@amirdaraee/namewise 0.5.5 → 0.6.1

package/CHANGELOG.md

@@ -7,233 +7,121 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.6.1] - 2026-04-02
+
+### Fixed
+- Movie, music, series, and book templates now produce correct filenames; the AI is explicitly instructed to include the release year, artist name, season/episode, or author in the output rather than leaving unfilled `{year}`, `{artist}`, `{author}`, `{season}`, `{episode}` placeholders that were silently stripped
+- `--date` flag for document and photo templates now uses the document's own creation date from file metadata when available, instead of always substituting today's date
+
+### Changed
+- `--model` flag now works for all providers (Claude, OpenAI, Ollama, LMStudio), not just local LLMs; defaults are `claude-sonnet-4-5-20250929` for Claude and `gpt-4o` for OpenAI
+- Content sent to AI increased from 2000 to 5000 characters for more accurate filename generation on longer documents
+
+## [0.6.0] - 2026-04-02
+
+### Added
+- `--recursive` / `-r` flag to scan subdirectories, with optional `--depth <n>` to limit depth
+- `namewise undo [session-id]` command to reverse any previous rename session
+- `namewise undo --list` to view recent sessions with their IDs
+- Cascading config file support: `~/.namewise.json` (user) and `<dir>/.namewise.json` (project); CLI flags take highest priority
+- `--concurrency <n>` flag to process files in parallel (default: 3)
+- `--output <path>` flag to save a full JSON rename report after each run
+
+### Changed
+- Conflict handling: instead of erroring when a target filename already exists, the tool now auto-numbers the new file (`report-2.pdf`, `report-3.pdf`, etc.)
+- Rename history is now saved to `~/.namewise/history.json` after every run
+
 ## [0.5.4] - 2026-03-31
 
-### Improved
-- **🧪 Test Coverage**: Achieved 100% coverage across all metrics (statements, branches, functions, lines)
-  - 325 tests across 26 test files
-  - Added tests for all error branches including non-Error exception paths
-  - Full coverage for scanned PDF handling, folder-based categorization, and naming convention truncation
-  - Added polyfill tests for `process.getBuiltinModule` in Node.js < 22.3.0 environments
+### Changed
+- Achieved 100% test coverage across all metrics (statements, branches, functions, lines)
+- Expanded test suite to 325 tests across 26 test files
 
 ## [0.5.0] - 2025-11-10
 
 ### Added
-- **📄 Scanned PDF Support**: Full support for image-only (scanned) PDFs with vision AI
-  - Automatic detection of scanned PDFs (documents with minimal or no text)
-  - Converts first page to image and sends to AI for content analysis
-  - Intelligent image optimization to stay under Claude's 5MB limit
-  - Progressive JPEG compression with multiple quality levels (0.85 → 0.3)
-  - Automatic dimension scaling if needed (100% → 70% → 50%)
-  - Smart size calculation to ensure API compatibility
-
-### Enhanced
-- **🔄 PDF Processing**: Replaced PDF.js with pdf-to-png-converter for better Node.js compatibility
-  - Resolves canvas rendering issues in Node.js environment
-  - More reliable PDF-to-image conversion
-  - Better error handling and debugging
-- **🤖 AI Model**: Updated to Claude Sonnet 4.5 (claude-sonnet-4-5-20250929)
-  - Latest Claude model with enhanced vision capabilities
-  - Improved accuracy for document analysis
-  - Better understanding of complex document layouts
-
-### Technical
-- Replaced `pdfjs-dist` with `pdf-to-png-converter` package
-- Enhanced PDFToImageConverter with size optimization algorithms
-- Added comprehensive test suite for PDF-to-image conversion (12 new tests)
-- Improved error logging with detailed stack traces
+- Scanned PDF support: image-only PDFs are automatically detected and sent to vision AI for analysis
+- Automatic image optimization to stay within the Claude 5MB API limit (progressive JPEG compression and dimension scaling)
+
+### Changed
+- Replaced `pdfjs-dist` with `pdf-to-png-converter` for more reliable PDF-to-image conversion in Node.js
+- Updated default Claude model to claude-sonnet-4-5-20250929
 
 ### Fixed
-- **🐛 Critical**: Fixed "Image or Canvas expected" errors when processing scanned PDFs
-- **🐛 Critical**: Fixed "image exceeds 5 MB maximum" errors with large scanned documents
-- **🐛 Model**: Fixed deprecated model warnings by updating to latest Claude API
-
-### Examples
-```bash
-# Process directory with scanned PDFs
-namewise rename ./documents --dry-run
-
-# Scanned PDFs are automatically detected and processed:
-# Input:  Iran-criminal-record-2.pdf (scanned, no text)
-# Output: iran-criminal-record-certificate.pdf
-#
-# Input:  Luxembourg-identity-2025.pdf (scanned, 14.7MB image)
-# Output: luxembourg-identity-card-2025.pdf (optimized to <5MB)
-```
+- "Image or Canvas expected" errors when processing scanned PDFs
+- "image exceeds 5 MB maximum" errors with large scanned documents
+- Deprecated model warnings from the Anthropic API
 
 ## [0.4.1] - 2025-09-15
 
-### Enhanced
-- **🔧 Reliability Improvements**: Enhanced stability and consistency across all AI providers
-- **📋 Documentation**: Improved examples and usage documentation
+### Changed
+- Improved stability and consistency across all AI providers
 
 ## [0.4.0] - 2025-09-15
 
 ### Added
-- **🤖 Local LLM Provider Support**: Full integration with local AI services
-  - **Ollama**: Local LLM support with customizable models (llama3.1, codellama, etc.)
-  - **LMStudio**: Local model hosting with OpenAI-compatible API
-  - Support for custom base URLs and model selection
-  - Availability checking and model listing for local providers
-
-### Enhanced
-- **🎯 Intelligent Person Name Detection**: Major AI prompting improvements
-  - AI now extracts person names from document content and places them at filename beginning
-  - Smart folder name filtering to ignore irrelevant names like "no", "temp", "downloads"
-  - Enhanced prompts that focus on document content rather than metadata
-  - Support for detecting visa applications, contracts, medical records, certificates
-
-### Architecture
-- **📋 Centralized Prompt System**: Single source of truth for all AI prompts
-  - Model-agnostic prompting that works across Claude, OpenAI, LMStudio, and Ollama
-  - Consolidated prompt building in `/src/utils/ai-prompts.ts`
-  - Consistent behavior across all AI providers
-  - Easier maintenance and updates
-
-### Examples
-```bash
-# Use local Ollama service
-namewise rename ./documents --provider ollama --dry-run
-
-# Use LMStudio with custom model
-namewise rename ./files --provider lmstudio --base-url http://localhost:1234 --model codellama
-
-# Enhanced person name detection (Sarah example)
-# Input: visitor-visa-application-for-family-in-canada.pdf (in folder "no")
-# Output: sarah-visitor-visa-application-for-family-members-in-canada.pdf
-```
-
-### Technical
-- Enhanced CLI with new provider options and base URL configuration
-- Added availability checking and model discovery for local providers
+- Ollama provider support for local LLM inference (`--provider ollama`)
+- LMStudio provider support with OpenAI-compatible API (`--provider lmstudio`)
+- `--base-url` flag to configure a custom endpoint for local LLM providers
+- `--model` flag to specify the model for local LLM providers
+
+### Changed
+- AI prompts now extract person names from document content and include them at the start of the filename
+- Folder name hints are filtered to ignore common irrelevant names (e.g., "no", "temp", "downloads")
+- Centralized all AI prompt logic into `src/utils/ai-prompts.ts` for consistency across providers
 
 ## [0.3.1] - 2025-09-05
 
 ### Security
-- **🔒 Vulnerability Fix**: Replaced vulnerable `xlsx` package with secure `exceljs`
-- Enhanced Excel file parsing with improved security and reliability
+- Replaced vulnerable `xlsx` package with `exceljs` for Excel file parsing
 
-### Infrastructure
-- **⚙️ CI/CD Improvements**: Enhanced GitHub Actions workflows
-- Updated Node.js versions in CI pipelines
-- Improved test workflow reliability and build process
+### Changed
+- Updated CI pipelines to test against Node.js 20, 22, and 24
 
 ## [0.3.0] - 2025-09-05
 
 ### Added
-- **🎯 Personal File Templates**: Customizable templates for different file categories
-  - `document`: Personal docs with name and date (e.g., `driving-license-amirhossein-20250213.pdf`)
-  - `movie`: Movies with release year (e.g., `the-dark-knight-2008.mkv`)
-  - `music`: Music with artist names (e.g., `the-beatles-hey-jude.mp3`)
-  - `series`: TV series with season/episode (e.g., `breaking-bad-s01e01.mkv`)
-  - `photo`: Photos with personal info (e.g., `vacation-paris-john-20240715.jpg`)
-  - `book`: Books with author names (e.g., `george-orwell-1984.pdf`)
-  - `general`: General files without special formatting
-- **🤖 Smart File Categorization**: Automatically detects file type based on extension and content
-- **👤 Personal Name Integration**: `-n, --name` option to include your name in documents
-- **📅 Flexible Date Formats**: `-d, --date` option with formats:
-  - `YYYY-MM-DD`: 2025-09-05
-  - `YYYY`: 2025
-  - `YYYYMMDD`: 20250905
-  - `none`: No date (default)
-- **📂 Category Templates**: `-t, --template` option to specify file category or use auto-detection
-
-### Enhanced
-- AI prompts now include category-specific instructions for better filename generation
-- File processing pipeline includes template application after AI generation
-- Comprehensive test coverage with 131 tests (23 new tests for templates)
-
-### Examples
-```bash
-# Personal documents with your name and date
-namewise rename ./documents -t document -n "amirhossein" -d "YYYYMMDD" --dry-run
-# Result: driving-license-amirhossein-20250905.pdf
-
-# Movies with auto-detection
-namewise rename ./movies --dry-run
-# Result: the-dark-knight-2008.mkv
-
-# Series with season/episode detection
-namewise rename ./shows --dry-run  
-# Result: breaking-bad-s01e01.mkv
-
-# Music with artist names
-namewise rename ./music -t music --dry-run
-# Result: the-beatles-hey-jude.mp3
-```
+- File category templates: `document`, `movie`, `music`, `series`, `photo`, `book`, `general`, `auto`
+- `-t, --template` flag to select a category template
+- `-n, --name` flag to include a personal name in filenames
+- `-d, --date` flag with format options: `YYYY-MM-DD`, `YYYY`, `YYYYMMDD`, `none`
+- Auto-categorization based on file extension, folder path, and content keywords
 
 ## [0.2.0] - 2025-09-05
 
 ### Added
-- **Naming Convention Customization**: Added `-c, --case` option to choose naming convention
-  - `kebab-case`: lowercase-with-hyphens (default)
-  - `snake_case`: lowercase_with_underscores  
-  - `camelCase`: camelCaseFormat
-  - `PascalCase`: PascalCaseFormat
-  - `lowercase`: lowercaseformat
-  - `UPPERCASE`: UPPERCASEFORMAT
-- AI services now receive naming convention instructions and generate appropriately formatted filenames
-- Enhanced filename sanitization with convention-aware processing
-
-### Example Usage
-```bash
-# Use snake_case naming
-namewise rename ./docs --case snake_case --dry-run
-
-# Use camelCase naming  
-namewise rename ./docs --case camelCase --provider openai
-```
+- `-c, --case` flag for naming convention: `kebab-case`, `snake_case`, `camelCase`, `PascalCase`, `lowercase`, `UPPERCASE`
 
 ## [0.1.5] - 2025-09-05
 
-### Improved
-- Enhanced CLI user experience with single-line progress display that updates in place
-- Improved results output format: clear `original-name → new-name` display instead of confusing double checkmarks
-- Added progress counter showing current file being processed `(3/7)`
-- Cleaner console output with proper line clearing after processing
+### Changed
+- Progress display now updates in place on a single line with a `[current/total]` counter
+- Results output changed to `original-name -> new-name` format for clarity
 
 ## [0.1.4] - 2025-09-05
 
 ### Fixed
-- Fixed console output formatting where `\n` was displayed as literal text instead of newlines
-- Console output now properly displays line breaks for better readability
+- Console output was printing literal `\n` instead of newlines
 
 ## [0.1.3] - 2025-09-05
 
 ### Changed
-- Package renamed from `ai-rename` to `namewise` (clearer branding and avoids confusion with existing ai-renamer package)
-- CLI binary name changed from `ai-rename` to `namewise`
-- All documentation and references updated to reflect new name
+- Renamed package from `ai-rename` to `namewise`
+- Renamed CLI binary from `ai-rename` to `namewise`
 
 ## [0.1.1] - 2025-09-05
 
 ### Changed
-- Package renamed from `smart-rename` to `ai-rename` (original name was taken on NPM)
-- CLI binary name changed from `smart-rename` to `ai-rename`
-- All documentation and references updated to reflect new name
+- Renamed package from `smart-rename` to `ai-rename` (original name was taken on npm)
+- Renamed CLI binary from `smart-rename` to `ai-rename`
 
 ## [0.1.0] - 2025-09-05
 
 ### Added
-- Initial beta release of AI Rename
+- Initial release
 - AI-powered file renaming using Claude or OpenAI
-- Support for PDF, Word, Excel, and text files
-- Dry-run mode for safe previewing
-- File conflict detection and prevention
-- Configurable file size limits
-- Interactive API key prompts
-- Comprehensive test suite (65 tests, 90%+ branch coverage)
-- CLI with intuitive commands and options
-
-### Features
-- **Document Parsers**: PDF, DOCX, DOC, XLSX, XLS, TXT, MD, RTF
-- **AI Providers**: Claude (Anthropic) and OpenAI support
-- **Safety Features**: Dry-run mode, conflict detection, error handling
-- **Configuration**: Flexible options for provider, API keys, and file sizes
-
-### Technical
-- TypeScript implementation with strict typing
-- ESM module support
-- Node.js 18+ compatibility
-- Vitest testing framework
-- Commander.js CLI framework
+- Support for PDF, DOCX, DOC, XLSX, XLS, TXT, MD, and RTF files
+- Dry-run mode to preview changes without renaming
+- File conflict detection
+- Configurable file size limit
+- Interactive API key prompt when no key is set in the environment