Develop Reference

Passwordless authentication flow using Cognito & API Gateway & Lambda (Python)

I've been trying to implement passwordless authentication using AWS Cognito & API Gateway & Lambda (Python)
I have followed these articles:
https://medium.com/digicred/password-less-authentication-in-cognito-cafa016d4db7
https://medium.com/#pjatocheseminario/passwordless-api-using-cognito-and-serverless-framework-7fa952191352
I have configured Cognito (to accept CUSTOM_AUTH), added the Lambdas, and created the API endpoints:
/sign-up
/initiate-auth (aka initiate login)
/respond-to-auth-challenge (aka (verify login)
When calling initiateAuth I receive the following response:
An error occurred (NotAuthorizedException) when calling the InitiateAuth operation: Incorrect username or password."
I'm using CUSTOM_AUTH which doesn't require password, and the user name is definitely correct because it actually initiates the authentication flow and I receive a code, however because boto3 doesn't respond with a session I can't continue the authentication.
This is how I call Cognito:
res = cognito.initiate_auth(
ClientId=client_id,
AuthFlow="CUSTOM_AUTH",
AuthParameters={
"USERNAME": email,
"PASSWORD": random_password
}
)
It's probably something small I'm missing but I can't figure out what.

Your client code looks OK, mine has ClientId param in it but if your code is not raising an exception then it should be fine. Unless you had Generate client secret option checked when you created your app client.
If that is the case then you have to pass in SECRET_HASH in AuthParameters like the following:
import hmac
import hashlib
import base64
def get_secret_hash(email, client_id, client_secret):
"""
A keyed-hash message authentication code (HMAC) calculated using
the secret key of a user pool client and username plus the client
ID in the message.
"""
message = email + client_id
client_secret = str.encode(client_secret)
dig = hmac.new(client_secret, msg=message.encode('UTF-8'), digestmod=hashlib.sha256).digest()
return base64.b64encode(dig).decode()
client.admin_initiate_auth(
UserPoolId=COGNITO_USER_POOL_ID,
ClientId=CLIENT_ID,
AuthFlow='CUSTOM_AUTH',
AuthParameters={
'USERNAME': email,
'SECRET_HASH': get_secret_hash(email, CLIENT_ID, CLIENT_SECRET) # Omit if secret key option is disabled.
},
)
Next, double check the following:
Under App clients > * > Auth Flows Configuration, is ALLOW_CUSTOM_AUTH option enabled for your client?
Under App integration > App client settings > * > Enabled Identity Providers, is your user pool selected?
If you have Cognito setup correctly and your code still doesn't work then it is probably the lambda code. You probably know this but for password-less custom auth you need to use 3 lambda triggers: Define Auth Challenge, Create Auth Challenge, and Verify Auth Challenge.
Custom auth lambdas events are triggered in the following order:
DefineAuthChallenge_Authentication:
Technically, issueTokens can be set to True here to return tokens without going through the rest of the steps.
def lambda_handler(event, context):
if event['triggerSource'] == 'DefineAuthChallenge_Authentication':
event['response']['challengeName'] = 'CUSTOM_CHALLENGE'
event['response']['issueTokens'] = False
event['response']['failAuthentication'] = False
if event['request']['session']: # Needed for step 4.
# If all of the challenges are answered, issue tokens.
event['response']['issueTokens'] = all(
answered_challenge['challengeResult'] for answered_challenge in event['request']['session'])
return event
CreateAuthChallenge_Authentication:
def lambda_handler(event, context):
if event['triggerSource'] == 'CreateAuthChallenge_Authentication':
if event['request']['challengeName'] == 'CUSTOM_CHALLENGE':
event['response']['privateChallengeParameters'] = {}
event['response']['privateChallengeParameters']['answer'] = 'YOUR CHALLENGE ANSWER HERE'
event['response']['challengeMetadata'] = 'AUTHENTICATE_AS_CHALLENGE'
return event
Then your client must respond to the challenge:
client.respond_to_auth_challenge(
ClientId=CLIENT_ID,
ChallengeName='CUSTOM_CHALLENGE',
Session=session,
ChallengeResponses={
'USERNAME': email,
'ANSWER': 'Extra Protection!',
'SECRET_HASH': get_secret_hash(email, CLIENT_ID, CLIENT_SECRET) # Omit if secret key option is disabled.
}
)
VerifyAuthChallengeResponse_Authentication:
def lambda_handler(event, context):
if event['triggerSource'] == 'VerifyAuthChallengeResponse_Authentication':
if event['request']['challengeAnswer'] == event['request']['privateChallengeParameters']['answer']:
event['response']['answerCorrect'] = True
return event
DefineAuthChallenge_Authentication:
Set event['response']['issueTokens'] to True to return tokens (code shown in step 1), or issue another challenge to keep repeating steps 1-3.
Finally, make sure that if case-insensitivity option is enabled for your user pool too. Also, I can't exactly recall if CUSTOM_AUTH flow worked if the user is in FORCE_CHANGE_PASSWORD status. If the user is in that state, then try settings a permanent password with the sdk to set the status to CONFIRMED.

I was facing the same error, and I think that the error message is misleading.
When you did not respond correctly in Create-Auth-Challenge lambda, you will get this error. So make sure everything is right in your lambda.

In flask, why request header is None in app.before_request()?

When I try to access the header of a request in a function adnotated with #app.before_request, it is always None. Anyone had this problem before?
Here is the function:
#app.before_request
def verifyToken():
if request.endpoint in ['myEndpoint']
auth = request.headers.get('Authorization')
if auth.startswith('Bearer '):
jwtToken = auth[7:]
try:
decoded = jwt.decode(jwtToken, 'secret_key', algorithms=['HS256'])
except jwt.ExpiredSignatureError as e:
responseObject = {
"status": "failed",
"message": "Token has expired. Please login again."
}
return jsonify(responseObject), 401

I was being 5 months in Flask, I discover Flask Embeded Middleware fires up 2 times, the first one got None value and the second one is the client request and each fire up has different request.method, the OPTIONS and <DEFAULT METHOD> of the client, why there is an OPTIONS method first in before...request and after...request befoe goes in client request?, I'm still searching for reference
and that is why I just catch it through conditional statement if request.method != 'OPTIONS': and here how it is look like
from flask import request
def my_before_request():
if request.method !== 'OPTIONS'
token = request.headers['<header var>']
# do something...
# same as after request

The OPTIONS method could have been triggered by the browser due to request being preflighted. This is done to determine if the actual request is safe to send. Refer CORS

MismatchingStateError: mismatching_state: CSRF Warning! State not equal in request and response

This is driving me absolutely crazy and preventing me from being able to do local dev/test.
I have a flask app that uses authlib (client capabilities only). When a user hits my home page, my flask backend redirects them to /login which in turn redirects to Google Auth. Google Auth then posts them back to my app's /auth endpoint.
For months, I have been experiencing ad-hoc issues with authlib.integrations.base_client.errors.MismatchingStateError: mismatching_state: CSRF Warning! State not equal in request and response. It feels like a cookie problem and most of the time, I just open a new browser window or incognito or try to clear cache and eventually, it sort of works.
However, I am now running the exact same application inside of a docker container and at one stage this was working. I have no idea what I have changed but whenever I browse to localhost/ or 127.0.0.1/ and go through the auth process (clearing cookies each time to ensure i'm not auto-logged in), I am constantly redirected back to localhost/auth?state=blah blah blah and I experience this issue:
authlib.integrations.base_client.errors.MismatchingStateError: mismatching_state: CSRF Warning! State not equal in request and response.
I think the relevant part of my code is:
#app.route("/", defaults={"path": ""})
#app.route("/<path:path>")
def catch_all(path: str) -> Union[flask.Response, werkzeug.Response]:
if flask.session.get("user"):
return app.send_static_file("index.html")
return flask.redirect("/login")
#app.route("/auth")
def auth() -> Union[Tuple[str, int], werkzeug.Response]:
token = oauth.google.authorize_access_token()
user = oauth.google.parse_id_token(token)
flask.session["user"] = user
return flask.redirect("/")
#app.route("/login")
def login() -> werkzeug.Response:
return oauth.google.authorize_redirect(flask.url_for("auth", _external=True))
I would hugely appreciate any help.
When I run locally, I start with:
export FLASK_APP=foo && flask run
When I run inside docker container, i start with:
.venv/bin/gunicorn -b :8080 --workers 16 foo

Issue was that SECRET_KEY was being populated using os.random which yielded different values for different workers and thus, couldn't access the session cookie.

#adamcunnington here is how you can debug it:
#app.route("/auth")
def auth() -> Union[Tuple[str, int], werkzeug.Response]:
# Check these two values
print(flask.request.args.get('state'), flask.session.get('_google_authlib_state_'))
token = oauth.google.authorize_access_token()
user = oauth.google.parse_id_token(token)
flask.session["user"] = user
return flask.redirect("/")
Check the values in request.args and session to see what's going on.
Maybe it is because Flask session not persistent across requests in Flask app with Gunicorn on Heroku

How I Fix My Issue
install old version of authlib it work fine with fastapi and flask
Authlib==0.14.3
For Fastapi
uvicorn==0.11.8
starlette==0.13.6
Authlib==0.14.3
fastapi==0.61.1
Imporantt if using local host for Google auth make sure get https certifcate
install chocolatey and setup https check this tutorial
https://dev.to/rajshirolkar/fastapi-over-https-for-development-on-windows-2p7d
ssl_keyfile="./localhost+2-key.pem" ,
ssl_certfile= "./localhost+2.pem"
--- My Code ---
from typing import Optional
from fastapi import FastAPI, Depends, HTTPException
from fastapi.openapi.docs import get_swagger_ui_html
from fastapi.openapi.utils import get_openapi
from starlette.config import Config
from starlette.requests import Request
from starlette.middleware.sessions import SessionMiddleware
from starlette.responses import HTMLResponse, JSONResponse, RedirectResponse
from authlib.integrations.starlette_client import OAuth
# Initialize FastAPI
app = FastAPI(docs_url=None, redoc_url=None)
app.add_middleware(SessionMiddleware, secret_key='!secret')
#app.get('/')
async def home(request: Request):
# Try to get the user
user = request.session.get('user')
if user is not None:
email = user['email']
html = (
f'<pre>Email: {email}</pre><br>'
'documentation<br>'
'logout'
)
return HTMLResponse(html)
# Show the login link
return HTMLResponse('login')
# --- Google OAuth ---
# Initialize our OAuth instance from the client ID and client secret specified in our .env file
config = Config('.env')
oauth = OAuth(config)
CONF_URL = 'https://accounts.google.com/.well-known/openid-configuration'
oauth.register(
name='google',
server_metadata_url=CONF_URL,
client_kwargs={
'scope': 'openid email profile'
}
)
#app.get('/login', tags=['authentication']) # Tag it as "authentication" for our docs
async def login(request: Request):
# Redirect Google OAuth back to our application
redirect_uri = request.url_for('auth')
print(redirect_uri)
return await oauth.google.authorize_redirect(request, redirect_uri)
#app.route('/auth/google')
async def auth(request: Request):
# Perform Google OAuth
token = await oauth.google.authorize_access_token(request)
user = await oauth.google.parse_id_token(request, token)
# Save the user
request.session['user'] = dict(user)
return RedirectResponse(url='/')
#app.get('/logout', tags=['authentication']) # Tag it as "authentication" for our docs
async def logout(request: Request):
# Remove the user
request.session.pop('user', None)
return RedirectResponse(url='/')
# --- Dependencies ---
# Try to get the logged in user
async def get_user(request: Request) -> Optional[dict]:
user = request.session.get('user')
if user is not None:
return user
else:
raise HTTPException(status_code=403, detail='Could not validate credentials.')
return None
# --- Documentation ---
#app.route('/openapi.json')
async def get_open_api_endpoint(request: Request, user: Optional[dict] = Depends(get_user)): # This dependency protects our endpoint!
response = JSONResponse(get_openapi(title='FastAPI', version=1, routes=app.routes))
return response
#app.get('/docs', tags=['documentation']) # Tag it as "documentation" for our docs
async def get_documentation(request: Request, user: Optional[dict] = Depends(get_user)): # This dependency protects our endpoint!
response = get_swagger_ui_html(openapi_url='/openapi.json', title='Documentation')
return response
if __name__ == '__main__':
import uvicorn
uvicorn.run(app, port=8000,
log_level='debug',
ssl_keyfile="./localhost+2-key.pem" ,
ssl_certfile= "./localhost+2.pem"
)
.env file
GOOGLE_CLIENT_ID=""
GOOGLE_CLIENT_SECRET=""
Google Console Setup

How to set up CORS in CherryPy

Overview
When creating a post request from my website to my Python server running CherryPy, I receive the error Access to XMLHttpRequest has been blocked by CORS policy: Request header field content-type is not allowed by Access-Control-Allow-Headers in preflight response. . I was able to get away with the problem temporarily with one of the "CORS Everywhere" browser extensions, but
Due to recent updates, the extensions have not yet been updated to be working again.
The website involved needs to eventually be used by many in my local complex without the browser extension, so once the extensions get updated, it does not really matter one way or another, as I cannot rely on these extensions, and force everyone to use them (when there is obviously a fix that would make an extension not necessary).
I figure that perhaps the solutions are outdated, but am not sure.
Here is the relevant code:
On the server side (CherryPy/Python):
The CherryPy Python function being called, from the website post request
#cherrypy.expose
#cherrypy.tools.json_in()
def add_meeting(self):
data = None
id = None
start_time = None
end_time = None
title = None
userlist = None
result = {"operation": "request", "result": "success"}
if cherrypy.request.method == "POST":
data = cherrypy.request.json
id = data["id"]
start_time = data["start_time"]
end_time = data["end_time"]
title = data["title"]
userlist = data["userlist"]
# Rest of relevant code in function is left out, to take up less
# space and not post irrelevant code. That being said, I am
# positive the logic is correct, as it originally ran smoothly
# with a "Cors Everywhere" Browser Extension.
return result
Here is the area where I set up and run CherryPy
def main():
# Create the configuration file parser object and start the CherryPy server
config = ConfigParser.ConfigParser()
config.read(CONFIG_FILE)
port = config.getint('Meta', 'port')
host = config.get('Meta', 'host')
cherrypy.config.update({'server.socket_port': port,
'server.socket_host': host,
'tools.CORS.on': True})
cherrypy.quickstart(Coordinator(config))
main()
Here is the config file mentioned in the code above (CONFIG_FILE)
[Meta]
host = 0.0.0.0
port = 3000
# Rest is left out, as it is irrelevant with problem
The solutions I have tried implementing
The inclusion of the following function above the main function:
def CORS():
cherrypy.response.headers["Access-Control-Allow-Origin"] = "*"
with cherrypy.tools.CORS = cherrypy.Tool('before_handler', CORS)
2. Adding " 'cors.expose.on': True " to cherrypy.config.update above
3. Using this cherrypy-cors Python library I found online: https://pypi.org/project/cherrypy-cors/
4. The inclusion of headers in the config.update portion of the Python file
5. Adding "#cherrypy.tools.accept(media='application/json')" before "def add_meeting"
Conclusion
I've tried the solutions above together, separately, some with and without the others, and I am still stuck. Maybe some of these solutions are partially correct, and there is something extra needed with my code. I am not sure; I just cannot get it working. I do not have much experience with web development before this, so maybe (and hopefully) the solution is extremely simple. I know the code works, I just cannot get it running without a working "Cors Everywhere" browser extension for every user.
As for the versions I am running: I am using CherryPy 14.2.0 and Python 2.7.6
Any help would mean the absolute world to me, thank you.

So first, you need to set pre-flight headers when processing OPTIONS request, you can list allowed methods there.
Then, you also need to enable the cors.expose tool.
There's some usage hints in the docstring of cherrypy-cors. For example, when using a MethodDispatcher, you could just decorate an OPTIONS handler method with #cherrypy_cors.tools.preflight() instead of doing this in every HTTP handler.
Here's a simple traversal example (without a method dispatcher). To test it, visit http://127.0.0.1/ and it will make requests against http://localhost:3333/add_meeting which is a different Origin in terms of CORS ('localhost' != '127.0.0.1').
"""Example of CORS setup using cherrypy-cors library."""
import cherrypy
import cherrypy_cors
# Python 2 compat: make all classes new-style by default
__metaclass__ = type # pylint: disable=invalid-name
class WebRoot:
"""Root node for HTTP handlers."""
#cherrypy.expose
def index(self): # pylint: disable=no-self-use
"""Render a web page handling request against ``/``.
Contains client JS snippet which will query the API endpoint.
It will be executed by the browser while loading the page.
"""
return """<html>
<script type="text/javascript">
async function addMeeting() {
/*
* Example coroutine for querying /add_meeing
* HTTP endpoint. It uses localhost as in the URL.
* For testing CORS, make sure to visit
* http://127.0.0.1/ which is a different origin
* from browser's perspective.
* /
const request_payload = {
some: 'data',
listed: ['h', 'er', 'e'],
}
try {
const resp = await fetch(
'http://localhost:3333/add_meeting',
{
method: 'POST',
mode: 'cors', // Required for customizing HTTP request headers
credentials: 'same-origin',
headers: {
'Content-Type': 'application/json; charset=UTF-8', // Required for ``cherrypy.tools.json_in`` to identify JSON payload and parse it automatically
},
body: JSON.stringify(request_payload),
},
)
const json_resp = await resp.json()
console.log(json_resp) // Will print: {"method": "POST", "payload": {"listed": ["h", "er", "e"], "some": "data"}}
} catch (e) {
console.warn('Exception: ' + e)
}
}
async function main() {
await addMeeting()
}
main() // Entry point
</script>
</html>""" # noqa: E501
#cherrypy.expose
#cherrypy.tools.json_in() # turn HTTP payload into an object; also checking the Content-Type header
#cherrypy.tools.json_out() # turn ``return``ed Python object into a JSON string; also setting corresponding Content-Type
def add_meeting(self):
"""Handle HTTP requests against ``/add_meeting`` URI."""
if cherrypy.request.method == 'OPTIONS':
# This is a request that browser sends in CORS prior to
# sending a real request.
# Set up extra headers for a pre-flight OPTIONS request.
cherrypy_cors.preflight(allowed_methods=['GET', 'POST'])
if cherrypy.request.method == 'POST':
return {'method': 'POST', 'payload': cherrypy.request.json}
return {'method': 'non-POST'}
def main():
"""Set up and run the web app.
Initializes CORS tools.
Sets up web server socket.
Enables the CORS tool.
"""
cherrypy_cors.install()
cherrypy.config.update({
'server.socket_host': '127.0.0.1',
'server.socket_port': 3333,
'cors.expose.on': True,
})
cherrypy.quickstart(WebRoot())
__name__ == '__main__' and main() # pylint: disable=expression-not-assigned

Trouble using bottle-jwt decorator (not works)

Having some trouble using this plugin https://github.com/agile4you/bottle-jwt/
It seems to not work as I expected, down below my code:
import bottle
from Py.engine import *
from bottle_jwt import (JWTProviderPlugin, jwt_auth_required)
class AuthBackend(object):
user = {'id': 1237832, 'username': 'pav', 'password': '123', 'data': {'sex': 'male', 'active': True}}
def authenticate_user(self, username, password):
"""Authenticate User by username and password.
Returns:
A dict representing User Record or None.
"""
if username == self.user['username'] and password == self.user['password']:
return self.user
return None
def get_user(self, user_id):
"""Retrieve User By ID.
Returns:
A dict representing User Record or None.
"""
if user_id == self.user['id']:
return {k: self.user[k] for k in self.user if k != 'password'}
return None
app = bottle.Bottle()
server_secret = 'secret'
provider_plugin = JWTProviderPlugin(
keyword='jwt',
auth_endpoint='/login',
backend=AuthBackend(),
fields=('username', 'password'),
secret=server_secret,
ttl=30
)
app.install(provider_plugin)
#app.route('/')
#jwt_auth_required
def index():
return open('Html/index.html', 'r').read()
#app.post('/login')
def login():
return open('Html/login.html', 'r').read()
#app.get('/login')
def login():
return open('Html/login.html', 'r').read()
def run_server():
bottle.run(app=app, host='localhost', port=8080, debug=True, reloader=True)
# Main
if __name__ == '__main__':
run_server()
Once running, if I open browser On 127.0.0.1/8080 i get back a blank page with the string "{"AuthError": ["Cannot access this resource!"]}"
Which is Fine, it means that I'm not allowed to open index.html file (Cool: #jwt_auth_required worked)
Digging in source file I found a function named validate_token() with:
if not token:
logger.debug("Forbidden access")
raise JWTForbiddenError('Cannot access this resource!')
Here is the exception
except JWTForbiddenError as error:
bottle.response.content_type = b('application/json')
bottle.response._status_line = b('403 Forbidden')
return {"AuthError": error.args}
So, is there any way to redirect me on my login.html page if token does not match or is absent?
Plugin includes some way to do that or is just an API pckg?

That's not how JWT concept is supposed to be used. JWT are for RESTFul.
You need to make the server as REST API and on the client use JS
libraries such as AngularJs / Vue.js etc.,
Coming to the question about the plugin:
provider_plugin = JWTProviderPlugin(
keyword='jwt',
auth_endpoint='/login',
backend=AuthBackend(),
fields=('username', 'password'),
secret=server_secret,
ttl=30
)
auth_endpoint='/login' is to give a custom endpoint for authorization where the Bottle_JWT methods are looking for credentials to validate and generate JWT for.
I created a mock just to construct a response and this is how it should be used.
Once you pass the correct credential, the plugin responds with the JWT and expire which you have to intercept in authorized calls and add as request headers
Hope this helps.