fix(crossseed): improve matching with size validation and relaxed audio checks

[s0up4200]

Summary

• Adds safety check to reject cross-seeds when main content file sizes don't match between source and candidate torrents
• Relaxes audio codec/channel matching to handle indexer metadata errors (e.g., BTN returning DDPA5.1 when actual file is DDP5.1)
• Prevents broken cross-seeds from season pack vs single episode mismatches
• Counts "rejected" status as skipped (not failed) in automation metrics

Problem 1: File Size Mismatches

When cross-seeding, if the incoming torrent matched an existing torrent by total size (within tolerance) but individual file sizes differed:

1. The file rename would fail (size-based bucket lookup in buildFileRenamePlan couldn't find matches)
2. This left the torrent in a broken state (folder renamed but file not renamed)
3. More critically, proceeding could corrupt existing good data

Example: A 12-episode season pack matched against a single episode candidate would pass initial filters but fail during alignment.

Problem 2: Indexer Metadata Errors

Indexers like BTN can return incorrect audio metadata in release names. For example:

• Actual file: Show.S01E01.DDP5.1.H.264-NTb.mkv
• BTN API returns: Show.S01E01.DDPA5.1.H.264-NTb

The strict audio matching would reject these valid matches.

Solution

1. Added hasContentFileSizeMismatch() function that:

   • Filters source and candidate files by ignore patterns
   • Checks if all content files can be matched by exact size
   • Returns mismatched files for logging
   • Rejects cross-seed with status "rejected" when mismatch detected

2. Relaxed audio matching in releasesMatch():

   • Removed audio codec and channel pre-filter checks
   • These checks are now handled downstream by file size matching
   • If files are truly different (different audio), sizes will differ and hasContentFileSizeMismatch() rejects
   • If indexer metadata was wrong but files are identical, cross-seed succeeds

3. Updated automation metrics: "rejected" status now counts as skipped rather than failed

Note on Ignore Patterns

The hasContentFileSizeMismatch() check uses ignore patterns to filter out sidecar files (NFO, samples, etc.) before comparing sizes. All automation flows (RSS, completion, seeded search, Autobrr webhooks) automatically populate ignore patterns from settings. Direct API callers should ensure they provide IgnorePatterns in the request to get consistent behavior.

Test plan

• Added comprehensive tests for hasContentFileSizeMismatch() covering:
  • Identical files (pass)
  • Size mismatch detection (reject)
  • Ignore patterns filtering (NFO, screens, etc.)
  • Season pack vs single episode (reject)
  • Audio metadata mismatch with same size (pass)
  • Audio metadata mismatch with different size (reject)
• Added TestReleasesMatch_AudioRelaxed proving DDP vs DDPA allowed through
• All existing tests pass
• Build succeeds
• Manually verified BTN cross-seed now works with incorrect DDPA metadata

[coderabbitai]

Walkthrough

Adds a size-based content-file comparison and enforces it in the cross-seed flow: candidates missing any non-ignored source file sizes are marked rejected and processing stops. Also relaxes audio/audio-channel checks in release matching and adds unit tests for both changes.

Changes

Cohort / File(s)	Change Summary
Content file size validation internal/services/crossseed/align.go	New unexported function hasContentFileSizeMismatch(sourceFiles, candidateFiles, ignorePatterns, normalizer) that filters files by ignore patterns, buckets candidate files by size, and reports any non-ignored source content files without a matching size. Returns (mismatchExists bool, mismatchedFileNames []string).
Cross-seed processing safety internal/services/crossseed/service.go	processCrossSeedCandidate now invokes the size-mismatch check and, on mismatch, sets status to "rejected", records a rejection message, logs details, and returns early to stop further processing. Also added "rejected" to skip/exists handling.
Matching rules relaxed internal/services/crossseed/matching.go	Removed strict audio/audio-channel checks from releasesMatch; audio codec/channel differences no longer block matches so downstream size/alignment handles real mismatches.
Unit tests internal/services/crossseed/align_test.go, internal/services/crossseed/matching_codec_test.go	Added TestHasContentFileSizeMismatch (multiple scenarios including ignores and cross-tracker cases) and TestReleasesMatch_AudioRelaxed to validate relaxed audio matching behavior.

Sequence Diagram(s)

sequenceDiagram
  participant Proc as processCrossSeedCandidate
  participant HasCheck as hasContentFileSizeMismatch
  participant Source as SourceTorrent
  participant Candidate as CandidateTorrent
  participant Log as Logger

  Proc->>HasCheck: call(sourceFiles, candidateFiles, ignorePatterns, normalizer)
  HasCheck->>Source: filter source files (exclude ignored)
  HasCheck->>Candidate: filter candidate files (exclude ignored) & build size buckets
  HasCheck-->>Proc: return (mismatchExists, mismatchedNames)
  alt mismatchExists == true
    Proc->>Log: warn with mismatchedNames
    Proc-->>Candidate: set status "rejected" and return
  else mismatchExists == false
    Proc->>Proc: continue cross-seed processing
  end

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• fix(crossseed): detect file name differences and fix hasExtraSourceFiles #741 — Overlaps crossseed alignment and file-size matching logic; touches similar paths in align/service.
• fix(crossseed): use matched torrent save path instead of category path #700 — Modifies cross-seed processing flow and candidate handling; related to rejection/skip logic.
• fix(cross): performance improvements #629 — Changes file-matching logic in crossseed helpers; related to ignore and matching behavior.

Suggested labels

bugfix

Poem

I hop through lists of files with care, 🐇
Counting bytes in burrows everywhere,
If sizes fail to line up true, I stop the race,
Mark the seed rejected, tidy up the place,
A rabbit’s check keeps swarms in safer space.

Pre-merge checks and finishing touches

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 42.86% which is insufficient. The required threshold is 80.00%.	You can run @coderabbitai generate docstrings to improve docstring coverage.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately describes the main changes: adding size validation to reject mismatched cross-seeds and removing audio codec/channel pre-filters for relaxed matching.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/crossseed-reject-file-size-mismatch

---

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.}

[coderabbitai]

Actionable comments posted: 0

🧹 Nitpick comments (1)

internal/services/crossseed/service.go (1)

2657-2673: Early rejection on content-size mismatch is well placed; confirm status semantics.

The new hasContentFileSizeMismatch guard is correctly positioned before any add/rename/recheck logic and logs enough context (instanceID, hashes, matchType, mismatchedFiles) to debug unsafe candidates. One thing to confirm: using Status: "rejected" means this path is counted under the default branch in processAutomationCandidate as a failure (TorrentsFailed++), not a skip. If you’d prefer these to show up as “rejected/skipped for safety” rather than hard failures in automation metrics, you might want to special‑case "rejected" alongside "no_match"/"skipped" there; otherwise the current behavior is consistent and safe.

📜 Review details

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 40982bc and 0dc8c3b.

📒 Files selected for processing (3)

• internal/services/crossseed/align.go
• internal/services/crossseed/align_test.go
• internal/services/crossseed/service.go

🧰 Additional context used

🧠 Learnings (3)

📓 Common learnings

Learnt from: s0up4200
Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go:209-212
Timestamp: 2025-11-28T20:32:30.126Z
Learning: Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go
Learning: The cross-seed recheck-resume worker intentionally runs for the process lifetime and keys pending entries by hash only. This is acceptable under the current constraint that background seeded-search runs operate on a single instance at a time; graceful shutdown and instanceID|hash keying are deferred by design.

📚 Learning: 2025-11-28T22:21:20.730Z

Learnt from: s0up4200
Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go:2415-2457
Timestamp: 2025-11-28T22:21:20.730Z
Learning: Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go
Learning: The determineSavePath function intentionally includes a contentLayout string parameter for future/content-layout branching and API consistency. Its presence is by design even if unused in the current body; do not flag as an issue in reviews.

Applied to files:

• internal/services/crossseed/align_test.go
• internal/services/crossseed/service.go

📚 Learning: 2025-11-28T20:32:30.126Z

Learnt from: s0up4200
Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go:209-212
Timestamp: 2025-11-28T20:32:30.126Z
Learning: Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go
Learning: The cross-seed recheck-resume worker intentionally runs for the process lifetime and keys pending entries by hash only. This is acceptable under the current constraint that background seeded-search runs operate on a single instance at a time; graceful shutdown and instanceID|hash keying are deferred by design.

Applied to files:

• internal/services/crossseed/service.go

🧬 Code graph analysis (1)

internal/services/crossseed/align_test.go (1)

pkg/stringutils/normalize.go (1)

• NewDefaultNormalizer (35-37)

🔇 Additional comments (2)

internal/services/crossseed/align_test.go (1)

9-11: Strong, scenario-driven coverage for hasContentFileSizeMismatch.

The new test harness exercises all the important behaviors of hasContentFileSizeMismatch: exact vs off‑by‑one sizes, multi-file sets, extra source files, directory/extension/suffix ignore patterns (NFO, samples, screens), empty/all-filtered sources, and multiple mismatches, while validating the exact set of mismatched source filenames. Using a shared stringutils.NewDefaultNormalizer() instance and ElementsMatch keeps the tests both realistic and order-agnostic. This is a solid, comprehensive test suite for the new helper.

Also applies to: 558-763

internal/services/crossseed/align.go (1)

553-600: Excellent implementation of the content file size safety check.

The logic correctly handles all edge cases:

• Proper filtering of ignored files (NFO, screens, etc.)
• Count-based size matching that handles duplicate sizes correctly
• Early return when no content files remain after filtering
• Asymmetric check (source must match in candidate) which is appropriate for the use case

The documentation clearly explains the rationale (scene releases should be byte-for-byte identical) and the safety implications. The O(n+m) complexity is optimal for this operation.

[coderabbitai]

Actionable comments posted: 0

🧹 Nitpick comments (1)

internal/services/crossseed/service.go (1)

2657-2673: New content‑file size‑mismatch guard is well‑placed; just sanity‑check ignore‑pattern coverage

The early hasContentFileSizeMismatch check here is in the right spot: it runs after choosing the best candidate but before any add/rename/recheck operations, so we avoid ever touching disk on season‑pack/single‑episode mismatches or genuinely different releases. Returning "rejected" and letting the automation side count that as a skipped item is consistent with the intent.

One thing to keep an eye on is how req.IgnorePatterns is populated across all CrossSeedRequest entrypoints: RSS automation, seeded search, completion, Autobrr/webhook flows all feed settings‑level ignore patterns through, but direct API/CLI CrossSeed calls that don’t set IgnorePatterns will now treat any extra non‑ignored files (e.g., NFOs, samples) as hard rejections rather than just “extra files → recheck”. If you have such callers, it may be worth either defaulting IgnorePatterns from automation settings in CrossSeed itself or documenting that they must supply the same ignore set used by automation to get consistent behavior.

📜 Review details

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 6ff0f62 and bb06821.

📒 Files selected for processing (1)

• internal/services/crossseed/service.go

🧰 Additional context used

🧠 Learnings (3)

📓 Common learnings

Learnt from: s0up4200
Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go:209-212
Timestamp: 2025-11-28T20:32:30.126Z
Learning: Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go
Learning: The cross-seed recheck-resume worker intentionally runs for the process lifetime and keys pending entries by hash only. This is acceptable under the current constraint that background seeded-search runs operate on a single instance at a time; graceful shutdown and instanceID|hash keying are deferred by design.

📚 Learning: 2025-11-28T20:32:30.126Z

Learnt from: s0up4200
Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go:209-212
Timestamp: 2025-11-28T20:32:30.126Z
Learning: Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go
Learning: The cross-seed recheck-resume worker intentionally runs for the process lifetime and keys pending entries by hash only. This is acceptable under the current constraint that background seeded-search runs operate on a single instance at a time; graceful shutdown and instanceID|hash keying are deferred by design.

Applied to files:

• internal/services/crossseed/service.go

📚 Learning: 2025-11-28T22:21:20.730Z

Learnt from: s0up4200
Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go:2415-2457
Timestamp: 2025-11-28T22:21:20.730Z
Learning: Repo: autobrr/qui PR: 641
File: internal/services/crossseed/service.go
Learning: The determineSavePath function intentionally includes a contentLayout string parameter for future/content-layout branching and API consistency. Its presence is by design even if unused in the current body; do not flag as an issue in reviews.

Applied to files:

• internal/services/crossseed/service.go

🔇 Additional comments (1)

internal/services/crossseed/service.go (1)

1895-1904: Classifying "rejected" as a skipped outcome is semantically correct

Treating "rejected" the same as "no_match" / "skipped" (incrementing TorrentsSkipped instead of TorrentsFailed) fits the new size‑mismatch behavior: these are intentional safety rejections of bad candidates, not automation errors, so they shouldn’t drag the run into a failed state.