Danila Poyarkov
|
f2ef9c976d
|
XPath query for node selection (#82)
* Add XPath query for node selection
- packages/core/src/xpath.ts: XPath engine using fontoxpath with custom
DOM facade over SceneGraph nodes
- RPC command: 'query' with selector, page, limit params
- CLI command: 'open-pencil query <file> <xpath>' with --json, --page, --limit
- AI/MCP tool: 'query_nodes' with XPath examples in description
- Queryable attributes: name, width, height, x, y, visible, opacity,
cornerRadius, fontSize, fontFamily, fontWeight, layoutMode, itemSpacing,
padding*, strokeWeight, rotation, locked, blendMode, text, lineHeight,
letterSpacing
* Add tests and docs for XPath query
- 9 unit tests for queryByXPath/matchByXPath core functions
- 6 tool tests for query_nodes AI/MCP tool
- README: query examples in CLI section
- VitePress docs: patterns, queryable attributes, example output
- CHANGELOG: feature entry
* Lazy-load fontoxpath to avoid bundling in desktop app
- Dynamic import() instead of static import for fontoxpath
- queryByXPath/matchByXPath now async (they were sync before)
- RpcCommand.execute allows R | Promise<R> for async commands
- Fix 2 lint errors (unnecessary optional chain, always-truthy)
- Update all tests and CLI to await results
|
2026-03-09 15:32:27 +03:00 |
|
Danila Poyarkov
|
661b83e749
|
Port analyze/diff tools from figma-use, split tools into domain files
Analyze: colors, typography, spacing, clusters
Diff: diff_create (tree diff), diff_show (preview changes)
Utility: get_components, get_current_page, arrange, node_to_component
Split schema.ts (2600 lines) into read, create, modify, structure,
variables, vector, analyze, registry — each under 600 lines.
Clean up inline types (use Color, Vector, SceneNode from existing defs).
Update AGENTS.md and CONTRIBUTING.md with code quality guidelines.
|
2026-03-05 19:16:23 +03:00 |
|