Fun with Python subprocess! Works on Python 2.7.x.

python.subprocess.py

import json
import logging
import os
import subprocess


logger = logging.getLogger(os.path.splitext(os.path.basename(__file__))[0])


# If you have ever used jq and tried to use it in Python but ran into
# issues with building a new jq instance rather than using the system-
# available jq binary then this may be beneficial to you!
# Please note that this has been tested on rather large JSON documents
# but it can be quite easy to overrun the buffers and cause the 
# operation to fail but this is a good start.
def execute_jq_cmd(data_str, jq_query):
    proc = subprocess.Popen(
        'jq \'{}\' -'.format(jq_query),
        shell=True,
        stdin=subprocess.PIPE,
        stdout=subprocess.PIPE,
        stderr=subprocess.PIPE
    )

    proc.stdin.write(data_str)
    proc_comm = proc.communicate()
    
    proc_stdout = proc_comm[0].strip()
    proc_stderr = proc_comm[1].strip()

    if proc_stdout:
        try:
            return json.loads(proc_stdout)
        except Exception as err:
            logger.exception(err)
            return []
    elif proc_stderr:
        raise Exception(proc_stdout)

 
# This assumes that you already have ffmpeg/ffprobe installed.
def ffprobe_json(file_path):
    """Execute ffprobe and return a dictionary object containing media medata.

    :param file_path: Path of the file to analyze.
    :type file_path: str
    :return: Dictionary structure containing media metadata.
    :rtype: dict
    """
    file_path = os.path.abspath(file_path)

    FFPROBE_CMD = [
        'ffprobe',
        # '/usr/local/bin/ffprobe',
        '-v quiet',
        '-show_format',
        '-show_streams',
        '-show_programs',
        '-show_chapters',
        '-show_private_data',
        '-show_pixel_formats',
        '-show_error',
        '-print_format json',
        '-i "{}"'.format(file_path)
    ]

    p = subprocess.Popen(
        ' '.join(FFPROBE_CMD),
        shell=True,
        stdout=subprocess.PIPE,
        stderr=subprocess.PIPE
    )

    # Better handling of large quanitties of data without overloading buffers.
    draineddata = []
    drainerthread = threading.Thread(target=draineddata.extend, args=(p.stdout,))
    drainerthread.daemon = True
    drainerthread.start()

    p.wait()
    drainerthread.join()

    out = ''.join(draineddata)

    data_dict = {}
    if out:
        data_dict.update(json.loads(out))
    return data_dict