Celldetective

Introducing helpers
================

Help buttons have invaded the Celldetective interface. These buttons trigger a series of questions that, we hope, will help the user choose the best option for her data, at critical analysis stages.

Notable changes
==============

- The survival plot and signal plot UIs have been completely remade to be more functional. It is now possible to export as high quality PNGs your results or make single-cell signals more transparent. The functionality to plot two mean signals in a phase space has been temporarily decommissioned. It will come back in a future version.
- In the signal annotator, you can now interact with the cell signal figure and export it.
- The classifier widget has been restructured to contain only GUI-related methods. In other words, it is now possible to use functions to perform the classification of cells and tracks in Jupyter Notebooks.
- Similarly, we now provide a `compute_survival` function to make survival functions more easily in notebooks.
- The multithreading mechanism has been replaced with a ThreadPoolExecutor for performance/stability.
- We provide a new function to replace missing segmentation labels with black frames when viewing a partial or bugged segmentation result.
- Expanding / shrinking the panels should look much better now with a smarter resizing function.
- Fix a bTrack bug when dummies are created, which were wrongfully attributed real labels.
- Package requirements have been relaxed to make Celldetective installable with Python 3.9 - 3.11 (included), on both Ubuntu and Windows.
- Bug fixes

Contributors
==========

- remyeltorro

**Full Changelog**: https://github.com/remyeltorro/celldetective/compare/v1.2.1...v1.2.2.post2

- Train segmentation models on samples with one or more channel missing (the missing channels are replaced with a black frame)
- Bug fixes

Contributors
ksenijader
remyeltorro

**Full Changelog**: https://github.com/remyeltorro/celldetective/compare/v1.2.0...v1.2.1

Introducing interactions
====================

With version 1.2.0, Celldetective provides a complete environment to study cell-cell interactions. The neighborhoods (isotropic or mask-contact) have been reorganized as a preliminary step to set the cell-pairs of interest.

A new cell-pair measurement feature allows the quantification of cell pairs (relative distance, velocity, angle, angular velocity), stored in a cell-pair table that can be explored like the trajectory tables.

An expanded signal annotator visualizer allows the interactive classification of cell pairs, complete with the possibility to export a dataset and train a classifier/regressor DL model to interpret the signals from pair, the reference and the neighbor populations.

Contributors
===========

- ksenijader
- remyeltorro


**Full Changelog**: https://github.com/remyeltorro/celldetective/compare/v1.1.1...v1.2.0

- Neighborhood:
The status of the reference/neighbor population can now be a "group" as well as "status" and "class"
Fix bugs with reading and merging previous pickle files
- Measurement annotator widget:
Show mask contours of the cells in the measurement annotator widget
Tracked cells remain selected as the frame is changed
A new export button allows the export of all the cell measurements from the current field of view (and frame). The user can thus produce a dataset of annotated cells and their respective measurements, that could be used to train a classifier model, later on
- A new plugin to revisit an exported annotation for segmentation, by selecting the annotated image. Napari will open with the associated mask that can be edited and reexported
- Segmentation annotation
The experiment metadata is now stored with the config file associated with the annotation
- Preprocessing:
Auto-detection of stack length to be able to pass stacks of different lengths across the positions
Bug fixes when the stack is not available
- Segmentation:
The segmentation function in segment.py now gives the same results as the segment_cells.py script
- Classifier widget:
Drop NaN values in any of the features called in the classifying condition to avoid them evaluating to False
- Table UI
A new scatterplot option, better separation of -x, -y and -hue
Subtable generated with a query now has the properties of a table of untracked data to allow plots

**Full Changelog**: https://github.com/remyeltorro/celldetective/compare/v1.1.1.post3...v1.1.1.post4

* Countplot available in plotting functions associated to the table widget
* New function to merge several binary classes (0 or 1) into a categorical class (0,1,2,3...)
* Write the image names from the train and validation set in the input_config.json file of a newly trained segmentation model
* Fix bug of "scale_model" missing when training a StarDist model
* Fix bug in the exclusion of NaN when optimizing the values of the background image to correct for the background using the model-free approach

See the **full Changelog**: https://github.com/remyeltorro/celldetective/compare/v1.1.1.post1...v1.1.1.post3

Minor increment for stability fixes and small features.

- Documentation: preprocessing functions
- Table widget: new option to plot measurement distributions directly with tracked data, without requiring to collapse the time axis
- Segmentation:
2D interpolation of NaNs in images before segmentation
Fix rescaling to reach a cell size of 30 px not applied when predicting with a Cellpose model
- Filtering:
Interpolate image before filtering to avoid the propagation of NaN values
- Threshold configuration wizard:
A new option to bypass marker detection and watershed, and instead label all non-touching objects post threshold
- Classifier widget:
fix a bug on the R2 computation
increased number of iterations for the sigmoid fit

**Full Changelog**: https://github.com/remyeltorro/celldetective/compare/v1.1.1...v1.1.1.post1