Changelog
- Bug fixes for Arena mask detection tool
- Bug fixes
- Updated Wiki and notebook documentations
- Bug fixes
- Updated Wiki and notebook documentations
- Removed speed and scalar threshold from transition graphs
- Changed graph layouts
- Removed linear classifier in the analysis and modeling notebook
- Bug fixes
- Updated Wiki and notebook documentations
- All figures now exportable
- Changed Interactive ROI Detection Tool to Interactive Arena Detection Tool
- Updated
daskand related packages inmoseq2-pca
- Bug fixes
- Updated Wiki and notebook documentations
- All figures now exportable
- Changed Interactive ROI Detection Tool to Interactive Arena Detection Tool
- Fix kernel death issues in in
moseq2-pca
- Update to moseq2-app == 1.0.2beta
- Update to moseq2-model == 1.0.1beta
- Update to moseq2-viz == 1.0.2beta
- Update to moseq2-extract == 1.0.1beta
- Update to moseq2-pca == 1.0.1beta
- Update to moseq2-viz == 1.0.1beta
- Update to moseq2-extract == 1.0.0beta
- Update to moseq2-pca == 1.0.0beta
- Update to moseq2-model == 1.0.0beta
- Update to moseq2-viz == 1.0.0beta
-
Update to moseq2-extract == 0.5.0
-
Update to moseq2-pca == 0.3.0
-
Update to moseq2-model == 0.4.0
-
Update to moseq2-viz == 0.3.0
-
Added complete codebase documentation.
-
Added unit testing+integration testing suites.
-
Added Makefile to download test data from S3 bucket, (used for TravisCI testing)
-
Updated travisCI testing protocol; created job matrices testing MoSeq2 on python versions >=3.6
-
Refactored codebase to use wrapper functions, merging functionality between CLI and GUI
-
Locked all setup.py dependency versions
-
Added installed-version checking cli command
- First release!
- Added complete codebase documentation.
- Added unit testing+integration testing suites.
- Added Makefile to download test data from S3 bucket, (used for TravisCI testing)
- Updated travisCI testing protocol; created job matrices testing MoSeq2 on python versions >=3.6
- Refactored codebase to use wrapper functions, merging functionality between CLI and GUI
- Locked all setup.py dependency versions
- Added installed-version checking cli command
- Total coverage=81%.
- Added Jupyter Notebook compatible functionality
- Added troubleshooting capabilities: sample extractions, rois + previews
- Added option to extract/re-extract select recordings
- Added index file generation utility (moved from moseq2-viz)
- Added aggregate_results functionality
- Added safer extract-overwriting policy
- Added support for Azure cameras
- Added support for concave-shaped arenas (\_/)
- Total coverage=83%.
- Added Jupyter Notebook compatible functionality
- Added safer PCA training-overwriting policy
- Re-added user-defined h5-path option, per Issue #55
- Bug fixes
- Total coverage=86%.
- Added Jupyter Notebook compatible functionality
- Added option to model select groups/sessions
- Updated model-training checkpoint functionality
- Added default validation partitioning via scikit-learn train_test_split
- Added verbose model training capability, allows for plotting log-likelihoods across training iterations.
- Total coverage=81%.
- Added Jupyter Notebook compatible functionality
- Added capability to use custom colors for almost all plotting functions.
- Updated usage plotting functionality
- Added safer group selection method for plotting
- Added capability to sort syllables by mutant-ness with respect to any given plotting statistic (usage, syllable duration, syllable speed).
- Added position heat-map graphing functionality (by group or by individual session)
- Added syllable speed compute and graphing capabilities (based on rodent's centroid speed)
- Cleanup of h5 file structure
- h5 files now include more useful metadata
- Added support for sorting ROIs
- Added the copy-slice utility
- Added support for extraction directly from tarballs
- Fixed statsmodel incompatibility with new scipy
- Upgraded joblib
- Simplified installation
- Fixed timestamp bug #43
- Upgraded joblib
- Updated regularization and normalization options
- Clean up setup.py
- Fixed jitter issue with crowd movies
- Fixed flickering in crowd movies
- Numerous feature improvements
- Numerous feature improvements
- Added hole-filling to ROI detection
- Switched to new OpenCV 4 API (note that you must re-run pip install so that the correct version of OpenCV is used)
- Addressed #45
- Added support for parallel processing
- Fixed packaging
- Added ffv1 conversion for data compression
- Made tracking more robust per issue #42
- Made initialization of tracking model more robust
- Added gradient filtering for ROI detection in noisy environments
- Fixed packaging
- Fix for slurm initialization
- Added bash completion scripts
- Addressed #38
- Now prints more useful information on Dask initialization
- By default kappa is set to the number of frames
- Added bash completion
- Added ability to center and rotate mouse for crowd movies
- Fixed issue with feedback paths
- Fixed packaging
- Added information metrics per #28
- Added graphviz layout algorithms
- Added lots of options for laying out transition graphs
- Addressed #33
- Fixed packaging
- Added ffv1 conversion support
- Pinned joblib version
- Updated ruamel.yaml
- Cleaned up yaml status reports for batch jobs
- Addressed issue #9
First release!
- More explicit metadata for extracted data
- Cleaned up handling of Dask computations for PCA
- Lots of features added to
moseq2-viz - Added ability to visualize differences between graphs
- Added scalar visualization
- Added better handling of add-group
- Updated format of
moseq2-index.yamlnote that this is a breaking change, users of older versions will need to regenerate the index file - Commented important functions to use in notebooks
Home | Changelog | Acquisition | Installation | Analysis Pipeline | Troubleshooting and Tips |
Tutorials |
Join our slack 
