def load_metadata(filename):
"""Load cluster metadata from a CSV file.
Return (field_name, dictionary).
"""
return _read_tsv_simple(filename)
def _load(path):
path = str(path)
if path.endswith('.npy'):
return np.load(path)
elif path.endswith(('.csv', '.tsv')):
return _read_tsv_simple(path)[1] # the function returns a tuple (field, data)
elif path.endswith('.bin'):
# TODO: configurable dtype
return np.fromfile(path, np.int16)
def write_cluster_data(self):
"""We load all cluster metadata from TSV files, renumber the clusters,
merge the dictionaries, and save in a new merged TSV file. """
cluster_data = [
'cluster_Amplitude.tsv', 'cluster_ContamPct.tsv',
'cluster_KSLabel.tsv'
]
for fn in cluster_data:
metadata = {}
for subdir, offset in zip(self.subdirs, self.cluster_offsets):
try:
field_name, metadata_loc = _read_tsv_simple(subdir / fn)
except ValueError:
# Skipping non-existing file.
continue
for k, v in metadata_loc.items():
metadata[k + offset] = v
if metadata:
_write_tsv_simple(self.out_dir / fn, field_name, metadata)