在Mac OS X上捕获Python子进程输出时出现问题 [英] Problems capturing Python subprocess output on Mac OS X

问题描述

我正在Mac OS 10.6.8上运行Python 3.3.我正在编写一个运行多个子进程的脚本，我想捕获每个子进程的输出并将其记录在一个文件中.我对此感到麻烦.

I'm running Python 3.3 on Mac OS 10.6.8. I am writing a script that runs several subprocesses, and I want to capture the output of each one and record it in a file. I'm having trouble with this.

我首先尝试了以下方法:

I first tried the following:

import subprocess
logFile = open("log.txt", 'w')
proc = subprocess.Popen(args, stdout=logFile, stderr=logFile)
proc.wait()

这产生了一个空的log.txt.在互联网上闲逛了一会之后，我尝试了一下

This produced an empty log.txt. After poking around on the internet for a bit, I tried this instead

import subprocess
proc = subprocess.Popen(args, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
output, err = proc.communicate()
logFile = open("log.txt", 'w')
logFile.write(output)

这也产生了一个空的log.txt.因此，我没有写文件，而是尝试将输出打印到命令行:

This, too, produced an empty log.txt. So instead of writing to the file, I tried to just print the output to the command line:

output, err = proc.communicate()
print(output)
print(err)

那产生了这个:

b''
b''

我要运行的进程是fastq_quality_trimmer.它获取一个输入文件，对其进行过滤，然后将结果保存到一个新文件中.它只向stdout写了几行，就像这样

The process I'm trying to run is fastq_quality_trimmer. It takes an input file, filters it, and saves the result to a new file. It only writes a few lines to stdout, like so

Minimum Quality Threshold: 20
Minimum Length: 20
Input: 750000 reads.
Output: 750000 reads.
discarded 0 (0%) too-short reads.

如果我从命令行运行它，然后像这样重定向输出

If I run it from the command line and redirect the output like this

fastq_quality_trimmer -Q 33 -v -t 50 -l 20 -i in.fq -o in_trimmed.fq > log.txt

输出已成功写入log.txt.

the output is successfully written to log.txt.

我想也许当我用Popen调用fastq_quality_trimmer时，它以某种方式无法运行，但是我的脚本所生成的过滤文件与我从命令行运行fastq_quality_trimmer时所生成的文件相同.所以这是有效；我只是无法捕获输出.为了使事情更加混乱，我可以使用与我发布的代码基本相同的代码成功捕获其他进程(echo，其他Python脚本)的输出.

I thought perhaps that fastq_quality_trimmer was somehow failing to run when I called it with Popen, but my script produces a filtered file that is identical to the one produced when I run fastq_quality_trimmer from the command line. So it's working; I just can't capture the output. To make matters more confusing, I can successfully capture the output of other processes (echo, other Python scripts) using code that is essentially identical to what I've posted.

有什么想法吗?我错过了令人眼花obvious乱的明显东西吗?

Any thoughts? Am I missing something blindingly obvious?

推荐答案

您忘记了逗号:

["fastq_quality_trimmer", "-Q", "33" "-v", "-t", "50", "-l", "20", "-i", leftInitial, "-o", leftTrimmed]

将其添加到"33"和"-v"之间.

您实质上是在传递参数-Q 33-v而不是-Q 33 -v.

You are essentially passing in the arguments -Q 33-v instead of -Q 33 -v.

如果两个相邻的字符串之间只有空格，则它们将串联两个字符串:

Python will concatenate two adjacent strings if there is only whitespace between them:

>>> "33", "-v"
('33', '-v')
>>> "33" "-v"
'33-v'

由于-v是使fastq_quality_trimmer完全产生输出所必需的冗长的开关，因此它将丢失而保持沉默.

Since -v is the verbose switch that is required to make fastq_quality_trimmer produce output at all, it'll remain silent with it missing.

每当调用子流程遇到问题时， triple 都会检查创建的命令行.在args之前加上['echo']可以帮助实现这一点:

Whenever you encounter problems with calling a subprocess, triple check the command line created. Pre-pending args with ['echo'] can help in that:

proc = subprocess.Popen(['echo'] + args, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
output, err = proc.communicate()
print(output)

这篇关于在Mac OS X上捕获Python子进程输出时出现问题的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！