httpx | a lightweight HTTP server | HTTP library

I am trying to learn and use AWS Cognito User Pools and integrate with and API implemented with Python FastAPI. So far I am using Authorisation Code Flow with my Cognito user pool redirecting to an endpoint on FastAPI to resolve the code challenge. The source code is appended at the end of this query.

The API has the following endpoints:

1. Root endpoint [ / ]: Redirects the browser to the sign-in page of my AWS Cognito user pool.
2. Redirect endpoint [ /aws_cognito_redirect ]: Activated after successful sign-in to user-pool. Receives a code challenge from the cognito user pool. In the code shown below, the aws_cognito_redirect endpoint resolves the code challenge by sending the code challenge, redirect_uri, client_id etc. through to the AWS Cognito user pool oauth2/token endpoint. I can see in the console log output that the identity, access and refresh tokens were successfully retrieved.

The FastAPI will additionally have some protected endpoints that will be called from a web application. Also there will be a web form to interact with the endpoints.

At this stage I could implement and host the webforms with FastAPI jinja2 templates. If I went with this option, presumably I could have the /aws_cognito_redirect endpoint return the tokens in a HTTP only session cookie. That way each subsequent client request would automatically include the cookie with no tokens exposed in the browser local storage. I am aware that I would have to deal with XSRF/CSRF with this option.

Alternatively I could implement the front end using Angular/React. Presumably, the recommended practice appears to be that I would have to reconfigure the authorisation flow to be Auth Code with PKCE? In that case, the Angular/React web client would then be communicating with AWS Cognito directly, to retrieve tokens that would be forwarded to the FastAPI endpoints. These tokens would be stored in the browser's local storage and then sent in an Authorisation Header for each subsequent request. I am aware that this approach is subject to XSS attacks.

Of the two, given my requirements, I think that I am leaning towards hosting the webapp on FastAPI using jinja2 templates and returning a HTTP Only session cookie on successfull sign in.

If I chose this implementation route, is there a FastAPI feature or Python library that allows an endpoint to be decorated/marked with auth required that would inspect the presence of the session cookie and perform token verification?

FastAPI

If I want to request a API 1000 times (send network request and handle response), it will begin to handle response after sending all 1000 request first, and then handle response.

Can I tell asyncio prefer to return back await position code if it's done?

I'm trying to figure out a way to limit the number of concurrent HTTP requests made to a server using Python's asyncio and httpx module. I came across this StackOverflow answer.

It proposes asyncio.Semaphore for stopping multiple consumers from making too many requests. While this answer works perfectly, it uses explicit loop construction, not asyncio.run. When I replace the explicit loop construction with asyncio.run, the behavior of the code changes. Instead of doing all 9 requests, now it just executes three requests and then stops.

I need to upload a big file to a third party service. This third party service gives me a list of urls and byteranges:

Valid Response:

I need to download a list of sites/URLs (which can vary over time) and I currently use multiprocessing.Manager().Queue() to submit and update said list.
I have to check each URL/task every second: hence each task will basically never ends (until a specific condition mets, like user interrupting). I thought that multiprocessing.Process() combined with asyncio and a good async HTTP client would solve the problem. Unfortunately, I still get a very high CPU usage after submitting 50 or more URLs. You will notice the difference yourselves when the tasks are not doing any requests - running mock_request() - and when they are - running do_request() -.

Here is an example to reproduce each case (press CTRL+C to end it gracefully at any time).

I have a very basic Django APM agent setup:

I have exactly the same problem as mentioned in PIP install rasa-x takes forever. In the Rasa installation guide they say, you have to create an environment first. Everytime I do: conda create --name rasa python==3.7.6 it automatically downloads pip-20.3.3. If I now try the pip install --upgrade pip==20.2 command it shows the following error: Error. What did I do wrong? Thanks for the help!

**Update: python -m pip install --upgrade pip==20.2 worked, but now there is another problem when trying to install Rasa-X:Rasa-X installation error

here is the code

I'm trying to query the following website

where Access Number will be a fixed value of 8778791867

And the PIN will be dynamic.

From the normal browser am able to check if it's valid or invalid pin.

But using my code below, I'm unable to get the exact answer as the browser get. as i keep getting invalid for all entries!