This PR introduces a number of changes, but the main one is the introduction of the PlotHelper class, which includes wrappers for several plotly graph functions.
New
PlotHelper class has been added. PlotHelper contains several different plotting functions and is designed to be able to easily use data indexed from ExperimentArrays. All functions are plotly based and return go.Figure objects, so customization after the plot is made is easy.
Evaluator has a basic function to overlay tracks over the input images.
Improved
Plots can handle transparency values in hexadecimal colors.
save_kept_cells is much faster now. Used a similar method that has provided speed ups before of using list comprehensions on a flattened array.
Update weights to use tensorflow format by default. This allows silencing of many unnecessary warnings.
Fixed
Percentile based filters can now handle nan in the input data. Additionally, bounds are set at 0 and 100, as opposed to 5 and 95.
requirements.txt now points to the correct umap package.
Fixed a bug in FluorUNetModel where images of certain dimensions would not be cropped correctly after padding.
This PR introduces a number of changes, but the main one is the introduction of the
PlotHelperclass, which includes wrappers for severalplotlygraph functions.New
PlotHelperclass has been added.PlotHelpercontains several different plotting functions and is designed to be able to easily use data indexed fromExperimentArrays. All functions areplotlybased and returngo.Figureobjects, so customization after the plot is made is easy.Evaluatorhas a basic function to overlay tracks over the input images.Improved
save_kept_cellsis much faster now. Used a similar method that has provided speed ups before of using list comprehensions on a flattened array.Fixed
nanin the input data. Additionally, bounds are set at 0 and 100, as opposed to 5 and 95.requirements.txtnow points to the correctumappackage.FluorUNetModelwhere images of certain dimensions would not be cropped correctly after padding.