def write_read(self, read):
"""Write a read.
:param read: either a `Read` object or an hdf group handle from a
source multi-read file.
"""
if self.closed:
raise RuntimeError('Cannot write after closed.')
if self.current_reads == 0:
# start a new file
self.close()
filename = '{}mreads_file{}.fast5'.format(self.prefix,
self.file_counter)
filename = os.path.join(self.out_path, filename)
self.current_file = h5py.File(filename, 'w')
self.current_file.attrs[_sanitize_data_for_writing(
'file_version')] = _sanitize_data_for_writing("2.0")
self.file_counter += 1
# write data
if isinstance(read, Read):
self._write_read(read)
elif isinstance(read, h5py.Group):
self._copy_read_group(read)
else:
raise TypeError("Cannot write type {} to output file.")
self.current_reads += 1
# update
if self.current_reads == self.reads_per_file:
self.current_reads = 0
def New(cls,
fname,
read='a',
tracking_id={},
context_tags={},
channel_id={}):
"""Construct a fresh bulk file, with meta data written to
standard locations. There is currently no checking this meta data.
TODO: Add meta data checking.
"""
# Start a new file, populate it with meta
with h5py.File(fname, 'w') as h:
h.attrs[_sanitize_data_for_writing(
'file_version')] = _sanitize_data_for_writing(1.0)
for data, location in zip(
[tracking_id, context_tags],
[cls.__tracking_path__, cls.__context_path__]):
# see cjw's comment in fast5.py:
# 'no idea why these must be str, just following ossetra'
cls.__add_attrs(h, data, location, convert=str)
# return instance from new file
return cls(fname, read)
def __add_attrs(self, data, location, convert=None):
"""Implementation of _add_attrs as staticmethod. This allows
functionality to be used in .New() constructor but is otherwise nasty!
"""
if location not in self:
self.create_group(location)
attrs = self[location].attrs
for k, v in data.items():
if convert is not None:
attrs[_sanitize_data_for_writing(k)] = _sanitize_data_for_writing(convert(v))
else:
attrs[_sanitize_data_for_writing(k)] = _sanitize_data_for_writing(v)
def _add_numpy_table(self, data, location):
data = _sanitize_data_for_writing(data)
self.create_dataset(location, data=data, compression=True)
