Transform tab separated matrix to gct file

tsv2gct.py

import pandas as pd
import sys
import glob
import os

# input / output directory
input_dir = sys.argv[1]
# input file extension
input_ext = sys.argv[2]
# cardinality of index columns (rownames)
# FIXME: by now it assumes the first two columns as the index
# n_index_columns = int(sys.argv[3])
n_index_columns = 2

index_columns = [i for i in range(n_index_columns)]

for f in glob.glob(os.path.join(input_dir, "*" + input_ext)):
    f_read = pd.read_table(f, sep='\t', index_col=index_columns)
    n_rows, n_cols = f_read.shape
    output_fn = os.path.splitext(f)[0] + '.gct'

    with open(f) as txt:
        data = txt.read()
    with open(output_fn, 'w') as gct:
        gct.write('#1.2\n' + str(n_rows) + '\t' + str(n_cols) + '\n' + data)