Releases: catalystneuro/roiextractors
v0.5.9
Deprecations
- Remove support for Python 3.8: PR #325
Features
- Add InscopixImagingExtractor: #276
- Updated testing workflows to include python 3.12, m1/intel macos, and dev tests to check neuroconv: PR #317
- Added API documentation: #337
- Optimized
get_streams
forBrukerTiffSinglePlaneImagingExtractor
by introducing a static functionget_available_channels
which uses lazy parsing of the XML to fetch the available channels: #344
Fixes
- Remove unnecessary
scipy
import error handling: #315 - Fixed the typing returned by the
InscopixImagingExtractor.get_dtype
method: #326 - Detect Changelog Updates was moved to its own dedicated workflow to avoid daily testing failures: #336
- Fixed the Daily testing workflows by passing along the appropriate secrets: #340
- Change the criteria of determining if Bruker data is volumetric #342
- Fixes a bug that assumes the channel name is is on the tiff file for
BrukerTiffSinglePlaneImagingExtractor
#343 - Including
packaging
explicitly in minimal requirements #347 - Updated requirements to include cv2 and update dev testing locations for neuroconv: #357
Improvements
- The
Suite2PSegmentationExtractor
now produces an error when a required sub-file is missin: #330 - Added
_image_mask
initialization inBaseSegmentationExtractor
; combinedabstractmethod
s into top of file: #327 - Optimize parsing of xml with
lxml
library for Burker extractors: #346 - Protect sima and dill export #351
- Improve error message when
TiffImagingExtractor
is not able to form memmap #353 - Updated Check Docstrings workflow to use new github action: #354
Testing
- Updated testing workflows to include python 3.12, m1/intel macos, and dev tests to check neuroconv: PR #317
- Added daily testing workflow and fixed bug with python 3.12 by upgrading scanimage-tiff-reader version: PR #321
- Remove wheel from requirements and move CI dependencies to test requirements PR #348
- Use Spikeinterface instead of Spikeextractors for toy_example PR #349
v0.5.8
Fixes
-
The triggering workflow name for update version was incorrectly set to
auto-publish
(the name of the yaml file). It has been renamed toUpload Package to PyPI
(the name field of the workflow): PR #304. -
Fixed bug with automatic changelog test that was causing it to fail for daily tests: PR #310
-
Updated zenodo to get a DOI on each release: No PR
v0.5.7
Features
-
Add support to get background components: add
get_background_ids()
,get_background_image_masks()
,get_background_pixel_masks()
toSegmentationExtractor
. PR #291 -
Add distinction for raw roi response and denoised roi response in
CaimanSegmentationExtractor
: PR #291 -
Bug fix for the
CaimanSegmentationExtractor
: correctly extract temporal and spatial background components PR #291 -
Added automatic version update workflow file that will run after publishing a new release to pypi: PR #290
-
Added
ScanImageTiffSinglePlaneMultiFileImagingExtractor
andScanImageTiffMultiPlaneMultiFileImagingExtractor
: PR #297 -
Added automatic changelog checking in the test workflow: PR #302
Fixes
v0.5.6
v0.5.5
v0.5.4
v0.5.3
v0.5.2
v0.5.1
Features
- Added a
has_time_vector
function for ImagingExtractors and SegmentationExtractors, similar to the SpikeInterface API for detecting if timestamps have been set. PR #216
Fixes
- Fixed two issues with the
SubFrameSegementation
class: (i) attempting to set the private attribute_image_masks
even when this was not present in the parent, and (ii) not calling the parent function forget_pixel_masks
and instead using the base method even in cases where this had been overridden by the parent. PR #215
v0.5.0
v0.5.0
Back-compatability break
- The orientation of traces in all
SegmentationExtractor
s has been standardized to have time (frames) as the first axis, and ROIs as the final axis. PR #200
Features
- Add support for newer versions of EXTRACT output files. PR #170
TheExtractSegmentationExtractor
class is now abstract and redirects to the newer or older
extractor depending on the version of the file. PR #170 - The
ExtractSegmentationExtractor.write_segmentation
method has now been deprecated. PR #170
Improvements
- Add
frame_to_time
toSegmentationExtractor
,get_roi_ids
is now a class method. PR #187 - Add
set_times
toSegmentationExtractor
. PR #188 - Updated the test for segmentation images to check all images for the given segmentation extractors. PR #190
- Refactored the
NwbSegmentationExtractor
to be more flexible with segmentation images and keep up
with the change in catalystneuro/neuoroconv#41
of trace names. PR #191 - Implemented a more efficient case of the base
ImagingExtractor.get_frames
throughget_video
when the indices are contiguous. PR #195 - Removed
max_frame
check onMultiImagingExtractor.get_video()
to adhere to upper-bound slicing semantics. PR #195 - Improved the
MultiImagingExtractor.get_video()
to no longer rely onget_frames
. PR #195 - Added
dtype
consistency check acrossMultiImaging
components as well as a direct override method. PR #195 - Added the
FrameSliceSegmentationExtractor
class and correspondingSegmentation.frame_slice(...)
method. PR #201 - Changed the
output_struct_name
argument to optional inExtractSegmentationExtractor
.
to allow more flexible usage for the user and better error message when it cannot be found in the file.
For consistency,output_struct_name
argument has been also added to the legacy extractor.
The orientation of segmentation images are transposed for consistency in image orientation (height x width). PR #210 - Relaxed rounding of
ImagingExtractor.frame_to_time(...)
andSegmentationExtractor.frame_to_time(...)
to be more consistent with SpikeInterface. PR #212