tests: benchmarks, xdist

[albertedwardson]

This pull request introduces:

• a CodSpeed-powered benchmarking workflow with addition of benchmark test suites
• parallelization of existing tests using pytest-xdist

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.97%. Comparing base (7576e18) to head (840415b).
⚠️ Report is 25 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #247   +/-   ##
=======================================
  Coverage   97.97%   97.97%           
=======================================
  Files           9        9           
  Lines         493      493           
  Branches       84       84           
=======================================
  Hits          483      483           
  Misses          6        6           
  Partials        4        4           

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[albertedwardson]

note to reviewers: the BENCHMARK_CASES dict is generated with help of AI, no need to validate it diligently, until you will be ok with enabling codspeed in ci. after that- I will do a self-review, disable supression of exceptions (and maybe do asserts with expected output, idk)

[codspeed-hq]

Congrats! CodSpeed is installed 🎉

🆕 31 new benchmarks were detected.

You will start to see performance impacts in the reports once the benchmarks are run from your default branch.

Detected benchmarks

• test_benchmark_encoding[html_body_declared_encoding]: 290.3 µs
• test_benchmark_encoding[read_bom]: 38.1 µs
• test_benchmark_url[path_to_file_uri]: 166 µs
• test_benchmark_encoding[html_to_unicode]: 338 µs
• test_benchmark_encoding[to_unicode]: 24.5 µs
• test_benchmark_url[parse_url]: 67.6 µs
• test_benchmark_encoding[http_content_type_encoding]: 59.5 µs
• test_benchmark_http[headers_raw_to_dict]: 40 µs
• test_benchmark_url[safe_url_string]: 4.5 ms
• test_benchmark_url[add_or_replace_parameter]: 636.5 µs
• test_benchmark_url[add_or_replace_parameters]: 316.8 µs
• test_benchmark_url[canonicalize_url]: 528.7 µs
• test_benchmark_url[parse_data_uri]: 604.2 µs
• test_benchmark_html[replace_entities]: 428.1 µs
• test_benchmark_url[url_query_cleaner]: 275.4 µs
• test_benchmark_url[is_url]: 26.8 µs
• test_benchmark_url[safe_download_url]: 562.2 µs
• test_benchmark_encoding[resolve_encoding]: 75.7 µs
• test_benchmark_url[url_query_parameter]: 91.2 µs
• test_benchmark_html[unquote_markup]: 143.2 µs
• ...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

---

[AdrianAtZyte]

I did notice the upstream docs mention the need for id-token: write permission in actions.

That said, I suspect the main reason https://codspeed.io/scrapy/w3lib/branches/albertedwardson%3Apytest-benchmark does not show much is because until the benchmarks run on the main branch there is no baseline to compare with.

[albertedwardson]

id-token: write wasn't the cause; the logs shows NOTICE: codspeed is enabled, but no performance measurement will be made since it's running in an unknown environment, although codspeed should definitely know that it's in CI

maybe it doesn't like being run through tox?

[albertedwardson]

yeah, tox was the cause... I'll try find a way to convince codspeed ​​that running with tox is ok, it only supports nox...

[albertedwardson]

ok, pypy is not supported: CodSpeedHQ/pytest-codspeed#34

[albertedwardson]

I thought everything was working locally because I didn't notice the notice log :)

NOTICE: codspeed is enabled, but performance measurements won't be performed because it's running in an unknown environment.

This means it doesn't even attempt to run the instrumentation, but only checks whether the tests pass

[albertedwardson]

idk how to do this nicely without excessive copypaste from real test cases

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.