v0.8.13
Graphify-Labs/graphifyv0.8.13May 18, 2026by safishamsi
AI Summary
Fixes node ID collisions and absolute path storage issues. Introduces a stat-based fastpath for file hashing and exposes merge logic as library-callable CLI commands.
Key Highlights
- Node ID collisions fixed using directory-qualified stems
- Absolute `source_file` paths relativized at build time
- Failed semantic chunks no longer freeze files permanently
- Stat-based mtime fastpath significantly speeds up incremental re-runs
- New CLI subcommands: `cache-check`, `merge-chunks`, `merge-semantic`
New Features
- New CLI subcommands for cache and merge operations
Full Release Notes
## What's fixed - **Node ID collisions** across same-named files in different directories — SQL extractor and Python import resolver now use directory-qualified stems (`dir_file_entity`) instead of bare filename stems, preventing silent node merging on repos with duplicate filenames - **Absolute `source_file` paths** from semantic subagents no longer stored in graph — `build_from_json`, `build`, and `build_merge` now accept a `root` param and relativize paths at build time, fixing broken MCP traversal on mixed corpora (#932) - **Failed semantic chunks permanently freezing files** — only files that appear in extraction output get `semantic_hash` stamped in the manifest; failed-chunk files keep an empty hash and are re-queued on the next run (#933) ## What's improved - **Stat-based mtime fastpath for `file_hash`** — skips full SHA256 read when file size+mtime_ns are unchanged (same trade-off as make); index flushed atomically via atexit, making incremental re-runs significantly faster on large corpora - **New CLI subcommands** — `graphify cache-check`, `graphify merge-chunks`, `graphify merge-semantic` expose cache and merge logic as library-callable commands for skill pipelines ## Upgrade ```bash pip install --upgrade graphifyy ```