Skip to content

cleanup.md

Latest commit

 

History

History
94 lines (71 loc) · 4.27 KB

cleanup.md

File metadata and controls

94 lines (71 loc) · 4.27 KB
Raw

Cleanup

I'd like to make this codebase easier to maintain and use.

  • Please make a todolist at the end of this file with steps to do in this format
  • Item 1
  • etc

The goal is to:

  • remove all audio recording and processing code, including required pip packages
  • remove any other code that isn't directly related to OCR capture, storage in chroma, and availability over NGROK

Please also make a TODO iteam for an example readme that is 200 or fewer lines. Want to make sure installation is easy for hacker news crowd

Also please move all of these .md files to a folder called workspace

TODO List

  • Remove audio recording Python files

    • Delete audio_recorder.py
    • Delete audio_background_recorder.py

  • Remove audio-related shell scripts

    • Delete setup_audio_recorder.sh
    • Delete start_audio_background.sh

  • Remove audio-related requirements file

    • Delete audio_requirements.txt

  • Remove audio-related directories

    • Delete audio_env/ virtual environment directory
    • Delete audio_sessions/ directory (if exists)
    • Delete refinery/data/audio/ directory (if exists)

  • Remove audio-related markdown documentation files

    • Delete AUDIO_BACKGROUND_README.md
    • Delete AUDIO_RECORDER_README.md
    • Delete AUDIO_SETUP_COMPLETE.md
    • Delete AUDIO_SETUP_GUIDE.md
    • Delete AUDIO_DUAL_CAPTURE_SUMMARY.md
    • Delete AUDIO_SEARCH_FIX.md

  • Remove audio references from code files

    • Remove audio-related code from mcp-server/tools/search.py (audio file reading, audio search logic)
    • Check and remove audio references from mcp-server/server.py
    • Update comments in refinery/run.py (remove "Tag to differentiate from audio" comments)
    • Update comments in refinery/load_ocr_data.py (remove "Tag to differentiate from audio" comments)

  • Remove audio references from README.md

    • Remove audio recording setup section
    • Remove audio transcription pipeline description
    • Remove audio data type references from search capabilities
    • Remove audio-related environment variables
    • Remove audio-related architecture diagrams and descriptions

  • Remove audio-related pip packages from requirements files

    • Check refinery/flow-requirements.txt for audio packages (pyaudio, whisper, openai if only used for audio)
    • Check mcp-server/requirements.txt for audio packages
    • Remove any audio-specific dependencies

  • Remove audio log files

    • Delete audio_background.log (if exists)

  • Remove non-core directories and folders

    • Delete dashboard/ directory (web dashboard not needed for core OCR/ChromaDB/NGROK)
    • Delete images/ directory (images not needed for core functionality)
    • Delete website-builder/ directory (website builder not needed for core functionality)
    • Delete refinery/data/test_images/ directory (test data not needed)
    • Delete refinery/test_screenshot.py (test file not needed for production)
    • Delete refinery/countdown_broadcaster.py (dashboard-related, not needed for core functionality)

  • Remove website-builder related code from MCP server

    • Check mcp-server/tools/website.py - remove or update if website builder functionality is not core
    • Check mcp-server/http_server.py - remove website-builder references if not needed for NGROK MCP server

  • Move all markdown files to workspace folder

    • Create workspace/ directory if it doesn't exist
    • Move all .md files from root directory to workspace/ (except README.md and cleanup.md)
    • Update any references to moved markdown files in code or documentation

  • Create simplified README (200 lines or fewer)

    • Create new simplified README focusing only on OCR capture, ChromaDB storage, and NGROK availability
    • Include quick start instructions
    • Include Claude Desktop integration steps
    • Include basic troubleshooting
    • Keep installation instructions simple and clear for hacker news audience
    • Remove all audio-related content
    • Remove team collaboration features (if not directly related to core functionality)
    • Remove dashboard references (if not core to OCR/ChromaDB/NGROK)
    • Focus on essential setup and usage only