refactor: resource fetcher

[chmjkb]

Description

• Extracted shared base class (BaseResourceFetcherClass) into the core package containing the fetch loop, source dispatching, and pause/resume/cancel plumbing — eliminating duplication between expo and bare
  implementations
• Extracted platform-specific handlers (handlers.ts) for both packages, separating download mechanics from the class structure
• Rewrote both fetchers as classes extending the base, implementing only platform-specific file system and download operations
• Replaced the linked-list download pattern with a simple sequential for loop
• Cancel now throws DownloadInterrupted instead of returning null, giving callers a consistent error-based API

Note: this PR still uses the legacy FS as the support for background / resumable downloads is WIP.

Introduces a breaking change?

• Yes
• No

Type of change

• Bug fix (change which fixes an issue)
• New feature (change which adds functionality)
• Documentation update (improves or adds clarity to existing documentation)
• Other (chores, tests, code style improvements etc.)

Tested on

• iOS
• Android

Testing instructions

1. Create a demo app trying to run all the functionalities exposed in Resource Fetchers
2. Fetch function with single and/or multiple resources
3. Cancel long fetching file, also verify fetch throws an error when its canceled
4. Check if a model with downloaded file works
5. Make sure deleting resources work
6. Try running a background download

Screenshots

Related issues

Checklist

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have updated the documentation accordingly
• My changes generate no new warnings

Additional notes

[msluszniak]

Please run yarn lint --fix in the root of the package and push changes

[mkopcins]

remove from here, i want those calls to be strictly inside core package so that they are mandatory.
Also there is a bug when we fire those calls even when file is already downloaded (see packages/react-native-executorch/src/utils/ResourceFetcher.ts:128

[mkopcins]

Two things here

1. why are we checking with (dl.task) and the other time with (dl.jobId !==...) instead of (!dl.jobId)
2. why are we using those fields (task/jobId) as a proxy for checking what platform we are using, especially since you are then writing comments explaining the platforms

[chmjkb]

regarding the first point, jobId can be 0 which is falsy, so it would incorrectly tread a valid job with id 0 as missing. Task is an object ref, so !downloadHandle.task is safe.

[chmjkb]

and regarding the second point, I could do something like this instead, but I'm not sure if its much cleaner though:

    if (Platform.OS === 'ios') {
      if (!downloadHandle.task) {
        throw new RnExecutorchError(
          RnExecutorchErrorCode.ResourceFetcherDownloadFailed,
          'Download task is missing. This should not happen on iOS.'
        );
      }
      downloadHandle.task.stop();
    } else if (Platform.OS === 'android') {
      if (downloadHandle.jobId === undefined) {
        throw new RnExecutorchError(
          RnExecutorchErrorCode.ResourceFetcherDownloadFailed,
          'Download job ID is missing. This should not happen on Android.'
        );
      }
      RNFS.stopDownload(downloadHandle.jobId);
      if (await ResourceFetcherUtils.checkFileExists(downloadHandle.cacheFileUri)) {
        await RNFS.unlink(downloadHandle.cacheFileUri);
      }
    }