feat: python server

[WinPlay02]

• every execution needs a unique id to identify calculated values
• executes compiled safe-ds code in memory (using a finder / loader)
• pipelines are executed in child processes
• communication from child processes to the main process uses a shared memory queue
• allows connections via websocket (and http)
• listens on a provided port (default 5000)
• uses flask and flask-sock for HTTP and WS
• uses gevent as WSGI implementation

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (1cb1ba1) 100.00% compared to head (e949194) 100.00%.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #6 +/- ## ========================================== Coverage 100.00% 100.00% ========================================== Files 2 6 +4 Lines 13 317 +304 ========================================== + Hits 13 317 +304 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[github-actions[bot]]

🦙 MegaLinter status: ✅ SUCCESS

Descriptor	Linter	Files	Fixed	Errors	Elapsed time
✅ PYTHON	black	8	0	0	0.62s
✅ PYTHON	mypy	8		0	1.7s
✅ PYTHON	ruff	8	0	0	0.03s
✅ REPOSITORY	git_diff	yes		no	0.02s

See detailed report in MegaLinter reports _Set VALIDATE_ALL_CODEBASE: true in mega-linter.yml to validate all sources, not only the diff_

_MegaLinter is graciously provided by _

[WinPlay02]

@lars-reimann Collecting coverage information across child processes (multiprocessing) does not seem to work. I tried adding this section to .coveragerc:

[run]
parallel = True
concurrency = multiprocessing, thread

This also does not work, it only causes the testing process to freeze when calling multiprocessing.Manager() during setup. Is there any (other) way to make coverage work with multiprocessing?

[lars-reimann]

I have no experience with this. Maybe the pytest-cov documentation helps.

[WinPlay02]

It seems to be a problem only on Windows, as the CI passes. I annotated the tests skip execution on Windows for now. Although they work when coverage calculation does not include multiprocessing (in .coveragerc, see above).

[lars-reimann]

It seems to be a problem only on Windows, as the CI passes. I annotated the tests skip execution on Windows for now. Although they work when coverage calculation does not include multiprocessing (in .coveragerc, see above).

I can't reproduce this on a Windows machine. pytest --cov=safeds_runner --cov-report=xml works fine even when the skipped tests are included.

[WinPlay02]

I think I addressed all the suggestions. Now, the PipelineManager class takes the place of most of the global variables. It is created at startup and used in the websocket session. But I couldn't get rid of the current_pipeline global in pipeline_manager.py as it is used for saving placeholders.

[lars-reimann]

:tada: This PR is included in version 0.2.0 :tada:

The release is available on:

• v0.2.0
• GitHub release

Your semantic-release bot :package::rocket: