Repository Exporter

Python scripts that gather metadata for all repositories in a provided GitHub or Hugging Face organization and automatically export the data into a desired Google Sheet (using a Google Cloud Console Service Account) for easy viewing and analysis.

Contents

• Features
• Usage
• Set up your own GitHub Actions workflow
  • Create a GitHub Personal Access Token
  • Create a Hugging Face Token
  • Set up Google Cloud Service Account Access
• Run repo exporter locally
• Environment Variables
• Testing

---

Features

• Fetches all repositories in the indicated GitHub or Hugging Face organization(s)
• Collects key details:
  • Repo visibility, name (hyperlinked to repo), and description
  • Date created and last updated, with a flag for inactivity (set after a year of no commits)
  • Creator and top 4 contributors (N/A creator means it was either a transferred repository or a forked repository and None (<GitHub Username>) means there was no full name attached to their GitHub account)
  • Number of stars and number of branches for GitHub repositories, for Hugging Face repos, the analogous number of likes and open Discussions/PRs are collected
  • Standard file/metadata checks:
    • For GitHub: README.md, license, .gitignore, package requirements (requirements.txt, environment.yaml, etc.), CITATION.cff, .zenodo.json, and CONTRIBUTING.md files
    • For Hugging Face: README.md (dataset or model card/Space README) and license (read from yaml)
    • DOI for the repository (HF generated or from Zenodo for GitHub repos)
  • Primary Programming Language (GitHub only)
  • Website Reference/Homepage, Associated Dataset(s), Model(s), or Paper(s), and associated [GitHub] repo for Hugging Face repositories
  • Supports configurable worksheet names with defaults:
    • GitHub: GH_SHEET_NAME defaults to GH-Repos
    • Hugging Face: HF_SHEET_NAME defaults to HF-Repos
• Exports everything to a given Google Sheet document that it will require Editor permission to on the sheet's sharing permissions list
• For Standard Files highlights No data cell values with red cell colors and for Recommended Files and Filters highlights No data cell values with orange cell colors

Usage

The workflow runs automatically each week (9am UTC on Mondays); however, you can also run the GitHub Actions workflow manually:

1. Go to the Actions tab
2. Click Update Metadata for GitHub Repository Sheet
3. Click Run workflow, with branch as Branch: main, with selection all and finally press Run workflow

Set up your own GitHub Actions workflow

To use this script within your own GitHub organization, first fork this repo, then follow the setup steps below to ensure proper access.

Create a GitHub Personal Access Token

To create one with permissions for both private and public repositories (public repository read-access only is enabled by default without adminstrator approval):

1. Go to github.com/settings/personal-access-tokens
2. Click Generate new token -> Fine-grained token
3. Under Resource owner, select the organization you want to access.
4. Under Repository access, choose All repositories.
5. Under Permissions select Repositories and set:
   • Metadata -> Read-only
   • Contents -> Read-only
   • Administration -> Read-only
6. Click Generate token and copy it (make sure to store it somewhere safe for future use).
7. Navigate to https://github.com/<gh-org-name>/repo-exporter/settings/secrets/actions and click New repository secret and name it GH_TOKEN and copy paste the token into the Secret section and click Add secret Note: The token must be approved by the organization administrator before accessing private repositories.

Create a Hugging Face Token

To create one with permissions for both private and public repositoriesL

1. Go to huggingface.co/settings/tokens
2. Click on New Token and name it repo-exporter
3. For permissions select Fine-grained:
   • Specify the desired organization (under Org permissions)
   • Under Repositories, select "Read access to contents of all repos in selected organizations"
4. Click Generate and copy it (make sure to store it somewhere safe for future use)
5. Navigate to https://github.com/<gh-org-name>/repo-exporter/settings/secrets/actions and click New repository secret and name it HF_TOKEN and copy paste the token into the Secret section and click Add secret

Set up Google Cloud Service Account Access

Instructions to create a Google Cloud Console Service Account and give it permission to use in the repository and in the Google sheet:

1. Go to https://console.cloud.google.com/
2. Under "IAM & Admin", create a new project and name it inventory
3. Go to https://console.cloud.google.com/iam-admin/serviceaccounts, if you have multiple projects you'll need to select the project that you just made if it hasn't already been selected
4. Create a service account, named Imageomics, with description: "Repo checklist automation account" and finally press Done (You do not need to add any Permissions or Principals with access)
5. Click on the service account email -> Keys -> Add key -> Create new key and select JSON then finally click Create
6. Go to https://github.com/<gh-org-name>/repo-exporter/settings/secrets/actions and click New repository secret and name it GOOGLE_SERVICE_ACCOUNT_JSON and copy paste the entire contents of the JSON file into the Secret section and click Add secret
7. Go to https://console.cloud.google.com/apis/library/sheets.googleapis.com and enable the Google Sheets API for the project you made
8. Go to your chosen Google Sheet and go to Share settings and add the new Service Account email you made and set it as an Editor

After cloning the repository, configure the environment variables required for the exporter(s) you plan to run.

Once configured, the workflow can be run by following the Usage Instructions.

Run repo exporter locally

1. Clone this repository:

   git clone https://github.com/Imageomics/repo-exporter.git
   cd repo-exporter
   

2. Create and activate the Conda environment:

   conda create -n repo-exporter python -y
   conda activate repo-exporter
   

3. Create a .env file in the root of the project to configure required environment variables. See .env.example for an example; the default sheet names are included and can be removed.

4. Install Python dependencies:

   pip install -r requirements.txt
   

5. Run the exporters

   You can run either exporter individually or both, depending on your needs:

   • Run only the GitHub repository exporter

     python gh_repo_exporter.py
     

   • Run only the Hugging Face repository exporter

     python hf_repo_exporter.py
     

   • Run both exporters (wait for one to finish before running the other)

     python hf_repo_exporter.py
     python gh_repo_exporter.py
     

Environment Variables

Note

SPREADSHEET_ID and Google service account credentials are required for both GitHub and Hugging Face exports, as well as the platform-prefaced variables (e.g., GH_ORG_NAME and HF_ORG_NAME)

• Both exporters support optional custom sheet name variables (GH_SHEET_NAME and HF_SHEET_NAME; these must be distinct).

Warning

Tokens and Google service account credentials must be saved as repository secrets. All other values may be saved as environment variables assuming your spreadsheet is appropriately protected.

GitHub exporter

• Set GH_ORG_NAME to your GitHub organization name (for API calls).
• Set SPREADSHEET_ID to the Google Sheet ID used by the exporter.
• GH_SHEET_NAME is optional. If not provided, the exporter uses "GH-Repos".
• GH_TOKEN is required to access GitHub repositories.

Hugging Face exporter

• Set HF_ORG_NAME to your Hugging Face organization name (case-sensitive, for API calls).
• Set SPREADSHEET_ID to the Google Sheet ID used by the exporter.
• HF_SHEET_NAME is optional. If not provided, the exporter uses "HF-Repos".
• HF_TOKEN is required to access Hugging Face repositories.

Shared configuration

• Both exporters require Google service account credentials created in the service account setup. Set GOOGLE_CREDENTIALS_PATH to the JSON key file path (defaults to service_account.json if omitted).
• The Google service account must have Editor access to the target spreadsheet.
• Ensure all required values are available as environment variables locally or as GitHub Actions secrets.

For example, if the spreadsheet URL is:

https://docs.google.com/spreadsheets/d/fake-long-alpha-numeric-id-1a2b3c4d/edit

then the spreadsheet ID is:

fake-long-alpha-numeric-id-1a2b3c4d

---

Testing

Follow the local install instructions, then run the following in your repo-exporter environment:

python -m pytest -q