Skip to content

GDrive Cataloguer

GDrive Cataloguer #3

Workflow file for this run

name: GDrive Cataloguer
on:
workflow_dispatch:
schedule:
- cron: "50 0 24 * *"
jobs:
gdrive:
runs-on: ubuntu-latest
steps:
- name: Checkout the Code
uses: actions/checkout@v4
- name: Install Dependencies
run: |
pip install titlecase pyyaml pypdf tqdm yaspin bs4 google google-api-python-client google_auth_oauthlib joblib youtube-transcript-api
printf '${{ secrets.GTOKEN }}' > ~/gtoken.json
printf '${{ secrets.LIBRARY_UTILS_CLIENT_SECRET }}' > ~/library-utils-client-secret.json
printf "${{ secrets.ARCHIVE_ORG_AUTH }}" > ~/archive.org.auth
- name: Run Cataloguer
run: python scripts/refresh_catalog_doc.py