Skip to content

notes.md

Latest commit

 

History

History
67 lines (52 loc) · 2.12 KB

notes.md

File metadata and controls

67 lines (52 loc) · 2.12 KB
Raw

  • Modes of offering

      1. Webapp - drop the jupyter notebook and forge the project you want
      1. CLI
      1. VS-Code extension

  • Milestones :

    1. convert notebook to python project -- done
    1. stream lit -- done
    • to provide stream lit desc create a markdown with the header NBForge_Streamlit
    • if there are multiple such markdown blocks we pick the latest one
    1. fast api support -- DONE
    1. proper structure - DONE
    • add description for each function -- DONE
    1. update toml file -- DONE
    • optimze (remove unnecessary dependencies)
    1. setup logging -- DONE
    • add proper logging to each file -- ADDED TO converter.py , rest not adding
    1. proper packaging (toml) -- DONE
    1. format the files -- DONE
    1. convert to package [ IMP ] -- DONE
    • understand difference between
      • a. publishing a package that one can import
      • b. publishing a pakcage that people install and use in cli
    1. cli -- DONE optimze (remove unnecessary dependencies)
    1. publish to pypi
    1. webapp
    1. vs code extension

  • MAYBE : we will provide an option to create test cases for the modules

  • add anotations :

    • through comments probably
    • user can state whether to add this markdown as comment or as as main readme

  • if we are offering as cli then how to use the groq api key?

Tasks

  • revamp the prompt

    • make sure llm does not add anything of its own
    • it just goes over the code and separates the code and comments
    • should we allow it to revamp the comments and readme?

  • check the streamlit app . check whether it is correctly picking the csv file .

  • test the demo MSE calculation program with all 3

    • default modules
    • stream lit
    • fast api

  • Sample image from streamlit app generated using this package

  • alt text

  • alt text

  • Sample endpoint from fastAPI app generated using this package

  • alt text

CLI tool

  • Run the following command to covnert the jupyter notebook file at NOTEBOOK_PATH into a python module created in directory OUTPUT_DIRECTORY.
$ nbforge convert <NOTEBOOK_PATH> <OUTPUT_DIRECTORY>