/supernote-llm

Consuming contents of a Supernote Notebook for use with LLMs

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

supernote-llm

Consuming contents of a Supernote Notebook for use with LLMs.

Bullet Journal + LLM

The bullet_journal_llm/ directory contains notebooks and libraries for parsing bullet journal entries using Gemini 1.5 Flash. Using the pipeline needs the following steps:

  • Download .note files and put them in notes/
  • Extract pngs form the notebooks (see below) in pages/
  • Run the pipeline which using a dynamic prompt based on the filename
  • The results are stored in docs/ which you can then index for searching

Tools

The https://github.com/jya-dev/supernote-tool can be used to extract metadata from a supernote file. We have a couple tools in this repo using the above as library to process content in a directory.

See extract-text/ and extract-png/. Populate notes/ with supernote files that contain text. The text will be extracted into texts/ or pages/.

$ python3 extract-text

Supernote index

See jupyter notebook following the LLama Index https://docs.llamaindex.ai/en/stable/getting_started/starter_example.html tutorial.