feat: connecteurs Google + Microsoft 365 (OAuth2 PKCE) — phase 6/8 (v7.56.0)

- app/services/oauth_connectors.py : flow OAuth2 complet PKCE (S256) pour
  2 fournisseurs décrits par 1 dict — Google (Drive/Gmail/Calendar en lecture
  seule) et Microsoft 365 (Graph Files.Read / Mail.Read / Calendars.Read) ;
  begin() = URL d'autorisation + state + code_verifier, complete() = échange du
  code, access_token() = refresh automatique (60 s de marge, refresh_token
  conservé si absent de la réponse), api_get() = path absolu refusé + validation
  SSRF + borne 20 000 car.
- Tokens chiffrés Fernet en réutilisant calendar_sync._encrypt_tokens (zéro
  dépendance) dans la table connector_tokens (migration 33, PK (kind, user_id)).
- 4 routes /api/agent/connectors/oauth/{kind}/… : status, authorize (cookies
  d'état HttpOnly 10 min, retour same-origin validé), callback (GET safe, state
  comparé en temps constant, tokens stockés puis cookies purgés, redirection
  ?oauth=connected / ?oauth_error=), disconnect. OpenAPI 523 chemins.
- Config + .env.example : GOOGLE_CLIENT_ID/SECRET, MS_CLIENT_ID/SECRET (vidés =
  « non configuré »), redirect URI dérivé d'APP_BASE_URL.
- Catalogue : google/ms365 en natifs avec badge connecté/non connecté ;
  connector_fetch et Tester passent par l'API du fournisseur avec le token de
  l'utilisateur (user_id transmis par l'outil LLM).
- Menu + : « Se connecter » / « Déconnecter » sur la fiche, toast au retour du
  flux (URL nettoyée par history.replaceState). État dans la fiche du menu
  plutôt qu'une page dédiée.
- Tests : tests/test_v756_oauth_connectors.py (13), 0 appel réseau réel
  (_post_form / _api_get monkeypatchés) — state forgé refusé sans échange,
  tokens chiffrés en base, refresh, URL absolue refusée, 401/404, câblage menu ;
  test_v755 adapté (5 natifs). Suite complète 1319 verts (-n auto), ruff 0,
  eslint 0 erreur (19 warnings préexistants hors fichiers touchés).

app/services/connectors.py

@@ -14,18 +14,30 @@ import logging
 
 from app.config import settings
 from app.db import get_conn
+from app.services import oauth_connectors
 from app.services.sso_provisioning import decrypt_secret, encrypt_secret
 
 logger = logging.getLogger(__name__)
 
-NATIVE_KINDS = ("gitea", "github", "web")
+NATIVE_KINDS = ("gitea", "github", "web", "google", "ms365")
+OAUTH_KINDS = oauth_connectors.OAUTH_KINDS
 FETCH_LIMIT = 20000   # caractères max renvoyés au LLM (ponytail : borne fixe)
 
 
 # ── Natifs (config, sans réseau) ─────────────────────────────────────────────
 
-def native_state(kind: str) -> tuple[str, str]:
+def native_state(kind: str, user_id: int | None = None) -> tuple[str, str]:
     """(status, detail) d'un connecteur natif — dérivé de la config, sans réseau."""
+    if kind in OAUTH_KINDS:
+        try:
+            oauth_connectors._client(kind)   # lève si client_id/secret absents
+        except ValueError as exc:
+            return ("missing", str(exc))
+        tok = oauth_connectors.tokens(kind, user_id)
+        if tok.get("access_token"):
+            scope = (tok.get("scope") or "").split()
+            return ("ok", f"connecté · {len(scope)} scope(s)")
+        return ("missing", "non connecté — lancer la connexion OAuth")
     if kind == "gitea":
         ok = bool(settings.gitea_url) and settings.gitea_token not in ("", "change-me")
         return ("ok", settings.gitea_url if ok else "GITEA_TOKEN non configuré")
@@ -36,15 +48,17 @@ def native_state(kind: str) -> tuple[str, str]:
     return ("ok", f"recherche web via {settings.web_search_provider or 'duckduckgo'}")
 
 
-def list_connectors() -> list[dict]:
-    """Catalogue : 3 natifs (toujours présents) + les connecteurs personnels."""
+def list_connectors(user_id: int | None = None) -> list[dict]:
+    """Catalogue : natifs (toujours présents) + connecteurs personnels."""
     out: list[dict] = []
     for kind in NATIVE_KINDS:
-        status, detail = native_state(kind)
+        status, detail = native_state(kind, user_id)
         out.append({
-            "id": None, "builtin": True, "kind": kind, "name": kind.capitalize(),
+            "id": None, "builtin": True, "kind": kind,
+            "name": (oauth_connectors.PROVIDERS[kind]["name"] if kind in OAUTH_KINDS
+                     else kind.capitalize()),
             "url": "", "enabled": True, "status": status, "detail": detail,
-            "has_secret": False,
+            "has_secret": False, "oauth": kind in OAUTH_KINDS,
         })
     with get_conn() as conn:
         rows = conn.execute(
@@ -183,11 +197,15 @@ def _resolve(connector: str) -> tuple[str, str | int]:
     raise ValueError(f"Connecteur inconnu: {token}")
 
 
-async def probe(connector: str) -> dict:
+async def probe(connector: str, user_id: int | None = None) -> dict:
     """Teste un connecteur et **persiste** le résultat (personnel uniquement)."""
     kind, ref = _resolve(connector)
     try:
-        if kind == "gitea":
+        if kind in OAUTH_KINDS:
+            res = await oauth_connectors.api_get(kind, user_id,
+                                                 oauth_connectors.PROVIDERS[kind]["probe_path"])
+            status, detail = res["status"], res["text"][:200]
+        elif kind == "gitea":
             code, _ = await _get(f"{settings.gitea_url.rstrip('/')}/api/v1/version",
                                  {"Authorization": f"token {settings.gitea_token}"})
             status, detail = ("ok" if code < 400 else "error"), f"HTTP {code}"
@@ -221,10 +239,13 @@ def _gh_headers() -> dict:
     return headers
 
 
-async def connector_fetch(connector: str, path: str = "", query: str = "") -> dict:
+async def connector_fetch(connector: str, path: str = "", query: str = "",
+                          user_id: int | None = None) -> dict:
     """Lit un connecteur pour l'LLM : natif = API, web = recherche, perso = GET."""
     kind, ref = _resolve(connector)
     path = (path or "").strip()
+    if kind in OAUTH_KINDS:
+        return await oauth_connectors.api_get(kind, user_id, path)
     if kind == "gitea":
         base = settings.gitea_url.rstrip("/")
         url = f"{base}/{path.lstrip('/')}" if path else f"{base}/api/v1/version"

app/services/oauth_connectors.py

@@ -0,0 +1,254 @@
+"""Connecteurs OAuth Google / Microsoft 365 (v7.56.0 — phase 6).
+
+Flow : ``begin()`` (URL d'autorisation PKCE S256 + state) → callback
+``complete()`` (échange du code) → tokens chiffrés Fernet → ``api_get()`` avec
+refresh automatique.
+
+Réutilise : `_encrypt_tokens` / `_decrypt_tokens` (`calendar_sync`, déjà Fernet)
+et `shared_client` (A42). Les scopes sont **figés en lecture seule**.
+
+ponytail : 2 providers décrits par 1 dict (pas de classe par provider) ; les
+scopes au choix et un écran de connexion dédié arriveront si le besoin se
+présente — l'état vit dans le catalogue du menu ＋.
+"""
+
+from __future__ import annotations
+
+import base64
+import hashlib
+import logging
+import secrets
+import time
+from urllib.parse import urlencode
+
+from app.config import settings
+from app.db import get_conn
+from app.services.calendar_sync import _decrypt_tokens, _encrypt_tokens
+
+logger = logging.getLogger(__name__)
+
+PROVIDERS: dict[str, dict] = {
+    "google": {
+        "name": "Google (Drive · Gmail · Calendar)",
+        "authorize": "https://accounts.google.com/o/oauth2/v2/auth",
+        "token": "https://oauth2.googleapis.com/token",
+        "api": "https://www.googleapis.com",
+        "scopes": [
+            "openid", "email", "profile",
+            "https://www.googleapis.com/auth/drive.readonly",
+            "https://www.googleapis.com/auth/gmail.readonly",
+            "https://www.googleapis.com/auth/calendar.readonly",
+        ],
+        "extra": {"access_type": "offline", "include_granted_scopes": "true"},
+        "probe_path": "/oauth2/v3/userinfo",
+    },
+    "ms365": {
+        "name": "Microsoft 365 (Outlook · OneDrive)",
+        "authorize": "https://login.microsoftonline.com/common/oauth2/v2.0/authorize",
+        "token": "https://login.microsoftonline.com/common/oauth2/v2.0/token",
+        "api": "https://graph.microsoft.com/v1.0",
+        "scopes": [
+            "openid", "email", "profile", "offline_access",
+            "User.Read", "Files.Read", "Mail.Read", "Calendars.Read",
+        ],
+        "extra": {"response_mode": "query", "prompt": "consent"},
+        "probe_path": "/me",
+    },
+}
+OAUTH_KINDS = tuple(PROVIDERS)
+# l'access token est renouvelé 60 s avant expiration
+_REFRESH_SKEW = 60
+
+
+# ── Config client ───────────────────────────────────────────────────────────
+
+def _client(kind: str) -> tuple[str, str, str]:
+    """(client_id, client_secret, redirect_uri) — lève si non configuré."""
+    if kind not in PROVIDERS:
+        raise ValueError(f"Connecteur OAuth inconnu: {kind}")
+    if kind == "google":
+        cid, secret = settings.google_client_id, settings.google_client_secret
+    else:
+        cid, secret = settings.ms_client_id, settings.ms_client_secret
+    if not cid or not secret:
+        raise ValueError(
+            f"Connecteur {kind} non configuré (GOOGLE_CLIENT_ID/SECRET ou MS_CLIENT_ID/SECRET)"
+        )
+    redirect = f"{settings.app_base_url.rstrip('/')}/api/agent/connectors/oauth/{kind}/callback"
+    return cid, secret, redirect
+
+
+def callback_path(kind: str) -> str:
+    return f"/api/agent/connectors/oauth/{kind}/callback"
+
+
+def _pkce_pair() -> tuple[str, str]:
+    verifier = secrets.token_urlsafe(48)
+    digest = hashlib.sha256(verifier.encode("ascii")).digest()
+    return verifier, base64.urlsafe_b64encode(digest).rstrip(b"=").decode("ascii")
+
+
+# ── Flow ────────────────────────────────────────────────────────────────────
+
+def begin(kind: str) -> dict:
+    """URL d'autorisation + state + code_verifier (le route met les cookies)."""
+    cid, _secret, redirect = _client(kind)
+    verifier, challenge = _pkce_pair()
+    state = secrets.token_urlsafe(24)
+    params = {
+        "client_id": cid,
+        "redirect_uri": redirect,
+        "response_type": "code",
+        "scope": " ".join(PROVIDERS[kind]["scopes"]),
+        "state": state,
+        "code_challenge": challenge,
+        "code_challenge_method": "S256",
+    }
+    params.update(PROVIDERS[kind].get("extra", {}))
+    return {
+        "url": f"{PROVIDERS[kind]['authorize']}?{urlencode(params)}",
+        "state": state,
+        "verifier": verifier,
+        "redirect_uri": redirect,
+    }
+
+
+def _normalize(data: dict) -> dict:
+    expires_in = int(data.get("expires_in") or 3600)
+    return {
+        "access_token": str(data.get("access_token") or ""),
+        "refresh_token": str(data.get("refresh_token") or ""),
+        "scope": str(data.get("scope") or ""),
+        "expires_at": int(time.time()) + expires_in,
+    }
+
+
+async def complete(kind: str, code: str, verifier: str) -> dict:
+    """Échange le code contre des tokens (aucun réseau ici hors `_post_form`)."""
+    cid, secret, redirect = _client(kind)
+    data = await _post_form(PROVIDERS[kind]["token"], {
+        "client_id": cid,
+        "client_secret": secret,
+        "grant_type": "authorization_code",
+        "code": code,
+        "redirect_uri": redirect,
+        "code_verifier": verifier,
+    })
+    tokens = _normalize(data)
+    if not tokens["access_token"]:
+        raise ValueError(str(data.get("error_description") or data.get("error") or "échec de l'échange du code"))
+    return tokens
+
+
+# ── Stockage (chiffré Fernet, comme calendar_sync) ──────────────────────────
+
+def save(kind: str, user_id: int, tokens: dict) -> None:
+    with get_conn() as conn:
+        conn.execute(
+            "INSERT INTO connector_tokens (kind, user_id, tokens_enc, updated_at) "
+            "VALUES (?,?,?,CURRENT_TIMESTAMP) "
+            "ON CONFLICT(kind, user_id) DO UPDATE SET "
+            "tokens_enc=excluded.tokens_enc, updated_at=CURRENT_TIMESTAMP",
+            (kind, user_id, _encrypt_tokens(tokens)),
+        )
+        conn.commit()
+
+
+def tokens(kind: str, user_id: int | None) -> dict:
+    if not user_id:
+        return {}
+    with get_conn() as conn:
+        row = conn.execute(
+            "SELECT tokens_enc FROM connector_tokens WHERE kind=? AND user_id=?",
+            (kind, user_id),
+        ).fetchone()
+    return _decrypt_tokens(row["tokens_enc"]) if row else {}
+
+
+def clear(kind: str, user_id: int) -> bool:
+    with get_conn() as conn:
+        cur = conn.execute(
+            "DELETE FROM connector_tokens WHERE kind=? AND user_id=?", (kind, user_id)
+        )
+        conn.commit()
+        return cur.rowcount > 0
+
+
+def is_connected(kind: str, user_id: int | None) -> bool:
+    return bool(tokens(kind, user_id).get("access_token"))
+
+
+# ── Réseau ──────────────────────────────────────────────────────────────────
+
+async def _post_form(url: str, data: dict) -> dict:
+    """POST x-www-form-urlencoded vers le token endpoint → dict JSON.
+
+    Point d'injection des tests (on monkeypatche ``oauth_connectors._post_form``).
+    """
+    from app.services.http_client import shared_client
+
+    async with shared_client(timeout=15, headers={"Accept": "application/json"}) as client:
+        resp = await client.post(url, data=data)
+        return resp.json()
+
+
+async def _api_get(url: str, headers: dict) -> tuple[int, str]:
+    """GET d'une API fournisseur — 2ᵉ point d'injection des tests."""
+    from app.services.http_client import shared_client
+
+    async with shared_client(timeout=15, headers=headers) as client:
+        resp = await client.get(url)
+        return resp.status_code, (resp.text or "")[:20000]
+
+
+async def access_token(kind: str, user_id: int | None) -> str:
+    """Access token valide, rafraîchi (refresh_token) si nécessaire."""
+    tok = tokens(kind, user_id)
+    if not tok.get("access_token"):
+        return ""
+    if tok.get("expires_at", 0) > time.time() + _REFRESH_SKEW:
+        return tok["access_token"]
+    if not tok.get("refresh_token"):
+        return ""   # expiré sans refresh → à reconnexion
+    try:
+        cid, secret, redirect = _client(kind)
+        data = await _post_form(PROVIDERS[kind]["token"], {
+            "client_id": cid,
+            "client_secret": secret,
+            "grant_type": "refresh_token",
+            "refresh_token": tok["refresh_token"],
+            "redirect_uri": redirect,
+        })
+        fresh = _normalize(data)
+        if not fresh["access_token"]:
+            raise ValueError(data.get("error_description") or "refresh refusé")
+        # Google ne renvoie parfois pas de nouveau refresh_token : on garde l'ancien
+        fresh["refresh_token"] = fresh["refresh_token"] or tok["refresh_token"]
+        save(kind, int(user_id), fresh)
+        return fresh["access_token"]
+    except Exception as exc:  # noqa: BLE001 — jamais d'exception qui casse un run
+        logger.warning("OAuth refresh failed (%s): %s", kind, exc)
+        return ""
+
+
+async def api_get(kind: str, user_id: int | None, path: str = "") -> dict:
+    """GET sur l'API du fournisseur avec le token de l'utilisateur."""
+    if kind not in PROVIDERS:
+        return {"status": "error", "text": f"Connecteur OAuth inconnu: {kind}"}
+    access = await access_token(kind, user_id)
+    if not access:
+        return {"status": "error", "text": f"{kind} non connecté (lancez la connexion OAuth)"}
+    if "://" in (path or ""):
+        return {"status": "error", "text": "path doit être relatif (pas d'URL absolue)"}
+    # base fixe + chemin : pas d'urljoin (une URL absolue ne peut pas dévier
+    # l'hôte), puis validation SSRF par principe.
+    url = PROVIDERS[kind]["api"].rstrip("/") + "/" + (path or "").lstrip("/")
+    from app.services.importers.url_fetch import _validate_url
+    try:
+        _validate_url(url)
+    except ValueError as exc:
+        return {"status": "error", "text": str(exc)}
+    code, text = await _api_get(url, {"Authorization": f"Bearer {access}"})
+    if code >= 400:
+        return {"status": "error", "text": f"HTTP {code} — {text[:400]}"}
+    return {"status": "ok", "text": text[:20000]}

app/services/tool_registry.py

@@ -1099,6 +1099,7 @@ class ConnectorFetch(Tool):
                 str(args.get("connector") or ""),
                 str(args.get("path") or ""),
                 str(args.get("query") or ""),
+                user_id=user_id,
             )
         except ValueError as exc:
             return ToolResult(status="error", tool=self.name, message=str(exc))