search

rmusser01 / tldw

tl/dw (Too Long, Didn't Watch): Your Personal Research Multi-Tool - a naive attempt at 'A Young Lady's Illustrated Primer'
Apache License 2.0
330 stars 11 forks source link

Improvement: Add support for user-defined chunking with API endpoints besides OpenAI #40

Closed rmusser01 closed 5 months ago

rmusser01 commented 5 months ago

As a user, I would like to have a choice of endpoints to use in order to be able to perform chunked summarization. I would also like the option of selecting my own API endpoint (Kobold/Llama/vLLM) for chunked summarization.

As a user I would like to be able to select one of the other API options besides OpenAI, and be able to select 'Enable Chunked Summarization' and have it work.

rmusser01 commented 5 months ago

Time-based chunking is in via CLI and GUI. Doesn't work with any API successfully Time-based chunking via word count ala how many words per sec, times seconds = chunks is in via CLI and GUI in testing. Breaks GUI though... 🤷

rmusser01 commented 5 months ago

(Simple) Baseline chunking code is implemented.

Chunking options are implemented properly in the GUI with options for changing the level of chunking.

Need to do: hook the two together, and write functions for performing the actual chunking and calling of summarization functions.