chore: 📦 Update dependencies

[PasteurBot]

This PR updates the lockfile to the latest versions of the dependencies.
Please review the changes and merge when ready.

[codecov]

Codecov Report

❌ Patch coverage is 94.11765% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 76.96%. Comparing base (685de1b) to head (fd87beb).

Files with missing lines	Patch %	Lines
tesseract_core/runtime/cli.py	88.88%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #613      +/-   ##
==========================================
+ Coverage   67.22%   76.96%   +9.73%     
==========================================
  Files          32       32              
  Lines        4519     4528       +9     
  Branches      743      743              
==========================================
+ Hits         3038     3485     +447     
+ Misses       1237      737     -500     
- Partials      244      306      +62     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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

[apaleyes]

typer no longer depends on click, with some breaking changes: https://typer.tiangolo.com/tutorial/click/#breaking-changes

[jpbrodrick89]

I'll try have a go at fixing this. I put some fixes in last week but either they were reverted or not robust enough.

[PasteurBot]

Benchmark Results

Benchmarks use a no-op Tesseract to measure pure framework overhead.

🚀 0 faster, ⚠️ 0 slower, ✅ 36 unchanged

✅ No significant performance changes detected.

Full results

Benchmark	Baseline	Current	Change	Status
api/apply_1,000	0.494ms	0.499ms	+0.8%	✅
api/apply_100,000	0.495ms	0.494ms	-0.1%	✅
api/apply_10,000,000	0.495ms	0.497ms	+0.4%	✅
cli/apply_1,000	1723.556ms	1690.671ms	-1.9%	✅
cli/apply_100,000	1704.208ms	1708.697ms	+0.3%	✅
cli/apply_10,000,000	1758.770ms	1754.779ms	-0.2%	✅
decoding/base64_1,000	0.026ms	0.028ms	+9.1%	✅
decoding/base64_100,000	0.609ms	0.605ms	-0.6%	✅
decoding/base64_10,000,000	70.468ms	70.080ms	-0.6%	✅
decoding/binref_1,000	0.185ms	0.185ms	-0.1%	✅
decoding/binref_100,000	0.236ms	0.240ms	+2.0%	✅
decoding/binref_10,000,000	8.763ms	8.675ms	-1.0%	✅
decoding/json_1,000	0.098ms	0.099ms	+0.2%	✅
decoding/json_100,000	9.322ms	9.363ms	+0.4%	✅
decoding/json_10,000,000	1121.222ms	1114.438ms	-0.6%	✅
encoding/base64_1,000	0.028ms	0.027ms	-0.7%	✅
encoding/base64_100,000	0.143ms	0.143ms	-0.6%	✅
encoding/base64_10,000,000	21.286ms	21.196ms	-0.4%	✅
encoding/binref_1,000	0.279ms	0.276ms	-0.9%	✅
encoding/binref_100,000	0.447ms	0.444ms	-0.8%	✅
encoding/binref_10,000,000	17.854ms	16.745ms	-6.2%	✅
encoding/json_1,000	0.143ms	0.142ms	-1.1%	✅
encoding/json_100,000	14.613ms	14.464ms	-1.0%	✅
encoding/json_10,000,000	1536.316ms	1536.569ms	+0.0%	✅
http/apply_1,000	2.908ms	2.919ms	+0.4%	✅
http/apply_100,000	8.662ms	8.187ms	-5.5%	✅
http/apply_10,000,000	661.902ms	665.194ms	+0.5%	✅
roundtrip/base64_1,000	0.059ms	0.057ms	-2.6%	✅
roundtrip/base64_100,000	0.766ms	0.764ms	-0.3%	✅
roundtrip/base64_10,000,000	91.831ms	91.604ms	-0.2%	✅
roundtrip/binref_1,000	0.481ms	0.478ms	-0.6%	✅
roundtrip/binref_100,000	0.685ms	0.692ms	+0.9%	✅
roundtrip/binref_10,000,000	25.702ms	25.657ms	-0.2%	✅
roundtrip/json_1,000	0.248ms	0.244ms	-1.5%	✅
roundtrip/json_100,000	21.765ms	21.757ms	-0.0%	✅
roundtrip/json_10,000,000	2645.884ms	2615.422ms	-1.2%	✅

• Runner: Linux 6.17.0-1015-azure x86_64

[jpbrodrick89]

@dionhaefner I think the latest commit should fix the failing "oldest" test that failed on my previous commit by bumping the lower bound of typer to 0.16 (0.15 used click 0.8.0 which had a different function signature for TyperArgument.make_metavar(), previously we got around this by pinning click but my changes remove any explicit click dependency).

Docs build will still fail, this is because we currently depend on sphinx_click, the ideal solution would be to switch to sphinxcontrib-typer, but this doesn't yet support typer 0.26, three options I see:

1. Keep using sphinx_click and pin typer<0.26 on docs build only
2. Migrate to sphinxcontrib-typer and pin typer<0.26 on docs build only until it supports typer 0.26
3. Migrate to sphinxcontrib-typer and (monkey)patch it to support typer 0.26 (by doing something like import typer._click as sphinxcontrib_typer.click.

What do you think?

[dionhaefner]

Thanks for looking into it @jpbrodrick89. Could try (2), and use (1) as a fallback if that doesn't work ootb?

[jpbrodrick89]

sphinxcontrib-typer now supports typer 0.26, but unfortunately its display isn't great. It just renders raw --help terminal output with a fixed width. This means that the final column (with the actual prose) is only ~four words per line and gets line wrapped to ~20 lines per option. As such I pinned typer<0.26 and stuck with sphinx_click for now while we take our time to deliberate on a more complete solution, the options I see long-term are:

1. monkeypatch sphinx_click
2. roll our own package
3. extend/make a PR to sphinxcontrib-typer to provide the proper format

I'm not really sure how to fix the 403 conda issue, does anyone have any insight into this? I think switching conda-forge instead of repo.anaconda might fix it (vectoradd test was flaky) or pinning to a different image than :latest for the containerised calls.