Skip to content

CHANGELOG.md

Latest commit

 

History

History
131 lines (101 loc) · 6.92 KB

CHANGELOG.md

File metadata and controls

131 lines (101 loc) · 6.92 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.1.0] — 2025-04-15

Added

Phase 1 — Core Scanning

  • Recursive directory scanning with depth limits and .forgeignore patterns
  • Content extraction from 30+ file formats (text, PDF, DOCX, XLSX, HTML, images)
  • SHA-256 exact duplicate detection with chunked hashing
  • AI classification via local Ollama (qwen3:4b model)
  • Rich terminal reports with category grouping and space estimates
  • SQLite session persistence for scan history

Phase 2 — Advanced Analysis

  • Near-duplicate detection via embeddings (nomic-embed-text + cosine similarity)
  • Stale file detection based on modification age and junk patterns
  • Version supersession detection (e.g., report_v1.docx → report_v2.docx)
  • --phase-2 flag to enable advanced features

Phase 3 — Interactive Actions

  • Interactive HTML reports with per-file action buttons
  • fileforge organize command for automatic file organization
  • fileforge dupes command for duplicate management
  • Dry-run mode for previewing actions
  • Trash staging with date-based organization
  • Action logging for undo support

Phase 4 — Background Automation

  • fileforge watch for continuous filesystem monitoring
  • fileforge schedule for cron-based automated scans
  • Systemd integration with service and timer units
  • Job audit trail for scheduled executions

Fixed

  • NameError bug in organize command (record.id scope issue)
  • Silent exception handling in extract_snippet (now logs specific errors)
  • Missing forgeignore error handling in organize and dupes commands
  • Broad exception catches replaced with specific handlers

Security

  • All file operations use safe path handling
  • No credentials or secrets in code
  • Graceful degradation when AI services unavailable

Documentation

  • Comprehensive docs/ folder with overview, architecture, setup, API reference
  • Tech stack documentation with dependencies and requirements
  • Updated README with quick start guide

[0.1.4] — 2026-05-17

Added

  • fileforge-server.service systemd unit for the long-running web UI
  • install.sh --scan|--server|--all modes in the systemd installer
  • Top-level scripts/install.sh now offers to register systemd services
  • scripts/release.sh verifies systemd templates are present before packaging

[0.1.5] — 2026-05-18

Added

  • CLAUDE.md — repository guidance for Claude Code sessions

Fixed

  • scanner.py: replace Path.is_dir/is_file(follow_symlinks=False) (Python 3.13+) with os.path equivalents for Python 3.11/3.12 compatibility

Changed

  • scripts/release.sh: stages CHANGELOG.md in the release commit and fails fast if the version entry is missing
  • scripts/install.sh: fixes missing blank line before the Development quick-start block

[0.1.6] — 2026-05-18

Fixed

  • api/server.py: /api/health endpoint returned hardcoded "version":"0.1.0"; now reflects __version__ from the package
  • systemd/fileforge-scan.timer: duplicate OnCalendar= directives caused the daily scan to fire at midnight AND 02:00; consolidated to OnCalendar=*-*-* 02:00:00

Changed

  • scripts/install.sh: systemd installer lookup now falls back to pip-installed package location when the dev-tree path is absent

[0.1.7] — 2026-05-19

Fixed

  • api/server.py: POST /api/scan rejected paths containing ~ (e.g. ~/Documents) with HTTP 400; now calls .expanduser() before validation and subprocess invocation
  • frontend/index.html: category filter dropdown compared raw LLM category paths (e.g. "Documents/Reports") against hardcoded keys ("document") — always false, hiding all rows; added normalizeCategoryKey() to map free-text paths to filter keys
  • frontend/index.html: category colour dot never rendered because getCategoryColor() was a literal string inside a CSS style attribute; now evaluated in JS before HTML injection
  • frontend/index.html: switching tabs did not sync the status <select> dropdown, leaving it showing a stale value

Added

  • frontend/index.html: "Quick scan (skip AI classification)" checkbox (checked by default) so web UI scans complete in seconds rather than hours on large directories
  • db.py: complete_session() sets completed_at timestamp on scan/organise/dupes finish; included in list_sessions() response

Changed

  • api/server.py: POST /api/scan is now non-blocking — returns immediately with a job_id; clients poll GET /api/job/{job_id} for status
  • api/server.py: added GET /api/browse?path= directory browser endpoint for the path-picker widget
  • packaging: upgraded packaging library to 26.2 to support Metadata-Version 2.4 (License-File field) in twine checks

[0.1.8] — 2026-05-19

Fixed

  • api/server.py + db.py: GET /api/session/{id} returned all records with no limit; sessions with millions of files sent a multi-GB JSON payload that froze the browser. Endpoint now accepts ?limit=500&offset=0 query params (default: first 500). Response includes total, limit, offset fields for pagination awareness.
  • frontend/index.html: page load fetches only first 500 records on large sessions; shows warning toast when session is truncated (e.g. "Showing first 500 of 2,417,064 files"). stat-total counter reflects session total, not truncated slice.

Added

  • db.get_session_file_count(session_id) — COUNT query returning total records without fetching rows
  • db.get_session_records() now accepts limit (default: 500) and offset (default: 0) parameters
  • Tests for get_session_records pagination and get_session_file_count in tests/test_db.py

Documentation

  • Comprehensive docs overhaul: all files updated to v0.1.7 (were stuck at 0.1.0)
  • docs/Blueprint_Overview.html: branded single-page HTML reference (open in browser)
  • docs/gap-analysis.md: tracks code vs. documentation discrepancies and planned features
  • docs/api.md: removed fictional actions/logger.py section; added complete REST API reference
  • docs/architecture.md: corrected module table (removed xlsx.py, html.py, image.py that don't exist); added web server diagram and async scan data flow
  • docs/tech-stack.md: added fastapi, uvicorn, Jinja2, packaging entries; clarified which formats have extractors vs. installed dependencies only
  • docs/setup.md: corrected fileforge[ui] → fileforge[web]; added fileforge server command reference
  • Removed stale docs/superpowers/plans/ planning docs (phases 2–4 are implemented)
  • src/fileforge/api/__init__.py: added missing package init file
  • scripts/install.sh: added Blueprint_Overview.html and gap-analysis.md to docs output

[Unreleased]

Planned

  • Phase 5: React PWA GUI
  • Phase 6: Cross-device sync
  • Additional extractor formats (RTF, EPUB, xlsx, OCR)
  • Custom classification models via config