feat(file-service): add model upload limits and the models_enabled flag

[tanishqgandhi1908]

What changes were proposed in this PR?

1. gui.tabs.models_enabled, default false — the flag the model UI will gate on. Nothing reads it yet; it hides the sidebar entry, not the route (same as datasets_enabled).
2. Models get their own upload ceiling: 2 GiB, up from the 20 MiB they inherited from datasets. The shared upload engine read the dataset's single_file_upload_max_size_mib key literally, so both types resolved the same site setting.
3. An UploadLimits descriptor — ties a resource type to its upload keys and exposes singleFileUploadMaxBytes. ResourceStorage gains an uploadLimits slot, so both call sites read the limit through the descriptor instead of naming a key. The same helper also existed in DatasetResource and ModelResource, where it was dead code; both copies removed.

Naming (review note). site_settings rows are keyed by a leaf's last path segment, so dataset { } / model { } are flattened away before storage and two bare leaves would collide — DefaultsConfig's uniqueness require fails at load. So the model leaves need a prefix, and per review the dataset leaves got one too:

dataset { dataset_single_file_upload_max_size_mib = 20   }
model   { model_single_file_upload_max_size_mib   = 2048 }

Clean rename, no alias. Also fixes MAX_NUMBER_OF_CONCURRENT_UPLOADING_FILE -> DATASET_MAX_....

That rename is why this touches the frontend and ships a migration:

• The key strings are hardcoded in admin-settings, files-uploader and dataset-detail.
• sql/updates/42.sql renames the four existing site_settings rows — otherwise the seeder recreates them at defaults and an admin's raised limit is silently lost. Idempotent.

ConfigResource derives its public whitelist from gui/dataset/model, which is why the pinned set in ConfigResourceSpec changed.

Any related issues, documentation, discussions?

Backend prerequisite for the model dashboard UI — #6499 and #6501.

How was this PR tested?

New UploadLimitsSpec (5 tests, no DB harness): pins each descriptor's keys against default.conf, asserts the dataset and model key sets are disjoint, and asserts the model ceiling resolves to 2 GiB.

42.sql run against a scratch Postgres: values preserved, unrelated rows untouched, second run a no-op, no PK violation when both old and new rows exist.

Existing suites pass unmodified except the two pinned config assertions:
DefaultsConfigSpec 5, ConfigResourceSpec 33, UploadLimitsSpec 5, DatasetResourceSpec 152, ModelUploadResourceSpec 13, ModelResourceSpec 22, admin-settings 45, dataset-detail 203, files-uploader 30. scalafmt + prettier clean.

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: @Yicong-Huang, @zyratlo, @Neilk1021
  You can notify them by mentioning @Yicong-Huang, @zyratlo, @Neilk1021 in a comment.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 93.33333% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 93.42%. Comparing base (646e460) to head (5e38344).

Files with missing lines	Patch %	Lines
.../apache/texera/service/resource/UploadLimits.scala	91.66%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff            @@
##               main    #8000   +/-   ##
=========================================
  Coverage     93.42%   93.42%           
- Complexity     4703     4706    +3     
=========================================
  Files          1180     1181    +1     
  Lines         47766    47787   +21     
  Branches       5321     5322    +1     
=========================================
+ Hits          44623    44647   +24     
+ Misses         1699     1697    -2     
+ Partials       1444     1443    -1     

Flag	Coverage Δ		*Carryforward flag
access-control-service	81.00% <ø> (ø)
agent-service	99.32% <ø> (ø)		Carriedforward from a6565fb
amber	89.58% <ø> (+<0.01%)	⬆️
computing-unit-managing-service	73.67% <ø> (ø)
config-service	86.73% <100.00%> (ø)
file-service	87.89% <92.59%> (+0.42%)	⬆️
frontend	95.89% <100.00%> (ø)
notebook-migration-service	79.31% <ø> (ø)
pyamber	97.78% <ø> (ø)		Carriedforward from a6565fb
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

🟢 2 better · 🔴 5 worse · ⚪ 8 noise (<±5%) · 0 without baseline

Compared against main 646e460 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	379	0.231	25,208/39,244/39,244 us	🔴 +24.0% / 🔴 +180.1%
🟢	bs=100 sw=10 sl=64	951	0.581	104,363/124,466/124,466 us	🟢 -6.6% / 🔴 +30.9%
⚪	bs=1000 sw=10 sl=64	1,086	0.663	922,409/997,301/997,301 us	⚪ within ±5% / 🔴 +8.7%

Baseline details

Latest main 646e460 from same runner

config	metric	PR	latest main	7d avg	Δ latest	Δ 7d
bs=10 sw=10 sl=64	throughput	379 tuples/sec	444 tuples/sec	893.99 tuples/sec	-14.6%	-57.6%
bs=10 sw=10 sl=64	MB/s	0.231 MB/s	0.271 MB/s	0.546 MB/s	-14.8%	-57.7%
bs=10 sw=10 sl=64	p50	25,208 us	20,322 us	11,176 us	+24.0%	+125.5%
bs=10 sw=10 sl=64	p95	39,244 us	31,641 us	14,009 us	+24.0%	+180.1%
bs=10 sw=10 sl=64	p99	39,244 us	31,641 us	16,898 us	+24.0%	+132.2%
bs=100 sw=10 sl=64	throughput	951 tuples/sec	931 tuples/sec	1,157 tuples/sec	+2.1%	-17.8%
bs=100 sw=10 sl=64	MB/s	0.581 MB/s	0.568 MB/s	0.706 MB/s	+2.3%	-17.8%
bs=100 sw=10 sl=64	p50	104,363 us	108,327 us	88,763 us	-3.7%	+17.6%
bs=100 sw=10 sl=64	p95	124,466 us	133,236 us	95,096 us	-6.6%	+30.9%
bs=100 sw=10 sl=64	p99	124,466 us	133,236 us	102,854 us	-6.6%	+21.0%
bs=1000 sw=10 sl=64	throughput	1,086 tuples/sec	1,089 tuples/sec	1,187 tuples/sec	-0.3%	-8.5%
bs=1000 sw=10 sl=64	MB/s	0.663 MB/s	0.665 MB/s	0.724 MB/s	-0.3%	-8.5%
bs=1000 sw=10 sl=64	p50	922,409 us	915,733 us	874,968 us	+0.7%	+5.4%
bs=1000 sw=10 sl=64	p95	997,301 us	1,015,283 us	917,815 us	-1.8%	+8.7%
bs=1000 sw=10 sl=64	p99	997,301 us	1,015,283 us	949,868 us	-1.8%	+5.0%

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
0,10,10,64,20,527.74,200,128000,379,0.231,25207.62,39243.89,39243.89
1,100,10,64,20,2102.39,2000,1280000,951,0.581,104362.56,124465.94,124465.94
2,1000,10,64,20,18419.04,20000,12800000,1086,0.663,922408.94,997300.72,997300.72

[tanishqgandhi1908]

/request-review @aicam

[aicam]

Add enabling model flag to this PR

[aicam]

Please follow the naming convention we have.

[aicam]

LGTM