agartland/pwseqdist

A small package that efficiently computes distances between genetic sequences. Can accommodate similarity matrices, sequences of different lengths and custom metrics.

pwseqdist's Issues

Is this supposed to be ``useqs = list(set(seqs))``

pwseqdist/pwseqdist/pairwise.py

Line 66 in 2219721

useqs = list(seqs)

Is this supposed to be useqs = list(set(seqs))?

fails on rhinos though because cpu_count is 56 and 6 // 56 is 0

Small example fails on large cluster:
fails on rhinos though because cpu_count is 56 and 6 // 56 is 0

If User submits an unrecognized character Numba does not allow seq2mat exception handling!

If the user does not prefilter CDR3:

pw.apply_pairwise_sq(seqs = ['CAS_GAF', 'CASRGAF'],
					 metric = pw.metrics.nb_editdistance, 
					 ncpus    = 1, 
					 use_numba= True)

~/TCRDIST/pwseqdist/pwseqdist/matrices.py in seqs2mat(seqs, alphabet, max_len)
    179             try:
    180                 mat[si, aai] = alphabet.index(s[aai])
--> 181             except ValueError('Unknown symbols given value for last column/row of matrix'):
    182                 """Unknown symbols given value for last column/row of matrix"""
    183                 mat[si, aai] = len(alphabet)

TypeError: catching classes that do not inherit from BaseException is not allowed

compared with standard behavior for recognizing unknown characters by non-Numba metric:

In [44]:
    ...: import Levenshtein
    ...: pw.apply_pairwise_sq(seqs = ['CAS_GAF', 'CASRGAF'],
    ...: ^I^I^I^I^I metric = Levenshtein.distance,
    ...: ^I^I^I^I^I ncpus    = 1,
    ...: ^I^I^I^I^I use_numba= False)
Out[44]: array([1])

In [45]: import Levenshtein
    ...: pw.apply_pairwise_sq(seqs = ['CAS8GAF', 'CASRGAF'],
    ...: ^I^I^I^I^I metric = Levenshtein.distance,
    ...: ^I^I^I^I^I ncpus    = 1,
    ...: ^I^I^I^I^I use_numba= False)
Out[45]: array([1])

tcr_dict_distance_matrix

Hello,

Thanks for the awesome package! However I do have a question about tcr_dict_distance_matrix on line 59 in matrices.py. How did you come up with the numbers for each amino acid pair? I tried to convert blosum62 matrix to distance matrix using kernel2dist function, but the result was not even close to those numbers in tcr_dict_distance_matrix.
Thanks so much in advance!

Replace multiprocessing with pathos

Currently we use python's built-in multiprocessing to spread computations across multiple CPUs. This requires that the metric function be globally importable, which is a limitation of pickle, which is used by multiprocessing. The package pathos offers a essentially drop-in replacement for multiprocessing which uses dill to serialize the objects that are passed to worker processes, which would allow for locally defined and more complex functions to be used as metrics. Currently this is not a major limitation, but if this issue crops up again in the future I think this would be an easy update to make.

Links with info:
https://github.com/uqfoundation/pathos
https://medium.com/@emlynoregan/serialising-all-the-functions-in-python-cd880a63b591