Besca

Major changes includes:
+ dependencies to rpy2 and R libraries
+ Rwrappers functions :
+ dsb_normalize
+ deviance for HVG selection
+ maxLikGlobalDimEst for data dimensionnality
+ valOutlier (Estimates and returns the thresholds to use for gene/cell filtering)
+ scTransform

+ auto-annot comparison reports is now generic and include a riverplot
+ R called into the notebooks + installation procedures

Bug / warning fixed:

+ if bbknn batch call contains only one category, waring is set up 112
+ exporting of a unique labelling is possible 105


Additional improvements:

+ default datasets is now Kotliarov for the different notebooks.
+ Small function to convert undirected signatures dictionary to signed one to use combined_signature_score as a routine;
+ Updating URLs for Granja datasets (new Zenodo upload).
+ some code refactoring in particular in export functions.

New implementations and bug fixes including:

New features:
* number of hvgs now take into account the batches
* geometric mean vs arithmetic mean
* riverplot display to compare 2 categories


Dependencies:
* Scanpy 1.6


Signatures and Nomenclatures
* update of the neuron nomenclature
* platelet signature added

Fixes:
* bug fix in the pseudo bulk function
* export fix for the FAIR export. cp10K exported to help upload in mongodb database (no dependency of the reformat C program)

Standard workflow:
* Standard workflow renamed (from 2.0 to 2)
* Standard workflow is automatically exported with the analysis name as an html file.

+ Batch correction default is BBKNN
+ Cite-Seq in standard workflow

+ auto_annot merged in
+ Bescape interface available

+ Sig-Annot with parameters output
+ Improved Sig Annot for epithelial like cells
+ DBLabel annotation for interoperability of the annotations
+ Making GEMS connection possible
+ Sig-Annot graph visualization

+ Versionneer
+ Link to all datasets in Zenodo and check function to output required URL