- change path to precurated output from the
sorting_precurated
tocurated
folder - pin to
aind-data-schema==0.19.1
- update to
spikeinterface==0.100.0
- accept
behavior_*
as session name - add `ArgumentParser`` to be able to pass discrete command line arguments
- extended available CLI arguments to control removal of out/bad channels, max fraction of bad channels, n_jobs, parameter file
- some renaming:
preprocessing_strategy
->denoising_strategy
max_bad_channel_fraction_to_remove
- >max_bad_channel_fraction
- better error handling
- better handling of spike sorting failures
- fix creation of DerivedDataDescription from scratch
- fix visualization of drift map for short recordings using peak pipeline
- update to
spikeinterface==0.99.1
- update to
aind-data-schema==0.17.1
- update to
wavpack-numcodecs==0.1.5
- add
min_processing_duration
(default 120), to skip processing for short (test) recordings - added
amplitude_median
,amplitude_cv
,firing_range
,synchrony_metrics
,nearest-neighbor
- include multi_channel template metrics (
velocity_above
,velocity_below
,spread
,exp_decay
) - change default correlograms params: 1ms bin, 50ms window
- add protection against recordings without channel locations
- dump preprocessed recording JSON to
preprocessed/
to easily reload processed object - use
UpgradeProcessing
to upgrade old processing schema models - Add nearest neighbor (and fix parameters) and silhouette metrics
- Update to
aind-data-schema==0.15.12
- Update to
spikeinterface==0.98.2
- Use new
Processing
schema
- Update data description name from 'Spike Sorting' to 'sorted'
- Upgrade to
wavpack-numcodecs==0.1.4
- Use
aind-data-schema.DataDescriptionUpgrade
- Update to
spikeinterface==0.98.1
- Fix bug in drift map path
- updates `Processing`` construction
- Fix bug in
detect_peaks
- Fix
matplotlib.Normalize
for drift plot
- clean up verbosity
- install fonts for sortingview
- first release,
spikeinterface==0.97.1