mirror of
https://github.com/exaloop/codon.git
synced 2025-06-03 15:03:52 +08:00
19 lines
436 B
Plaintext
19 lines
436 B
Plaintext
# SeqMap
|
|
# Seq workshop -- Section 2
|
|
# Reads and constructs a hash table index from an input
|
|
# FASTA file.
|
|
# Usage: seqc run section2.seq <FASTA path>
|
|
from sys import argv
|
|
from bio import *
|
|
import pickle
|
|
import gzip
|
|
|
|
index = {}
|
|
|
|
for record in FASTA(argv[1]):
|
|
for pos,kmer in record.seq.kmers_with_pos(k=32, step=1):
|
|
index[min(kmer, ~kmer)] = pos
|
|
|
|
with gzip.open(argv[1] + '.index', 'wb') as jar:
|
|
pickle.dump(index, jar)
|