TechEmpower benchmarks

[tomchristie]

This’d be a good one for a contributor to jump on.

Adding TechEmpower benchmarks for Responder. I’d suggest copying the starlette case, and adapting it accordingly. https://github.com/TechEmpower/FrameworkBenchmarks/tree/master/frameworks/Python/starlette

PR would be against the TechEmpower repo, not here. They run continuos benchmarks against the suite, so the performance section could be updated once there’s a run that includes Responder.

You’ll want to use an async DB connector, as per the starlette case.

[5hirish]

Working on it, except I am getting 404s for requests defined following the first API request.

import responder

api = responder.API()

@api.route("/hello")
async def greet_world(req, resp):
    resp.text = f"Hello, world!"

@api.route("/bye")
async def bye_world(req, resp):
    resp.media = {"data": "Bye, world!"}

if __name__ == '__main__':
    api.run()

Here /hello returns 200 but /bye returns 404. If I change the sequence of declaration then /bye works.

[5hirish]

https://github.com/5hirish/FrameworkBenchmarks/blob/responder/frameworks/Python/responder/app.py

[kennethreitz]

that's because you named your function the same thing twice

[5hirish]

@kennethreitz sorry, that was copy-paste typo, fixed it, but still, the issue is reproducible, can try running the app.py from my fork. Or am I declaring something wrong here?

[5hirish]

@tomchristie I have created a PR on the TechEmpower repo: (Still, haven't resolved the issue mentioned above. Maybe it is happening only for my system...?) TechEmpower/FrameworkBenchmarks#4122

[sheb]

@5hirish I had the same problem with v.0.0.3, but seems to be fixed on master

[5hirish]

Thanks for informing... BTW I am getting this responder: ERROR: Framework is not accepting requests from client machine in the CI console might be due to the above issue I guess...In my local, at least the first API was working...Will have to wait for the next release.

[5hirish]

Tried running the docker file in local, to pin point the error in docker run. I executed the the gunicorn task separately
gunicorn app:app --log-level debug -k uvicorn.workers.UvicornWorker -c responder_conf.py
I am getting the following error:

Exception in 'lifespan' protocol
Traceback (most recent call last):
  File "/home/felix/.pyenv/versions/3.6.5/envs/pyw365/lib/python3.6/site-packages/uvicorn/lifespan.py", line 29, in run
    await self.asgi(self.receive, self.send)
  File "/home/felix/.pyenv/versions/3.6.5/envs/pyw365/lib/python3.6/site-packages/uvicorn/middleware/message_logger.py", line 51, in __call__
    inner = self.app(self.scope)
  File "/home/felix/.pyenv/versions/3.6.5/envs/pyw365/lib/python3.6/site-packages/responder/api.py", line 100, in __call__
    path = scope["path"]
KeyError: 'path'

[tomchristie]

Make sure to update to the latest uvicorn.

You’re seeing an integration issue there as uvicorn is sending ASGI lifespan messages that responder doesn’t yet handle. However the latest version of uvicorn should deal with that gracefully.

[5hirish]

Will update the benchmarks when the PR is approved and merged. @tomchristie responder doesn't need Jinja dependency but the api.template() call was not taking template file path as an argument, hence reverted to external Jinja dependency and usage. If this will affect the performance any form then let me know.

[tomchristie]

@5hirish good work. They have a continuous benchmarking suite running, tho it has a turnover time of 2 or 3 days. Check https://tfb-status.techempower.com/ and you should see their latest results soon enough.

[therumbler]

the Tech Empower benchmark has completed, and includes Responder's results now.
https://www.techempower.com/benchmarks/#section=test&runid=a0d523de-091b-4008-b15d-bd4c8aa25066

[5hirish]

@tomchristie can close this issue. PR: Update ReadMe #130

[kennethreitz]

@therumbler we were included in the last batch -- any idea why?

[tomchristie]

I guess it must have just missed the round 17 cutoff.

[waghanza]

@tomchristie I have tested tfb

• Cloning https://github.com/TechEmpower/FrameworkBenchmarks
• Updating deps -> pip3 install -r requirements.txt
• Running -> gunicorn app:app -k uvicorn.workers.UvicornWorker -c responder_conf.py

leads me to a socket.gaierror: [Errno -2] Name or service not known