Using select_streams argument traverses file twice about pyxdf

[cbrnr]

I agree that duplicating chunk reading is not ideal, but at that time it was the quickest solution, because I could use functions from an alternative XDF reader I'd written. Also, these functions were mainly intended for querying the XDF file (resolve_streams, match_streaminfos), which I think would be nice to keep.

If you want to put in the effort to de-duplicate chunk reading code paths this would be great, provided that there are no regressions in functionality (i.e. keeping the two functions I've mentioned before).

[cboulay]

@cbrnr - I'm doing a little cleanup of pyxdf, deduplicating code, eliminating the double-traversal when using select_streams, adding some other shortcuts, etc.

Am I right in assuming that resolve_streams only needs stream headers? Indeed the functions it calls (parse_xdf, parse_chunks, _read_chunks) only operate on stream headers so the answer seems like an obvious "yes", but those functions' doc strings don't state this explicitly so I was just wondering if you were planning on building these out in the future?

In my deduplication effort, I'm reusing your _read_chunks method, but this means adding a lot of other functionality to it, only to disable that functionality with stream_headers_only=True when being called via parse_xdf (via resolve_streams). Is this compatible with your use-case?

[cbrnr]

I'll answer in #74.