def possible_lims():
concs = '/Users/berriz/Work/attachments/MF/JWGray_drug_conc_v1.csv'
with open(concs) as fh:
allrows = tuple(csv.reader(fh))
record = co.namedtuple('Record', allrows[0])
table = (record(*(row[:3] + list(convert_conc_row(row[3:-1])) + row[-1:]))
for row in allrows[1:])
possible_lims = dict()
for row in mu.unique(
tuple((row.drug, row.c1, row.c9, row.units) for row in table)):
possible_lims.setdefault(row[0], []).append(tuple(row[1:3]))
return possible_lims
def possible_lims():
concs = '/Users/berriz/Work/attachments/MF/JWGray_drug_conc_v1.csv'
with open(concs) as fh:
allrows = tuple(csv.reader(fh))
record = co.namedtuple('Record', allrows[0])
table = (record(*(row[:3] +
list(convert_conc_row(row[3:-1])) +
row[-1:]))
for row in allrows[1:])
possible_lims = dict()
for row in mu.unique(tuple((row.drug, row.c1, row.c9, row.units)
for row in table)):
possible_lims.setdefault(row[0], []).append(tuple(row[1:3]))
return possible_lims
def main():
allrows = tuple(csv.reader(open(DATASRC), delimiter=FIELDDELIMITER))
allcols = zip(*allrows)
colheaders = tuple(allrows[COLHEADERSROW][FIRSTDATACOL:])
colsubheaders = tuple(allrows[COLSUBHEADERSROW][FIRSTDATACOL:])
rowheaders = tuple(allcols[ROWHEADERSCOL][FIRSTDATAROW:])
alldrugs = mu.unique(colheaders)
drugs = tuple((dr for dr in alldrugs
if not dr.startswith('private_company_drug_')))
drugset = set(drugs)
metrics = 'lc50 tgi gi50'.split()
assert set(colsubheaders) == set(metrics)
nmetrics = len(metrics)
assert len(colheaders) == len(alldrugs) * nmetrics
cell_lines = rowheaders
alldata = [list(col[FIRSTDATAROW:]) for col in allcols[FIRSTDATACOL:]]
col_lookup = dict()
for drug, subheader, data in zip(colheaders, colsubheaders, alldata):
if drug not in drugset:
continue
d = col_lookup.setdefault(subheader, ncd.NoClobberDict())
d[drug] = data
nrows = len(allrows[FIRSTDATAROW:])
assert nrows == len(cell_lines)
maxvals = read_maxvals(MAXVALSFILE)
for metric, drug in it.product(metrics, drugs):
col = col_lookup[metric][drug]
mv = maxvals[drug]
for j, v in enumerate(col):
assert isna(v) or float(v) <= float(mv)
if v == mv:
col[j] = MAXVALMARKER
for metric, lkp in col_lookup.items():
table0 = zip(*(lkp[dr] for dr in drugs))
drug_scores = score_cols(table0)
cols = zip(*table0)
sorted_cols = sort_by_score(cols, drug_scores)
table1 = zip(*sorted_cols)
cell_line_scores = score_rows(table1)
table2 = sort_by_score(table1, cell_line_scores)
sorted_cell_lines = sort_by_score(cell_lines, cell_line_scores)
sorted_drugs = sort_by_score(drugs, drug_scores)
print_table(table2, sorted_cell_lines, sorted_drugs, metric)
