Feature: Add support for `.lrcignore` to exclude specific files from AI review

[himnishag]

Problem Statement

Currently, git-lrc uses standard .gitignore configurations. However, developers often need to commit certain files (like package-lock.json, auto-generated code, database migrations, or translation files) but want to skip sending them to the AI review.

Sending these files to the AI connector:

1. Wastes valuable API tokens.
2. Increases review latency unnecessarily.
3. Clutters the Review UI with noisy comments.

---

Planned Tech Stack & Implementation

• Language: Go (Golang)
• Libraries: Leverage Go's standard library or github.com/gobwas/glob for pattern matching.
• Location: Update the internal file collection/staged diff processing logic under storage/ or internal/appcore/.

---

Proposed Solution

Add support for a local .lrcignore file in the repository's root directory:

1. On running git lrc review or pre-commit checks, read .lrcignore if present.
2. Parse glob patterns (e.g., package-lock.json, **/generated/**).
3. Exclude matching staged diffs before compiling the payload sent to the AI connector.

Planned Tech Stack & Implementation

• Language: Go (Golang)
• Pattern Matching: Use standard Go libraries (like path/filepath or standard glob libraries) to handle file pattern matching.
• Integration Points: Integrate the exclusion check inside the staged diff gathering system (typically under the storage/ or internal/appcore/ modules).

---

Expected Features

• Local .lrcignore File Parsing: Detects and parses a .lrcignore file at the root of the active Git repository.
• Glob Pattern Matching: Supports standard wildcard pattern matching (e.g., *.lock, **/vendor/**, temp_file.txt).
• Comments and Whitespace Support: Gracefully skips commented lines (starting with #) and empty lines in the ignore file.
• Fallbacks: If no .lrcignore is present, it should fallback to default behavior (reviewing all staged files not in .gitignore).

---

Expected Outcome

• Zero Token Waste: AI Connectors (Gemini, Claude, OpenAI) will only receive relevant diff files, drastically saving API usage.
• Faster Reviews: Less payload size means much faster response times from the AI endpoints.
• Noisy-Free UI: The local Review UI will only display critical, human-written code diffs, making the review process highly productive.

---

[himnishag]

Please assign this to me . i would love to work on this as a GSSOC'26 contributor