import sys
sys.path.append('/nfs1/Koslicki_Lab/koslickd/Repositories/MinHashMetagenomics/src/')
import os, timeit, h5py
import MinHash as MH
import numpy as np
fid = open('/nfs1/Koslicki_Lab/koslickd/MinHash/Data/FileNames.txt', 'r')
file_names = fid.readlines()
fid.close()
file_names = [name.strip() for name in file_names]
###############################
# Compute the hashes for all the training genomes
n = 500
CEs = MH.compute_multiple(n=n, max_prime=9999999999971., ksize=31, input_files_list=file_names, save_kmers='y', num_threads=48)
# Export
MH.export_multiple_hdf5(CEs, '/nfs1/Koslicki_Lab/koslickd/MinHash/Out/N'+str(n)+'k31/')
# Save the hashes in the training genomes
hash_list = set()
for CE in CEs:
hash_list.update(CE._mins)
fid = h5py.File('/nfs1/Koslicki_Lab/koslickd/MinHash/Out/N'+str(n)+'k31_mins.h5', 'w')
fid.create_dataset("hash_list", data=list(hash_list))
fid.close()
# If I need to read it back in
#fid = h5py.File('/nfs1/Koslicki_Lab/koslickd/MinHash/Out/N500k31_mins.h5','r')
#hash_list = set(fid["hash_list"][:])
n = 5000
CEs = MH.compute_multiple(n=n, max_prime=9999999999971., ksize=31, input_files_list=file_names, save_kmers='y', num_threads=48)
'/nfs1/Koslicki_Lab/koslickd/Repositories/MinHashMetagenomics/src/')
import os, timeit, h5py
import MinHash as MH
import numpy as np
fid = open('/nfs1/Koslicki_Lab/koslickd/MinHash/Data/FileNames.txt', 'r')
file_names = fid.readlines()
fid.close()
file_names = [name.strip() for name in file_names]
###############################
# Compute the hashes for all the training genomes
n = 500
CEs = MH.compute_multiple(n=n,
max_prime=9999999999971.,
ksize=31,
input_files_list=file_names,
save_kmers='y',
num_threads=48)
# Export
MH.export_multiple_hdf5(
CEs, '/nfs1/Koslicki_Lab/koslickd/MinHash/Out/N' + str(n) + 'k31/')
# Save the hashes in the training genomes
hash_list = set()
for CE in CEs:
hash_list.update(CE._mins)
fid = h5py.File(
'/nfs1/Koslicki_Lab/koslickd/MinHash/Out/N' + str(n) + 'k31_mins.h5', 'w')
fid.create_dataset("hash_list", data=list(hash_list))
fid.close()
# If I need to read it back in
CS.jaccard(CS2)
CS2.jaccard(CS)
i = 0
for record in screed.open(file_names[0]):
for kmer in MH.kmers(record.sequence,10):
if kmer == 'TGGAATTCCA':
i += 1
print(i)
CE = MH.CountEstimator(n=500, ksize=20, input_file_name=file_names[0], save_kmers='y')
#Compute a bunch of them and save to a single, big HDF5 file
CEs = MH.compute_multiple(n=500,ksize=11,input_files_list=file_names,save_kmers='y')
MH.export_multiple_to_single_hdf5(CEs, 'test_big.h5')
#Load them back in
CEs = MH.import_multiple_from_single_hdf5('test_big.h5')
#load just a few in
CEs = MH.import_multiple_from_single_hdf5('test_big.h5', file_names[0:2])
# Let's look at forming the Y vector
CEs = MH.import_multiple_hdf5(out_file_names)
MCE = MH.import_single_hdf5('/nfs1/Koslicki_Lab/koslickd/MinHash/Out/SRR172902.fastq.CE_N500_k31_inComparison.h5')
Y = np.zeros(len(CEs))
for i in range(len(CEs)):
Y[i] = CEs[i].jaccard_count(MCE)[1]
