def test_open_file_xz(tmpdir):
test_string = b'test\n'
file_name = tmpdir.join('test-open.xz').strpath
handle = open_file(file_name, mode='w')
handle.write(test_string)
handle.close()
assert open_file(file_name, mode='r').read() == test_string
def test_open_file_text(tmpdir):
test_string = 'test\n'
file_name = tmpdir.join('test-open').strpath
handle = open_file(file_name, mode='wb')
handle.write(test_string.encode('ascii'))
handle.close()
assert open_file(file_name,
mode='rb').read().decode('ascii') == test_string
def test_write_fasta_sequence1(nucseq, tmpdir):
seq_id, seq = next(fasta.load_fasta(nucseq))
file_name = (tmpdir / 'test.fa').strpath
file_handle = open_file(file_name, 'w')
fasta.write_fasta_sequence(file_handle, seq_id, seq)
file_handle.close()
seq_idw, seqw = next(fasta.load_fasta(file_name))
assert (seq_id, seq) == (seq_idw, seqw)
def test_write_fasta_sequence2(nucseq, tmpdir):
file_name = (tmpdir / 'test.fa').strpath
file_handle = open_file(file_name, 'w')
for seq_id, seq in fasta.load_fasta(nucseq):
fasta.write_fasta_sequence(file_handle, seq_id, seq)
file_handle.close()
count1 = sum(1 for x in fasta.load_fasta(nucseq))
count2 = sum(1 for x in fasta.load_fasta(file_name))
assert count1 == count2
def read_samtools_depth(file_handle, num_seqs=10000, seq_ids=None):
"""
.. versionchanged:: 0.4.2
the function returns **lists** instead of numpy arrays for speed (at
least in my tests it seems ~4x increase)
.. versionchanged:: 0.4.0
now returns 3 array, instead of 2. Also added *seq_ids* to skip lines
.. versionchanged:: 0.3.4
*num_seqs* can be None to avoid a log message
.. versionadded:: 0.3.0
Reads a samtools *depth* file, returning a generator that yields the
array of each base coverage on a per-sequence base.
.. note::
There's no need anymore to use `samtools depth -aa`, because the
function returns the position array and this can be used to create a
Pandas SparseArray which can be reindexed to include missing positions
(with values of 0)
**Valid for version < 0.4.0**:
The information on position is not used, to use numpy and save memory.
samtools *depth* should be called with the `-aa` option::
`samtools depth -aa bamfile`
This options will output both base position with 0 coverage and
sequneces with no aligned reads
Arguments:
file_handle (file): file handle of the coverage file
num_seqs (int or None): number of sequence that fires a log message. If
None, no message is triggered
seq_ids (dict, set): a hashed container like a dictionary or set with
the sequences to return
Yields:
tuple: the first element is the sequence identifier, the second one
is the list with the positions, the third element is the list with the
coverages
"""
curr_key = ''
curr_pos = []
curr_cov = []
file_handle = open_file(file_handle, 'rb')
LOG.info(
'Reading coverage from file (%s)',
getattr(file_handle, 'name', repr(file_handle))
)
line_no = 0
for line in file_handle:
line = line.decode('ascii')
# From Python3 the default is Universal newlines, and it's not expected
# to have more than '\n' at the end of the line - increases speed
# slightly
name, pos, cov = line[:-1].split('\t')
if (seq_ids is not None) and (name not in seq_ids):
continue
# only converts if sequence is to be used
pos = int(pos)
cov = int(cov)
if curr_key == name:
curr_pos.append(pos)
curr_cov.append(cov)
else:
if curr_key == '':
curr_cov.append(cov)
curr_pos.append(pos)
curr_key = name
else:
line_no += 1
if (num_seqs is not None) and (line_no % num_seqs == 0):
LOG.info('Read %d sequence coverage', line_no)
yield curr_key, curr_pos, curr_cov
curr_key = name
curr_cov = [cov]
curr_cov = [pos]
else:
yield curr_key, curr_pos, curr_cov
LOG.info('Read a total of %d sequence coverage', line_no + 1)