scrape-service-lib

Generated TypeScript client wrapping the marketing scrape service API.

Project structure

• api/: OpenAPI definitions (api.cleaned.yaml drives the generator).
• scripts/: helper scripts for the generator (custom request + wrapper CLI).
• src/: TypeScript source generated from the OpenAPI spec plus lightweight wrappers.
• dist/: compiled output produced by pnpm run build.

Key scripts

• pnpm run generate: regenerates the TypeScript client from the OpenAPI spec.
• pnpm run build: compiles the client into dist/.
• pnpm run prepare: convenience script that runs both commands; executed automatically before publishing.

Publishing to npm

1. Store the npm automation token as the NPM_TOKEN secret in the GitHub repository (Settings → Secrets and variables → Actions).
2. Push changes to main or production. The Publish library to npm workflow builds the client and runs npm publish --access public.
3. Ensure package.json#version is bumped before pushing, otherwise the publish step will fail because npm rejects duplicate versions.

The package is published under the scoped name @viechoice/scrape-service-lib with public access. Install directly via:

pnpm add @viechoice/scrape-service-lib

How to consume the client

1. Install normally in your application:

   pnpm add @viechoice/scrape-service-lib

2. Instantiate the generated client (simplified example). If you pass a TOKEN (or per-request headers.Authorization), the client skips the browser localStorage refresh logic automatically:

   import { AppClient, DefaultService } from '@viechoice/scrape-service-lib';
   
   const client = new AppClient({
     BASE: 'https://your-api-host',
     TOKEN: process.env.SCRAPE_SERVICE_TOKEN,
     HEADERS: {
       'x-api-key-service': process.env.SCRAPE_SERVICE_API_KEY ?? '',
     },
   });
   
   const data = await DefaultService.scrapeControllerScrape({
     body: {
       socialNetwork: 'FACEBOOK',
       contentType: 'POST',
       url: 'https://facebook.com/some-post',
     },
   });
   
   console.log(data);

3. For CI, you only need npm auth if the registry requires it globally (public packages do not). Keep NPM_TOKEN handy for publishing workflows.