https://github.com/arokem/pubmed-gpt

Use gpt to summarize abstracts for a pubmed query
https://github.com/arokem/pubmed-gpt

Last synced: 12 days ago
JSON representation

Use gpt to summarize abstracts for a pubmed query

• Host: GitHub
• URL: https://github.com/arokem/pubmed-gpt
• Owner: arokem
• License: bsd-3-clause
• Created: 2023-03-26T03:50:22.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-03-26T14:31:02.000Z (over 1 year ago)
• Last Synced: 2024-08-02T01:22:48.616Z (3 months ago)
• Language: Python
• Homepage:
• Size: 9.77 KB
• Stars: 36
• Watchers: 1
• Forks: 5
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-ChatGPT-repositories - pubmed-gpt - Use gpt to summarize abstracts for a pubmed query (NLP)

README

        
# pubmed-gpt

Use gpt to summarize articles from pubmed as a slide presentation. The goal of

this software is to find the top 8 articles on a topic in pubmed (based on

pubmed's "relevance" criterion -- you'll need my fork of [pymed](https://github.com/arokem/pymed) for that) and ask GPT to summarize the abstract of each

article as a single slide, providing also a title slide and a summary slide,

which should hopefully synthesize the mess.

## Install

To install, download the source code and run:

    pip install .

In the top-level directory.

## Run

The software can be run by issuing:

    pmgpt

On the command line. To run this, you will need to get an [API key from OpenAI](https://platform.openai.com/account/api-keys).

The software will prompt you for a pubmed query (try "Glymphatic system" or

"Maiken Nedergaard", for example) and will then generate a (hopefully functioning) file

called "presentation.md" that should contain slides according to the

[remark](https://github.com/gnab/remark) format.

To view these slides, you can run:

    python -m http.server

And open a browser to http://localhost:8000/