fix(workflow-operator): validate zero-shot-image-classification labels before running

[PG1204]

What changes were proposed in this PR?

When the zero-shot-image-classification task ran without at least 2 candidate labels, the operator raised the error from inside the per-row loop, so it crashed mid-run instead of failing cleanly. This moves the check up front, alongside the operator's other config validations, so it fails fast with a clear message before any rows are processed.

Labels now come from the Candidate Labels field only (matching the text zero-shot-classification task); the old fallback that read labels from the prompt column is removed.

Any related issues?

Closes #7199 (the Part B follow-up to #7297, which addressed the same issue).

How was this PR tested?

Unit tests + the generated-Python compile check, plus a quick headless run of the generated operator: with 0 labels it now raises a clear "requires at least 2 Candidate Labels" error before the loop; with 2+ labels it proceeds normally.

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

Yes, this PR was co-authored with Claude in compliance with ASF policy.

[PG1204]

/request-review @xuang7

[github-actions]

Backport auto-label report

This fix: PR was checked against each actively-supported release branch. release/* labels drive the post-merge backport, so add or remove one to change where this fix lands.

Release branch

Analysis

⚠️ release/v1.2

Not labeled automatically — none of the files this PR modifies exist on this branch (common/workflow-operator/src/main/scala/org/apache/texera/amber/operator/huggingFace/codegen/HuggingFaceCodegenBase.scala, common/workflow-operator/src/main/scala/org/apache/texera/amber/operator/huggingFace/codegen/ImageTaskCodegen.scala, common/workflow-operator/src/test/scala/org/apache/texera/amber/operator/huggingFace/HuggingFaceInferenceOpDescSpec.scala, common/workflow-operator/src/test/scala/org/apache/texera/amber/operator/huggingFace/codegen/ImageTaskCodegenSpec.scala). The fix may target code that isn't on this release, or the files were moved/renamed after the branch was cut. Please check and add release/v1.2 by hand if this fix should be backported here.

Auto-label run.

[github-actions]

Automated Reviewer Suggestions

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

• No candidates found from git blame history.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.66%. Comparing base (12169c2) to head (2c323b4).

Additional details and impacted files

@@            Coverage Diff            @@
##               main    #7585   +/-   ##
=========================================
  Coverage     89.65%   89.66%           
- Complexity     4397     4401    +4     
=========================================
  Files          1177     1177           
  Lines         46996    46990    -6     
  Branches       5268     5268           
=========================================
- Hits          42136    42135    -1     
  Misses         3094     3094           
+ Partials       1766     1761    -5     

Flag	Coverage Δ		*Carryforward flag
access-control-service	70.00% <ø> (ø)
agent-service	98.62% <ø> (-0.01%)	⬇️	Carriedforward from a09260a
amber	86.38% <ø> (+0.02%)	⬆️	Carriedforward from a09260a
computing-unit-managing-service	72.46% <ø> (ø)
config-service	77.31% <ø> (ø)
file-service	68.90% <ø> (ø)
frontend	90.67% <ø> (ø)		Carriedforward from a09260a
notebook-migration-service	78.89% <ø> (ø)
pyamber	97.57% <ø> (ø)		Carriedforward from a09260a
workflow-compiling-service	57.89% <ø> (ø)

*This pull request uses carry forward flags. Click here to find out more.

☔ 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

🟢 2 better · 🔴 7 worse · ⚪ 6 noise (<±5%) · 0 without baseline

Compared against main 12169c2 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	366	0.223	25,411/42,565/42,565 us	🔴 +17.3% / 🔴 +160.4%
🟢	bs=100 sw=10 sl=64	808	0.493	122,348/135,227/135,227 us	🟢 -15.2% / 🔴 +25.7%
🔴	bs=1000 sw=10 sl=64	929	0.567	1,066,366/1,204,723/1,204,723 us	🔴 +5.8% / 🔴 +16.2%

Baseline details

Latest main 12169c2 from same runner

config	metric	PR	latest main	7d avg	Δ latest	Δ 7d
bs=10 sw=10 sl=64	throughput	366 tuples/sec	436 tuples/sec	774.89 tuples/sec	-16.1%	-52.8%
bs=10 sw=10 sl=64	MB/s	0.223 MB/s	0.266 MB/s	0.473 MB/s	-16.2%	-52.8%
bs=10 sw=10 sl=64	p50	25,411 us	21,655 us	12,738 us	+17.3%	+99.5%
bs=10 sw=10 sl=64	p95	42,565 us	36,855 us	16,348 us	+15.5%	+160.4%
bs=10 sw=10 sl=64	p99	42,565 us	36,855 us	18,848 us	+15.5%	+125.8%
bs=100 sw=10 sl=64	throughput	808 tuples/sec	804 tuples/sec	1,005 tuples/sec	+0.5%	-19.6%
bs=100 sw=10 sl=64	MB/s	0.493 MB/s	0.491 MB/s	0.613 MB/s	+0.4%	-19.6%
bs=100 sw=10 sl=64	p50	122,348 us	119,228 us	100,970 us	+2.6%	+21.2%
bs=100 sw=10 sl=64	p95	135,227 us	159,416 us	107,605 us	-15.2%	+25.7%
bs=100 sw=10 sl=64	p99	135,227 us	159,416 us	116,429 us	-15.2%	+16.1%
bs=1000 sw=10 sl=64	throughput	929 tuples/sec	928 tuples/sec	1,030 tuples/sec	+0.1%	-9.8%
bs=1000 sw=10 sl=64	MB/s	0.567 MB/s	0.566 MB/s	0.629 MB/s	+0.2%	-9.8%
bs=1000 sw=10 sl=64	p50	1,066,366 us	1,071,717 us	991,433 us	-0.5%	+7.6%
bs=1000 sw=10 sl=64	p95	1,204,723 us	1,138,683 us	1,036,668 us	+5.8%	+16.2%
bs=1000 sw=10 sl=64	p99	1,204,723 us	1,138,683 us	1,070,470 us	+5.8%	+12.5%

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
0,10,10,64,20,546.88,200,128000,366,0.223,25411.09,42565.07,42565.07
1,100,10,64,20,2475.64,2000,1280000,808,0.493,122348.44,135226.63,135226.63
2,1000,10,64,20,21525.19,20000,12800000,929,0.567,1066366.15,1204723.11,1204723.11

[xuang7]

LGTM. One thing worth a quick ack since it goes beyond what #7199 asked for: this also drops the prompt-column fallback for candidate labels, so any workflow still supplying labels through the prompt column will now fail with the config error instead. Given the fallback was a pre-#5574 transition shim and the error message points users to the right field, this seems fine.

[PG1204]

LGTM. One thing worth a quick ack since it goes beyond what #7199 asked for: this also drops the prompt-column fallback for candidate labels, so any workflow still supplying labels through the prompt column will now fail with the config error instead. Given the fallback was a pre-#5574 transition shim and the error message points users to the right field, this seems fine.

@xuang7 thanks! yep, deliberate. It makes zero-shot-image-classification consistent with text zero-shot-classification (labels come only from the candidate labels field), and the pre-loop check turns any prompt-column usage into a clear error instead of silently-wrong results. Agreed that it's fine to drop.