Add process_seqstr in orca_predict.py

[gr-grey]

prcoess_seqstr(seqstr_input, file=None, mpos=None, custom_models=None, use_cuda=True)

• The function extracts the sequence string (seqstrout) from input, such as '[hg38]chr9:94904000-126904000 +'
• Sequence length needs to be at least 32Mb, if longer than 32Mb, only the middle 32Mb will be used.
• Zoom in middle (16Mb) if mpos is not specified
• Use selene_sdk.sequences Genome.sequence_to_encoding(sequence_str) to get encoding of sequence
• Use genomepredict to predict multi-scale interaction
• Use genomeplot to plot the results

Need Seqstr package to use the function.

Try import seqstr in the beginning, if package not installed, calling the function will invoke ImportError.

Tested 3 seqstrs with 1, 2, and 3 segments, all generated pdf files

teststr = '[hg38]chr9:94904000-126904000 +'
testout = process_region_seqstr(teststr, file="seqstr_test1")
teststr = '[hg38]chr9:94904000-100904000 +; chr7:5480600-32480600 -'
testout = process_region_seqstr(teststr, file="seqstr_test2")
teststr = '[hg38]chr9:94904000-100904000 +; chr7:5480600-13480600 -; chr1:10904000-75904000 +'
testout = process_region_seqstr(teststr, file="seqstr_test3")