Releases: RatInABox-Lab/RatInABox
v1.4.0
v1.4.0 major changes
Two primary upgrades have been made in this release:
FieldOfViewNeurons
: arranged manifolds of egocentric BVC or OVCsNeurons
to make field-of-view tiling encodings.
from ratinabox.contribs.FieldOfViewNeurons import FieldOfViewNeurons
FoV_BVCs = FieldOfViewNeurons(Ag)
FoV_OVCs = FieldOfViewNeurons(Ag,params={
'cell_type':'OVC',
})
FoV_whiskers = FieldOfViewNeurons(Ag,params={
"FoV_angles":[75,105],
"FoV_distance":[0.1,0.2],
"spatial_resolution":0.02,})
-
Updates to how RatInABox figures are saved.
- User defined
RatInABox.figure_directory
sets where figures will be saved into. - By setting
RatInABox.autosave_plots = True
RiaB will automatically attempt to save figures into a figure directory. - Alternatively,
RatInABox.utils.save_figure(fig, figure_name)
can be used to manually save figures. - Figures saved by either of the above two methods will be put into a dated folder and appended with the current time (so no overwriting). This is a more sustainable way to make scientific figures, easing the load of saving each one by hand.
- User defined
-
Retired simultaneously having both a
1.x
and adev
branch in favour of just adev
branch. Easier to maintain. Users warned thatdev
may be unstable and that pip is the stable version.
Other "What's Changed"
- Minor changes by @TomGeorge1234 in #27
- Fix bug when sampling points in a non-rectangular environment by @frederikrogge in #26
- Bug fix for concave Environments by @TomGeorge1234 in #28
- Various merges made into 1.x by @TomGeorge1234 in #29
New Contributors
- @frederikrogge made their first contribution in #26
Full Changelog: v1.2.0...v1.4.0
v1.3.3 <-- v1.2.0
My release notes
Lots of new features, most important ones are listed on the README
1.3.3 because three minor bugs were caught after I published 1.3.0.
Major features:
-
FieldOfViewNeurons class for egocentric representations !!
-
Polygon environments and environments with holes
-
Generalised value function to multidimensional rewards
-
Object vector cells!
- Objects are inherent to the Environment (
Env.add_object()
) and can be multiple types. - OVCs are selective to one type (at a preferred distance and angle)
- Objects are inherent to the Environment (
-
Lots of figure plotting improvements
- Default figure formatting and saving functions
- Animations show inline on jupyter
- Trajectories can be coloured by time
- Rate timeseries figure changes (defaults look better now)
- Curved environments (just by virtue of allowing N-polygon envs)
- Plotting bug fixes and make defaults look prettier
- New
ratinabox.stylize()
function sets rcParams on import to make plots better - New
utils.save_figure()
saves figures and animation in dated folder for rapid access - Default colormaps changed (inferno for rate maps)
-
Optional interpolation of trajectories when imported
-
save_history() optional
-
reset_history() method
-
Objects inherent to environment (Env.add_object()), OVCs read off these
-
Lots of smaller bug fixes
-
Updates to the README
=====
Autogenerated Release Notes
What's Changed
- Minor changes by @TomGeorge1234 in #27
- Fix bug when sampling points in a non-rectangular environment by @frederikrogge in #26
- Bug fix for concave Environments by @TomGeorge1234 in #28
- Various merges made into 1.x by @TomGeorge1234 in #29
- Merge dev into 1.x by @TomGeorge1234 in #31
- animation save bug fix by @TomGeorge1234 in #32
- matplotlib install bug fix by @TomGeorge1234 in #33
New Contributors
- @frederikrogge made their first contribution in #26
Full Changelog: v1.2.0...v1.3.0
v1.2.0
This release add two new features:
ThetaSequenceAgent()
-- A newAgent()
subclass for generating hippocampal theta sequences. Class can be found incontribs
folder. Demonstrative video is on the README.- Support for any-polygon-shaped environments and adding holes into environments. Demonstrations on the README. Plotting functions have been updated to support this aesthetically.
Full release notes...What's Changed
- v1.2 <-- v1.1 by @TomGeorge1234 in #25
Full Changelog: v1.1.0...v1.2.0
v1.1.0
Some new features include:
some minor updates:
- Made
VelocityCells
subclass ofHeadDirectionCells
- Changes to
ratinabox/__init__.py
so it is possible to importEnvironment
,Agent
, andNeurons
, with:
import ratinabox as riab
Env = riab.Environment()
Ag = riab.Agent(Env)
PCs = riab.PlaceCells(Ag)
- README update
- Shift from
setup.py
tosetup.cfg
&pyproject.toml
style installs - Citation bibtex file
CITATION.bib
added.
v1.0.1
What's Changed
Only minor unnoticeable changes
- rate timerseries plotting sped-up
- Install metainfo shifted from legacy setup.py to setup.cfg, added pyproject.toml
Change log
- 1.x --> dev by @TomGeorge1234 in #14
- Merge dev into 1.x by @TomGeorge1234 in #15
Full Changelog
v1.0.0
v1.0.0, the initial release!
It is recommended to install this from PyPI (https://pypi.org/project/ratinabox/1.0.0/) via pip
.
Very minor incompatibility may be observed with pre-release versions, primarily how modules are imported.
Built and source binaries provided.