def read_metrics_csv(self, cndata):
"""
read the input file
"""
samples = cndata.index
data = {}
numread_data = {}
reads_per_bin_data = {}
sepdata = defaultdict(list)
colordata = {}
header, dtypes, columns = csvutils.get_metadata(self.metrics)
idxs = self.build_label_indices(columns)
color_col = self.color_by_col
sep_col = self.plot_by_col
with helpers.getFileHandle(self.metrics) as freader:
if header:
assert freader.readline().strip().split(',') == columns
for line in freader:
line = line.strip().split(self.sep)
sample_id = line[idxs['cell_id']]
# skip samples that are just na or inf
if sample_id not in samples:
continue
val = line[idxs["mad_neutral_state"]]
val = float('nan') if val == "NA" else float(val)
ec = 'all' if sep_col == 'all' else line[idxs[sep_col]]
cc = line[idxs[color_col]]
numreads = float(line[idxs['total_mapped_reads_hmmcopy']])
reads_per_bin = line[idxs['median_hmmcopy_reads_per_bin']]
reads_per_bin = 0 if reads_per_bin == "NA" else float(
reads_per_bin)
if self.cellcalls and cc not in self.cellcalls:
continue
numread_data[sample_id] = numreads
data[sample_id] = val
reads_per_bin_data[sample_id] = reads_per_bin
colordata[sample_id] = cc
sepdata[ec].append(sample_id)
return data, sepdata, colordata, numread_data, reads_per_bin_data
def parse_segs(self, segs, metrics):
"""parses hmmcopy segments data
:param segs: path to hmmcopy segs file
"""
header_flag, dtypes, columns = csvutils.get_metadata(segs)
header = {v: i for i, v in enumerate(columns)}
segs_data = {}
with helpers.getFileHandle(segs) as segfile:
if header_flag:
assert segfile.readline().strip().split(',') == columns
for row in segfile:
row = row.strip().split(',')
chrom = row[header["chr"]]
start = row[header["start"]]
end = row[header["end"]]
cell_id = row[header["cell_id"]]
state = row[header["state"]]
# float to handle scientific notation
segment_length = int(float(end)) - int(float(start)) + 1
if metrics[cell_id] > self.quality_threshold:
continue
segs_data[cell_id] = [
cell_id, chrom, start, end, segment_length, state
]
return segs_data
def read_segs_csv(self):
"""
read the input file
"""
data = {}
bins = {}
header, dtypes, columns = csvutils.get_metadata(self.input)
with helpers.getFileHandle(self.input, 'rt') as freader:
idxs = self.build_label_indices(columns)
if header:
assert freader.readline().strip().split(',') == columns
for line in freader:
line = line.strip().split(self.sep)
sample_id = line[idxs['cell_id']]
val = line[idxs[self.column_name]]
val = float('nan') if val == "NA" else float(val)
chrom = line[idxs['chr']]
start = int(line[idxs['start']])
end = int(line[idxs['end']])
seg = (chrom, start, end)
if self.mappability_threshold and float(
line[idxs["map"]]) <= self.mappability_threshold:
val = float("nan")
if chrom not in bins:
bins[chrom] = set()
bins[chrom].add((start, end))
# just a sanity check, not required
if sample_id in data and seg in data[sample_id]:
raise Exception("repeated val")
if sample_id not in data:
data[sample_id] = {}
data[sample_id][seg] = val
samples = sorted(data.keys())
bins = self.sort_bins_csv(bins)
data = self.conv_to_matrix(data, bins, samples)
data = self.get_pandas_dataframe(data, bins)
return data