Open/read/close API for WFDB format #100
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Implement the open_waveforms / close_waveforms / read_opened_waveforms API as defined in BaseFormat. Since wfdb-python doesn't have any API for doing this, we implement I/O for formats 16 and 516 "by hand". This isn't a complete or conformant implementation of the WFDB format and not recommended for applications to use as-is - it should work for the data files produced by write_waveforms, but not for arbitrary WFDB records. But it should be useful in better quantifying the performance overhead of the wfdb-python API. For each signal format, a custom reader class is implemented. Format 16 is implemented using numpy, much like wfdb-python does; this class could be adapted to handle formats 80, 61, 160, 24, or 32 if needed. Format 516 is implemented using soundfile, again much like wfdb-python does (but lacks the workaround for libsndfile bug #431), and could be adapted to handle formats 508 or 524 if needed.
|
@bemoody , thanks, this looks good to me! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Here's a (basic) implementation of the new open/read/close API for the WFDB format classes.
As we've discussed elsewhere, the wfdb-python package doesn't currently have any API for doing this sort of thing (which is certainly on the wishlist!)
So this is a quick re-implementation that just handles the specific (single-segment, format 16/516) formats that we're using for benchmarking.
Since it is a re-implementation, it avoids a lot of the overhead of the wfdb.rdrecord function and consequently the performance numbers are not going to be directly comparable to what you get when using rdrecord - I expect the performance will be better, because there are a number of things in wfdb-python that are not optimally implemented. But let's try it and see. :)
You might also notice that the array returned by
read_opened_waveformsis slightly different from the array returned byread_waveforms, due to FP rounding errors. Probably because the former multiplies by 1/gain where the latter divides by gain.This does rely on having a fairly recent version of libsndfile. I think (but I'm not 100% sure) that the soundfile packages on PyPI are sufficiently recent.