Fix: parser.int() silently drops multi-seed strings in release task request builder

[ChuxiJ]

Fixes #782

Summary

This PR fixes a bug where multi-seed batching was silently broken in the FastAPI server. When users provided comma-separated seeds (e.g., seed="1024,2048") with batch_size > 1, only the first sample remained deterministic due to incorrect parameter parsing.

Problem

The FastAPI request builder in acestep/api/http/release_task_request_builder.py was using parser.int("seed", -1) which attempted to convert comma-separated seed strings like "1024,2048" into an integer. This conversion silently failed and returned the default value -1, breaking the multi-seed feature.

Solution

Changed line 57 from:

seed=parser.int("seed", -1),

To:

seed=parser.get("seed", -1),

This preserves the raw seed value (whether string or int), allowing downstream code to parse it correctly.

Why This Works

The downstream prepare_seeds() function in acestep/core/generation/handler/task_utils.py already handles:

• ✅ Comma-separated strings: "1024,2048" → parsed as list
• ✅ Single integers: 1024 → treated as single seed
• ✅ -1 or None → generates random seeds

The multi-seed pipeline was fully functional end-to-end, but the API request builder was breaking the chain by forcing int conversion.

Testing

Before fix: All batch samples got seed=-1 (non-deterministic)
After fix: Each sample gets its specified seed

Test command:

curl -X POST http://localhost:8001/release_task \
  -F "prompt=upbeat electronic music" \
  -F "batch_size=3" \
  -F "seed=1024,2048,4096"

Expected: Three deterministic outputs with different seeds.

Impact

• Bug severity: Medium (breaks documented feature)
• Backward compatibility: ✅ Maintained (single int seeds still work)
• Lines changed: 1
• Risk: Minimal (only affects seed parameter handling)

Checklist

• Code change made
• Commit message follows format
• Backward compatibility verified
• Tests run locally (requires full setup)
• Documentation updated (if needed)

Summary by CodeRabbit

• Bug Fixes
  • Improved seed parameter handling in music generation requests to ensure proper processing of missing or undefined values.

[coderabbitai]

Caution

Review failed

Pull request was closed or merged during review

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 7e506f11-c9a2-4b9e-acb7-a4c10fa47d75

📥 Commits

Reviewing files that changed from the base of the PR and between 746eec4 and 6d027f9.

📒 Files selected for processing (1)

• acestep/api/http/release_task_request_builder.py

---

📝 Walkthrough

Walkthrough

Changed seed field retrieval in build_generate_music_request from parser.int() to parser.get() to preserve comma-separated seed strings for multi-seed batch support, enabling downstream pipeline handling of both integer and string seed formats.

Changes

Cohort / File(s)	Summary
Seed Parsing Fix acestep/api/http/release_task_request_builder.py	Switched seed field population from parser.int("seed", -1) to parser.get("seed", -1) to preserve comma-separated seed strings instead of silently coercing them to the default value, unblocking multi-seed batch functionality.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~5 minutes

Possibly related PRs

• PR #355: Adds robust seed parsing and preservation through the generation pipeline, complementing this fix by ensuring downstream components properly handle both single integer and comma-separated multi-seed formats.

Poem

🐰 A seed, once lost to parser's might,
Now flows through batches, pure and bright!
From one small change, a bloom unfolds—
Multi-seeds dance, as the string still holds! 🌱

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title directly describes the specific bug fix: replacing parser.int() with parser.get() to prevent silent loss of multi-seed strings.
Linked Issues check	✅ Passed	The code change directly addresses all requirements from issue #782: replacing parser.int() with parser.get() preserves comma-separated seed strings for downstream processing.
Out of Scope Changes check	✅ Passed	The change is a single-line edit in the seed parsing logic, directly addressing the linked issue with no additional or unrelated modifications.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

• 📝 Generate docstrings (stacked PR)
• 📝 Generate docstrings (commit on current branch)

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/issue-782-multi-seed-batch-parsing

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}