Accounts Phase 2: Google sign-in (OAuth 2.0 / OIDC)

- oauth_google.py (stdlib): PKCE, auth URL, code exchange, ID-token claim
  validation (iss/aud/exp/email_verified — token comes straight from Google's
  token endpoint over TLS, so no signature re-verify / JWKS needed).
- API: GET /api/auth/google/start (302 to Google, PKCE + signed state cookie
  binding the flow to the browser) and /callback (CSRF-checked state, exchange,
  find-or-create by verified email → links to an existing magic-link account,
  session cookie, redirect home). Errors land on /auth/verify?error=google.
- SignIn modal: "Continue with Google" + an "or email link" divider.
- 112 tests pass.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

goodnews/api.py

@@ -13,9 +13,12 @@ so the API and CLI always read the same file.
 
 from __future__ import annotations
 
+import hashlib
+import hmac
 import json
 import os
 import re
+import secrets
 import sqlite3
 from collections import Counter
 from contextlib import contextmanager
@@ -24,10 +27,11 @@ from pathlib import Path
 
 from fastapi import BackgroundTasks, FastAPI, HTTPException, Query, Request, Response
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.responses import RedirectResponse
 from fastapi.staticfiles import StaticFiles
 from pydantic import BaseModel
 
-from . import auth, email_send, feeds, queries
+from . import auth, email_send, feeds, oauth_google, queries
 from .db import connect
 from .filters import filter_articles, prefs_from_json
 from .hero import safe_to_lead
@@ -52,12 +56,31 @@ def db_path() -> Path:
 
 PUBLIC_BASE_URL = os.environ.get("GOODNEWS_PUBLIC_BASE_URL", "https://upbeatbytes.com").rstrip("/")
 SESSION_COOKIE = "ub_session"
+OAUTH_COOKIE = "ub_oauth"
 SESSION_MAX_AGE = int(auth.SESSION_TTL.total_seconds())
+SESSION_SECRET = os.environ.get("GOODNEWS_SESSION_SECRET", "dev-insecure-secret")
 # Secure cookies in production (https); off for http (local/test) so they round-trip.
 _COOKIE_SECURE = PUBLIC_BASE_URL.startswith("https")
 _EMAIL_RE = re.compile(r"^[^@\s]+@[^@\s]+\.[^@\s]+$")
 
 
+def _sign(value: str) -> str:
+    sig = hmac.new(SESSION_SECRET.encode(), value.encode(), hashlib.sha256).hexdigest()
+    return f"{value}.{sig}"
+
+
+def _unsign(signed: str | None) -> str | None:
+    if not signed or "." not in signed:
+        return None
+    value, _, sig = signed.rpartition(".")
+    expected = hmac.new(SESSION_SECRET.encode(), value.encode(), hashlib.sha256).hexdigest()
+    return value if hmac.compare_digest(sig, expected) else None
+
+
+def _google_redirect_uri() -> str:
+    return f"{PUBLIC_BASE_URL}/api/auth/google/callback"
+
+
 def _session_token_from_request(request: Request) -> str | None:
     """Web sends the session as an httpOnly cookie; the app sends a bearer token."""
     cookie = request.cookies.get(SESSION_COOKIE)
@@ -348,6 +371,55 @@ def create_app() -> FastAPI:
         response.delete_cookie(SESSION_COOKIE, path="/")
         return {"ok": True}
 
+    # --- Auth: Google (OAuth 2.0 / OIDC) ----------------------------------
+
+    @app.get("/api/auth/google/start")
+    def google_start() -> RedirectResponse:
+        if not oauth_google.configured():
+            raise HTTPException(status_code=503, detail="Google sign-in isn't configured.")
+        state = secrets.token_urlsafe(24)
+        verifier, challenge = oauth_google.new_pkce()
+        url = oauth_google.auth_url(_google_redirect_uri(), state, challenge)
+        resp = RedirectResponse(url, status_code=302)
+        # Bind the flow to this browser; read back (and CSRF-checked) on callback.
+        resp.set_cookie(
+            OAUTH_COOKIE, _sign(f"{state}:{verifier}"), max_age=600,
+            httponly=True, secure=_COOKIE_SECURE, samesite="lax", path="/",
+        )
+        return resp
+
+    @app.get("/api/auth/google/callback")
+    def google_callback(
+        request: Request,
+        code: str | None = None,
+        state: str | None = None,
+        error: str | None = None,
+    ) -> RedirectResponse:
+        fail = RedirectResponse(f"{PUBLIC_BASE_URL}/auth/verify?error=google", status_code=302)
+        if error or not code or not state:
+            return fail
+        saved = _unsign(request.cookies.get(OAUTH_COOKIE))
+        if not saved:
+            return fail
+        saved_state, _, verifier = saved.partition(":")
+        if not hmac.compare_digest(saved_state, state):
+            return fail
+        try:
+            tokens = oauth_google.exchange_code(code, _google_redirect_uri(), verifier)
+            info = oauth_google.verify_id_token(tokens["id_token"])
+        except Exception:
+            return fail
+        with get_conn() as conn:
+            user_id = auth.find_or_create_user(
+                conn, info["email"], "google", info["sub"], display_name=info.get("name")
+            )
+            token = auth.create_session(conn, user_id, user_agent=request.headers.get("User-Agent"))
+            conn.commit()
+        ok = RedirectResponse(f"{PUBLIC_BASE_URL}/", status_code=302)
+        _set_session_cookie(ok, token)
+        ok.delete_cookie(OAUTH_COOKIE, path="/")
+        return ok
+
     @app.get("/api/categories", response_model=CategoriesResponse)
     def categories() -> CategoriesResponse:
         return CategoriesResponse(

goodnews/oauth_google.py

@@ -0,0 +1,101 @@
+"""Google Sign-In (OAuth 2.0 / OpenID Connect) — stdlib only.
+
+Authorization-code flow with PKCE. The ID token is received directly from
+Google's token endpoint over TLS (server-to-server), so per Google's own
+guidance we validate its claims (iss/aud/exp/email_verified) without re-verifying
+the RS256 signature — no JWT/JWKS dependency required.
+"""
+
+from __future__ import annotations
+
+import base64
+import hashlib
+import json
+import os
+import secrets
+import urllib.parse
+import urllib.request
+from datetime import datetime, timezone
+
+AUTH_URL = "https://accounts.google.com/o/oauth2/v2/auth"
+TOKEN_URL = "https://oauth2.googleapis.com/token"
+_VALID_ISS = {"accounts.google.com", "https://accounts.google.com"}
+
+
+def configured() -> bool:
+    return bool(
+        os.environ.get("GOODNEWS_GOOGLE_CLIENT_ID")
+        and os.environ.get("GOODNEWS_GOOGLE_CLIENT_SECRET")
+    )
+
+
+def _b64url(raw: bytes) -> str:
+    return base64.urlsafe_b64encode(raw).rstrip(b"=").decode("ascii")
+
+
+def new_pkce() -> tuple[str, str]:
+    """Return (code_verifier, code_challenge) for S256 PKCE."""
+    verifier = _b64url(secrets.token_bytes(48))
+    challenge = _b64url(hashlib.sha256(verifier.encode("ascii")).digest())
+    return verifier, challenge
+
+
+def auth_url(redirect_uri: str, state: str, code_challenge: str) -> str:
+    params = {
+        "client_id": os.environ["GOODNEWS_GOOGLE_CLIENT_ID"],
+        "redirect_uri": redirect_uri,
+        "response_type": "code",
+        "scope": "openid email profile",
+        "state": state,
+        "code_challenge": code_challenge,
+        "code_challenge_method": "S256",
+        "access_type": "online",
+        "prompt": "select_account",
+    }
+    return AUTH_URL + "?" + urllib.parse.urlencode(params)
+
+
+def exchange_code(code: str, redirect_uri: str, code_verifier: str) -> dict:
+    data = urllib.parse.urlencode(
+        {
+            "code": code,
+            "client_id": os.environ["GOODNEWS_GOOGLE_CLIENT_ID"],
+            "client_secret": os.environ["GOODNEWS_GOOGLE_CLIENT_SECRET"],
+            "redirect_uri": redirect_uri,
+            "grant_type": "authorization_code",
+            "code_verifier": code_verifier,
+        }
+    ).encode("ascii")
+    request = urllib.request.Request(
+        TOKEN_URL,
+        data=data,
+        headers={"Content-Type": "application/x-www-form-urlencoded", "Accept": "application/json"},
+    )
+    with urllib.request.urlopen(request, timeout=15) as response:
+        return json.loads(response.read())
+
+
+def _decode_jwt_payload(token: str) -> dict:
+    parts = token.split(".")
+    if len(parts) != 3:
+        raise ValueError("malformed id_token")
+    payload = parts[1]
+    payload += "=" * (-len(payload) % 4)
+    return json.loads(base64.urlsafe_b64decode(payload))
+
+
+def verify_id_token(id_token: str) -> dict:
+    """Validate an ID token's claims (not its signature — see module docstring).
+
+    Returns {"sub", "email", "name"} or raises ValueError.
+    """
+    claims = _decode_jwt_payload(id_token)
+    if claims.get("iss") not in _VALID_ISS:
+        raise ValueError("unexpected issuer")
+    if claims.get("aud") != os.environ.get("GOODNEWS_GOOGLE_CLIENT_ID"):
+        raise ValueError("audience mismatch")
+    if datetime.now(timezone.utc).timestamp() > float(claims.get("exp", 0)):
+        raise ValueError("id_token expired")
+    if not claims.get("email") or claims.get("email_verified") not in (True, "true"):
+        raise ValueError("email not verified")
+    return {"sub": str(claims["sub"]), "email": claims["email"], "name": claims.get("name")}