Annovar的注释结果，如果输出制表符分割的VCF格式，显得混乱。如果输出为csv格式，方便windows下的用户用excel打开，但不方便数据处理，比如某一列的注释信息中包含了逗号，这种情况就要特别注意。python中有csv模块可以方便的读取csv，推荐使用。

本文写的小脚本主要处理简单的csv格式，亮点在于支持接收标准输入和标准输出，方便生物信息多命令之间通过管道进行处理。如果没有指定输入文件，则读取管道流数据，如果没有指定输出文件，则可以用管道接收数据进行下一步处理。

比如 cat xxx.csv | python convert.py | grep "xx" > result.txt

或者 python convert.py -i input.csv | grep "xx" > result.txt

或者 python convert.py -i input.csv - o result.txt

查看用法 python convert.py --help

详细代码如下

#! /usr/bin/python

# -*- coding: utf-8 -*-

__author__ = 'Jason'

# Create: 20160114

# Modified on: 20160115

# Simple python script that convert csv to tab (-c or --csv option) or convert tab to csv (-t or --tab option).

# Support stdin and stdout

from optparse import OptionParser

import sys

def getOptions():

parser = OptionParser(usage="%prog [-i INPUT] [-csv|-tab] [-o OUT]", version="%prog 1.0")

parser.add_option("-i", "--input", type="string", dest="input", action="store", help="Input file. File should be seperated by tab or comma. If not specify, use stdin", metavar="file")

parser.add_option("-o", "--output", type="string", dest="output", action="store", help="Output file path. If not specify, use stdout.", metavar="file")

parser.add_option("-t", "--tab", default=False, dest="tab", action="store_true", help="Line field is tab splitted" )

parser.add_option("-c", "--csv", default=True, dest="csv", action="store_true", help="Line field is ',' splitted. Default True.")

(options, args) = parser.parse_args()

return options

def readLine(path):

"""

Yeild a line for processing.

:param path: Input file path. If path not specified, use stdin instead.

:return: line string

"""

if path:

for line in open(path):

yield line

else:

for line in sys.stdin.readlines():

yield line

def csv2tab(line):

"""

Just simplely convert csv to tab. Do not concert csv format with complex condition

:param line:

:return:

"""

line=line.replace(',"','"').replace('",','"')

t=line.split('"')#if , is enclosed by " , then , in odd number

newt=[]

if len(t) <=2 :

return line.replace(',','t')

else:

for i,e in enumerate(t):

if i%2==1:

newt.append(e)

else:

if e=='': continue

newt.extend(e.split(','))

return 't'.join(newt)

def tab2csv(line):

temp=[]

for t in line.split('t'):

if ',' in t:

temp.append('"%s"' % t) # if comma in field, use double quotation marks to enclose comma

else:

temp.append(t)

return ','.join(temp)

options=getOptions()

if options.output:

out=open(options.output,'w')

else: out=sys.stdout # if output path not specified, use stdout instead

for line in readLine(options.input):

if options.tab:

newLine=tab2csv(line)

else:

newLine=csv2tab(line)

out.write(newLine)

out.flush()

out.close()