Project Vectorizer Release v0.1.4

[0.1.4] - 2025-10-13

Fixed

• **Hardcoded value ** – Replaced hardcoded configuration with dynamic variable lookup

Notes

• This is a minor bugfix release with no API or CLI changes.

Project Vectorizer Release v0.1.3

[0.1.3] - 2025-10-13

Fixed

• Hardcoded value in work module – Replaced hardcoded configuration with dynamic variable lookup
  • Prevents unexpected behavior when running with custom configs

Notes

• This is a minor bugfix release with no API or CLI changes.

Project Vectorizer Release v0.1.2

[0.1.2] - 2025-10-13

Added

Performance & Optimization

• Optimized Config Generation - New Config.create_optimized() method that auto-detects optimal settings based on system resources
  • Auto-detects CPU cores and sets optimal max_workers
  • Calculates safe batch sizes based on available RAM
  • Dynamically adjusts memory thresholds and GC intervals
  • Use with pv init --optimize for permanent optimization
• Max Resources Flag - New --max-resources flag for index and index-git commands
  • Temporarily overrides config with optimized settings
  • Perfect for one-time performance boosts
  • Example: pv index . --max-resources
• psutil Integration - Added psutil library for system resource detection
  • Automatic CPU core detection
  • Memory availability monitoring
  • Smart resource allocation

Progress & UI Improvements

• Unified Progress Tracking - Clean, single-line progress bar for all indexing operations
  • Replaces cluttered per-file logging
  • Real-time progress updates with file names
  • Shows file status tags: [New], [Modified], [Deleted]
  • Callback-based architecture for flexible progress reporting
• Library Progress Bar Suppression - Suppressed sentence-transformers batch progress bars
  • Added show_progress_bar=False parameter to embedding generation
  • Eliminates "Batches: 0%|..." clutter during indexing
• Timing Information - All indexing operations now display elapsed time
  • Shows duration in seconds (< 1 min) or minutes+seconds (≥ 1 min)
  • Displayed in completion panels
  • Example: "Time taken: 2m 17s"
• Clean Terminal Output - Professional, easy-to-read output
  • Automatic log suppression when progress callback is active
  • Warnings and errors still shown
  • Verbose mode available with --verbose flag

Documentation

• Comprehensive Documentation Index - New docs/README.md with organized navigation
  • Links to all feature guides
  • Quick reference section
  • CLI command reference
  • Configuration reference
  • Troubleshooting guide
• Performance Guides - Three detailed guides added:
  • MAX_RESOURCES_GUIDE.md - Using maximum system resources
  • OPTIMIZED_CONFIG.md - Auto-optimization features
  • CLEAN_PROGRESS_OUTPUT.md - Progress tracking system
• Changelog - This file! Track all version changes

Changed

Configuration

• Chunk Size Enforcement - Engine now enforces max 128-token chunks (line 35 in engine.py)
  • Improved precision for searches
  • Better granularity for code matching
  • Configurable chunk size still respected for larger values
• Default Chunk Overlap - Optimized overlap settings
  • Better balance between precision and performance
  • Reduced from 32 to 16 tokens for 128-chunk configs

Performance

• Batch Processing - Memory-based batch size validation
  • Safe batch sizes calculated based on available RAM
  • Prevents out-of-memory errors on large projects
• Parallel Processing - Improved worker management
  • Auto-detection of optimal worker count
  • Better CPU utilization
  • Configurable via max_workers setting

CLI

• Progress Output - All index commands now use unified progress bar
  • pv index - Shows clean progress with timing
  • pv index-git - Shows git reference and timing
  • pv sync --watch - Real-time progress updates

Fixed

• Progress Bar Clutter - Fixed multiple progress bars appearing during indexing
• Sentence-Transformers Output - Suppressed library-internal batch progress bars
• Memory Usage - Better memory management with GC intervals and monitoring
• Log Suppression - Informational logs now properly suppressed when progress callback is active

Performance

Benchmarks (with --max-resources)

• Indexing Speed: ~2m 16s for 48 files (9,222 chunks) on 8-core system
• Chunk Size Impact: Virtually identical performance between 128 and 512 token chunks
• Optimal Settings: Auto-detected based on system (16 workers, 400 batch size on 8-core/16GB RAM)

Improvements

• Smart Incremental: 60-70% faster than full re-indexing
• Git-Aware: 80-90% faster for recent changes only
• Parallel Processing: Linear scaling with CPU cores

Project Vectorizer Release 0.1.0

Refactor GitHub Actions workflow for PyPI publishing

- Simplified the workflow by removing the TestPyPI publishing step and associated conditions.
- Streamlined the environment input options for manual deployment.
- Ensured consistent formatting and indentation across steps for better readability.
- Maintained the core functionality for building and publishing to PyPI.