Release Notes: v0.7.0 - 12/06/23
You may download the latest release with:
git clone -b v0.7.0 https://github.com/angelolab/ark-analysis.git
What's new?
ezSegmenter support
Starting with v0.7.0
, support for ezSegmenter masks has been added. The Jupyter notebook for this can be found in templates/ez_segmenter.ipynb
.
Please note that this feature may experience frequent updates. Updates may be required for Pixie integration and further downstream analysis.
Updated mask visualization
v0.7.0
introduces significant updates to the visualization process of masks. Shared functionality (ex. colormap generation, plotting clusters) has been abstracted away into internal functions and dataclasses. Erosion support has been added to decrease the size of the borders between cells, and interpolation
has been turned off to prevent interpolation of cell borders.
In the past, users reported problems of incorrectly sorted colors in the cmap file. This has been addressed. Additionally, users may now specify a dpi
and fig_size
, as well as multi-compatibility support for saving matplotlib
figures.
Mantis saving updates
Starting with v0.7.0
, only a single top-level Mantis directory will be created, and support for subfolders (via the img_sub_folder
param) has been added.
Pixie updates
During Pixie SOM assignment for pixel clustering, some users reported it taking unusually long. This was caused by the batch size for cluster assignment being too small. With this update, the batch size can now be set by the user to speed the process up, and defaults to a more reasonable value.
Additionally, in some Pixie applications, certain cells won't have any pixel clusters associated with them. These cells now get dropped prior to running the cell clustering portion.
Previously, the overwrite
flag did not work for certain functions because it wasn't propagated correctly. This has also been addressed by v0.7.0
.
Versioning Updates
ark-analysis
now uses scikit-image==0.19.3
and alpineer==v0.1.10
.
Licensing
ark-analysis
has migrated from the Modified Apache 2.0 License to the MIT license. This was made due to the cell segmentation algorithm, Mesmer, having a more restrictive license.
🚀 Features
- ezSegmentation @srivarra (#958)
- Mask Generation Refactoring @srivarra (#1030)
- Include
img_sub_folder
as parameter in Mantis calls @alex-l-kong (#1050) - Remove cells that don't have any pixel clusters expressed prior to Pixie @alex-l-kong (#1051)
- Single mantis directory @camisowers (#1061)
- Allow user to dynamically set batch size of pixel SOM assignments @alex-l-kong (#1069)
🐛 Bug Fixes
- Make sure overwrite functionality for pixel clustering propagates into helper functions @alex-l-kong (#1058)
- Adjustments to Mass Plotting Workflow @srivarra (#1067)
🧰 Maintenance
- Update README from v0.6.5 to v0.6.6 @alex-l-kong (#1048)
- Update HuggingFace to main @alex-l-kong (#1053)
- Pin to scikit-image to v0.19.3 @camisowers (#1060)
- Alpineer v0.1.10 @camisowers (#1057)
@alex-l-kong, @camisowers and @srivarra - MIT license @srivarra (#1080)
- Bump to v0.7.0 @alex-l-kong (#1093)