Gstlal I/O fixes, part 2
This merge request addresses a few issues:
- Handles the case of only a single trigger for a given channel in a file by also checking for this when preallocating rows and setting to
np.nan
. - Fix filtering by segments that I think could've been introduced when adding a patch to work with the new gstlal feature schema.
- Greatly simplify a bunch of redundant code given that there are no nans now. I believe the mask was in fact used to filter segments properly before, but now it's kinda pointless since I'm doing all the filtering at the end when filtering by bounds happens.
- Removed filtering by nans at the end. It won't ever be needed anymore unless there's a bug somewhere upstream and trying to do this preemptively could hide that behavior leading to problems that may be hard to catch.
- Fixed a typo in the directory structure in the comments.