1. Add batch effect correct method by set integration_type="BatchCorrect"; 2. Add more distances (e.g. cosine, cityblock, see sklearn.metrics.pairwise) to formulate distance matrices. 3. Fix some bugs;
0.3.0
[unioncom-0.3.0.tar.gz](https://github.com/caokai1073/UnionCom/files/5769046/unioncom-0.3.0.tar.gz) + Add more comments to make the software easier to understand + Fix some bugs
0.2.2
+ Fix some bugs; + Change function "PCA_visualize" to "Visualize", and provide PCA, TSNE and UMAP for visulalization; + Add a function to find maximum connected subgraph;
0.2.1
+ Split function train into functions Match and Project. + Use Kuhn-Munkres to find optimal pairs between datasets instead of parbabilistic matrix matching. + Add a new parameter "project" to provide options for barycentric projection. + Separate "test_label_transfer_accuracy" function from "fit_transform" function + fix some bugs