Releases: kavli-ntnu/opexebo
Opexebo v0.7.2
Update python-publish.yml
Opexebo 0.7.1
Merge pull request #45 from simon-ball/future Update build version to minimum
Opexebo v0.7.0
Update target Python and dependency versions
Opexebo v0.6.1
- Fix a problem with the default behaviour of
accumulate_spatial
when the default bin_width is specified- The default range of bins was selected wrongly in this case
Opexebo v0.6.0
** Signature breaking changes **
spatial_occupancy
andratemap
no longer provide a built in walk-filter.spatial_occupancy
no longeracceptsspeeds
as an argumentratemap
now acceptsspikes_tracking
as a 2 or 3 element array, not 3 or 4 -speeds
no longer accepted
- A stand-alone function has been provided instead:
walk_filter
- Provides high-pass filtering of an arbitrary number of arrays in parallel with
speed
(or other source)
- Provides high-pass filtering of an arbitrary number of arrays in parallel with
shuffle
restructured to better handle edge cases. No longer accepts keywordtracking_range
. If the shuffling time boundaries are not equal to the (min, max) oftimes
,t_start
andt_stop
may be specified- Function has been rewritten to improve legibility and performance.
** other changes **
- Expanded use of opexebo-specific errors
- fixed an issue in power_spectrum wth fft
Opexebo v0.5.5
Publication on PyPi
Version 0.5.2
- Support 1D data in spatial occupancy and ratemap
(Changes from previous, not officially released versions)
- Changed generic Python errors to Opexebo specific exceptions
- Add continuous integration testing
- Fix edge-case error in
accumulate_spatial
- Add
calculate_speed
- Add
egocentric_occupancy
Development version 0.4.3
- Cleaned up
accumulate_spatial
for clearer use and modification- NumPy convention is based on row-major: always invoked as (y, x) or (z, y, x)
- Opexebo uses standard mathematical notation, i.e. (x, y) or (x, y, z)
- The transofmrations actually cancelled each other out so that this was done correctly, but it was highly nonintuitive. Now modified to be easier to work with
- Added
upsampling
andcircular_mask
general functions - Mandatory keyword arguments (principally
arena_shape
andarena_size
) are now positional arguments. This is backwards compatible. peak_search
with thedefault
search method now correctly handles masking of all arrays, not just positive-only arrays. This seems to have been the source of continued errors ingrid_score_stats
- Documentation (mostly) upgraded to Sphinx standards
- Further development work on
border_score
- Mostly functional for rectangular arenas
- Circular arenas are WIP, may still require substantial re-working
Development version 0.4.2
-
Speed score changed substantially
- Adaptive filter fixed to behave correctly in the case of very small bin size
- Alternative filtering is available
- speeds are now smoothed as well - this was the source of the previous ~zero correlations
-
spatial cross correlation added
- Method to correlate two 1d or 2d arrays of the same size
-
code to get bin_number from bin_width grouped as a single function
-
Unit testing added for speed_score, accumulate_spatial
-
Unit testing for Ratemate and Placefield expanded
-
Unit testing sped up by disabling run-once code on import
-
Suppress errors where MaskedArrays use ">" symbol or similar
-
Code quality improvements
Development version 0.4.0
RateMap signature updated to include speeds
RateMap now filters spikes by speed
AngularOccupancy coverage now calculated correctly
SpatialOccupancy coverage now accounts for circular arenas
PopulvatioNVectorCorrelation added, in alpha status