Honor ignore rules outside Git repos during indexing

[Copilot]

In non-Git directories, Arbor was effectively indexing without exclusions, so vendored/dependency trees (e.g., venv, node_modules, site-packages) could dominate graph size and ranking. This change makes ignore behavior consistent between Git and non-Git workspaces.

• Ignore semantics in non-Git directories

  • Updated watcher traversal to apply .gitignore even when .git/ is absent (require_git(false)).
  • Enabled custom ignore file support for .arborignore.

• Arbor-native and built-in excludes

  • Added ignore matcher composition for:
    • built-in heavy-tree excludes (venv/, .venv/, node_modules/, site-packages/, __pycache__/, target/, dist/, build/, out/)
    • .arbor/config.json ignore entries
  • Applied this matcher to both full indexing and staleness checks (sources_newer_than), so behavior is consistent across index/read paths.

• Coverage for regression-prone paths

  • Added focused tests validating exclusion from:
    • .gitignore outside Git repos
    • .arborignore
    • .arbor/config.json ignore list

let walker = WalkBuilder::new(root)
    .git_ignore(true)
    .ignore(true)
    .require_git(false)                  // honor .gitignore outside git repos
    .add_custom_ignore_filename(".arborignore")
    .follow_links(options.follow_symlinks)
    .build();

• Fixes Ignore rules (.gitignore / .arbor config) not honored outside a git repo — venv/vendored trees bloat the graph #167

[github-actions]

🌳 Arbor Impact Report

Risk Level: 🟠 High | Blast Radius: 50 nodes | Changed Symbols: 21

Changed Files

File	Status
crates/arbor-watcher/src/indexer.rs	Modified

📊 Visual Impact Graph

graph TD
  classDef changed fill:#ef4444,stroke:#333,stroke-width:2px,color:#fff;
  classDef caller fill:#f59e0b,stroke:#333,stroke-width:1px,color:#fff;
  index_directory[index_directory] --> build_ignore_matcher[build_ignore_matcher]
  sources_newer_than[sources_newer_than] --> build_ignore_matcher[build_ignore_matcher]
  class build_ignore_matcher changed;
  class IndexResult changed;
  class ArborConfig changed;
  class DEFAULT_EXCLUDE_PATTERNS changed;
  class IndexOptions changed;
  class sources_newer_than caller;
  class index_directory caller;

Loading

Impact Summary

Metric	Count
Direct callers affected	23
Indirect callers affected	41
API entrypoints impacted	40
Files likely requiring updates	6
Total blast radius	50

⚠️ Warning: 40 API entrypoints are affected. Integration tests recommended.

🔍 Suggestion: Consider breaking this change into smaller PRs.

---

Powered by Arbor v2.6.0 — graph-native code intelligence