filter_transcripts.py

filename = snakemake.input[0]
file = open(filename, "r")
lines = file.readlines()
file.close()

output = []
current = []
fake = True
ips = False

for line in lines:
    if (line.startswith(">")):
        if (not fake):
            output += current
        current = []    
        fake = False
        
    if (line.startswith("#####InterProScan")):
        ips = True
        count = -1
        
    if (line.startswith("#####BrewerySS8 Analysis")):
        if (count < 1):
            fake = True
        ips = False
        
    if (ips == True):
        count += 1
        
    current.append(line.strip())
        
output_file = open(snakemake.output[0],"w+")
output_file.write("\n".join(output))
output_file.close()