fix: use word boundaries for issue label matching

[oritwoen]

labelMatchesAny was using String.includes() to check if a label contained a keyword, which caused substring false positives. Labels like debug matched keyword bug, docker matched doc, prefix matched fix - all getting misclassified in the type quota system and stealing slots from actual bugs/docs issues.

Switched to a single pre-compiled regex with \b word boundaries per keyword set (cached via WeakMap so sets that don't change only compile once). Emoji-prefixed labels like :lady_beetle: bug still match correctly since \b triggers at the colon/space boundary.

Added 15 unit tests covering the edge cases.

Summary by CodeRabbit

• Bug Fixes

  • Improved label matching behavior to reduce false positives by using word boundary detection instead of substring matching when filtering issues by labels.

• Tests

  • Added unit test coverage for label matching and issue classification to validate accuracy and ensure proper handling of edge cases.

[coderabbitai]

📝 Walkthrough

Walkthrough

The changes refactor label matching logic in issue handling to use word-boundary regex patterns instead of substring matching, improving accuracy for keyword identification. The labelMatchesAny function is exported and backed by a regex cache for performance. A comprehensive test suite validates the updated matching behavior and issue classification logic.

Changes

Cohort / File(s)	Summary
Label Matching Logic src/sources/issues.ts	Implemented regex-based word-boundary matching with caching via labelRegexCache and getLabelRegex(). Replaced substring matching in labelMatchesAny with exact or word-boundary pattern matching. Exported the function and updated JSDoc to reflect the new behavior and elimination of substring false positives.
Test Coverage test/unit/issues.test.ts	Added comprehensive unit tests for labelMatchesAny and classifyIssue. Tests validate exact matches, emoji-prefixed labels, hyphen-separated forms, case-sensitivity, keyword boundary detection (e.g., doc vs docker), and correct issue classification priorities (bug > question > other).

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Poem

🐰 Hop, hop—the labels now align!
Word boundaries drawn with regex fine,
No more false whiskers from partial strings,
Just exact matches—oh, the clarity it brings!
With tests as thorough as a burrow deep,
These labels will truth forever keep. ✨

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% 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 directly and accurately summarizes the main change: implementing word boundaries for issue label matching to fix substring false positives.

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

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/issue-label-word-boundary

---

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

[coderabbitai]

🧹 Nitpick comments (1)

test/unit/issues.test.ts (1)

20-23: Verify emoji shortcode matching is intentional behavior.

Line 22 tests that :bug: report matches the keyword bug. This works because the colons in :bug: create word boundaries, causing "bug" inside the shortcode to match. If the intent is only to match labels like :lady_beetle: bug (emoji followed by keyword), the current regex would also match shortcodes that happen to contain keywords (e.g., :bug:, :firefix: would match "fix").

If this is the desired behavior, consider adding a brief comment in the test explaining that emoji shortcodes containing keywords are intentionally matched.

🤖 Prompt for AI Agents

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

In `@test/unit/issues.test.ts` around lines 20 - 23, Test currently asserts that
labelMatchesAny(':bug: report', BUG_KEYWORDS) returns true because the regex
treats colons as word boundaries and matches "bug" inside an emoji shortcode; to
address the review, either (preferred) add a short inline comment above the test
case explaining that emoji shortcodes containing keywords (e.g., ':bug:') are
intentionally considered matches for labelMatchesAny against BUG_KEYWORDS, or
(if unintended) update the matching logic in labelMatchesAny to ignore text
inside colon-delimited shortcodes (i.e., skip or strip /:[^:]+:/ sequences
before applying BUG_KEYWORDS) so ':bug:' no longer triggers a match.

🤖 Prompt for all review comments with AI agents

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

Nitpick comments:
In `@test/unit/issues.test.ts`:
- Around line 20-23: Test currently asserts that labelMatchesAny(':bug: report',
BUG_KEYWORDS) returns true because the regex treats colons as word boundaries
and matches "bug" inside an emoji shortcode; to address the review, either
(preferred) add a short inline comment above the test case explaining that emoji
shortcodes containing keywords (e.g., ':bug:') are intentionally considered
matches for labelMatchesAny against BUG_KEYWORDS, or (if unintended) update the
matching logic in labelMatchesAny to ignore text inside colon-delimited
shortcodes (i.e., skip or strip /:[^:]+:/ sequences before applying
BUG_KEYWORDS) so ':bug:' no longer triggers a match.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 658e565b-e837-4bf9-a526-2685f3483c6e

📥 Commits

Reviewing files that changed from the base of the PR and between 8b54a9a and c84ca54.

📒 Files selected for processing (2)

• src/sources/issues.ts
• test/unit/issues.test.ts

[harlan-zw]

Nice one