match/mismatch accuracy on simulations

[emdann]

Added notebook with:

• functions to split dataset in reference and query simulating mismatch (deletions/insertions) in the query, of different lengths and starts
• Code to compute "match prediction accuracy" (i.e. in which fraction of bins does the alignment correctly identify a match or a mismatch?)
• Test on pancreas dataset using 5 genes picked from the CellRank paper.

[emdann]

Insertion/deletion accuracy results are saved in:

/nfs/team205/ed6/bin/Genes2Genes/notebooks/data/match_accuracy_pancreas.insertion.csv
/nfs/team205/ed6/bin/Genes2Genes/notebooks/data/match_accuracy_pancreas.deletion.csv