https://github.com/daveshap/literaturereviewbot
Experiment to use GPT-3 to help write grant proposals.
https://github.com/daveshap/literaturereviewbot
Last synced: 10 months ago
JSON representation
Experiment to use GPT-3 to help write grant proposals.
- Host: GitHub
- URL: https://github.com/daveshap/literaturereviewbot
- Owner: daveshap
- License: mit
- Created: 2022-09-25T13:08:36.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2022-09-27T10:41:48.000Z (over 3 years ago)
- Last Synced: 2025-09-08T18:32:09.119Z (10 months ago)
- Language: Python
- Size: 7.42 MB
- Stars: 48
- Watchers: 2
- Forks: 13
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Literature Review Bot
## Setup
### Docker and Qdrant
1. Download and setup Docker and/or Docker Desktop
2. Pull the Qdrant container and run it: `docker run -p 6333:6333 -v ./qdrant_storage:/qdrant/storage qdrant/qdrant`
### Download arXiv metadata from Kaggle
1. https://www.kaggle.com/datasets/Cornell-University/arxiv
Or download my processed data directly (but it is out of date)
1. https://www.kaggle.com/datasets/ltcmdrdata/arxiv-embeddings (I will not be maintaining this)
### Process data
1. Run `generate_embeddings.py` to fill up `embeddings` folder (you may need to create this folder first)
2. Fire up Qdrant if its not already running
3. Run `index_arxiv_metadata.py` to upload embeddings to Qdrant
4. Run `search_server.py` and go to http://127.0.0.1 to search for your articles
5. Download the PDFs you want from arXiv into th `PDFs` folder
6. Run `generate_literature_review.py` to create your final literature review