fix(crossseed): normalize hdr aliases

[s0up4200]

Description:

Normalize HDR aliases in cross-seed parsing and matching so HDR10+, HDR10P, and separated HDR10 PLUS forms compare consistently across /check, /apply, and shared candidate discovery. Tighten raw-name HDR enrichment to avoid false positives from scene groups like -DV, and add regressions for the End of Watch repro plus HDR alias edge cases.

ref: #1277

Summary by CodeRabbit

• New Features

  • Parsing now detects and normalizes more HDR tag variants (HDR10+, HDR10P, DV, HLG) and enriches release HDR tags for more accurate comparisons.
  • Cross-seed matching improved to handle HDR10+ alias variations across different naming and filename formats.
  • Parsing trims exposed trailing group/site tokens for cleaner release names.

• Tests

  • Expanded coverage for HDR variant extraction and matching, including HDR10+/HDR10/DV edge cases and cross-format candidate resolution.

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

Walkthrough

Adds HDR-aware extraction and normalization to the release parser and updates cross-seed matching to compare HDR sets using HDR-specific normalization (resolving HDR10+ vs HDR10). Tests added/expanded to cover HDR10+/DV and HDR10+/HDR10 alias and format edge cases.

Changes

Cohort / File(s)	Summary
Cross-seed HDR Matching internal/services/crossseed/matching.go, internal/services/crossseed/matching_hdr_test.go	Replaces generic HDR comparison with HDR-aware normalization (new joinNormalizedHDRSlice, normalizeHDRVariant); adjusts matching to handle HDR10+ vs HDR10 and DV interactions; adds integration tests for HDR edge cases.
Cross-seed Tests internal/services/crossseed/crossseed_test.go	Adds test case to cover HDR10+ alias in discussion title and a new test verifying cross-format HDR10+ alias matching across filename variants.
Release Parser HDR Enrichment pkg/releases/parser.go, pkg/releases/parser_test.go	Adds hdrTagMatchers, enrichReleaseHDR, normalization helpers (normalizeHDRTags, canonicalHDRTag, etc.), trimming helpers, and integrates enrichment into Parse; Parser gains keyNormalizer and new constructors; tests validate HDR enrichment and trailing-token trimming.

Sequence Diagram(s)

sequenceDiagram
    participant Client as Client
    participant Parser as Release Parser
    participant Enricher as HDR Enricher
    participant Normalizer as HDR Normalizer
    participant Matcher as Cross-seed Matcher

    Client->>Parser: submit release name / parse request
    Parser->>Enricher: enrichReleaseHDR(release)
    Enricher->>Enricher: scan raw name with hdrTagMatchers
    Enricher->>Normalizer: normalizeHDRTags(extractedTags)
    Normalizer->>Normalizer: normalizeHDRVariant, dedupe (remove HDR10 when HDR10+ present)
    Normalizer-->>Enricher: normalized HDR set
    Enricher-->>Parser: release with enriched HDR tags
    Parser->>Matcher: FindCandidates(source, candidates)
    Matcher->>Normalizer: joinNormalizedHDRSlice(sourceHDR, candidateHDR)
    Normalizer-->>Matcher: comparison result
    Matcher-->>Client: match / no-match decision

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• fix(crossseed): require strict HDR and Collection matching #799: HDR comparison changes and HDR-focused tests — directly related to HDR normalization updates.
• fix(crossseed): improve matching with size validation and relaxed audio checks #845: Modifies cross-seed matching logic in the same matching file (overlaps with matching behavior changes).
• fix(crossseed): enforce resolution and language matching with sensible defaults #855: Also updates cross-seed matching logic and test coverage in the same area.

Suggested labels

bugfix, cross-seed, tests, backend

Poem

🐰 I hopped through names both bright and odd,

HDR10+ and HDR10 I trod,
I trimmed the tails and tamed the tags,
Now matches find what once was jagged,
A small rabbit, joyous and glad.

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 18.75% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main change: normalizing HDR aliases in cross-seed matching and parsing to ensure consistent comparison across different HDR format variants (HDR10+, HDR10P, etc.).

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch fix/crossseed-hdr-alias

---

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

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@pkg/releases/parser.go`:
- Around line 86-91: enrichReleaseHDR currently calls shouldScanRawHDR and then
runs hdrTagMatchers against the full rawName, which lets terminal
scene-group/site tokens (e.g., "-DV") be picked up; update enrichReleaseHDR to
strip the parsed trailing Group/Site token from rawName before running the
hdrTagMatchers loop (use the same parsing logic used elsewhere to identify the
trailing group/site token), and ensure shouldScanRawHDR still gates video-shaped
releases but the matcher scan uses the truncated name; apply the same change to
the other scan locations referenced (the code block around the other hdr scan
loop) so no trailing group/site token is included when matching HDR tags.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 6c419aac-321a-4bd2-aebb-c06d2736df4f

📥 Commits

Reviewing files that changed from the base of the PR and between 7aadde7 and edbdca1.

📒 Files selected for processing (5)

• internal/services/crossseed/crossseed_test.go
• internal/services/crossseed/matching.go
• internal/services/crossseed/matching_hdr_test.go
• pkg/releases/parser.go
• pkg/releases/parser_test.go

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@pkg/releases/parser.go`:
- Around line 154-166: trimTrailingGroupOrSite currently trims release.Group and
release.Site only once each, leaving a leftover trailing token if removing one
token exposes another (e.g., "...-GROUP-SITE" becomes "...-GROUP"); update
trimTrailingGroupOrSite to repeatedly call trimTrailingParsedToken for the
tokens release.Group and release.Site until trimmed no longer changes (use a
loop that tracks previous value), ensuring you reference the existing function
trimTrailingParsedToken and the variables release.Group, release.Site and
trimmed so trailing tokens are removed until the suffix stabilizes.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 6c340a55-978c-4fe4-b214-f7a9766f75e4

📥 Commits

Reviewing files that changed from the base of the PR and between edbdca1 and b09e5b1.

📒 Files selected for processing (2)

• pkg/releases/parser.go
• pkg/releases/parser_test.go

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@pkg/releases/parser.go`:
- Around line 177-199: trimTrailingParsedToken currently only strips tokens that
sit at the literal end of rawName, so filenames with an extension (e.g.,
Movie.2024.2160p.BluRay.x265-DV.mkv) keep trailing group/site tokens before the
extension; update trimTrailingParsedToken to also match the token when followed
by an optional file extension (e.g., allow patterns that permit `(\.[^./\\]+)?$`
after the quoted token) and then trim the matched portion the same way, update
the regex patterns used in trimTrailingParsedToken to accept an optional
trailing extension, and add a regression unit test case that asserts a filename
like `Movie.2024.2160p.BluRay.x265-DV.mkv` yields the expected trimmed result
without preserving `DV`.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 22c88f8a-2a89-4d30-8a1d-e51b6a271fb1

📥 Commits

Reviewing files that changed from the base of the PR and between b09e5b1 and 388f91c.

📒 Files selected for processing (2)

• pkg/releases/parser.go
• pkg/releases/parser_test.go

[coderabbitai]

🧹 Nitpick comments (1)

pkg/releases/parser.go (1)

126-151: Preserve nil when normalization finds no HDR tags.

normalizeHDRTags now returns a non-nil empty slice for non-HDR releases, so enrichReleaseHDR changes the zero-value behavior from nil to []. If any caller serializes or compares release.HDR, that is a visible behavior change unrelated to the alias fix.

Suggested adjustment

 func normalizeHDRTags(tags []string) []string {
 	seen := make(map[string]struct{}, len(tags))
 	hasHDR10Plus := false
@@
 	if hasHDR10Plus {
 		delete(seen, "HDR10")
 	}
+
+	if len(seen) == 0 {
+		return nil
+	}
 
 	normalized := make([]string, 0, len(seen))
 	for tag := range seen {
 		normalized = append(normalized, tag)
 	}

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@pkg/releases/parser.go` around lines 126 - 151, normalizeHDRTags currently
returns a non-nil empty slice when no canonical HDR tags are found, changing
zero-value behavior of fields like release.HDR; modify normalizeHDRTags so it
returns nil when there are no normalized tags (e.g., avoid allocating normalized
until you know seen is non-empty or return nil if len(seen)==0), keeping the
existing canonicalization, HDR10+ handling, and sorting logic intact; ensure
callers such as enrichReleaseHDR that assign to release.HDR still observe nil
when no tags are present.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@pkg/releases/parser.go`:
- Around line 126-151: normalizeHDRTags currently returns a non-nil empty slice
when no canonical HDR tags are found, changing zero-value behavior of fields
like release.HDR; modify normalizeHDRTags so it returns nil when there are no
normalized tags (e.g., avoid allocating normalized until you know seen is
non-empty or return nil if len(seen)==0), keeping the existing canonicalization,
HDR10+ handling, and sorting logic intact; ensure callers such as
enrichReleaseHDR that assign to release.HDR still observe nil when no tags are
present.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 3aa54119-517d-4e06-8ff5-1326f1379673

📥 Commits

Reviewing files that changed from the base of the PR and between 388f91c and 40f72b5.

📒 Files selected for processing (2)

• pkg/releases/parser.go
• pkg/releases/parser_test.go

[coderabbitai]

🧹 Nitpick comments (1)

pkg/releases/parser.go (1)

196-200: Regex compilation inside loop may impact performance on cache misses.

regexp.MustCompile is called for each pattern per invocation, and trimTrailingParsedToken is called iteratively from trimTrailingGroupOrSite. While the Parser-level cache mitigates repeated parsing of the same release names, high-throughput scenarios with many unique names could see overhead.

Consider caching compiled regexes per token or using string operations for simple suffix matching where feasible.

♻️ Potential optimization using string-based matching for simple cases

 func trimTrailingParsedToken(rawName, token string) string {
 	token = strings.TrimSpace(token)
 	if rawName == "" || token == "" {
 		return rawName
 	}

+	// Fast path: check if token appears near end before regex
+	lowerName := strings.ToLower(rawName)
+	lowerToken := strings.ToLower(token)
+	if !strings.Contains(lowerName[max(0, len(lowerName)-len(token)-10):], lowerToken) {
+		return rawName
+	}
+
 	quoted := regexp.QuoteMeta(token)

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@pkg/releases/parser.go` around lines 196 - 200, The loop in
trimTrailingParsedToken repeatedly calls regexp.MustCompile(pattern) for each
pattern causing unnecessary compilation overhead; modify trimTrailingParsedToken
so regexes are precompiled and cached (e.g., a map[string]*regexp.Regexp at
parser level or package init) and reused across invocations, and where patterns
are simple suffixes replace the regex with faster string operations
(strings.HasSuffix/TrimSuffix) to avoid regex overhead; ensure the change
updates any callers such as trimTrailingGroupOrSite to use the
cached/precompiled regexes.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@pkg/releases/parser.go`:
- Around line 196-200: The loop in trimTrailingParsedToken repeatedly calls
regexp.MustCompile(pattern) for each pattern causing unnecessary compilation
overhead; modify trimTrailingParsedToken so regexes are precompiled and cached
(e.g., a map[string]*regexp.Regexp at parser level or package init) and reused
across invocations, and where patterns are simple suffixes replace the regex
with faster string operations (strings.HasSuffix/TrimSuffix) to avoid regex
overhead; ensure the change updates any callers such as trimTrailingGroupOrSite
to use the cached/precompiled regexes.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 548c89ac-cba6-4969-9d1d-0fa3aaa2f314

📥 Commits

Reviewing files that changed from the base of the PR and between 40f72b5 and 631831f.

📒 Files selected for processing (2)

• pkg/releases/parser.go
• pkg/releases/parser_test.go