Add extract manifests command

[mjpitz]

The idea here would be to take the extraction logic from the indexer process and port it to the CLI as a dedicated command.

1. walk the file system and create a match request using the file names
2. call extractor.match
3. using the returned matches, read the contents from disk and construct an extract requests
4. call extractor.extract
5. output the result to the command line

[kendallm]

It seems like the extractor requires a repo url to extract dependencies.

Reading through the code for that though, it doesn't seem like the URL is actually used. Is there a reason why the repo URL is passed here?

With this functionality it seems like we wont have a repo URL as it will be run from a repo's root dir. Is it okay that the URL is not set or will that break something planned for the future?

[mjpitz]

The URL used to be / is used to infer a module name. I didn't persist this model in the v1beta model: https://github.com/depscloud/api/pull/70/files#diff-2d509e895e6903a5a290a5f41c6fba49R50-R53

[mjpitz]

You could pull the url from a .git/config file.

[abitrolly]

I guess there is no way to make it a deps subcommand, because deps is written in Go and all extraction logic is in JavaScript https://github.com/depscloud/depscloud/tree/main/services/extractor

Still is there any way to extract dependencies standalone from command line?

[mjpitz]

The extraction logic is a gRPC service so calling it from Go should be doable. Most of the logic is already implemented in the indexer process. Here's the related block of code: https://github.com/depscloud/depscloud/blob/main/services/indexer/internal/consumer/consumer.go#L155-L192