chore(deps): remove unused Python dependencies

[Yicong-Huang]

What changes were proposed in this PR?

Removes three Python packages that nothing uses and moves ruff to the dev-only requirements file:

Package	File	Why it can go
Deprecated==1.2.14	requirements.txt	Last import removed in #2847; no installed package requires it
protobuf==7.34.1	requirements.txt	No google.protobuf import anywhere; betterproto 2.0 runtime does not depend on it (only its optional extras do)
pybase64==1.3.2	operator-requirements.txt	Added in #2167, but the operator templates have always used the stdlib base64 (see ImageUtility.scala in that PR) — zero references to pybase64 in the entire git history
ruff==0.14.7	requirements.txt → dev-requirements.txt	Lint tool, not a runtime dependency; keeping it in requirements.txt wrongly puts it in LICENSE-binary-python

Follow-up changes required by the move/removals:

• .github/workflows/build.yml: the "Lint with Ruff" step now runs after "Install dev dependencies", since ruff comes from dev-requirements.txt.
• amber/LICENSE-binary-python: dropped the four matching entries. wrapt stays — it is still a hard dependency of aiobotocore, not just of Deprecated.
• Added # Not imported directly: ... comments to the transitive pins (setuptools, python-dateutil, s3fs/aiobotocore, SQLAlchemy/pg8000, scikit-image) so they don't look unused in future audits.

Any related issues, documentation, discussions?

Closes #6101

How was this PR tested?

Uninstalled the three packages from the local venv (pip uninstall -y protobuf Deprecated pybase64), then ran the existing test suite and the same lint commands CI runs:

cd amber && pytest -m "not integration" -q
# 551 passed, 1 deselected in 64s
cd amber && ruff check src/main/python src/test/python && ruff format --check src/main/python src/test/python
# All checks passed! 190 files already formatted

No new tests: dependency-manifest-only change; CI's check_binary_deps.py license gate verifies the LICENSE-binary-python entries match the installed set.

Was this PR authored or co-authored using generative AI tooling?

Generated-by: Claude Code (claude-fable-5)

[github-actions]

Automated Reviewer Suggestions

Based on the git blame history of the changed files, we recommend the following reviewers:

• Contributors with relevant context: @Ma77Ball
  You can notify them by mentioning @Ma77Ball in a comment.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.09%. Comparing base (4c9d30a) to head (d4f5714).

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #6103      +/-   ##
============================================
- Coverage     59.11%   59.09%   -0.03%     
- Complexity     3197     3199       +2     
============================================
  Files          1132     1132              
  Lines         43674    43674              
  Branches       4734     4734              
============================================
- Hits          25820    25810      -10     
- Misses        16420    16432      +12     
+ Partials       1434     1432       -2     

Flag	Coverage Δ
access-control-service	70.00% <ø> (ø)
agent-service	44.59% <ø> (ø)
amber	63.10% <ø> (+0.01%)	⬆️
computing-unit-managing-service	0.00% <ø> (ø)
config-service	52.30% <ø> (ø)
file-service	62.81% <ø> (ø)
frontend	51.31% <ø> (-0.07%)	⬇️
notebook-migration-service	78.57% <ø> (ø)
pyamber	91.15% <ø> (ø)
workflow-compiling-service	55.14% <ø> (ø)

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

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[github-actions]

⚠️ Benchmark changes need a look

🟢 4 better · 🔴 2 worse · ⚪ 9 noise (<±5%) · 0 without baseline

Compared against main 4c9d30a benchmarked on this same runner, so the delta is largely free of cross-runner hardware noise. The "7d avg" column still reflects the gh-pages dashboard. Treat <±5% as noise unless repeated.

Dashboard · Run

	config	throughput	MB/s	latency	max Δ latest / 7d
🔴	bs=10 sw=10 sl=64	404	0.247	23,933/28,705/28,705 us	🟢 -15.4% / 🔴 +89.6%
🔴	bs=100 sw=10 sl=64	811	0.495	122,651/132,276/132,276 us	🟢 -9.4% / 🔴 +22.5%
⚪	bs=1000 sw=10 sl=64	947	0.578	1,056,931/1,104,341/1,104,341 us	⚪ within ±5% / 🔴 +7.4%

Baseline details

Latest main 4c9d30a from same runner

config	metric	PR	latest main	7d avg	Δ latest	Δ 7d
bs=10 sw=10 sl=64	throughput	404 tuples/sec	422 tuples/sec	776.36 tuples/sec	-4.3%	-48.0%
bs=10 sw=10 sl=64	MB/s	0.247 MB/s	0.258 MB/s	0.474 MB/s	-4.3%	-47.9%
bs=10 sw=10 sl=64	p50	23,933 us	21,748 us	12,636 us	+10.0%	+89.4%
bs=10 sw=10 sl=64	p95	28,705 us	33,941 us	15,143 us	-15.4%	+89.6%
bs=10 sw=10 sl=64	p99	28,705 us	33,941 us	18,954 us	-15.4%	+51.4%
bs=100 sw=10 sl=64	throughput	811 tuples/sec	851 tuples/sec	985.33 tuples/sec	-4.7%	-17.7%
bs=100 sw=10 sl=64	MB/s	0.495 MB/s	0.519 MB/s	0.601 MB/s	-4.6%	-17.7%
bs=100 sw=10 sl=64	p50	122,651 us	116,109 us	101,671 us	+5.6%	+20.6%
bs=100 sw=10 sl=64	p95	132,276 us	146,044 us	107,939 us	-9.4%	+22.5%
bs=100 sw=10 sl=64	p99	132,276 us	146,044 us	113,798 us	-9.4%	+16.2%
bs=1000 sw=10 sl=64	throughput	947 tuples/sec	954 tuples/sec	1,016 tuples/sec	-0.7%	-6.8%
bs=1000 sw=10 sl=64	MB/s	0.578 MB/s	0.582 MB/s	0.62 MB/s	-0.7%	-6.8%
bs=1000 sw=10 sl=64	p50	1,056,931 us	1,045,746 us	989,693 us	+1.1%	+6.8%
bs=1000 sw=10 sl=64	p95	1,104,341 us	1,105,873 us	1,028,327 us	-0.1%	+7.4%
bs=1000 sw=10 sl=64	p99	1,104,341 us	1,105,873 us	1,059,969 us	-0.1%	+4.2%

Raw CSV

config_idx,batch_size,schema_width,string_len,num_batches,total_ms,total_tuples,total_bytes,tuples_per_sec,mb_per_sec,lat_p50_us,lat_p95_us,lat_p99_us
0,10,10,64,20,494.79,200,128000,404,0.247,23932.72,28705.45,28705.45
1,100,10,64,20,2466.53,2000,1280000,811,0.495,122651.30,132276.28,132276.28
2,1000,10,64,20,21123.55,20000,12800000,947,0.578,1056930.71,1104340.92,1104340.92

[xuang7]

Per our discussion, this unused dependency removal will not be backported to the release branch for now, since it would require changes to the CI file and requirements.txt, which may introduce conflicts. We can leave the unused dependency as is for now.