├── .github
    └── workflows
    │   └── scrape.yml
├── README.md
├── TheBloke.json
└── requirements.txt


/.github/workflows/scrape.yml:
--------------------------------------------------------------------------------
 1 | name: Scrape Hugging Face models
 2 | 
 3 | on:
 4 |   push:
 5 |   workflow_dispatch:
 6 |   schedule:
 7 |   - cron:  '33 * * * *'
 8 | 
 9 | permissions:
10 |   contents: write
11 | 
12 | jobs:
13 |   fetch:
14 |     runs-on: ubuntu-latest
15 |     steps:
16 |     - uses: actions/checkout@v3
17 |       with:
18 |         fetch-depth: 0
19 |     - uses: actions/setup-python@v4
20 |       with:
21 |         python-version: '3.11'
22 |         cache: pip
23 |     - name: Install requirements
24 |       run: pip install -r requirements.txt
25 |     - name: Scrape
26 |       run: |        
27 |         curl --request GET \
28 |           --url 'https://huggingface.co/api/models?author=TheBloke' \
29 |           | jq > TheBloke.json
30 |     - name: Commit and push
31 |       run: |-
32 |         git config user.name "Automated"
33 |         git config user.email "actions@users.noreply.github.com"
34 |         git add -A
35 |         timestamp=$(date -u)
36 |         git commit -m "${timestamp}" || exit 0
37 |         git pull --rebase
38 |         git push
39 |     - name: Generate history
40 |       run: |
41 |         git-history file history.db TheBloke.json --id id
42 |     - name: Upload as artifact
43 |       uses: actions/upload-artifact@v3
44 |       with:
45 |         name: history
46 |         path: history.db
47 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # scrape-huggingface-models
2 | 
3 | This repository mainly exists to make it easy to [browse models by TheBloke using Datasette Lite](https://lite.datasette.io/?json=https://github.com/simonw/scrape-huggingface-models/blob/main/TheBloke.json#/data/TheBloke?_filter_column=id&_filter_op=contains&_filter_value=%2F&_sort=downloads&_sort_by_desc=on&_facet_array=tags).
4 | 
5 | **Archived April 21st 2024**.
6 | 


--------------------------------------------------------------------------------
/requirements.txt:
--------------------------------------------------------------------------------
1 | git-history
2 | 


--------------------------------------------------------------------------------