Tag Index Engine

[ElioNeto]

Tag Index Engine

Build a tag indexing system that extracts #tags from note content and maintains a searchable index using the LSM prefix scan.

Storage Design

Use column family tag: with prefix-based lookups:

cf "tag":
  tag:{tagname}  -> JSON array of note paths that contain this tag

Components

TagIndex::index_tags(note_path, tags: Vec<String>)

• Diff old vs new tags for the note
• Add note path to tag:{tag} for new tags
• Remove note path from tag:{tag} for removed tags

TagIndex::remove_note_tags(note_path)

• Get all tags for the note from metadata
• Remove note path from each tag:{tag}

TagIndex::search_by_tag(tag, cursor, limit) -> (Vec<String>, Option<String>)

• Use Engine::search_prefix("tag:{tag}") with cursor pagination
• Return note paths sorted lexicographically

TagIndex::list_all_tags() -> Vec<String>

• Scan tag: prefix, collect unique tag names
• Support pagination via cursor

Tag Parsing Rules (from markdown)

• #tag — at word boundary
• #tag/subtag — nested tags (store as full path)
• #tag with spaces — not valid (must be alphanumeric)
• #tag# — trailing hash is not part of tag
• Ignore tags inside code blocks, inline code, and HTML comments
• Max tag length: 100 chars
• Allowed chars: [a-zA-Z0-9_/-]

Acceptance Criteria

• Tags extracted correctly from markdown content
• Tags inside code blocks are ignored
• Tag index is updated atomically on note write
• search_by_tag returns correct notes with pagination
• Nested tags (tag/subtag) work as expected
• Performance: 10K tags indexed in < 50ms
• Unit tests for all operations

Parent Epic

#275

[ElioNeto]

Resolved by commit 48dd088 in branch feat/notes-engine-v2.

Implementation:

• Tag Index Engine #279 — Integrated into the notes engine and REST API layer.

To create a PR manually:

gh pr create --base main --head feat/notes-engine-v2

Or via the web: https://github.com/ElioNeto/ApexStore/pull/new/feat/notes-engine-v2