Is it possible to output to and monitor streams other than stdin, stdout & stderr? (python)

Question

This is a python question, but also a linux/BSD question.

I have a python script with two threads, one downloading data from the web and the other sending data to a device over a serial port. Both of these threads print a lot of status information to stdout using python's logging module.

What I would like is to have two terminal windows open, side by side, and have each terminal window show the output from one thread, rather than have the messages from both interleaved in a single window.

Are there file descriptors other than stdin, stdout & stderr to write to and connect to other terminal windows? Perhaps this wish is better fulfilled with a GUI?

I'm not sure how to get started with this.

edit: I've tried writing status messages to two different files instead of printing them to stdout, and then monitoring these two files with tail -f in other terminal windows, but this doesn't work for live monitoring because the files aren't written to until you call close() on them.

Answer 1

First, customize your logging formatter to include a Thread id field (https://docs.python.org/2/library/logging.html#logrecord-attributes). Then change your logging destination to some file instead of stdout.

# A simple logger as print
import logging
import logging.handlers

hdr = logging.FileHandler(filename='output.log')
hdr.setFormatter(logging.Formatter('[%(asctime)s] thread=%(thread)s:%(levelname)s: %(message)s'))
logger = logging.getLogger(__name__)
logger.addHandler(hdr)
logger.setLevel(logging.DEBUG)

import threading


def func():
    logger.info('test message')


for i in range(2):
    threading.Thread(target=func).start()

Your logging output may look like this now:

% tail -f output.log
[2015-09-28 15:14:49,782] thread=4344852480:INFO: test message
[2015-09-28 15:14:49,782] thread=4349059072:INFO: test message

Run your script, open two individual terminals, and use the command tail -f output.log | grep thread=<THREAD_ID> to monitor logs by thread ID.