bash 4.4 inside python with os.system

Question

I have problems with running a bash script inside a python script script.py:

import os
bashCommand = """
sed "s/) \['/1, color=\"#ffcccc\", label=\"/g" list.txt | sed 's/\[/    GraphicFeature(start=/g' | sed 's/\:/, end=/g' | sed 's/>//g' | sed 's/\](/, strand=/g' | sed "s/'\]/\"),/g" >list2.txt"""

os.system("bash %s" % bashCommand)

When I run this as python script.py, no list2.txt is written, but on the terminal I see that I am inside bash-4.4 instead of the native macOS bash.

Any ideas what could cause this?

The script I posted above is part of a bigger script, where first it reads in some file and outputs list.txt.

edit: here comes some more description In a first python script, I parsed a file (genbank file, to be specific), to write out a list with items (location, strand, name) into list.txt. This list.txt has to be transformed to be parsable by a second python script, therefore the sed.

list.txt

[0:2463](+) ['bifunctional aspartokinase/homoserine dehydrogenase I']
[2464:3397](+) ['Homoserine kinase']
[3397:4684](+) ['Threonine synthase']

all the brackets, :, ' have to be replaced to look like desired output list2.txt

    GraphicFeature(start=0, end=2463, strand=+1, color="#ffcccc", label="bifunctional aspartokinase/homoserine dehydrogenase I"),
    GraphicFeature(start=2464, end=3397, strand=+1, color="#ffcccc", label="Homoserine kinase"),
    GraphicFeature(start=3397, end=4684, strand=+1, color="#ffcccc", label="Threonine synthase"),

This looks much more complicated than it needs to be. What are you actually trying to do? At the very least, you should be able to reduce this to a single call to `sed`, using multiple `-e` arguments, and handling the output redirection with `subprocess.call`. — chepner, May 28 '17 at 19:45
`with open("list2.txt", "w") as fh: subprocess.call(["sed", ...], stdout=fh)` — chepner, May 28 '17 at 19:52
@chepner I know I should be able to reduce that or even implement in python, but I am not unfortunately... — rororo, May 28 '17 at 19:59
Hence, me asking for a description of what you are trying to accomplish, so that I can suggest a pure Python solution. What's the input and expected output? — chepner, May 28 '17 at 20:02

chepner · Accepted Answer · 2017-05-28T20:27:42.733

Read the file in Python, parse each line with a single regular expression, and output an appropriate line constructed from the captured pieces.

import re
import sys

#                         1     2                3
#                        ---   ---              --
regex = re.compile(r"^\[(\d+):(\d+)\]\(\+\) \['(.*)'\]$")
# 1 - start value
# 2 - end value
# 3 - text value
with open("list2.txt", "w") as out:
    for line in sys.stdin:
        line = line.strip()
        m = regex.match(line)
        if m is None:
            print(line, file=out)
        else:
            print('GraphicFeature(start={}, end={}, strand=+1, color="#ffcccc", label="{}"),'.format(*m.groups()), file=out)

I output lines that don't match the regular expression unmodified; you may want to ignore them altogether or report an error instead.

I assumed Python 3; add `from __future__ import print_function` to the top of your file to make it work in Python 2.x (assuming you aren't using a truly ancient version of Python). — chepner, May 28 '17 at 20:35

bash 4.4 inside python with os.system

1 Answers1