Add partial file retrieval when using GridFS stream download

[laleksiunas]

Implemented support for partial file retrieval according to GridFS spec for GridFsBucket::open_download_stream and GridFsBucket::open_download_stream_by_name.
Unfortunately, there is a breaking change in GridFsBucket::open_download_stream interface - additional options argument to accept start/end range.

• start and end values are non negative (ensured by design using u64)
• validate whether start is less than or equal to end
• validate whether start and end are less than or equal to file length
• file chunk size is being used to skip and limit chunks when performing partial reads

[isabelatkinson]

Hi @laleksiunas, thanks for creating this pull request! The team is currently focused on other priorities, but we will take a look at this when we have more bandwidth in the future. One piece of immediate feedback I have is that making a breaking change here means that we would not be able to merge this work until we do a major release, which we don't currently have a time frame for. One way to avoid this breaking change would be to add a GridFsBucket::open_download_stream_with_range method rather than change the signature for the existing open_download_stream method.

[laleksiunas]

@isabelatkinson Thanks for review, I will add GridFsBucket::open_download_stream_with_range in the upcoming days

[nathanfranke]

I rebased this to upstream/main: nathanfranke@bed9f9c (two commits)
Tests aren't passing for me even on upstream/v3.5.1 or upstream/main; perhaps my fault.
Some code was moved between files and I made some members public, notably actions::gridfs::download::DownloadRange and those modules.

Otherwise, the async by_name API seems to work well; I did not test with IDs or sync.

---

One way to avoid this breaking change would be to add a GridFsBucket::open_download_stream_with_range method rather than change the signature for the existing open_download_stream method.

This should be resolved now. open_download_stream and open_download_stream_by_name use the builder pattern (e.g. .open_download_stream_by_name("test").start(0).end(100))

---

Test project:
rust-mongodb.zip

Note: The mongodb dependency path is hardcoded in this project.

rust mongodb: 100 x 1MiB
    WRITE: 534ms
    READ FULL: 320ms
    READ PARTIAL: 3ms