Sankey diagram of id mapping, data retention, and embeddings

[mezarque]

Would be really nice to have a visual summary of various steps in the pipeline using Sankey diagrams, e.g. via plotly. Some diagram ideas:

• Flowchart of common names between different files associated with a given dataset (e.g. protein ids, gene names, uniprot ids, etc.) to see how much information is lost during file download and initial read mapping per species
• Sankey diagram of mapping from different species to orthogroups (flow volume proportional to number of orthogroups). Show e.g. orthogroups with 3 species, each combination of 2 species, and unique to 1 species.

[mezarque]

I've started building one of these functions - visualizing gene vs. orthogroup Leiden clusters. Need to convert to a function though; currently implemented in 3_ prefix notebooks just as repeated code.