- Created
cell2cell.analysis.tensor_pipelines
with functions to easily run Tensor-cell2cell.
- Deleted
cell2cell.external.tensorly_nn_cp
since it is not used anymore. Tensorly is directly used instead. - Renamed
cell2cell.analysis.pipelines
tocell2cell.analysis.cell2cell_pipelines
- Added
elbow_metric
,elbow_metric_mean
andelbow_metric_raw
attributes tocell2cell.tensor.tensor.BaseTensor
for storing the curve generated from the elbow analysis. - Removed parameter
loc_zeros
fromcell2cell.tensor.tensor.PreBuiltTensor
- Converted factors to numpy in
cell2cell.tensor.factorization._multiple_runs_elbow_analysis()
whenmetric='similarity'
to avoid errors when using GPU. - Fixed error obtained with functions
sparsity_fraction()
andmissing_fraction()
in acell2cell.tensor.tensor.BaseTensor
when tensorly backend is different to numpy and the device is nto a CPU. This error was fixed with making loc_nans and loc_zeros attributes ofcell2cell.tensor.tensor.InteractionTensor
andcell2cell.tensor.tensor.PreBuiltTensor
to be now a tensorly.tensor object.