feat(huggingface): keep the row on an empty petal measurement

[kz930]

What changes were proposed in this PR?

The operator puts the two petal values straight into a numpy array and standardizes them, so an empty cell reaches the subtraction as a None and ends the run with unsupported operand type(s) for -: 'NoneType' and 'float', naming neither the column nor a way out. Nothing in the configuration can exclude the row.

It now answers the empty cell the way the three other operators in the group already do: keep the row, leave the prediction columns empty. Those three read a text column, which is why #7569 did not reach this one, and it was left as the only operator in the Hugging Face group that ends the run on a value the model was never given.

Any related issues, documentation, discussions?

Closes #8056. The same answer the three text operators were given in #7569.

How was this PR tested?

HuggingFaceIrisLogisticRegressionOpDescSpec pins the guard in the generated code; removing it turns that test red. WorkflowOperator/test passes: 2391 tests.

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

Generated-by: Claude Code (Claude Opus 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: @carloea2, @aglinxinyuan
  You can notify them by mentioning @carloea2, @aglinxinyuan in a comment.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.42%. Comparing base (604f109) to head (569b089).

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #8057      +/-   ##
============================================
- Coverage     93.52%   93.42%   -0.10%     
+ Complexity     4727     4725       -2     
============================================
  Files          1195     1181      -14     
  Lines         48492    47703     -789     
  Branches       5379     5282      -97     
============================================
- Hits          45352    44568     -784     
- Misses         1693     1698       +5     
+ Partials       1447     1437      -10     

Flag	Coverage Δ		*Carryforward flag
access-control-service	81.00% <ø> (ø)
agent-service	99.32% <ø> (ø)		Carriedforward from 646e460
amber	89.58% <100.00%> (-0.02%)	⬇️
computing-unit-managing-service	73.67% <ø> (ø)
config-service	86.86% <ø> (ø)
file-service	87.91% <ø> (ø)
frontend	95.91% <ø> (+0.05%)	⬆️	Carriedforward from 646e460
notebook-migration-service	79.31% <ø> (ø)
pyamber	97.78% <ø> (-0.76%)	⬇️	Carriedforward from 646e460
workflow-compiling-service	77.19% <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

🟢 4 better · 🔴 1 worse · ⚪ 10 noise (<±5%) · 0 without baseline

Compared against main 604f109 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	658	0.401	14,464/19,510/19,510 us	🟢 -16.3% / 🔴 +37.2%
🔴	bs=100 sw=10 sl=64	1,236	0.755	79,773/91,480/91,480 us	🟢 -23.5% / 🟢 -10.6%
⚪	bs=1000 sw=10 sl=64	1,552	0.947	632,895/706,766/706,766 us	⚪ within ±5% / 🟢 +30.7%

Baseline details

Latest main 604f109 from same runner

config	metric	PR	latest main	7d avg	Δ latest	Δ 7d
bs=10 sw=10 sl=64	throughput	658 tuples/sec	648 tuples/sec	894.27 tuples/sec	+1.5%	-26.4%
bs=10 sw=10 sl=64	MB/s	0.401 MB/s	0.395 MB/s	0.546 MB/s	+1.5%	-26.5%
bs=10 sw=10 sl=64	p50	14,464 us	14,252 us	11,164 us	+1.5%	+29.6%
bs=10 sw=10 sl=64	p95	19,510 us	23,322 us	14,219 us	-16.3%	+37.2%
bs=10 sw=10 sl=64	p99	19,510 us	23,322 us	16,863 us	-16.3%	+15.7%
bs=100 sw=10 sl=64	throughput	1,236 tuples/sec	1,240 tuples/sec	1,158 tuples/sec	-0.3%	+6.7%
bs=100 sw=10 sl=64	MB/s	0.755 MB/s	0.757 MB/s	0.707 MB/s	-0.3%	+6.8%
bs=100 sw=10 sl=64	p50	79,773 us	75,600 us	88,711 us	+5.5%	-10.1%
bs=100 sw=10 sl=64	p95	91,480 us	119,575 us	95,062 us	-23.5%	-3.8%
bs=100 sw=10 sl=64	p99	91,480 us	119,575 us	102,275 us	-23.5%	-10.6%
bs=1000 sw=10 sl=64	throughput	1,552 tuples/sec	1,543 tuples/sec	1,188 tuples/sec	+0.6%	+30.7%
bs=1000 sw=10 sl=64	MB/s	0.947 MB/s	0.941 MB/s	0.725 MB/s	+0.6%	+30.6%
bs=1000 sw=10 sl=64	p50	632,895 us	640,601 us	874,204 us	-1.2%	-27.6%
bs=1000 sw=10 sl=64	p95	706,766 us	722,448 us	916,845 us	-2.2%	-22.9%
bs=1000 sw=10 sl=64	p99	706,766 us	722,448 us	947,186 us	-2.2%	-25.4%

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
0,10,10,64,20,304.04,200,128000,658,0.401,14463.91,19510.49,19510.49
1,100,10,64,20,1617.79,2000,1280000,1236,0.755,79772.86,91480.15,91480.15
2,1000,10,64,20,12883.91,20000,12800000,1552,0.947,632894.88,706766.47,706766.47

[kz930]

/request-review @carloea2 @aglinxinyuan

[carloea2]

Looks good. A missing measurement now keeps the row and leaves both prediction fields empty.

[kz930]

@aglinxinyuan May you take a look at this PR please?