If login and caching are not required, you can directly deploy to platforms like Cloudflare Pages or Vercel. Just fork the repository and import it into the respective platform.
For Cloudflare Pages, you need to set the build command to `pnpm run build` and the build output directory to `dist/output/public`.
For login, which involves GitHub OAuth, you only need to [create a GitHub App](https://github.com/settings/applications/new). No special permissions are required. The callback URL should be `https://your-domain.com/api/oauth/github` (replace `your-domain` with your actual domain).
After creating the app, you will get a Client ID and Client Secret. Different platforms have different places to set environment variables; refer to the `example.env.server` file. If running locally, rename it to `.env.server` and add the necessary values.
This project primarily supports deployment on Cloudflare Pages and Docker. For Vercel, you need to set up your own database. Supported databases can be found at https://db0.unjs.io/connectors .
The Cloudflare D1 database can be used for free. To set it up, go to the Cloudflare Worker control panel and manually create a D1 database. Then, add the `database_id` and `database_name` to the corresponding fields in your `wrangler.toml` file.
If you don't have a `wrangler.toml` file, you can rename `example.wrangler.toml` to `wrangler.toml` and modify it with your configuration. The changes will take effect on your next deployment.
If you receive an error indicating that the database is empty during deployment, you can create a table arbitrarily. The first execution will automatically initialize the required tables. Alternatively, you can use the following SQL statement to create a Cache Table:
If you want to add data sources, refer to the `shared/metadata`, `shared/sources`, and `server/sources` directories. The project has complete types and a simple structure; feel free to explore.