def test_precompute_gencode(self):
k = 1
file_name = get_unzipped_file_name(skr_config.GENCODE_HUMAN)
with open('../cache/' + file_name, mode='r') as infasta:
(mean, std, unnormalized_frequency,
names) = compute_normalization_and_frequency(infasta, k, False)
print('For k=', k, 'length is', len(mean))
assert len(mean) == 4**k
def build_cache_files():
if not os.path.exists(CACHE_DIR):
os.mkdir(CACHE_DIR)
fasta_sets = get_precomputed_fasta_sets()
for fasta_set in fasta_sets:
if VERBOSE:
print('Getting ', fasta_set)
fasta_file = getGenCode.get_unzipped_file_name(fasta_set)
dir_name = pathlib.PurePath(fasta_file).stem
path_to_dir = os.path.join(CACHE_DIR, dir_name)
if not os.path.exists(path_to_dir):
os.mkdir(path_to_dir)
names_written = False
tsave = 0
if VERBOSE:
print(dir_name + ' computing normalization took\t', end='')
for kmer_length in range(1, skr_config.MAX_KMER_LENGTH_PRECOMPUTE + 1):
fasta_path = os.path.join(CACHE_DIR, fasta_file)
with open(fasta_path, mode='r') as infasta:
t1 = time.perf_counter()
(mean, std, unnormalized_frequency,
names) = compute_normalization_and_frequency(
infasta, kmer_length, return_normalized=False)
t2 = time.perf_counter()
if VERBOSE:
print('k=' + str(kmer_length) + ',%.3fs;\t' % (t2 - t1),
end='')
t1 = time.perf_counter()
np.save(
get_file_path_for(fasta_file, kmer_length,
CACHE_FILE_TYPES.get('mean')), mean)
np.save(
get_file_path_for(fasta_file, kmer_length,
CACHE_FILE_TYPES.get('std')), std)
np.save(
get_file_path_for(
fasta_file, kmer_length,
CACHE_FILE_TYPES.get('unnormalized_frequency')),
unnormalized_frequency)
if not names_written:
with open(
get_file_path_for(fasta_file, kmer_length,
CACHE_FILE_TYPES.get('names')),
'wb') as names_file:
pickle.dump(names, names_file)
names_written = True
t2 = time.perf_counter()
tsave += (t2 - t1)
if VERBOSE:
print('\nAggregate save time for ' + dir_name +
' was %.3fs' % tsave)
def get_precomputed_frequency_path(comparison_set, kmer_length):
if comparison_set is None or len(comparison_set) <= 0:
return None
fasta_sets = get_precomputed_fasta_sets()
for fasta_set in fasta_sets:
if comparison_set == fasta_set.server_name:
fasta_file = getGenCode.get_unzipped_file_name(fasta_set)
unnormalized_frequency_path = get_file_path_for(fasta_file, kmer_length, CACHE_FILE_TYPES.get('unnormalized_frequency'))
names_path = get_file_path_for(fasta_file, kmer_length, CACHE_FILE_TYPES.get('names'))
if os.path.exists(unnormalized_frequency_path) and os.path.exists(names_path):
return unnormalized_frequency_path, names_path
return None
def get_precomputed_normalization_path(parameters):
normal_set = parameters['normal_set']
if normal_set is None or len(normal_set) <= 0:
return None
fasta_sets = get_precomputed_fasta_sets()
for fasta_set in fasta_sets:
if normal_set == fasta_set.server_name:
fasta_file = getGenCode.get_unzipped_file_name(fasta_set)
mean_path = get_file_path_for(fasta_file, parameters['kmer_length'], CACHE_FILE_TYPES.get('mean'))
std_path = get_file_path_for(fasta_file, parameters['kmer_length'], CACHE_FILE_TYPES.get('std'))
if os.path.exists(mean_path) and os.path.exists(std_path):
return (mean_path, std_path)
else:
raise SeekrServerError('Fasta file <' + fasta_file + '> not found for kmer_length=' + str(parameters['kmer_length']) )
return None
return None