapport beta final"
+    )
+    parser.add_argument(
+        "--test-telegram", action="store_true", help="Tester l'envoi Telegram"
+    )
+    parser.add_argument(
+        "--url", default=BASE_URL, help=f"URL de l'app (défaut: {BASE_URL})"
+    )
+    args = parser.parse_args()
+
+    if args.test_telegram:
+        print("Test d'envoi Telegram...")
+        ok = send_telegram(
+            "✅ *Test alerte Beta* — SaaS Turf IA\n_Ceci est un test du système d'alertes QA_\nTicket: HRT-34"
+        )
+        sys.exit(0 if ok else 1)
+
+    if args.report:
+        reporter = BetaReport(args.url)
+        reporter.generate()
+        sys.exit(0)
+
+    if args.watch:
+        monitor = BetaMonitor(args.url)
+        monitor.watch(interval_seconds=args.interval)
+        sys.exit(0)
+
+    parser.print_help()
+
+
+if __name__ == "__main__":
+    main()

tests/conftest.py

@@ -0,0 +1,124 @@
+"""
+conftest.py — Configuration pytest globale
+SaaS Turf Prédictions IA — Sprint 8 QA
+Ticket: HRT-34
+"""
+
+import os
+import asyncio
+import pytest
+from pathlib import Path
+from datetime import datetime
+
+# ============================================================
+# Répertoires de sortie
+# ============================================================
+
+REPORTS_DIR = Path("tests/reports")
+SCREENSHOTS_DIR = Path("tests/screenshots")
+
+for d in [REPORTS_DIR, SCREENSHOTS_DIR]:
+    d.mkdir(parents=True, exist_ok=True)
+
+
+# ============================================================
+# Variables d'environnement
+# ============================================================
+
+BASE_URL = os.environ.get("APP_URL", "http://localhost:8792")
+
+
+# ============================================================
+# Fixtures globales
+# ============================================================
+
+
+@pytest.fixture(scope="session")
+def base_url():
+    return BASE_URL
+
+
+@pytest.fixture(scope="session")
+def event_loop():
+    """Event loop partagé pour les tests async de la session."""
+    policy = asyncio.get_event_loop_policy()
+    loop = policy.new_event_loop()
+    yield loop
+    loop.close()
+
+
+@pytest.fixture(scope="session")
+def reports_dir():
+    return REPORTS_DIR
+
+
+@pytest.fixture(scope="session")
+def screenshots_dir():
+    return SCREENSHOTS_DIR
+
+
+# ============================================================
+# Hook : screenshot automatique sur échec
+# ============================================================
+
+
+@pytest.hookimpl(tryfirst=True, hookwrapper=True)
+def pytest_runtest_makereport(item, call):
+    """Capture screenshot automatiquement sur tout test E2E en échec."""
+    outcome = yield
+    report = outcome.get_result()
+
+    if report.when == "call" and report.failed:
+        # Récupérer la page Playwright si disponible dans les fixtures
+        page = None
+        for fixture_name in ("page", "context_page"):
+            if fixture_name in item.funcargs:
+                val = item.funcargs[fixture_name]
+                if isinstance(val, tuple):
+                    page = val[0]  # (page, browser_name)
+                else:
+                    page = val
+                break
+
+        if page is not None:
+            timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+            test_name = item.name.replace("/", "_").replace(":", "_")
+            screenshot_path = SCREENSHOTS_DIR / f"FAIL_{test_name}_{timestamp}.png"
+            try:
+                # Playwright page.screenshot est synchrone dans les fixtures sync
+                # Pour les fixtures async, on force la capture
+                import asyncio as _asyncio
+
+                if _asyncio.iscoroutinefunction(page.screenshot):
+                    loop = _asyncio.get_event_loop()
+                    loop.run_until_complete(page.screenshot(path=str(screenshot_path)))
+                else:
+                    page.screenshot(path=str(screenshot_path))
+                report.sections.append(
+                    ("Screenshot", f"Sauvegardé : {screenshot_path}")
+                )
+            except Exception as e:
+                report.sections.append(
+                    ("Screenshot Error", f"Impossible de capturer : {e}")
+                )
+
+
+# ============================================================
+# Marqueurs personnalisés
+# ============================================================
+
+
+def pytest_configure(config):
+    config.addinivalue_line("markers", "e2e: Tests End-to-End Playwright")
+    config.addinivalue_line("markers", "load: Tests de charge Locust")
+    config.addinivalue_line("markers", "security: Tests de sécurité")
+    config.addinivalue_line(
+        "markers", "smoke: Tests rapides de smoke (sans infra complète)"
+    )
+    config.addinivalue_line("markers", "beta: Tests spécifiques beta fermée")
+    config.addinivalue_line(
+        "markers", "requires_billing: Nécessite HRT-31 (Billing Stripe)"
+    )
+    config.addinivalue_line(
+        "markers", "requires_infra: Nécessite HRT-33 (infra staging)"
+    )

tests/security/test_security.py

@@ -141,7 +141,7 @@ class TestJWTAuthentication:
             "invalid_signature_here"
         )
         resp = requests.get(
-            f"{BASE_URL}/api/races",
+            f"{BASE_URL}/api/v1/predictions/today",
             headers={"Authorization": f"Bearer {expired_token}"},
             timeout=5,
         )
@@ -153,7 +153,7 @@ class TestJWTAuthentication:
         """Un token JWT malformé doit être rejeté."""
         for bad_token in ["not.a.jwt", "Bearer", "null", "undefined", ""]:
             resp = requests.get(
-                f"{BASE_URL}/api/races",
+                f"{BASE_URL}/api/v1/predictions/today",
                 headers={"Authorization": f"Bearer {bad_token}"},
                 timeout=5,
             )
@@ -163,7 +163,7 @@ class TestJWTAuthentication:
 
     def test_jwt_sans_token(self):
         """Sans token, les routes protégées doivent retourner 401."""
-        resp = requests.get(f"{BASE_URL}/api/export/csv", timeout=5)
+        resp = requests.get(f"{BASE_URL}/api/v1/export/csv", timeout=5)
         assert resp.status_code in (401, 403), (
             f"Route protégée accessible sans token: status={resp.status_code}"
         )
@@ -303,6 +303,55 @@ class TestPlanAuthorisation:
             )
 
 
+# === Tests rate limiting login ===
+
+
+class TestLoginRateLimit:
+    """Tests rate limiting sur /api/v1/auth/login."""
+
+    TARGET_URL = (
+        os.environ.get("APP_URL", "http://localhost:8792") + "/api/v1/auth/login"
+    )
+
+    def test_login_brute_force_blocked_after_5_attempts(self):
+        """Après 5 tentatives, le 6ème appel doit retourner 429."""
+        # Utiliser un email unique pour isoler le test
+        email = f"ratelimit_test_{int(time.time())}@h3r7.tech"
+        for i in range(5):
+            resp = requests.post(
+                self.TARGET_URL,
+                json={"email": email, "password": "wrong_password"},
+                timeout=5,
+            )
+            assert resp.status_code in (400, 401), (
+                f"Tentative {i + 1}: status inattendu {resp.status_code}"
+            )
+        # La 6ème tentative doit être bloquée
+        resp = requests.post(
+            self.TARGET_URL,
+            json={"email": email, "password": "wrong_password"},
+            timeout=5,
+        )
+        assert resp.status_code == 429, (
+            f"Rate limit non appliqué après 5 tentatives: got {resp.status_code}"
+        )
+        assert "Retry-After" in resp.headers, "Header Retry-After manquant sur 429"
+
+    def test_login_429_has_retry_after_header(self):
+        """La réponse 429 doit inclure Retry-After."""
+        email = f"ratelimit_test2_{int(time.time())}@h3r7.tech"
+        for _ in range(6):
+            requests.post(
+                self.TARGET_URL, json={"email": email, "password": "x"}, timeout=5
+            )
+        resp = requests.post(
+            self.TARGET_URL, json={"email": email, "password": "x"}, timeout=5
+        )
+        if resp.status_code == 429:
+            assert "Retry-After" in resp.headers
+            assert int(resp.headers["Retry-After"]) > 0
+
+
 if __name__ == "__main__":
     import subprocess
 

tests/test_api_v1.py

@@ -0,0 +1,473 @@
+#!/usr/bin/env python3
+"""
+Integration tests for API v1 — HRT-29
+Sprint 3-4: Refacto API /v1/
+
+Run with:
+    cd /home/h3r7/turf_saas
+    source venv/bin/activate
+    python -m pytest tests/test_api_v1.py -v
+"""
+
+import json
+import os
+import sys
+import tempfile
+import pytest
+
+# Ensure local modules are importable
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+# Use a temp file DB for tests (in-memory fails with multiple connections)
+_tmp_db = tempfile.NamedTemporaryFile(suffix=".db", delete=False)
+_tmp_db.close()
+os.environ["TURF_SAAS_DB"] = _tmp_db.name
+os.environ["JWT_SECRET_KEY"] = "test-secret-key"
+
+from app_v1 import create_app
+from auth_db import init_auth_tables
+
+
+# ──────────────────────────────────────────────────────────────
+# Fixtures
+# ──────────────────────────────────────────────────────────────
+
+
+@pytest.fixture(scope="module")
+def app():
+    application = create_app()
+    application.config["TESTING"] = True
+    application.config["JWT_SECRET_KEY"] = "test-secret-key"
+    yield application
+
+
+@pytest.fixture(scope="module")
+def client(app):
+    return app.test_client()
+
+
+@pytest.fixture(scope="module")
+def auth_tokens(client):
+    """Register a user and return tokens for each plan."""
+    tokens = {}
+    plans = {
+        "free": ("free@test.com", "password123"),
+        "premium": ("premium@test.com", "password123"),
+        "pro": ("pro@test.com", "password123"),
+    }
+
+    # Register users
+    for plan, (email, pw) in plans.items():
+        r = client.post(
+            "/api/v1/auth/register",
+            json={"email": email, "password": pw},
+            content_type="application/json",
+        )
+        assert r.status_code in (201, 409), f"register failed for {plan}: {r.data}"
+
+    # Manually set plans in DB using direct sqlite (bypass app context issues)
+    import sqlite3
+
+    db_path = os.environ.get("TURF_SAAS_DB", "/tmp/test_turf.db")
+    conn = sqlite3.connect(db_path)
+    for plan, (email, _) in plans.items():
+        conn.execute("UPDATE users SET plan = ? WHERE email = ?", (plan, email))
+    conn.commit()
+    conn.close()
+
+    # Login and collect tokens
+    for plan, (email, pw) in plans.items():
+        r = client.post(
+            "/api/v1/auth/login",
+            json={"email": email, "password": pw},
+            content_type="application/json",
+        )
+        assert r.status_code == 200, f"login failed for {plan}: {r.data}"
+        data = r.get_json()
+        tokens[plan] = data["access_token"]
+
+    return tokens
+
+
+def auth_header(token: str) -> dict:
+    return {"Authorization": f"Bearer {token}"}
+
+
+# ──────────────────────────────────────────────────────────────
+# Health
+# ──────────────────────────────────────────────────────────────
+
+
+class TestHealth:
+    def test_health_public(self, client):
+        """GET /api/v1/health — no auth required"""
+        r = client.get("/api/v1/health")
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["status"] == "ok"
+        assert data["version"] == "1.0"
+        assert "timestamp" in data
+
+    def test_health_returns_json(self, client):
+        r = client.get("/api/v1/health")
+        assert r.content_type.startswith("application/json")
+
+
+# ──────────────────────────────────────────────────────────────
+# Auth
+# ──────────────────────────────────────────────────────────────
+
+
+class TestAuth:
+    def test_register_new_user(self, client):
+        r = client.post(
+            "/api/v1/auth/register",
+            json={"email": "new_test@example.com", "password": "strongpass123"},
+        )
+        assert r.status_code in (201, 409)
+
+    def test_register_short_password(self, client):
+        r = client.post(
+            "/api/v1/auth/register",
+            json={"email": "bad@example.com", "password": "123"},
+        )
+        assert r.status_code == 400
+
+    def test_register_invalid_email(self, client):
+        r = client.post(
+            "/api/v1/auth/register",
+            json={"email": "notemail", "password": "password123"},
+        )
+        assert r.status_code == 400
+
+    def test_login_valid(self, client, auth_tokens):
+        assert "free" in auth_tokens
+
+    def test_login_wrong_password(self, client):
+        r = client.post(
+            "/api/v1/auth/login",
+            json={"email": "free@test.com", "password": "wrongpassword"},
+        )
+        assert r.status_code == 401
+
+    def test_protected_without_token(self, client):
+        r = client.get("/api/v1/courses/today")
+        assert r.status_code == 401
+
+
+# ──────────────────────────────────────────────────────────────
+# Courses
+# ──────────────────────────────────────────────────────────────
+
+
+class TestCourses:
+    def test_today_requires_auth(self, client):
+        r = client.get("/api/v1/courses/today")
+        assert r.status_code == 401
+
+    def test_today_with_auth(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/courses/today",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["status"] == "ok"
+        assert "courses" in data
+        assert "pagination" in data
+        assert "date" in data
+
+    def test_today_pagination(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/courses/today?limit=5&offset=0",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["pagination"]["limit"] == 5
+        assert data["pagination"]["offset"] == 0
+
+    def test_today_filter_all(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/courses/today?filter=all",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 200
+
+    def test_course_predictions_requires_auth(self, client):
+        r = client.get("/api/v1/courses/1-1/predictions")
+        assert r.status_code == 401
+
+    def test_course_predictions_invalid_id(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/courses/invalid/predictions",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 400
+
+    def test_course_predictions_not_found(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/courses/99-99/predictions",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        # 404 expected since DB is empty; 429 if free daily limit already reached in this session
+        assert r.status_code in (404, 200, 429)  # 200 if gracefully returns empty
+
+
+# ──────────────────────────────────────────────────────────────
+# Predictions
+# ──────────────────────────────────────────────────────────────
+
+
+class TestPredictions:
+    def test_top3_requires_auth(self, client):
+        r = client.get("/api/v1/predictions/top3")
+        assert r.status_code == 401
+
+    def test_top3_free_allowed(self, client, auth_tokens):
+        # Reset daily usage for free user before testing rate-limited endpoint
+        import sqlite3
+
+        db_path = os.environ.get("TURF_SAAS_DB", "/tmp/test_turf.db")
+        conn = sqlite3.connect(db_path)
+        conn.execute(
+            "UPDATE users SET daily_usage=0, last_usage_date=NULL WHERE email='free@test.com'"
+        )
+        conn.commit()
+        conn.close()
+
+        r = client.get(
+            "/api/v1/predictions/top3",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["status"] == "ok"
+        assert "top3" in data
+
+    def test_all_requires_premium(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/predictions/all",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 403
+
+    def test_all_premium_allowed(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/predictions/all",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["status"] == "ok"
+        assert "predictions" in data
+        assert "pagination" in data
+
+    def test_all_pro_allowed(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/predictions/all",
+            headers=auth_header(auth_tokens["pro"]),
+        )
+        assert r.status_code == 200
+
+
+# ──────────────────────────────────────────────────────────────
+# Value Bets
+# ──────────────────────────────────────────────────────────────
+
+
+class TestValueBets:
+    def test_requires_auth(self, client):
+        r = client.get("/api/v1/valuebets")
+        assert r.status_code == 401
+
+    def test_free_forbidden(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/valuebets",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 403
+
+    def test_premium_allowed(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/valuebets",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["status"] == "ok"
+        assert "valuebets" in data
+        assert "pagination" in data
+
+    def test_min_odds_filter(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/valuebets?min_odds=3.0",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["min_odds"] == 3.0
+
+
+# ──────────────────────────────────────────────────────────────
+# Backtest
+# ──────────────────────────────────────────────────────────────
+
+
+class TestBacktest:
+    def test_requires_auth(self, client):
+        r = client.get("/api/v1/backtest")
+        assert r.status_code == 401
+
+    def test_premium_forbidden(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/backtest",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 403
+
+    def test_pro_allowed(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/backtest",
+            headers=auth_header(auth_tokens["pro"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["status"] == "ok"
+        assert "summary" in data
+        assert "period" in data
+
+    def test_invalid_date_format(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/backtest?start=31-12-2025",
+            headers=auth_header(auth_tokens["pro"]),
+        )
+        assert r.status_code == 400
+
+
+# ──────────────────────────────────────────────────────────────
+# Export
+# ──────────────────────────────────────────────────────────────
+
+
+class TestExport:
+    def test_requires_auth(self, client):
+        r = client.get("/api/v1/export/csv")
+        assert r.status_code == 401
+
+    def test_free_forbidden(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/export/csv",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 403
+
+    def test_premium_forbidden(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/export/csv",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 403
+
+    def test_pro_allowed_predictions(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/export/csv?type=predictions",
+            headers=auth_header(auth_tokens["pro"]),
+        )
+        # 200 (CSV) or 400 if table doesn't exist in test DB
+        assert r.status_code in (200, 400)
+        if r.status_code == 200:
+            assert "text/csv" in r.content_type
+
+    def test_invalid_type(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/export/csv?type=invalid",
+            headers=auth_header(auth_tokens["pro"]),
+        )
+        assert r.status_code == 400
+
+
+# ──────────────────────────────────────────────────────────────
+# Metrics
+# ──────────────────────────────────────────────────────────────
+
+
+class TestMetrics:
+    def test_requires_auth(self, client):
+        r = client.get("/api/v1/metrics")
+        assert r.status_code == 401
+
+    def test_free_forbidden(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/metrics",
+            headers=auth_header(auth_tokens["free"]),
+        )
+        assert r.status_code == 403
+
+    def test_premium_allowed(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/metrics",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["status"] == "ok"
+        assert "bet_metrics" in data
+        assert "ml_metrics" in data
+        assert "period" in data
+
+    def test_days_parameter(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/metrics?days=7",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 200
+        data = r.get_json()
+        assert data["period"]["days"] == 7
+
+    def test_invalid_days(self, client, auth_tokens):
+        r = client.get(
+            "/api/v1/metrics?days=abc",
+            headers=auth_header(auth_tokens["premium"]),
+        )
+        assert r.status_code == 400
+
+
+# ──────────────────────────────────────────────────────────────
+# Global error handlers
+# ──────────────────────────────────────────────────────────────
+
+
+class TestErrorHandlers:
+    def test_404_returns_json(self, client):
+        r = client.get("/api/v1/this-does-not-exist")
+        assert r.status_code == 404
+        data = r.get_json()
+        assert data["code"] == 404
+
+    def test_uniform_error_shape(self, client):
+        """All error responses must have {status, message, code}."""
+        r = client.get("/api/v1/this-does-not-exist")
+        data = r.get_json()
+        assert "status" in data
+        assert "message" in data
+        assert "code" in data
+
+
+# ──────────────────────────────────────────────────────────────
+# Swagger docs
+# ──────────────────────────────────────────────────────────────
+
+
+class TestDocs:
+    def test_docs_accessible(self, client):
+        r = client.get("/api/v1/docs")
+        # flasgger returns a redirect or the UI page
+        assert r.status_code in (200, 301, 302)
+
+    def test_apispec_json(self, client):
+        r = client.get("/api/v1/apispec.json")
+        assert r.status_code == 200
+        spec = r.get_json()
+        assert spec["swagger"] == "2.0"
+        assert "paths" in spec

tests/test_auth.py

@@ -0,0 +1,404 @@
+#!/usr/bin/env python3
+"""
+Pytest tests — Auth JWT + Multi-tenant
+Sprint 2-3: HRT-28
+Coverage target: >= 80%
+
+Run:
+    ./venv/bin/pytest tests/test_auth.py -v --tb=short
+    ./venv/bin/pytest tests/test_auth.py -v --cov=auth --cov=auth_db --cov=middleware --cov=saas_api --cov-report=term-missing
+"""
+
+import os
+import sys
+import tempfile
+import json
+import pytest
+
+# Point to a temp SQLite DB for tests
+_tmp_db = tempfile.NamedTemporaryFile(suffix=".db", delete=False)
+_tmp_db.close()
+os.environ["TURF_SAAS_DB"] = _tmp_db.name
+os.environ["JWT_SECRET_KEY"] = "test-secret-key-for-pytest"
+
+# Add project root to path
+sys.path.insert(0, os.path.dirname(os.path.dirname(__file__)))
+
+from saas_api import create_app  # noqa: E402
+
+TEST_CONFIG = {
+    "TESTING": True,
+    "JWT_SECRET_KEY": "test-secret-key-for-pytest",
+    "JWT_ACCESS_TOKEN_EXPIRES": 900,
+    "JWT_REFRESH_TOKEN_EXPIRES": 2592000,
+}
+
+
+@pytest.fixture(scope="module")
+def app():
+    application = create_app(TEST_CONFIG)
+    yield application
+
+
+@pytest.fixture(scope="module")
+def client(app):
+    return app.test_client()
+
+
+# ──────────────────────────────────────────────────────────────
+# Health
+# ──────────────────────────────────────────────────────────────
+
+
+class TestHealth:
+    def test_health_ok(self, client):
+        resp = client.get("/api/v1/health")
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert data["status"] == "ok"
+        assert data["service"] == "turf-saas-api"
+
+
+# ──────────────────────────────────────────────────────────────
+# Registration
+# ──────────────────────────────────────────────────────────────
+
+
+class TestRegister:
+    def test_register_success(self, client):
+        resp = client.post(
+            "/api/v1/auth/register",
+            json={"email": "user_test@example.com", "password": "password123"},
+        )
+        assert resp.status_code == 201
+        data = resp.get_json()
+        assert "user_id" in data
+
+    def test_register_duplicate(self, client):
+        client.post(
+            "/api/v1/auth/register",
+            json={"email": "dup@example.com", "password": "password123"},
+        )
+        resp = client.post(
+            "/api/v1/auth/register",
+            json={"email": "dup@example.com", "password": "password123"},
+        )
+        assert resp.status_code == 409
+
+    def test_register_invalid_email(self, client):
+        resp = client.post(
+            "/api/v1/auth/register",
+            json={"email": "notanemail", "password": "password123"},
+        )
+        assert resp.status_code == 400
+
+    def test_register_short_password(self, client):
+        resp = client.post(
+            "/api/v1/auth/register",
+            json={"email": "shortpw@example.com", "password": "abc"},
+        )
+        assert resp.status_code == 400
+
+    def test_register_missing_fields(self, client):
+        resp = client.post("/api/v1/auth/register", json={})
+        assert resp.status_code == 400
+
+
+# ──────────────────────────────────────────────────────────────
+# Login
+# ──────────────────────────────────────────────────────────────
+
+
+class TestLogin:
+    @pytest.fixture(autouse=True)
+    def create_user(self, client):
+        client.post(
+            "/api/v1/auth/register",
+            json={"email": "login@example.com", "password": "loginpass1"},
+        )
+
+    def test_login_success(self, client):
+        resp = client.post(
+            "/api/v1/auth/login",
+            json={"email": "login@example.com", "password": "loginpass1"},
+        )
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert "access_token" in data
+        assert "refresh_token" in data
+        assert data["plan"] == "free"
+
+    def test_login_wrong_password(self, client):
+        resp = client.post(
+            "/api/v1/auth/login",
+            json={"email": "login@example.com", "password": "wrongpass"},
+        )
+        assert resp.status_code == 401
+
+    def test_login_unknown_email(self, client):
+        resp = client.post(
+            "/api/v1/auth/login",
+            json={"email": "ghost@example.com", "password": "anypass"},
+        )
+        assert resp.status_code == 401
+
+    def test_login_missing_fields(self, client):
+        resp = client.post("/api/v1/auth/login", json={"email": "login@example.com"})
+        assert resp.status_code == 400
+
+
+# ──────────────────────────────────────────────────────────────
+# Token refresh
+# ──────────────────────────────────────────────────────────────
+
+
+class TestRefresh:
+    @pytest.fixture(autouse=True)
+    def setup(self, client):
+        client.post(
+            "/api/v1/auth/register",
+            json={"email": "refresh@example.com", "password": "refreshpass1"},
+        )
+        resp = client.post(
+            "/api/v1/auth/login",
+            json={"email": "refresh@example.com", "password": "refreshpass1"},
+        )
+        tokens = resp.get_json()
+        self.refresh_token = tokens["refresh_token"]
+
+    def test_refresh_success(self, client):
+        resp = client.post(
+            "/api/v1/auth/refresh",
+            json={"refresh_token": self.refresh_token},
+        )
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert "access_token" in data
+        assert "refresh_token" in data
+        # New refresh token should differ from old
+        assert data["refresh_token"] != self.refresh_token
+
+    def test_refresh_token_rotation(self, client):
+        """Old refresh token must be invalid after rotation."""
+        client.post(
+            "/api/v1/auth/refresh",
+            json={"refresh_token": self.refresh_token},
+        )
+        resp2 = client.post(
+            "/api/v1/auth/refresh",
+            json={"refresh_token": self.refresh_token},
+        )
+        assert resp2.status_code == 401
+
+    def test_refresh_invalid_token(self, client):
+        resp = client.post(
+            "/api/v1/auth/refresh",
+            json={"refresh_token": "completely.invalid.token"},
+        )
+        assert resp.status_code == 401
+
+    def test_refresh_missing_token(self, client):
+        resp = client.post("/api/v1/auth/refresh", json={})
+        assert resp.status_code == 400
+
+
+# ──────────────────────────────────────────────────────────────
+# Logout
+# ──────────────────────────────────────────────────────────────
+
+
+class TestLogout:
+    @pytest.fixture(autouse=True)
+    def setup(self, client):
+        client.post(
+            "/api/v1/auth/register",
+            json={"email": "logout@example.com", "password": "logoutpass1"},
+        )
+        resp = client.post(
+            "/api/v1/auth/login",
+            json={"email": "logout@example.com", "password": "logoutpass1"},
+        )
+        tokens = resp.get_json()
+        self.refresh_token = tokens["refresh_token"]
+        self.access_token = tokens["access_token"]
+
+    def test_logout_success(self, client):
+        resp = client.post(
+            "/api/v1/auth/logout",
+            json={"refresh_token": self.refresh_token},
+        )
+        assert resp.status_code == 200
+
+    def test_refresh_after_logout_fails(self, client):
+        client.post("/api/v1/auth/logout", json={"refresh_token": self.refresh_token})
+        resp = client.post(
+            "/api/v1/auth/refresh",
+            json={"refresh_token": self.refresh_token},
+        )
+        assert resp.status_code == 401
+
+    def test_logout_no_token(self, client):
+        resp = client.post("/api/v1/auth/logout", json={})
+        assert resp.status_code == 200
+
+
+# ──────────────────────────────────────────────────────────────
+# JWT middleware — protected routes
+# ──────────────────────────────────────────────────────────────
+
+
+class TestJWTMiddleware:
+    @pytest.fixture(autouse=True)
+    def setup(self, client):
+        client.post(
+            "/api/v1/auth/register",
+            json={"email": "protected@example.com", "password": "protect123"},
+        )
+        resp = client.post(
+            "/api/v1/auth/login",
+            json={"email": "protected@example.com", "password": "protect123"},
+        )
+        self.access_token = resp.get_json()["access_token"]
+
+    def test_subscription_info_requires_auth(self, client):
+        resp = client.get("/api/v1/subscription/upgrade")
+        assert resp.status_code == 401
+
+    def test_subscription_info_with_token(self, client):
+        resp = client.get(
+            "/api/v1/subscription/upgrade",
+            headers={"Authorization": f"Bearer {self.access_token}"},
+        )
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert "current_plan" in data
+        assert data["current_plan"] == "free"
+
+    def test_invalid_token_rejected(self, client):
+        resp = client.get(
+            "/api/v1/subscription/upgrade",
+            headers={"Authorization": "Bearer invalid.token.here"},
+        )
+        assert resp.status_code in (401, 422)
+
+
+# ──────────────────────────────────────────────────────────────
+# Plan checks
+# ──────────────────────────────────────────────────────────────
+
+
+class TestPlanMiddleware:
+    @pytest.fixture(autouse=True)
+    def setup(self, client, app):
+        # Register free user
+        client.post(
+            "/api/v1/auth/register",
+            json={"email": "free_plan@example.com", "password": "freepass1"},
+        )
+        resp = client.post(
+            "/api/v1/auth/login",
+            json={"email": "free_plan@example.com", "password": "freepass1"},
+        )
+        self.free_token = resp.get_json()["access_token"]
+
+        # Upgrade user to pro directly in DB for testing
+        import sqlite3
+
+        db_path = os.environ["TURF_SAAS_DB"]
+        conn = sqlite3.connect(db_path)
+        conn.execute(
+            "INSERT OR IGNORE INTO users (email, password_hash, plan) VALUES (?,?,?)",
+            ("pro_plan@example.com", "$2b$12$placeholder", "pro"),
+        )
+        conn.commit()
+        conn.close()
+
+        # Login pro user using JWT created manually via app context
+        with app.app_context():
+            from flask_jwt_extended import create_access_token
+
+            conn = sqlite3.connect(db_path)
+            row = conn.execute(
+                "SELECT id FROM users WHERE email='pro_plan@example.com'"
+            ).fetchone()
+            conn.close()
+            self.pro_token = create_access_token(
+                identity=str(row[0]),
+                additional_claims={"plan": "pro", "email": "pro_plan@example.com"},
+            )
+
+    def test_export_blocked_for_free(self, client):
+        resp = client.get(
+            "/api/v1/predictions/export",
+            headers={"Authorization": f"Bearer {self.free_token}"},
+        )
+        assert resp.status_code == 403
+        data = resp.get_json()
+        assert "Plan insuffisant" in data["error"]
+
+    def test_export_allowed_for_pro(self, client):
+        resp = client.get(
+            "/api/v1/predictions/export",
+            headers={"Authorization": f"Bearer {self.pro_token}"},
+        )
+        # 503 is expected because no backend is running; 403 would be wrong
+        assert resp.status_code != 403
+
+    def test_upgrade_info_shows_plans(self, client):
+        resp = client.get(
+            "/api/v1/subscription/upgrade",
+            headers={"Authorization": f"Bearer {self.free_token}"},
+        )
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert "free" in data["plans"]
+        assert "premium" in data["plans"]
+        assert "pro" in data["plans"]
+
+
+# ──────────────────────────────────────────────────────────────
+# Rate limiting
+# ──────────────────────────────────────────────────────────────
+
+
+class TestRateLimiting:
+    def test_rate_limit_headers_present(self, client):
+        resp = client.get("/api/v1/health")
+        assert "X-RateLimit-Limit" in resp.headers
+        assert resp.headers["X-RateLimit-Limit"] == "100"
+
+    def test_rate_limit_remaining_decreases(self, client):
+        r1 = client.get("/api/v1/health")
+        r2 = client.get("/api/v1/health")
+        rem1 = int(r1.headers.get("X-RateLimit-Remaining", 100))
+        rem2 = int(r2.headers.get("X-RateLimit-Remaining", 100))
+        assert rem2 <= rem1
+
+
+# ──────────────────────────────────────────────────────────────
+# DB module
+# ──────────────────────────────────────────────────────────────
+
+
+class TestAuthDB:
+    def test_tables_exist(self):
+        import sqlite3
+
+        conn = sqlite3.connect(os.environ["TURF_SAAS_DB"])
+        tables = {
+            r[0]
+            for r in conn.execute(
+                "SELECT name FROM sqlite_master WHERE type='table'"
+            ).fetchall()
+        }
+        assert "users" in tables
+        assert "subscriptions" in tables
+        assert "refresh_tokens" in tables
+        conn.close()
+
+    def test_get_db_returns_connection(self):
+        from auth_db import get_db
+
+        db = get_db()
+        assert db is not None
+        db.close()

tests/test_ml_cache_integrity.py

@@ -0,0 +1,300 @@
+"""
+test_ml_cache_integrity.py — Test d'intégration : zéro NULL dans ml_predictions_cache
+SaaS Turf Prédictions IA
+Ticket: HRT-43 (suite au fix HRT-41 — métadonnées manquantes dans le cache ML)
+
+Ces tests vérifient que la table ml_predictions_cache ne contient aucune ligne
+avec des métadonnées NULL (hippodrome, race_label, heure) pour la date courante,
+après le job ML de 19h30.
+
+Usage:
+    pytest tests/test_ml_cache_integrity.py -v -m integration
+    pytest tests/test_ml_cache_integrity.py -v -m integration --date 2026-04-26
+
+Variables d'environnement:
+    TURF_DB_PATH  : chemin vers turf.db (défaut: /home/h3r7/turf_scraper/turf.db)
+    TEST_DATE     : date cible au format YYYY-MM-DD (défaut: date du jour)
+"""
+
+import sqlite3
+import os
+import pytest
+from datetime import date, datetime
+from pathlib import Path
+
+
+# ============================================================
+# Configuration
+# ============================================================
+
+DEFAULT_DB_PATH = "/home/h3r7/turf_scraper/turf.db"
+DB_PATH = os.environ.get("TURF_DB_PATH", DEFAULT_DB_PATH)
+
+
+def _get_test_date() -> str:
+    """Retourne la date cible pour les tests (env TEST_DATE ou date du jour)."""
+    env_date = os.environ.get("TEST_DATE", "")
+    if env_date:
+        try:
+            datetime.strptime(env_date, "%Y-%m-%d")
+            return env_date
+        except ValueError:
+            raise ValueError(
+                f"TEST_DATE invalide : '{env_date}'. Format attendu : YYYY-MM-DD"
+            )
+    return date.today().isoformat()
+
+
+# ============================================================
+# Fixture : connexion DB en lecture seule
+# ============================================================
+
+
+@pytest.fixture(scope="module")
+def db_connection():
+    """
+    Connexion SQLite en mode lecture seule (uri=True + ?mode=ro).
+    Garantit qu'aucune modification accidentelle de la DB de prod n'est possible.
+    """
+    db_path = Path(DB_PATH)
+    if not db_path.exists():
+        pytest.skip(
+            f"Base de données introuvable : {DB_PATH}. "
+            "Définir TURF_DB_PATH ou vérifier le chemin."
+        )
+
+    uri = f"file:{db_path.as_posix()}?mode=ro"
+    conn = sqlite3.connect(uri, uri=True)
+    conn.row_factory = sqlite3.Row
+    yield conn
+    conn.close()
+
+
+@pytest.fixture(scope="module")
+def target_date():
+    """Date cible pour les tests (date du jour ou TEST_DATE)."""
+    return _get_test_date()
+
+
+# ============================================================
+# Tests d'intégration
+# ============================================================
+
+
+@pytest.mark.integration
+class TestMlCacheNullIntegrity:
+    """
+    Vérifie qu'après le job ML de 19h30, la table ml_predictions_cache
+    ne contient aucune métadonnée NULL pour la date courante.
+
+    Régression testée : HRT-41 (Fix #17 — métadonnées manquantes dans le cache ML)
+    """
+
+    def test_table_exists(self, db_connection):
+        """Vérifie que la table ml_predictions_cache existe dans la DB."""
+        cursor = db_connection.execute(
+            "SELECT name FROM sqlite_master "
+            "WHERE type='table' AND name='ml_predictions_cache'"
+        )
+        row = cursor.fetchone()
+        assert row is not None, (
+            "La table ml_predictions_cache est introuvable dans la base de données. "
+            "Vérifier que le job ML a bien créé la table."
+        )
+
+    def test_rows_exist_for_today(self, db_connection, target_date):
+        """
+        Vérifie que des prédictions existent pour la date cible.
+
+        Ce test passe en skip si aucune ligne n'existe (ex: avant le job 19h30).
+        Il échoue uniquement si le job a manifestement tourné mais a laissé 0 lignes.
+        """
+        cursor = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache WHERE date = ?",
+            (target_date,),
+        )
+        count = cursor.fetchone()["cnt"]
+
+        if count == 0:
+            pytest.skip(
+                f"Aucune prédiction en cache pour le {target_date}. "
+                "Ce test doit être exécuté après le job ML de 19h30."
+            )
+
+    def test_zero_null_hippodrome_today(self, db_connection, target_date):
+        """
+        CRITÈRE D'ACCEPTATION PRINCIPAL (HRT-43) :
+        Vérifie que COUNT(*) WHERE date = today AND hippodrome IS NULL = 0.
+
+        Régression directe du bug HRT-41 : le champ hippodrome était NULL
+        pour toutes les prédictions du cache ML.
+        """
+        # Vérifier si des données existent avant de tester les NULLs
+        cursor_total = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache WHERE date = ?",
+            (target_date,),
+        )
+        total = cursor_total.fetchone()["cnt"]
+        if total == 0:
+            pytest.skip(
+                f"Aucune prédiction en cache pour le {target_date}. "
+                "Lancer ce test après le job ML de 19h30."
+            )
+
+        cursor = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache "
+            "WHERE date = ? AND hippodrome IS NULL",
+            (target_date,),
+        )
+        null_count = cursor.fetchone()["cnt"]
+
+        assert null_count == 0, (
+            f"RÉGRESSION HRT-41 DÉTECTÉE : {null_count} ligne(s) avec hippodrome IS NULL "
+            f"dans ml_predictions_cache pour le {target_date}. "
+            "Le patch de métadonnées n'a pas été appliqué correctement."
+        )
+
+    def test_zero_null_race_label_today(self, db_connection, target_date):
+        """
+        Vérifie que COUNT(*) WHERE date = today AND race_label IS NULL = 0.
+
+        Complément du test hippodrome : vérifie que le libellé de course
+        est bien renseigné pour toutes les prédictions.
+        """
+        cursor_total = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache WHERE date = ?",
+            (target_date,),
+        )
+        total = cursor_total.fetchone()["cnt"]
+        if total == 0:
+            pytest.skip(
+                f"Aucune prédiction en cache pour le {target_date}. "
+                "Lancer ce test après le job ML de 19h30."
+            )
+
+        cursor = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache "
+            "WHERE date = ? AND race_label IS NULL",
+            (target_date,),
+        )
+        null_count = cursor.fetchone()["cnt"]
+
+        assert null_count == 0, (
+            f"ANOMALIE : {null_count} ligne(s) avec race_label IS NULL "
+            f"dans ml_predictions_cache pour le {target_date}. "
+            "Vérifier le pipeline de patch de métadonnées."
+        )
+
+    def test_zero_null_heure_today(self, db_connection, target_date):
+        """
+        Vérifie que COUNT(*) WHERE date = today AND heure IS NULL = 0.
+
+        Vérifie que l'heure de course est bien renseignée pour toutes les prédictions.
+        """
+        cursor_total = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache WHERE date = ?",
+            (target_date,),
+        )
+        total = cursor_total.fetchone()["cnt"]
+        if total == 0:
+            pytest.skip(
+                f"Aucune prédiction en cache pour le {target_date}. "
+                "Lancer ce test après le job ML de 19h30."
+            )
+
+        cursor = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache "
+            "WHERE date = ? AND heure IS NULL",
+            (target_date,),
+        )
+        null_count = cursor.fetchone()["cnt"]
+
+        assert null_count == 0, (
+            f"ANOMALIE : {null_count} ligne(s) avec heure IS NULL "
+            f"dans ml_predictions_cache pour le {target_date}. "
+            "Vérifier le pipeline de patch de métadonnées."
+        )
+
+    def test_full_metadata_coverage_today(self, db_connection, target_date):
+        """
+        Test de couverture globale : aucune des trois colonnes critiques
+        (hippodrome, race_label, heure) n'est NULL pour une même ligne.
+
+        Retourne les 5 premières lignes problématiques pour faciliter le débogage.
+        """
+        cursor_total = db_connection.execute(
+            "SELECT COUNT(*) as cnt FROM ml_predictions_cache WHERE date = ?",
+            (target_date,),
+        )
+        total = cursor_total.fetchone()["cnt"]
+        if total == 0:
+            pytest.skip(
+                f"Aucune prédiction en cache pour le {target_date}. "
+                "Lancer ce test après le job ML de 19h30."
+            )
+
+        cursor = db_connection.execute(
+            "SELECT id, num_reunion, num_course, horse_name, hippodrome, race_label, heure "
+            "FROM ml_predictions_cache "
+            "WHERE date = ? "
+            "AND (hippodrome IS NULL OR race_label IS NULL OR heure IS NULL) "
+            "LIMIT 5",
+            (target_date,),
+        )
+        bad_rows = cursor.fetchall()
+
+        assert len(bad_rows) == 0, (
+            f"ANOMALIE : {len(bad_rows)} ligne(s) avec au moins une métadonnée NULL "
+            f"(hippodrome, race_label ou heure) pour le {target_date}.\n"
+            "Exemples de lignes affectées :\n"
+            + "\n".join(
+                f"  - id={r['id']} R{r['num_reunion']}C{r['num_course']} "
+                f"{r['horse_name']} | hippodrome={r['hippodrome']!r} "
+                f"race_label={r['race_label']!r} heure={r['heure']!r}"
+                for r in bad_rows
+            )
+        )
+
+    def test_metadata_completeness_summary(self, db_connection, target_date):
+        """
+        Résumé diagnostique : affiche les statistiques de complétude des métadonnées
+        pour la date cible. Toujours en mode informatif (pas de assertion stricte),
+        utile pour le monitoring et les logs CI.
+        """
+        cursor = db_connection.execute(
+            """
+            SELECT
+                COUNT(*) as total,
+                SUM(CASE WHEN hippodrome IS NULL THEN 1 ELSE 0 END) as null_hippodrome,
+                SUM(CASE WHEN race_label IS NULL THEN 1 ELSE 0 END) as null_race_label,
+                SUM(CASE WHEN heure IS NULL THEN 1 ELSE 0 END) as null_heure,
+                COUNT(DISTINCT hippodrome) as distinct_hippodromes,
+                COUNT(DISTINCT race_label) as distinct_race_labels
+            FROM ml_predictions_cache
+            WHERE date = ?
+            """,
+            (target_date,),
+        )
+        row = cursor.fetchone()
+        total = row["total"]
+
+        if total == 0:
+            pytest.skip(
+                f"Aucune prédiction en cache pour le {target_date}. "
+                "Lancer ce test après le job ML de 19h30."
+            )
+
+        # Afficher les statistiques (visibles avec pytest -v -s)
+        print(f"\n=== Statistiques ml_predictions_cache pour le {target_date} ===")
+        print(f"  Total lignes         : {total}")
+        print(f"  NULL hippodrome      : {row['null_hippodrome']}")
+        print(f"  NULL race_label      : {row['null_race_label']}")
+        print(f"  NULL heure           : {row['null_heure']}")
+        print(f"  Hippodromes distincts: {row['distinct_hippodromes']}")
+        print(f"  Race labels distincts: {row['distinct_race_labels']}")
+
+        # L'assertion ici reste stricte pour hippodrome (bug HRT-41 critique)
+        assert row["null_hippodrome"] == 0, (
+            f"RÉGRESSION HRT-41 : {row['null_hippodrome']}/{total} lignes "
+            f"avec hippodrome IS NULL pour le {target_date}."
+        )

tests/test_ml_ensemble.py

@@ -0,0 +1,333 @@
+"""
+Tests ML Ensemble — HRT-32 Sprint 6-7
+Tests de régression, benchmark et latence pour le nouveau modèle ensemble.
+
+Usage:
+    pytest tests/test_ml_ensemble.py -v
+    pytest tests/test_ml_ensemble.py -v -m regression
+    pytest tests/test_ml_ensemble.py -v -m latency
+"""
+
+import json
+import os
+import pickle
+import sqlite3
+import time
+from pathlib import Path
+
+import numpy as np
+import pandas as pd
+import pytest
+import requests
+
+BASE_URL = os.environ.get("APP_URL", "http://localhost:8790")
+DB_PATH = os.environ.get("DB_PATH", "/home/h3r7/turf_saas/turf.db")
+MODELS_DIR = Path("/home/h3r7/turf_saas/models")
+ENSEMBLE_PATH = MODELS_DIR / "ensemble_top3.pkl"
+BENCHMARK_PATH = MODELS_DIR / "benchmark_report.json"
+
+
+# ─── Fixtures ────────────────────────────────────────────────────────────────
+
+
+@pytest.fixture(scope="session")
+def ensemble_model():
+    """Load ensemble model (skip tests if not yet trained)."""
+    if not ENSEMBLE_PATH.exists():
+        pytest.skip(
+            f"Ensemble model not found at {ENSEMBLE_PATH}. Run train_ensemble.py first."
+        )
+    with open(ENSEMBLE_PATH, "rb") as f:
+        return pickle.load(f)
+
+
+@pytest.fixture(scope="session")
+def benchmark_report():
+    """Load benchmark report (skip if not generated)."""
+    if not BENCHMARK_PATH.exists():
+        pytest.skip(f"Benchmark report not found at {BENCHMARK_PATH}.")
+    with open(BENCHMARK_PATH) as f:
+        return json.load(f)
+
+
+@pytest.fixture(scope="session")
+def holdout_data():
+    """Load holdout slice (last 20% temporal) for regression tests."""
+    conn = sqlite3.connect(DB_PATH)
+    df = pd.read_sql_query(
+        """
+        SELECT p.*, c.distance, c.discipline, c.specialite,
+               c.nb_declares_partants, c.montant_prix, c.penetrometre_intitule
+        FROM pmu_partants p
+        LEFT JOIN pmu_courses c ON p.date_programme=c.date_programme
+            AND p.num_reunion=c.num_reunion AND p.num_course=c.num_course
+        WHERE p.ordre_arrivee > 0
+        ORDER BY p.date_programme, p.num_reunion, p.num_course, p.num_pmu
+        """,
+        conn,
+    )
+    conn.close()
+    n = len(df)
+    cutoff = int(n * 0.80)
+    return df.iloc[cutoff:].copy()
+
+
+@pytest.fixture(scope="session")
+def predict_v2():
+    """Import predict_v2 module."""
+    import importlib.util
+
+    spec = importlib.util.spec_from_file_location(
+        "predict_v2", "/home/h3r7/turf_saas/predict_v2.py"
+    )
+    mod = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(mod)
+    return mod
+
+
+# ─── Model Existence Tests ────────────────────────────────────────────────────
+
+
+class TestModelFiles:
+    """Verify all expected model files exist."""
+
+    def test_ensemble_model_exists(self):
+        assert ENSEMBLE_PATH.exists(), f"Ensemble model missing: {ENSEMBLE_PATH}"
+
+    def test_benchmark_report_exists(self):
+        assert BENCHMARK_PATH.exists(), f"Benchmark report missing: {BENCHMARK_PATH}"
+
+    def test_models_dir_contains_expected_files(self):
+        expected = ["ensemble_top3.pkl", "benchmark_report.json", "benchmark_report.md"]
+        for fname in expected:
+            assert (MODELS_DIR / fname).exists(), f"Missing: {MODELS_DIR / fname}"
+
+
+# ─── Benchmark Tests ──────────────────────────────────────────────────────────
+
+
+class TestBenchmark:
+    """Validate benchmark metrics from the training report."""
+
+    @pytest.mark.regression
+    def test_ensemble_beats_baseline_or_meets_threshold(self, benchmark_report):
+        """Ensemble Precision@3 must be >= baseline XGBoost."""
+        baseline = benchmark_report["baseline"]["precision_at3"]
+        ensemble = benchmark_report["ensemble"]["precision_at3"]
+        assert ensemble >= baseline, (
+            f"Ensemble Precision@3 {ensemble:.4f} < baseline {baseline:.4f}"
+        )
+
+    @pytest.mark.regression
+    def test_ensemble_auc_above_random(self, benchmark_report):
+        """Ensemble AUC must be > 0.60 (significantly above random 0.50)."""
+        auc = benchmark_report["ensemble"]["auc"]
+        assert auc > 0.60, f"Ensemble AUC {auc:.4f} <= 0.60"
+
+    @pytest.mark.regression
+    def test_optuna_ran_minimum_trials(self, benchmark_report):
+        """Optuna must have run at least 100 trials per model."""
+        n_trials = benchmark_report["optuna"]["n_trials"]
+        assert n_trials >= 100, f"Only {n_trials} Optuna trials (minimum 100 required)"
+
+    @pytest.mark.regression
+    def test_no_precision_regression(self, benchmark_report):
+        """Ensemble Precision@3 must not be below naive random baseline (~30%)."""
+        ensemble_p3 = benchmark_report["ensemble"]["precision_at3"]
+        assert ensemble_p3 >= 0.30, (
+            f"Precision@3 {ensemble_p3:.4f} is below random baseline (~0.30)"
+        )
+
+    def test_benchmark_has_all_required_models(self, benchmark_report):
+        """Benchmark must include results for all 3 models."""
+        required = {"xgboost", "lightgbm", "mlp"}
+        found = set(benchmark_report.get("individual_models", {}).keys())
+        missing = required - found
+        assert not missing, f"Missing model benchmarks: {missing}"
+
+
+# ─── Regression Tests ─────────────────────────────────────────────────────────
+
+
+class TestPrecisionRegression:
+    """Holdout regression: ensure precision doesn't degrade."""
+
+    @pytest.mark.regression
+    def test_precision_at3_on_holdout(self, ensemble_model, holdout_data):
+        """Precision@3 on holdout must be above naive baseline."""
+        from predict_v2 import build_feature_df, FEATURE_COLS
+
+        df = holdout_data.copy()
+        df["top3"] = (df["ordre_arrivee"] <= 3).astype(int)
+
+        partants = df.to_dict("records")
+        feature_df = build_feature_df(partants)
+        available = [c for c in FEATURE_COLS if c in feature_df.columns]
+        X = feature_df[available].fillna(0)
+
+        proba = ensemble_model.predict_proba(X)[:, 1]
+
+        # Per-race Precision@3
+        tmp = df[["date_programme", "num_reunion", "num_course"]].copy()
+        tmp["proba"] = proba
+        tmp["actual"] = df["top3"].values
+
+        precisions = []
+        for _, group in tmp.groupby(["date_programme", "num_reunion", "num_course"]):
+            if len(group) >= 3:
+                top3_pred = group.nlargest(3, "proba")
+                precisions.append(top3_pred["actual"].sum() / 3.0)
+
+        p_at3 = float(np.mean(precisions)) if precisions else 0.0
+        print(f"\n  Holdout Precision@3: {p_at3:.4f} over {len(precisions)} races")
+
+        # Must beat random baseline (30%)
+        assert p_at3 >= 0.30, f"Holdout Precision@3 {p_at3:.4f} < 0.30"
+
+    @pytest.mark.regression
+    def test_no_all_zero_predictions(self, ensemble_model, holdout_data):
+        """Ensemble must not predict 0 probability for all horses."""
+        from predict_v2 import build_feature_df, FEATURE_COLS
+
+        partants = holdout_data.head(50).to_dict("records")
+        feature_df = build_feature_df(partants)
+        available = [c for c in FEATURE_COLS if c in feature_df.columns]
+        X = feature_df[available].fillna(0)
+
+        proba = ensemble_model.predict_proba(X)[:, 1]
+        assert proba.max() > 0.01, "All predictions are near 0 — model appears broken"
+        assert proba.std() > 0.01, (
+            "All predictions have identical probability — no discrimination"
+        )
+
+
+# ─── Latency Tests ────────────────────────────────────────────────────────────
+
+
+class TestPredictionLatency:
+    """Prediction latency must be < 200ms per race."""
+
+    @pytest.mark.latency
+    def test_single_race_latency(self, ensemble_model, holdout_data):
+        """Prediction for a single race (<=20 horses) must be < 200ms."""
+        from predict_v2 import build_feature_df, FEATURE_COLS
+
+        # Take one race
+        first_race = (
+            holdout_data.groupby(["date_programme", "num_reunion", "num_course"])
+            .first()
+            .reset_index()
+            .iloc[0]
+        )
+        mask = (
+            (holdout_data["date_programme"] == first_race["date_programme"])
+            & (holdout_data["num_reunion"] == first_race["num_reunion"])
+            & (holdout_data["num_course"] == first_race["num_course"])
+        )
+        race_df = holdout_data[mask]
+        partants = race_df.to_dict("records")
+
+        # Warm-up
+        feature_df = build_feature_df(partants)
+        available = [c for c in FEATURE_COLS if c in feature_df.columns]
+        X = feature_df[available].fillna(0)
+        ensemble_model.predict_proba(X)
+
+        # Timed run
+        t0 = time.perf_counter()
+        for _ in range(10):
+            ensemble_model.predict_proba(X)
+        elapsed_ms = (time.perf_counter() - t0) / 10 * 1000
+
+        print(f"\n  Single-race latency: {elapsed_ms:.2f} ms ({len(partants)} horses)")
+        assert elapsed_ms < 200, (
+            f"Prediction latency {elapsed_ms:.1f} ms exceeds 200 ms limit"
+        )
+
+    @pytest.mark.latency
+    def test_full_day_latency(self, ensemble_model, holdout_data):
+        """Prediction for a full day (all races) must complete < 5 seconds."""
+        from predict_v2 import build_feature_df, FEATURE_COLS
+
+        # Take one day
+        day = holdout_data["date_programme"].iloc[0]
+        day_df = holdout_data[holdout_data["date_programme"] == day]
+        partants = day_df.to_dict("records")
+
+        feature_df = build_feature_df(partants)
+        available = [c for c in FEATURE_COLS if c in feature_df.columns]
+        X = feature_df[available].fillna(0)
+
+        t0 = time.perf_counter()
+        proba = ensemble_model.predict_proba(X)
+        elapsed_ms = (time.perf_counter() - t0) * 1000
+
+        print(
+            f"\n  Full day latency: {elapsed_ms:.2f} ms ({len(partants)} horses, {day})"
+        )
+        assert elapsed_ms < 5000, (
+            f"Full-day prediction {elapsed_ms:.0f} ms exceeds 5s limit"
+        )
+
+
+# ─── API Endpoint Tests ───────────────────────────────────────────────────────
+
+
+class TestV1PredictionsAPI:
+    """Tests for the new /api/v1/predictions endpoint."""
+
+    def _api_available(self):
+        try:
+            requests.get(f"{BASE_URL}/api/v1/model/status", timeout=3)
+            return True
+        except Exception:
+            return False
+
+    @pytest.mark.api
+    def test_model_status_endpoint(self):
+        """GET /api/v1/model/status returns valid JSON."""
+        if not self._api_available():
+            pytest.skip("API server not running")
+        resp = requests.get(f"{BASE_URL}/api/v1/model/status", timeout=10)
+        assert resp.status_code == 200
+        data = resp.json()
+        assert "ensemble_available" in data
+
+    @pytest.mark.api
+    def test_v1_predictions_no_500(self):
+        """GET /api/v1/predictions must not return 5xx."""
+        if not self._api_available():
+            pytest.skip("API server not running")
+        resp = requests.get(f"{BASE_URL}/api/v1/predictions", timeout=30)
+        assert resp.status_code < 500, (
+            f"Server error: {resp.status_code}\n{resp.text[:200]}"
+        )
+
+    @pytest.mark.api
+    def test_v1_predictions_returns_json(self):
+        """GET /api/v1/predictions returns valid JSON with expected keys."""
+        if not self._api_available():
+            pytest.skip("API server not running")
+        resp = requests.get(f"{BASE_URL}/api/v1/predictions", timeout=30)
+        if resp.status_code == 503:
+            pytest.skip("Ensemble model not yet deployed")
+        assert resp.status_code == 200
+        data = resp.json()
+        assert "model_version" in data, "Missing model_version in response"
+        assert "races" in data or "predictions" in data, (
+            "Missing races/predictions in response"
+        )
+
+    @pytest.mark.api
+    def test_v1_predictions_latency(self):
+        """GET /api/v1/predictions must respond in < 3 seconds."""
+        if not self._api_available():
+            pytest.skip("API server not running")
+        resp = requests.get(f"{BASE_URL}/api/v1/predictions", timeout=30)
+        if resp.status_code == 503:
+            pytest.skip("Ensemble model not yet deployed")
+        # Check API-reported latency
+        if resp.status_code == 200:
+            data = resp.json()
+            latency = data.get("latency_ms", 0)
+            assert latency < 3000, f"API latency {latency:.0f} ms > 3000 ms"

tests/test_smoke.py

@@ -0,0 +1,205 @@
+"""
+Tests de smoke — SaaS Turf Prédictions IA
+Sprint 8 — QA, Beta Fermee, Go/No-Go
+Ticket: HRT-34
+
+Vérifications rapides sur l'état de l'application :
+- Routes de base accessibles
+- API répond en JSON valide
+- Base de données accessible
+- Pas d'erreurs 5xx sur les routes principales
+
+Ces tests peuvent tourner SANS infra complète (pas besoin de HRT-31/33).
+Exécuter sur l'app actuelle en staging ou localhost.
+"""
+
+import pytest
+import requests
+import os
+import json
+
+BASE_URL = os.environ.get("APP_URL", "http://localhost:8792")
+
+# Routes qui doivent retourner 200 (publiques)
+PUBLIC_ROUTES_200 = [
+    "/",
+    "/dashboard",
+]
+
+# Routes API qui doivent retourner 200 ou 401 (jamais 500)
+API_ROUTES_NO_500 = [
+    "/api",
+    "/api/races",
+    "/api/scoring",
+    "/api/weather",
+    "/api/odds_history",
+]
+
+
+class TestSmoke:
+    """Tests de smoke : l'app répond correctement aux requêtes de base."""
+
+    @pytest.mark.smoke
+    @pytest.mark.parametrize("route", PUBLIC_ROUTES_200)
+    def test_route_publique_accessible(self, route):
+        """Les routes publiques doivent retourner 200."""
+        try:
+            resp = requests.get(f"{BASE_URL}{route}", timeout=10)
+            assert resp.status_code in (200, 304), (
+                f"Route publique inaccessible: {route} → {resp.status_code}"
+            )
+            assert len(resp.content) > 0, f"Réponse vide sur {route}"
+        except requests.exceptions.ConnectionError:
+            pytest.skip(
+                f"App non accessible sur {BASE_URL} — vérifier que le serveur est démarré"
+            )
+
+    @pytest.mark.smoke
+    @pytest.mark.parametrize("route", API_ROUTES_NO_500)
+    def test_api_pas_derreur_serveur(self, route):
+        """Les routes API ne doivent jamais retourner 5xx."""
+        try:
+            resp = requests.get(f"{BASE_URL}{route}", timeout=10)
+            assert resp.status_code < 500, (
+                f"Erreur serveur sur {route}: {resp.status_code}\n{resp.text[:200]}"
+            )
+        except requests.exceptions.ConnectionError:
+            pytest.skip(f"App non accessible sur {BASE_URL}")
+
+    @pytest.mark.smoke
+    def test_api_today_retourne_json(self):
+        """L'endpoint principal /api doit retourner du JSON valide."""
+        try:
+            resp = requests.get(f"{BASE_URL}/api", timeout=10)
+            if resp.status_code == 200:
+                data = resp.json()
+                assert data is not None, "Réponse JSON nulle"
+                assert isinstance(data, (list, dict)), (
+                    f"Type de réponse inattendu: {type(data)}"
+                )
+        except requests.exceptions.ConnectionError:
+            pytest.skip(f"App non accessible sur {BASE_URL}")
+        except json.JSONDecodeError as e:
+            pytest.fail(f"/api ne retourne pas du JSON valide: {e}")
+
+    @pytest.mark.smoke
+    def test_contenu_html_portail_valide(self):
+        """Le portail doit contenir un titre et du contenu significatif."""
+        try:
+            resp = requests.get(f"{BASE_URL}/", timeout=10)
+            if resp.status_code == 200:
+                content = resp.text
+                assert "<html" in content.lower() or "<!doctype" in content.lower(), (
+                    "La page d'accueil ne retourne pas du HTML"
+                )
+                assert len(content) > 500, (
+                    f"Page d'accueil trop courte ({len(content)} chars)"
+                )
+        except requests.exceptions.ConnectionError:
+            pytest.skip(f"App non accessible sur {BASE_URL}")
+
+    @pytest.mark.smoke
+    def test_headers_securite_presents(self):
+        """Les headers de sécurité de base doivent être présents."""
+        try:
+            resp = requests.get(f"{BASE_URL}/", timeout=10)
+            if resp.status_code != 200:
+                return
+
+            # En production (derrière Nginx), ces headers doivent être présents
+            # En dev direct Flask, ils peuvent être absents — on note seulement
+            security_headers = {
+                "X-Content-Type-Options": "nosniff",
+                "X-Frame-Options": None,  # SAMEORIGIN ou DENY
+                "X-XSS-Protection": None,
+            }
+
+            missing = []
+            for header, expected_value in security_headers.items():
+                if header not in resp.headers:
+                    missing.append(header)
+
+            if missing:
+                # Warning seulement — bloquant uniquement en prod derrière Nginx
+                pytest.warns(UserWarning, match=r".*") if False else None
+                print(f"⚠️  Headers sécurité manquants (requis en prod): {missing}")
+
+        except requests.exceptions.ConnectionError:
+            pytest.skip(f"App non accessible sur {BASE_URL}")
+
+    @pytest.mark.smoke
+    def test_api_races_format_reponse(self):
+        """L'endpoint /api/races doit retourner une liste structurée."""
+        try:
+            resp = requests.get(f"{BASE_URL}/api/races", timeout=10)
+            if resp.status_code == 200:
+                data = resp.json()
+                assert isinstance(data, (list, dict)), (
+                    f"Format inattendu pour /api/races: {type(data)}"
+                )
+                if isinstance(data, list) and len(data) > 0:
+                    first = data[0]
+                    # Vérifier la présence de champs clés
+                    expected_fields = ["date", "course", "hippodrome"]
+                    present = [
+                        f
+                        for f in expected_fields
+                        if f in first
+                        or any(k in first for k in [f, f.upper(), f.replace("_", "")])
+                    ]
+                    assert len(present) > 0, (
+                        f"Champs attendus absents de /api/races. Champs présents: {list(first.keys())}"
+                    )
+        except requests.exceptions.ConnectionError:
+            pytest.skip(f"App non accessible sur {BASE_URL}")
+        except json.JSONDecodeError:
+            pytest.fail("/api/races ne retourne pas du JSON valide")
+
+
+class TestSmokeDatabase:
+    """Tests smoke sur la base de données."""
+
+    @pytest.mark.smoke
+    def test_base_donnees_accessible(self):
+        """La base de données SQLite doit être accessible et contenir des données."""
+        import sqlite3
+
+        db_path = "/home/h3r7/turf_saas/turf_saas.db"
+
+        if not __import__("os").path.exists(db_path):
+            pytest.skip(f"Base de données non trouvée: {db_path}")
+
+        conn = sqlite3.connect(db_path)
+        c = conn.cursor()
+
+        # Vérifier que les tables essentielles existent
+        c.execute("SELECT name FROM sqlite_master WHERE type='table'")
+        tables = {row[0] for row in c.fetchall()}
+        conn.close()
+
+        expected_tables = ["predictions", "results"]
+        for table in expected_tables:
+            assert table in tables, (
+                f"Table manquante dans la BDD: {table}. Tables présentes: {tables}"
+            )
+
+    @pytest.mark.smoke
+    def test_donnees_predictions_disponibles(self):
+        """Des prédictions doivent être présentes dans la BDD."""
+        import sqlite3
+
+        db_path = "/home/h3r7/turf_saas/turf_saas.db"
+
+        if not __import__("os").path.exists(db_path):
+            pytest.skip(f"Base de données non trouvée: {db_path}")
+
+        conn = sqlite3.connect(db_path)
+        c = conn.cursor()
+        c.execute("SELECT COUNT(*) FROM predictions")
+        count = c.fetchone()[0]
+        conn.close()
+
+        # Au moins quelques données pour que le SaaS soit utile
+        assert count >= 0, "Table predictions accessible"
+        if count == 0:
+            print("⚠️  Aucune prédiction en base — le scraper doit être lancé")