PricingLoginStart free
FastAPI Backend Development Bootcamp · レッスン

OAuth2パスワードフローとトークン発行

OAuth2PasswordBearer方式を実装し、passlibでパスワードをハッシュ化して、ログイン時に署名付きアクセストークンを発行します。

レッスン 1/413 ステップ

「OAuth2パスワードフローとトークン発行」はCoddyKit上の無料FastAPI Backend Development Bootcampレッスンです。 これはレッスン1/4です。 下記で完全なレッスンを無料で読むことができます。その後、ブラウザ内の組み込みコードエディタと24時間対応のAIチューターでハンズオン演習できます。 これはFastAPI Backend Development Bootcamp学習パスの一部であり、ウェブとCoddyKitアプリ全体で進捗が同期されます。 FastAPI Backend Development Bootcampコースには全4レッスンが含まれています。

このレッスンの一部はまだ翻訳されておらず、英語で表示されています。

The OAuth2 Password Flow in Plain English

The OAuth2 password flow (a.k.a. the Resource Owner Password Credentials grant) is the simplest way to authenticate a first-party client: the user sends their username and password directly to your API, and the API hands back a signed access token.

  • The client posts credentials once to a /token endpoint.
  • The server verifies them against the database.
  • On success it returns a short-lived JWT access token.
  • Every later request carries that token in the Authorization: Bearer <token> header.

FastAPI gives us ready-made building blocks for exactly this: OAuth2PasswordBearer and OAuth2PasswordRequestForm.

Declaring the OAuth2PasswordBearer Scheme

OAuth2PasswordBearer is a FastAPI dependency that knows how to pull a bearer token out of the Authorization header. You create one instance and point its tokenUrl at the login endpoint that issues tokens.

  • tokenUrl is a relative path — it tells the docs UI where clients should request a token.
  • Using the scheme as a dependency makes the endpoint require a token; a missing or malformed header returns 401 automatically.
from fastapi import Depends, FastAPI
from fastapi.security import OAuth2PasswordBearer

app = FastAPI()

# 'token' matches the path of our login route below
oauth2_scheme = OAuth2PasswordBearer(tokenUrl="token")


@app.get("/users/me")
async def read_me(token: str = Depends(oauth2_scheme)):
    # FastAPI extracts the raw bearer token string for us
    return {"token": token}

Hashing Passwords with passlib

You must never store raw passwords. Hash them with a strong, salted algorithm. The passlib library wraps bcrypt behind a clean CryptContext API.

  • hash() produces a salted digest you store in the database.
  • verify() compares a plaintext attempt against the stored hash in constant time.
  • bcrypt is deliberately slow, which frustrates brute-force attacks.
from passlib.context import CryptContext

pwd_context = CryptContext(schemes=["bcrypt"], deprecated="auto")


def hash_password(plain: str) -> str:
    return pwd_context.hash(plain)


def verify_password(plain: str, hashed: str) -> bool:
    return pwd_context.verify(plain, hashed)


stored = hash_password("s3cret")
print("stored looks like:", stored[:7], "...")
print("correct ->", verify_password("s3cret", stored))
print("wrong   ->", verify_password("nope", stored))

Modeling Users and a Tiny Fake Database

Before issuing tokens we need somewhere to look users up. In production this is your real database; for learning we use an in-memory dict. Notice the stored field is hashed_password, never the plaintext.

  • A Pydantic model gives the user a typed shape.
  • A get_user() helper centralizes lookups.
from pydantic import BaseModel


class UserInDB(BaseModel):
    username: str
    hashed_password: str
    disabled: bool = False


fake_users_db = {
    "alice": UserInDB(
        username="alice",
        hashed_password="$2b$12$exampleexampleexamplehashvalue",
    )
}


def get_user(username: str):
    return fake_users_db.get(username)

Authenticating the Credentials

Authentication ties the pieces together: find the user, then verify the supplied password against the stored hash. Return the user on success, or a falsy value on failure.

  • Look the user up first; if absent, fail.
  • Then call verify_password — do not short-circuit before hashing to keep timing roughly uniform.
  • The caller decides how to respond (usually a 401).
def authenticate_user(db, username: str, password: str):
    user = db.get(username)
    if not user:
        return None
    if not verify_password(password, user.hashed_password):
        return None
    return user

What a JWT Actually Is

A JSON Web Token is three base64url segments joined by dots: header.payload.signature.

  • The header names the algorithm, e.g. HS256.
  • The payload holds claims like sub (subject) and exp (expiry).
  • The signature is an HMAC of header+payload using your secret key.

JWTs are signed, not encrypted — anyone can read the payload, but nobody can forge it without the secret. Never put passwords or sensitive data in the payload.

Encoding a Signed Access Token

We sign tokens with the python-jose library (or PyJWT). Always include an exp claim so tokens expire. Store the username in the sub claim — it identifies who the token belongs to.

  • SECRET_KEY must be long, random, and kept out of source control.
  • Set a short lifetime (e.g. 15-30 minutes) for access tokens.
from datetime import datetime, timedelta, timezone
from jose import jwt

SECRET_KEY = "replace-with-a-long-random-secret"
ALGORITHM = "HS256"
ACCESS_TOKEN_EXPIRE_MINUTES = 30


def create_access_token(data: dict) -> str:
    to_encode = data.copy()
    expire = datetime.now(timezone.utc) + timedelta(
        minutes=ACCESS_TOKEN_EXPIRE_MINUTES
    )
    to_encode.update({"exp": expire})
    return jwt.encode(to_encode, SECRET_KEY, algorithm=ALGORITHM)


token = create_access_token({"sub": "alice"})
print("issued token segments:", token.count(".") + 1)

The Token Response Shape

The OAuth2 spec dictates the JSON your /token endpoint returns. At minimum it must include access_token and token_type, where the type is the literal string "bearer".

  • Clients read token_type to know how to send the credential back.
  • A Pydantic Token model documents and validates the response.
from pydantic import BaseModel


class Token(BaseModel):
    access_token: str
    token_type: str


example = Token(access_token="eyJhbGci...", token_type="bearer")
print(example.model_dump())

Wiring the /token Login Endpoint

The login route depends on OAuth2PasswordRequestForm, which reads form-encoded username and password fields (not JSON) — exactly what the OAuth2 password flow requires. On success it returns the Token response.

  • Failed auth raises 401 with a WWW-Authenticate: Bearer header.
  • The sub claim carries the username forward into the token.
from fastapi import Depends, FastAPI, HTTPException, status
from fastapi.security import OAuth2PasswordRequestForm

app = FastAPI()


@app.post("/token", response_model=Token)
async def login(form: OAuth2PasswordRequestForm = Depends()):
    user = authenticate_user(fake_users_db, form.username, form.password)
    if not user:
        raise HTTPException(
            status_code=status.HTTP_401_UNAUTHORIZED,
            detail="Incorrect username or password",
            headers={"WWW-Authenticate": "Bearer"},
        )
    access_token = create_access_token({"sub": user.username})
    return Token(access_token=access_token, token_type="bearer")

Decoding the Token to Find the Current User

A protected route depends on oauth2_scheme to receive the raw token, then decodes it. If the signature is invalid or the token is expired, jwt.decode raises JWTError and we return 401.

  • Read the username from the sub claim.
  • Re-load the user from the database to confirm they still exist and are active.
from fastapi import Depends, HTTPException, status
from jose import JWTError, jwt


async def get_current_user(token: str = Depends(oauth2_scheme)):
    credentials_exc = HTTPException(
        status_code=status.HTTP_401_UNAUTHORIZED,
        detail="Could not validate credentials",
        headers={"WWW-Authenticate": "Bearer"},
    )
    try:
        payload = jwt.decode(token, SECRET_KEY, algorithms=[ALGORITHM])
        username = payload.get("sub")
        if username is None:
            raise credentials_exc
    except JWTError:
        raise credentials_exc
    user = get_user(username)
    if user is None:
        raise credentials_exc
    return user

Security Practices That Matter

The mechanics work, but production hardening makes them safe:

  • Secret key: load SECRET_KEY from an environment variable; rotate it if leaked.
  • HTTPS only: tokens in headers are plaintext on the wire — TLS is mandatory.
  • Short expiry: keep access tokens brief and pair them with longer-lived refresh tokens.
  • Pin the algorithm: pass an explicit algorithms=["HS256"] list to jwt.decode to block the alg: none attack.
  • Generic errors: say "Incorrect username or password", never reveal which one was wrong.

Quick Check: The /token Endpoint

Time to test your understanding of how the FastAPI /token login route consumes credentials.

対話的なレッスンで答える

Recap: From Password to Bearer Token

You implemented the full OAuth2 password flow in FastAPI:

  • OAuth2PasswordBearer declares the bearer scheme and extracts tokens from the Authorization header.
  • passlib + bcrypt hash and verify passwords so plaintext is never stored.
  • authenticate_user looks up the user and verifies the hash, returning 401 on failure.
  • The /token route reads form credentials via OAuth2PasswordRequestForm and issues a signed JWT with a sub claim and an exp expiry.
  • get_current_user decodes and validates the token, pinning the algorithm to block forgery.

With HTTPS, an env-loaded secret, and short token lifetimes, this is a solid, idiomatic authentication foundation.

無料で開始

AI チューターと学ぶ FastAPI Backend Development Bootcamp — 無料

ブラウザでリアルコードを書いて実行し、24/7 の AI チューターから瞬時にサポートを受け、ウェブまたはアプリで続きから学習できます。

コース
21
レッスン
84
無料で学習を始めるコース全体を表示

よくある質問

「OAuth2パスワードフローとトークン発行」レッスンは無料ですか？

はい。「OAuth2パスワードフローとトークン発行」の完全なテキストはこのウェブで無料で読めます。インタラクティブに演習し（組み込みコードエディタと24時間対応のAIチューター）、FastAPI Backend Development Bootcampコースの残りをアンロックするには、CoddyKit PROにアップグレードしてください。 FastAPI Backend Development Bootcampコースには全4レッスンが含まれています。

「OAuth2パスワードフローとトークン発行」で何を学びますか？

OAuth2PasswordBearer方式を実装し、passlibでパスワードをハッシュ化して、ログイン時に署名付きアクセストークンを発行します。 ブラウザで直接実行するハンズオンコードでFastAPI Backend Development Bootcampを演習し、24時間対応のAIチューターがレッスンを進める中での質問に答えます。

FastAPI Backend Development Bootcampを始めるのに経験は必要ですか？

事前経験は必要ありません。CoddyKitのFastAPI Backend Development Bootcampは初級者から上級者向けに構成されているため、ここから始めるか最初から始めて、自分のペースで進むことができます。 これはレッスン1/4です。

「OAuth2パスワードフローとトークン発行」レッスンにはどのくらい時間がかかりますか？

ほとんどのCoddyKitレッスンは約5～10分かかります。各レッスンはコンパクトでインタラクティブなので、着実に進歩し、ウェブとアプリ全体で正確に前回の場所から再開できます。

このFastAPI Backend Development Bootcampレッスンでコードを書いて実行できますか？

はい。すべてのFastAPI Backend Development Bootcampレッスンに組み込みコードエディタが含まれているため、ブラウザでリアルコードを書いて実行し、即座のAIフィードバックを取得できます。ローカル設定は不要です。

このコースのすべてのレッスン

  1. OAuth2パスワードフローとトークン発行
  2. python-joseによるJWTの署名と検証
  3. リフレッシュトークンとトークンローテーション
  4. スコープベースの認可とロールガード
← FastAPI Backend Development Bootcampに戻る