Migrate from sklearn.manifold TSNE to openTSNE for visualizing node embeddings

[JohT]

⚙️ Optimization

• Migrate from sklearn to openTSNE. This has several advantages:
• Speed improvements
• The perplexity parameter doesn't need to be limited to the number of rows in the data for small graphs with less than 30 nodes.
• It doesn't lead to a Kernel crash in rare cases with specific data constellations. In that case it produces NaN (not a number) embedding values, which also might need to be investigated.
• Ignore local conda environment data