import pandas as pd
import os
import sys

sampleFile = sys.argv[1]
transFile = sys.argv[2]
dataDir = sys.argv[3]
outFile = sys.argv[4]

sampleList = pd.read_csv(sampleFile, header=None, sep="\t")
transList = pd.read_csv(transFile, header=None, sep="\t")
#all gtf have the same gene order
out = pd.DataFrame(0, index=transList[0], columns=sampleList[0])

for sample in sampleList[0]:
    print("load " + sample)
    data = pd.read_csv(os.path.join(dataDir, sample+".gtf"), sep="\t", header=None, comment="#")
    data_tmp1 = data.loc[data[2]=="transcript",8]
    trans_id = [x.split("\"")[3] for x in data_tmp1.values]
    trans_fpkm = [x.split("\"")[7] for x in data_tmp1.values]
    data_tmp2 = pd.DataFrame({"fpkm":trans_fpkm}, index=trans_id)
    idx = set(data_tmp2.index).intersection(set(out.index))
    data_tmp3 = data_tmp2.loc[idx,]
    out.loc[data_tmp3.index, sample] = data_tmp3['fpkm']
out.to_csv(outFile, sep="\t", header=True, index=True)