podcaster/README.md

Front-end test for INDITEX

Instructions

1. Clone this repository
2. Run pnpm install. If you don't have pnpm installed, you can install it by running npm install -g pnpm.
3. For development, run pnpm dev. This will start the development server and open the browser. The assets served are not optimized for production (not minimized)
4. For production, run pnpm build. This will generate the optimized assets in the dist folder. To serve the assets, you can run pnpm serve or any other server of your choice.

About the solution

Caching requests

Requests are persisted for 24 hours in IndexedDB. It was chosen over localstorage due to its asynchronous nature and better performance, as well as local storage's 5MB limit.

By default, all requests are cached, since it aligns with the requirements of the test. However, this behaviour can be modified on per-request basis if needed.

RSS Parsing instead of iTunes API

The requirements specify that the description field should be rendered as HTML. The problem with using ITunes API is that it strips all HTML tags from the description field.

To circumvent this, the ITunes lookup API is only used to fetch the podcast details, from which the RSS feed URL is extracted. The RSS feed is then fetched and parsed to get the description field with HTML tags.

The description is later sanitized to remove any potentially harmful tags.

Known issues

• The RSS feed is fetched and parsed on the client side. This can be a performance bottleneck if the feed is large.
• Due to RSS feed being fetch on the client side, CORS issues can arise if the feed is not properly configured. The current solution tries to fetch the feed directly first, and if it fails, it tries to fetch it through a proxy. The proxy in question is often times very slow and is not a good solution for production. A better solution would be to fetch and parse the feed on the server side.
• RSS feed in not very well standardized, so some feeds might not be parsed correctly. In a real-world scenario, this would be handled on a case-by-case basis and would require a lot of testing.