[Best Practices Violation] Do not hard-code an embeddings file batch size

Validations

• I believe this is a way to improve. I'll try to join the Continue Discord for questions
• I'm not able to find an open issue that requests the same enhancement

Problem

https://github.com/continuedev/continue/blob/f29e1849f3169b7dc8f9586f0b258dd82626a3c8/core/indexing/CodebaseIndexer.ts#L34-L40

The above code sets a hard-coded limit on the number of batched files processed by the Continue extension. There is no way to see this number or to override it by the end-user.

This has a secondary impact on what's shown to the end user while embeddings are being calculated. Confusingly, the progress bar seems to be operating off a different number? Or maybe it isn't at all.

That string is actually built in a completely different place:

https://github.com/continuedev/continue/blob/f29e1849f3169b7dc8f9586f0b258dd82626a3c8/core/indexing/LanceDbIndex.ts#L300-L306

Whereas the progress bar is built here:

https://github.com/continuedev/continue/blob/f29e1849f3169b7dc8f9586f0b258dd82626a3c8/core/indexing/CodebaseIndexer.ts#L418-L423

which is then used below to calculate progress percentage:

https://github.com/continuedev/continue/blob/f29e1849f3169b7dc8f9586f0b258dd82626a3c8/core/indexing/CodebaseIndexer.ts#L445-L448

Solution

• Make it user configurable.
• In the UI, change the phrase "Computing embeddings for X files" such that X is totalOps instead of LanceDb.results.compute.length, which at that point is actually being called from within a batch