-- it requires python 3.10 --
git clone [email protected]:mlionello/MySciLibManager_via_OpenAI.git
conda create -n lib python==3.10
conda activate lib
cd MySciLibManager_via_OpenAI
pip install -r requirements.txt
modify config.py by inserting your openai api key, and the prompt for GPT models modify main.py at line 59:
truncated_text = truncate_text(text, max_tokens=4000)
in case you want more inputs to be read from your library. Do change max_tokens higher than 10-15k to be sure all the content of the article is read
python app.py
if not done automatically, please navigate with your most favourite web browser to http://localhost:5000 or http://127.0.0.1:5000
Generate a CSV by reading all the pdfs in the subdirs of a specific parent folder
if not done automatically convert the csv into a db from the GUI
if not done automatically load the generated db directly from the GUI
You can play around by adding specific questions in RESULT_LABELS in config.py. For each new question be sure to:
- add a keyword for the database in DB_LABELS in config.py
- add the variable name in static_fields in app.py by keeping the same order of DB_LABELS