Closed davmlaw closed 2 years ago
Implemented using a defaultdict of key = contig and value being an IntervalTree
Storing everything per exon made the interval tree very big and took a long time to start up.
Instead to keep the RAM and startup time lowis (170mb) I store the transcript ID across tx start/end then at lookup time loop through the exons doing a range check
Requested as a feature
This is part of the biocommons API which is not implemented yet.
The smart thing to do in local memory would be an interval tree you build lazily from JSON I the API is called.
Not sure about rest service maybe redis has something