fq_smarttrim

#!/bin/env python2.7

def exit(str):
	print str
	parser.print_help()
	sys.exit()


def trim(string, length, overlap):
	trimmedlines = []

	while True:
		firstchunk = string[:length]
		string = string[(length - overlap):]

		trimmedlines.append(firstchunk)

		if len(string) < length:
			trimmedlines.append(string[:-1])
			break
 	
		
	return trimmedlines


if __name__ == "__main__":

	import sys
	import argparse

	parser = argparse.ArgumentParser(description = 'Trims a fasta/fastq file by length, splitting too long reads into smaller ones',
						prog='fq_smarttrim',
						add_help = False)
	
	required = parser.add_argument_group('Required Arguments')
	required.add_argument('-i', '--input', default = '', help = 'input file')
	required.add_argument('-l', '--max-length', type = int, default = 0, help = 'maximum read length')

	optional = parser.add_argument_group('Optional Arguments')
	optional.add_argument('-o', '--output', default = '', help = 'output file (def: stdout)')
	optional.add_argument('-Q', '--fastq', action='store_true', help = 'input is fastq')
	optional.add_argument('-ovl', '--overlap', default = 0, type = int, help = 'overlap when splitting reads (0)')
	optional.add_argument('-m', '--min-length', default = 1, type = int, help = 'minimum read lenght')

	args = vars( parser.parse_args() )

	# Verify input validity
	if args['input'] is '':
		exit('Must provide input file')
	if args['max_length'] == 0:
		exit('Must provide valid maximum read length')

	
	# Define output
	if args['output'] is '':
		out = sys.stdout
	else:
		out = open(args['output'], 'w')
	
	
	# Open file
	try:
		infile = open(args['input'], 'r')
		lines = infile.readlines()
		
		if not args['fastq']:
			i = 0
			for line in lines:
				# out.write(line)
				if (len(line) > args['max_length']):
					newlines = trim(line, args['max_length'], args['overlap'])
					for newline in newlines:
						if len(line) >= args['min_length']:

							header = ">" + args['input'] + "_" + str(i)
							i +=1
							out.write(header + '\n')
							out.write(newline + '\n')
						else:
							# read too short
							pass			
				else:
					if not (line[0] in ['>', '@', '+']) and (len(line) >= args['min_length']):
						header = '>' + args['input'] + '_' + str(i)
						i+=1
						out.write(header + '\n')
						out.write(line)	

		else:
			print 'Fastq not supported yet'
		
	
	
	except IOError as e:
		print 'Error: File: ' + args['input'] + ' not found'
		print e
	finally:
		if not args['output'] is '':
			out.close()