def test_060_construct_new_file_checks(self):
tmp_file = os.path.join(tempfile.gettempdir(), str(uuid4()))
with self.assertRaises(IOError):
fh = Fast5.New(tmp_file, 'r')
fh = Fast5.New(tmp_file, 'a', channel_id=self.tmp_channel_id)
fh = Fast5.New(tmp_file, 'a', tracking_id=self.tmp_tracking_id)
# This should be fine
with Fast5.New(tmp_file,
'a',
channel_id=self.tmp_channel_id,
tracking_id=self.tmp_tracking_id) as h:
h.set_read(self.tmp_events_float, self.tmp_read_id)
def test_067_write_raw_data(self):
tmp_file = os.path.join(tempfile.gettempdir(), str(uuid4()))
with Fast5.New(tmp_file,
'a',
channel_id=self.tmp_channel_id,
tracking_id=self.tmp_tracking_id) as h:
h.set_raw(self.tmp_raw, meta=self.tmp_read_id, read_number=1)
with self.assertRaises(TypeError):
with Fast5.New(tmp_file,
'a',
channel_id=self.tmp_channel_id,
tracking_id=self.tmp_tracking_id) as h:
h.set_raw(self.tmp_raw.astype(float),
meta=self.tmp_read_id,
read_number=1)
def test_061_write_read_float_data(self):
tmp_file = os.path.join(tempfile.gettempdir(), str(uuid4()))
with Fast5.New(tmp_file,
'a',
channel_id=self.tmp_channel_id,
tracking_id=self.tmp_tracking_id) as h:
h.set_read(self.tmp_events_float, self.tmp_read_id)
# Metadata duration and start_time should be integers, not floats
with Fast5(tmp_file, 'r') as h:
for key in ['duration', 'start_time']:
self.assertIsInstance(h.attributes[key], int)
with Fast5(tmp_file) as h:
events = h.get_read()
self.assertEqual(
events['start'].dtype.descr[0][1], '<f8',
'Writing float data did not give float data on read.')
actual = events['start'][0]
expected = self.tmp_events_float['start'][0]
self.assertEqual(
actual, expected,
'Write float, data on read not scaled correctly, got {} not {}'
.format(actual, expected))
os.unlink(tmp_file)
def setUpClass(self):
"""Create a read fast5 from scratch with previously simulated mapping and basecall 1D data"""
print('* Fast5 Basecaller and Mapper')
self.seq = 'CATTACGCATTTACCGAAACCTGGGCAAA'
self.qstring = '!'*len(self.seq)
self.model_file = 'example_template.model'
self.events_file = 'example_template.events'
self.model_file = 'example_template.model'
self.bc_scale_file = 'example_template.bc_scale'
self.bc_path_file = 'example_template.bc_path'
self.map_scale_file = 'example_template.map_scale'
self.map_path_file = 'example_template.map_path'
self.map_post_file = 'example_template.map_post'
self.ref_name = 'test_seq'
# Open new file
header = ['channel_number', 'offset', 'range', 'digitisation', 'sampling_rate']
channel_id = {x:0 for x in header}
tracking_id = tracking_id = {
'exp_start_time': '1970-01-00T00:00:00Z',
'run_id': 'a'*32,
'flow_cell_id': 'FAH00000',
}
fakefile = tempfile.NamedTemporaryFile()
self.fh = Fast5.New(fakefile.name, channel_id=channel_id, tracking_id=tracking_id, read='a')
# load data to set within fast5 file
self.model = np.genfromtxt(self.get_file_path(self.model_file), dtype=None, delimiter='\t', names=True)
self.model['kmer'] = self.model['kmer'].astype(str)
self.events = np.genfromtxt(self.get_file_path(self.events_file), dtype=None, delimiter='\t', names=True)
# use namedtuple to imitate a Scale object
Scale = namedtuple('Scale', ['shift', 'scale', 'drift', 'var', 'scale_sd', 'var_sd'])
bc_scale = Scale(*np.genfromtxt(self.get_file_path(self.bc_scale_file), dtype=None, delimiter='\t'))
bc_path = np.genfromtxt(self.get_file_path(self.bc_path_file), dtype=np.int32, delimiter='\t')
self.fh.set_basecall_data(self.events, bc_scale, bc_path, self.model, self.seq)
map_scale = Scale(*np.genfromtxt(self.get_file_path(self.map_scale_file), dtype=None, delimiter='\t'))
map_path = np.genfromtxt(self.get_file_path(self.map_path_file), dtype=np.int32, delimiter='\t')
map_post = np.genfromtxt(self.get_file_path(self.map_post_file), delimiter='\t')
n_states = len(self.seq) - len(self.model['kmer'][0]) + 1
self.fh.set_mapping_data(self.events, map_scale, map_path, self.model, self.seq, self.ref_name)
self.fh.set_mapping_data(self.events, map_scale, map_path, self.model, self.seq, self.ref_name, post=map_post)
def test_065_write_int_read_float_data(self):
tmp_file = os.path.join(tempfile.gettempdir(), str(uuid4()))
with Fast5.New(tmp_file, 'a', channel_id = self.tmp_channel_id) as h:
h.set_read(self.tmp_events_int, self.tmp_read_id)
with Fast5(tmp_file) as h:
events = h.get_read()
self.assertEqual(events['start'].dtype.descr[0][1], '<f8',
'Writing uint data did not give float data on read.'
)
actual = events['start'][0]
expected = self.tmp_events_float['start'][0]
self.assertEqual(actual, expected,
'Write unit, data on read not scaled correctly, got {} not {}'.format(
actual, expected
)
)
os.unlink(tmp_file)
def create_fast5(raw_data, fast5_filename):
raw_data = np.array(raw_data)
# create fast5 (from https://nanoporetech.github.io/fast5_research/examples.html)
# example of how to digitize data
start, stop = int(min(raw_data - 1)), int(max(raw_data + 1))
rng = stop - start
digitisation = 8192.0
bins = np.arange(start, stop, rng / digitisation)
# np.int16 is required, the library will refuse to write anything other
raw_data_binned = np.digitize(raw_data, bins).astype(np.int16)
# The following are required meta data
channel_id = {
'digitisation': digitisation,
'offset': 0,
'range': rng,
'sampling_rate': 4000,
'channel_number': 1,
}
read_id = {
'start_time': 0,
'duration': len(raw_data),
'read_number': 1,
'start_mux': 1,
'read_id': str(uuid4()),
'scaling_used': 1,
'median_before': 0,
}
tracking_id = {
'exp_start_time': '1970-01-01T00:00:00Z',
'run_id': str(uuid4()).replace('-', ''),
'flow_cell_id': 'FAH00000',
}
context_tags = {}
with Fast5.New(fast5_filename,
'w',
tracking_id=tracking_id,
context_tags=context_tags,
channel_id=channel_id) as h:
h.set_raw(raw_data_binned, meta=read_id, read_number=1)
def digitize_write(raw_data, read_id, params):
digitisation = 8192.0
start, stop = int(min(raw_data - 1)), int(max(raw_data + 1))
rng = stop - start
bins = np.arange(start, stop, rng / digitisation)
# np.int16 is required, the library will refuse to write anything other
#raw_data = np.digitize(raw_data, bins).astype(np.int16)
raw_data = np.round(raw_data)
raw_data = raw_data.astype(np.int16)
filename = params.fast5_path + read_id + '.fast5'
# The following are required meta data
channel_id = {
'digitisation': digitisation,
'offset': 0,
'range': rng,
'sampling_rate': 4000,
'channel_number': 1,
}
read_id = {
'start_time': 0,
'duration': len(raw_data),
'read_number': 1,
'start_mux': 1,
'read_id': str(read_id),
'scaling_used': 1,
'median_before': 0,
}
tracking_id = {
'exp_start_time': '1970-01-01T00:00:00Z',
'run_id': str(uuid4()).replace('-', ''),
'flow_cell_id': 'FAH00000',
}
context_tags = {}
with Fast5.New(filename,
'w',
tracking_id=tracking_id,
context_tags=context_tags,
channel_id=channel_id) as h:
h.set_raw(raw_data, meta=read_id, read_number=1)