Merge branch 'master' of https://github.com/nstanger/process

nigel.stanger / process_podcast

Browse code Merge branch 'master' of https://github.com/nstanger/process_podcast Python2.7 master LAST_2.7_RELEASE
commit 896025181de57365b8231794289aef464f0897e5 2 parents 3b8ff99 + 1ef51f2 Nigel Stanger authored on 23 Jul 2018

Patch

Showing 2 changed files

                Ignore Space
               Show notes
              View
            
          
          56 ■■■■■
          process_podcast.py
             "See config_help.md details on the file "
            "format.".format(p=globals.PROGRAM))
    
    parser.add_argument(
        "--copy-audio", dest="process_audio", action="store_false",
        default=True,
        help="Disable additional processing of the source audio. The audio "
            "will still be re-encoded using the specified audio codec "
            "because the concatenation filter requires it, but extra "
            "processing such as reduction of channels or normalisation "
            "will not be carried out. (Implies --no-normalise.)")
    
    parser.add_argument(
        "--copy-video", dest="process_video", action="store_false",
        default=True,
        help="Disable additional processing of the source video. The video "
            "will still be re-encoded using the specified video codec "
            "because the concatenation filter requires it, but extra "
            "processing such as remapping of colours will not be "
            "carried out.")
    
    parser.add_argument(
        "--no-normalise", dest="normalise", action="store_false",
        default=True,
        help="Disable normalisation of the source audio level (implied by "
            "--copy-audio).")
    
    parser.add_argument(
        "--audio-codec", dest="audio_codec", metavar="CODEC",
        default="pcm_s16le",
        help="Specify ffmpeg audio codec for output (default pcm_s16le). "
            "See the output of ffmpeg -codecs for possible codecs.")
    
    parser.add_argument(
        "--video-codec", dest="video_codec", metavar="CODEC", default="h264",
        help="Specify ffmpeg video codec for output (default h264). "
            "See the output of ffmpeg -codecs for possible codecs.")
    
    parser.add_argument(
        "--input-prefix", "-i", dest="prefix", metavar="PATH", default=".",
        help="Path to be prefixed to all INPUT files. This includes the "
            "configuration file, if applicable, and any files specified "
            "within the configuration file. Input files that already "
        prefix_path, [args.prefix] * 3, [args.audio, args.video, args.config])
 
    if args.quiet:
        globals.log.setLevel(logging.WARNING)
    
    # --copy-audio implies --no-normalise.
    if not args.process_audio:
        args.normalise = False
        
    # --debug overrides --quiet.
    if args.debug:
        globals.log.setLevel(logging.DEBUG)
    
    # Odd number of timestamps: punch in at last timestamp,
    # out at stream duration.
    if (len(time_list) % 2 != 0):
        globals.log.debug("{fn}(): odd number of timestamps".format(fn=fn))
        punch_in, _ = process_timestamp_pair(args, [time_list[-1], None])
        punch_out = stream_duration - punch_in
        punch_out = stream_duration
        segments.append(make_new_segment(type, filename, punch_in,
                                         punch_out, num))
    return segments
 
    globals.log.info("Rendering final podcast...")
    command = FFmpegConcatCommand(has_audio=len(audio_segments) > 0,
                                  has_video=len(video_segments) > 0,
                                  max_progress=duration,
                                  quiet=args.quiet and not args.debug)
                                  quiet=args.quiet and not args.debug,
                                  process_audio=args.process_audio,
                                  process_video=args.process_video,
                                  audio_codec=args.audio_codec,
                                  video_codec=args.video_codec)
    input_files = Segment.input_files()
    for f in input_files:
        if (input_files[f]):
            command.append_input_options(input_files[f])
        command.append_input_options(["-i", f])
    for s in (audio_segments + video_segments):
        command.append_filter(s.trim_filter())
    command.append_concat_filter("a", [s for s in audio_segments])
    command.append_normalisation_filter()
    if (args.normalise):
        command.append_normalisation_filter()
    command.append_concat_filter("v", [s for s in video_segments])
    if args.preview:
        globals.log.info("PREVIEW MODE: {fps} fps".format(fps=args.preview))
        command.append_output_options(["-r", args.preview])

            #!/usr/bin/env python

import argparse
import datetime
import logging
import os.path
import sys

from pyparsing import ParseResults

import globals
from config_parser import (
    parse_configuration_file, parse_configuration_string)
from progress_bar import (ProgressBar)
from segment import (Segment, AudioSegment, VideoSegment,
                     FrameSegment, SegmentException)
from shell_command import (FFprobeCommand, FFmpegConcatCommand)


class InputStreamAction(argparse.Action):
    def __call__(self, parser, namespace, values, option_string=None):
        input = values.split(":")
        file = input[0]
        stream = None if (len(input) == 1) else input[1]
        setattr(namespace, self.dest, file)
        if (option_string in ["--audio", "-a"]):
            setattr(namespace, 'audio_stream_number', stream)
        elif (option_string in ["--video", "-v"]):
            setattr(namespace, 'video_stream_number', stream)


def parse_command_line():
    """Parse command line arguments."""
    parser = argparse.ArgumentParser(
        usage="%(prog)s [options] <output>",
        description="where: <output> is the name of the output file "
            "(note: .mov seems generally best)",
        epilog="Default input files can be specified using either of "
            "--audio or --video. If neither of these are specified, "
            "then you must supply a configuration file using --config. "
            "(Of course, you can always supply a configuration file "
            "regardless.)\n\n"
            "Input streams can be taken from the same input file.\n\n"
            "If no segments are specified, the entire input stream is "
            "processed as one segment. The number and duration of segments "
            "can differ, but the total duration across all input streams "
            "should ideally be the same.")
    
    parser.set_defaults(audio_stream_number=None, video_stream_number=None)

    parser.add_argument(
        "output",
        help="name of the output file (note: .mov is best)")
    
    parser.add_argument(
        "--audio", "-a", metavar="FILE[:STREAM]", action=InputStreamAction,
        help="File name for the default audio input stream (can be the "
            "same as for other input streams). You can optionally specify "
            "the default audio stream number to use if the file contains "
            "more than one (this can be overidden in a configuration "
            "file). If you don't specify a stream number, it defaults "
            "to 0 (i.e., the first audio stream in the file).")
    
    parser.add_argument(
        "--video", "-v", metavar="FILE[:STREAM]", action=InputStreamAction,
        help="File name for the default video input stream (can be the "
            "same as for other input streams). You can optionally specify "
            "the default video stream number to use if the file contains "
            "more than one (this can be overidden in a configuration "
            "file). If you don't specify a stream number, it defaults "
            "to 0 (i.e., the first video stream in the file).")
    
    parser.add_argument(
        "--configuration", "--config", "-c", dest="config", metavar="FILE",
        help="File name for the podcast segment configuration (plain text). "
            "See config_help.md details on the file "
            "format.".format(p=globals.PROGRAM))
    
    parser.add_argument(
        "--copy-audio", dest="process_audio", action="store_false",
        default=True,
        help="Disable additional processing of the source audio. The audio "
            "will still be re-encoded using the specified audio codec "
            "because the concatenation filter requires it, but extra "
            "processing such as reduction of channels or normalisation "
            "will not be carried out. (Implies --no-normalise.)")
    
    parser.add_argument(
        "--copy-video", dest="process_video", action="store_false",
        default=True,
        help="Disable additional processing of the source video. The video "
            "will still be re-encoded using the specified video codec "
            "because the concatenation filter requires it, but extra "
            "processing such as remapping of colours will not be "
            "carried out.")
    
    parser.add_argument(
        "--no-normalise", dest="normalise", action="store_false",
        default=True,
        help="Disable normalisation of the source audio level (implied by "
            "--copy-audio).")
    
    parser.add_argument(
        "--audio-codec", dest="audio_codec", metavar="CODEC",
        default="pcm_s16le",
        help="Specify ffmpeg audio codec for output (default pcm_s16le). "
            "See the output of ffmpeg -codecs for possible codecs.")
    
    parser.add_argument(
        "--video-codec", dest="video_codec", metavar="CODEC", default="h264",
        help="Specify ffmpeg video codec for output (default h264). "
            "See the output of ffmpeg -codecs for possible codecs.")
    
    parser.add_argument(
        "--input-prefix", "-i", dest="prefix", metavar="PATH", default=".",
        help="Path to be prefixed to all INPUT files. This includes the "
            "configuration file, if applicable, and any files specified "
            "within the configuration file. Input files that already "
            "include the input prefix will not have it added again.")
    
    parser.add_argument(
        "--preview", "-p", metavar="RATE", nargs="?", const="1",
        help="Generate a preview of the podcast by only rendering a "
            "subset of the video frames. RATE is the number of frames "
            "per second to render (default 1 fps). You can specify "
            "fractions (e.g., 1/10 for one frame every 10 seconds). "
            'NOTE: if you get a "too few arguments" error when using '
            "this option, you've probably not provided an fps value and "
            "placed the option just before the output filename. "
            "Either move --preview earlier in the option list, add "
            'a "--" between it and the output filename, or provide a '
            "fps value.")
    
    parser.add_argument(
        "--debug", "-d", action="store_true",
        help="Print debugging output (overrides --quiet).")
    
    parser.add_argument(
        "--keep", "-k", action="store_true",
        help="Don't delete any generated temporary files.")
    
    parser.add_argument(
        "--quiet", "-q", action="store_true",
        help="Mute all console output (overridden by --debug).")

    args = parser.parse_args()
    
    return args


def prefix_path(prefix, path):
    """Prepend prefix to path, unless path already starts with prefix"""
    assert(prefix is not None)
    if not path:
        return None
    elif path == "." or path.startswith(prefix):
        return path
    else:
        return os.path.join(prefix, path)


def check_arguments(args):
    """Sanity check the command line arguments."""
    fn = "check_arguments"
    # Prepend input files with --input-prefix where applicable.
    args.audio, args.video, args.config = map(
        prefix_path, [args.prefix] * 3, [args.audio, args.video, args.config])

    if args.quiet:
        globals.log.setLevel(logging.WARNING)
    
    # --copy-audio implies --no-normalise.
    if not args.process_audio:
        args.normalise = False
        
    # --debug overrides --quiet.
    if args.debug:
        globals.log.setLevel(logging.DEBUG)
        globals.log.debug("{fn}(): args = {a}".format(fn=fn, a=args))
    
    # Must specify at least one of --audio, --video, --config.
    if not any([args.audio, args.video, args.config]):
        globals.log.error("must specify at least one of --audio, --video, "
                          "or --config")
        sys.exit(1)
    
    if not os.path.exists(args.prefix):
        globals.log.error('input prefix "{p}" does not '
                          "exist".format(p=args.prefix))
        sys.exit(1)
    

def get_configuration(args):
    """Load podcast configuration."""
    # Fill in missing file names for default input streams.
    fn = "get_configuration"
    # These types can have default input files and streams.
    type_mapping = {
        "audio": {"file": args.audio, "stream": args.audio_stream_number},
        "video": {"file": args.video, "stream": args.video_stream_number}}
    globals.log.info("Processing configuration...")
    if (args.config):
        config = parse_configuration_file(args.config)
        
        # Check that applicable default input streams have been specified.
        for i, c in enumerate(config):
            type = c["type"]
            
            # Add prefix to filename, if applicable.
            if c["filename"] and (c["filename"] != "^"):
                config[i]["filename"] = prefix_path(args.prefix,
                                                    config[i]["filename"])
            
            if (type in type_mapping):
                default_file = type_mapping[type]["file"]
                default_stream = type_mapping[type]["stream"]
                error_string = ("attempting to use default {s} input file, "
                    "but --{s} hasn't been specified".format(s=type))
            else:
                default_file = None
                default_stream = 0
                error_string = ("attempting to use a default input file, "
                    "but the {s} type doesn't support this".format(s=type))

            # No filename in configuration.
            if (not c["filename"]):
                if (default_file):
                    config[i]["filename"] = default_file
                # No filename on command line either.
                else:
                    globals.log.error(error_string)
                    sys.exit(1)
            
            # No stream number in configuration. Note: 0 is a valid
            # stream number, so explicitly check for None.
            if (c["num"] is None):
                # Assume 0 if no stream on command line either.
                if (default_stream is None):
                    config[i]["num"] = 0
                else:
                    config[i]["num"] = default_stream
    
    # No configuration file.
    else:
        conf_list = []
        for m in type_mapping:
            default_file = type_mapping[m]["file"]
            default_stream = type_mapping[m]["stream"]
            if (default_file and default_stream is not None):
                conf_list += ["[{type}:{file}:{stream}]".format(
                    type=m, file=default_file, stream=default_stream)]
        globals.log.debug("{fn}(): default config = "
                          "{c}".format(fn=fn, c=conf_list))
        config = parse_configuration_string("\n".join(conf_list))
    
    return config


def get_file_duration(file):
    """Calculate the duration a media file as a timedelta object."""
    fn = "get_file_duration"
    command = FFprobeCommand([file])
    globals.log.debug("{fn}(): {cmd}".format(fn=fn, cmd=command))
    # Only consider the first stream. If it's the only stream in the
    # file, great; otherwise it seems reasonable to assume that all
    # streams in the same file will have the same duration.
    ss, ms = command.get_entries(
        section="format", find_list=["duration"])[0].split(".")
    ms = ms[:3].ljust(3, "0")
    globals.log.debug("{fn}(): ss = {ss}, ms = {ms}".format(fn=fn, ss=ss,
                                                            ms=ms))
    return datetime.timedelta(seconds=int(ss), milliseconds=int(ms))


def make_new_segment(type, filename, punch_in, punch_out, num):
    """Make a new segment instance of the correct class."""
    fn = "make_new_segment"
    globals.log.debug("{fn}(): type = {t}".format(fn=fn, t=type))
    globals.log.debug("{fn}(): filename = {f}".format(fn=fn, f=filename))
    globals.log.debug("{fn}(): punch in = {i}".format(fn=fn, i=punch_in))
    globals.log.debug("{fn}(): punch out = {o}".format(fn=fn, o=punch_out))
    globals.log.debug("{fn}(): num = {n}".format(fn=fn, n=num))
    
    if (type == "audio"):
        return AudioSegment(file=filename, punch_in=punch_in,
                            punch_out=punch_out, input_stream=num)
    elif (type == "video"):
        return VideoSegment(file=filename, punch_in=punch_in,
                            punch_out=punch_out, input_stream=num)
    elif (type == "frame"):
        return FrameSegment(file=filename, punch_in=punch_in,
                            punch_out=punch_out, frame_number=num)
    else:
        return None


def process_timestamp_pair(args, times):
    """Constructs timedelta instances from a pair of config timestamps."""
    fn = "process_timestamp_pair"
    globals.log.debug("{fn}(): times[0] = {t}".format(fn=fn, t=times[0]))
    globals.log.debug("{fn}(): times[1] = {t}".format(fn=fn, t=times[1]))
    
    # If the first item in the timestamp list in the configuration file
    # is a filename, the parser inserts a zero timestamp before it. We
    # can therefore guarantee that the first item of the pair will
    # always be a timestamp.
    t0 = datetime.timedelta(
        hours=times[0]["hh"], minutes=times[0]["mm"],
        seconds=times[0]["ss"], milliseconds=times[0]["ms"])
    if (times[1]):
        if (len(times[1]) == 1): # filename
            t1 = t0 + get_file_duration(
                prefix_path(args.prefix, times[1]["filename"]))
        elif (len(times[1]) == 4): # normal timestamp
            t1 = datetime.timedelta(
                hours=times[1]["hh"], minutes=times[1]["mm"],
                seconds=times[1]["ss"], milliseconds=times[1]["ms"])
        else:
            globals.log.error("{fn}():unreadable timestamp {t}".format(
                fn=fn, t=times[1]))
            t1 = None
    else:
        t1 = None
    
    globals.log.debug("{fn}(): t0 = {t}".format(fn=fn, t=t0))
    globals.log.debug("{fn}(): t1 = {t}".format(fn=fn, t=t1))
    return t0, t1


def process_time_list(args, type, filename, num, time_list):
    """Process an audio or video stream and build a list of segments."""
    fn = "process_time_list"
    if (os.path.exists(filename) and type in ["audio", "video"]):
        stream_duration = get_file_duration(filename)
    else:
        stream_duration = 0
    segments = []
    globals.log.debug("{fn}(): stream duration = {d}".format(
        fn=fn, d=stream_duration))
    
    # No timestamps: punch in at 0, out at stream duration.
    if (len(time_list) == 0):
        punch_in = datetime.timedelta()
        punch_out = stream_duration
        segments.append(make_new_segment(type, filename, punch_in,
                                         punch_out, num))
    else:
        # Process each pair of timestamps as punch in, out. If there's
        # an odd number of items, the last one is processed separately.
        for t in zip(time_list[::2], time_list[1::2]):
            punch_in, punch_out = process_timestamp_pair(args, t)
            if (punch_in == punch_out):
                globals.log.warning(
                    "punch in ({i}s) and punch out ({o}s) times are "
                    "equal; no segment will be "
                    "generated".format(i=punch_in.total_seconds(),
                                       o=punch_out.total_seconds()))
                continue
            elif (punch_out < punch_in):
                globals.log.error(
                    "punch out time ({i}s) falls before punch in time "
                    "({o}s); can't generate a valid "
                    "segment".format(i=punch_in.total_seconds(),
                                     o=punch_out.total_seconds()))
                sys.exit(1)
            segments.append(make_new_segment(type, filename, punch_in,
                                             punch_out, num))
    
    # Odd number of timestamps: punch in at last timestamp,
    # out at stream duration.
    if (len(time_list) % 2 != 0):
        globals.log.debug("{fn}(): odd number of timestamps".format(fn=fn))
        punch_in, _ = process_timestamp_pair(args, [time_list[-1], None])
        punch_out = stream_duration
        segments.append(make_new_segment(type, filename, punch_in,
                                         punch_out, num))
    return segments


def process_input_streams(args, config):
    """Process a list of stream specifications and build a list of segments."""
    fn = "process_input_streams"
    globals.log.info("Processing input streams...")
    segments = []
    for cnf in config:
        globals.log.debug("{fn}(): type = {t}".format(fn=fn, t=cnf["type"]))
        globals.log.debug(
            "{fn}(): filename = {f}".format(fn=fn, f=cnf["filename"]))
        globals.log.debug("{fn}(): num = {n}".format(fn=fn, n=cnf["num"]))
        globals.log.debug("{fn}(): times = {t}".format(fn=fn, t=cnf["times"]))
    
        segments += process_time_list(args, cnf["type"], cnf["filename"],
                                      cnf["num"], cnf["times"])  
    return segments


def smallest_video_dimensions(args, segments):
    """Compute the smallest frame dimensions across all video inputs."""
    fn = "smallest_video_dimensions"
    width = 2048
    height = 1536
    for s in segments:
        if isinstance(s, FrameSegment):
            continue
        command = FFprobeCommand([s.input_file])
        globals.log.debug("{fn}(): {cmd}".format(fn=fn, cmd=command))
        w, h = command.get_entries(find_list=["width", "height"])
        if (w * h) < (width * height):
            width, height = w, h
    return width, height


def process_frame_segments(args, segments, width, height):
    """Post-process frame segments to set frame images, etc."""
    fn = "process_frame_segments"
    globals.log.info("Processing frames...")
    frame_segments = [s for s in segments if isinstance(s, FrameSegment)]
    n = len(frame_segments)
    globals.log.debug("{fn}(): num frames = {n}".format(fn=fn, n=n))
    progress = ProgressBar(max_value=n,
                           quiet=args.quiet or args.debug or n == 0)
    progress.update(0)
    for i, f in enumerate(frame_segments):
        try:
            globals.log.debug(
                "{fn}(): frame (before) = {b}".format(fn=fn, b=f))
            # Frame segments that use a frame from the previous segment.
            if (f.input_file == "^"):
                if (f.segment_number > 0):
                    prev = segments[f.segment_number - 1]
                    globals.log.debug(
                        "{fn}(): prev = {p}".format(fn=fn, p=prev))
                    prev.generate_temp_file(args.output, width=width,
                                            height=height)
                    f.use_frame(
                        prev.generate_frame(f.frame_number, args.output,
                                            width=width, height=height))
                else:
                    globals.log.error(
                        "frame segment {s} is attempting to use the last "
                        "frame of a non-existent previous "
                        "segment".format(s=f.segment_number))
                    sys.exit(1)
            # Frame segments whose frame comes from a PDF file.
            else:
                _, suffix = os.path.splitext(f.input_file)
                if (suffix.lower() == ".pdf"):
                    f.use_frame(f.generate_temp_file(args.output, width=width,
                                            height=height))
                else:
                    globals.log.error(
                        'unexpected input file type "{s}" for frame segment '
                        "{f}".format(s=suffix, f=f.segment_number))
                    sys.exit(1)
            progress.update(i)
            globals.log.debug("{fn}(): frame (after) = ""{a}".format(fn=fn, a=f))
        except SegmentException as e:
            progress.finish()
            globals.log.error(e.message)
            sys.exit(1)
    else:
        progress.finish()


def render_podcast(args, audio_segments, video_segments, output, duration):
    """Stitch together the various input components into the final podcast."""
    fn = "render_podcast"
    globals.log.info("Rendering final podcast...")
    command = FFmpegConcatCommand(has_audio=len(audio_segments) > 0,
                                  has_video=len(video_segments) > 0,
                                  max_progress=duration,
                                  quiet=args.quiet and not args.debug,
                                  process_audio=args.process_audio,
                                  process_video=args.process_video,
                                  audio_codec=args.audio_codec,
                                  video_codec=args.video_codec)
    input_files = Segment.input_files()
    for f in input_files:
        if (input_files[f]):
            command.append_input_options(input_files[f])
        command.append_input_options(["-i", f])
    for s in (audio_segments + video_segments):
        command.append_filter(s.trim_filter())
    command.append_concat_filter("a", [s for s in audio_segments])
    if (args.normalise):
        command.append_normalisation_filter()
    command.append_concat_filter("v", [s for s in video_segments])
    if args.preview:
        globals.log.info("PREVIEW MODE: {fps} fps".format(fps=args.preview))
        command.append_output_options(["-r", args.preview])
    command.append_output_options([output])
    globals.log.debug("{fn}(): {c}".format(fn=fn, c=command))
    if (command.run() != 0):
        globals.log.error("Failed to render final podcast")


def cleanup(segments):
    """Clean up generated temporary files."""
    globals.log.info("Cleaning up...")
    for s in segments:
        s.delete_temp_files()


def main():
    fn = "main"
    logging.basicConfig(
        level=logging.INFO,
        format="%(levelname)s: {p}: %(message)s".format(p=globals.PROGRAM))
    segments = None
    
    try:
        args = parse_command_line()
        check_arguments(args)
    
        config = get_configuration(args)
    
        segments = process_input_streams(args, config)
        globals.log.debug("{fn}(): audio segments = {a}".format(
            fn=fn, a=[s for s in segments if isinstance(s, AudioSegment)]))
        globals.log.debug("{fn}(): video segments = {v}".format(
            fn=fn, v=[s for s in segments if isinstance(s, VideoSegment)]))
    
        audio_segments = [s for s in segments if isinstance(s, AudioSegment)]
        video_segments = [s for s in segments if isinstance(s, VideoSegment)]
    
        audio_duration = sum([s.get_duration() for s in audio_segments])
        video_duration = sum([s.get_duration() for s in video_segments])
        globals.log.debug("{fn}(): audio duration = "
                          "{a}".format(fn=fn, a=audio_duration))
        globals.log.debug("{fn}(): video duration = "
                          "{v}".format(fn=fn, v=video_duration))
    
        if (len(audio_segments) and len(video_segments)):
            if (audio_duration != video_duration):
                globals.log.warning("total video duration ({v}s) doesn't match "
                            "total audio duration "
                            "({a}s)".format(v=video_duration, a=audio_duration))
        
        width, height = smallest_video_dimensions(args, video_segments)
        globals.log.debug("{fn}(): width = {w}, height = "
                          "{h}".format(fn=fn, w=width, h=height))
        
        process_frame_segments(args, segments, width, height)
    
        globals.log.debug("{fn}(): input files = "
                          "{i}".format(fn=fn, i=Segment.input_files()))
    
        render_podcast(args, audio_segments, video_segments, args.output,
                       max(audio_duration, video_duration))

    except (KeyboardInterrupt):
        pass
    finally:
        if segments and not args.keep:
            cleanup(segments)


if (__name__ == "__main__"):
    main()

            #!/usr/bin/env python

import argparse
import datetime
import logging
import os.path
import sys

from pyparsing import ParseResults

import globals
from config_parser import (
    parse_configuration_file, parse_configuration_string)
from progress_bar import (ProgressBar)
from segment import (Segment, AudioSegment, VideoSegment,
                     FrameSegment, SegmentException)
from shell_command import (FFprobeCommand, FFmpegConcatCommand)


class InputStreamAction(argparse.Action):
    def __call__(self, parser, namespace, values, option_string=None):
        input = values.split(":")
        file = input[0]
        stream = None if (len(input) == 1) else input[1]
        setattr(namespace, self.dest, file)
        if (option_string in ["--audio", "-a"]):
            setattr(namespace, 'audio_stream_number', stream)
        elif (option_string in ["--video", "-v"]):
            setattr(namespace, 'video_stream_number', stream)


def parse_command_line():
    """Parse command line arguments."""
    parser = argparse.ArgumentParser(
        usage="%(prog)s [options] <output>",
        description="where: <output> is the name of the output file "
            "(note: .mov seems generally best)",
        epilog="Default input files can be specified using either of "
            "--audio or --video. If neither of these are specified, "
            "then you must supply a configuration file using --config. "
            "(Of course, you can always supply a configuration file "
            "regardless.)\n\n"
            "Input streams can be taken from the same input file.\n\n"
            "If no segments are specified, the entire input stream is "
            "processed as one segment. The number and duration of segments "
            "can differ, but the total duration across all input streams "
            "should ideally be the same.")
    
    parser.set_defaults(audio_stream_number=None, video_stream_number=None)

    parser.add_argument(
        "output",
        help="name of the output file (note: .mov is best)")
    
    parser.add_argument(
        "--audio", "-a", metavar="FILE[:STREAM]", action=InputStreamAction,
        help="File name for the default audio input stream (can be the "
            "same as for other input streams). You can optionally specify "
            "the default audio stream number to use if the file contains "
            "more than one (this can be overidden in a configuration "
            "file). If you don't specify a stream number, it defaults "
            "to 0 (i.e., the first audio stream in the file).")
    
    parser.add_argument(
        "--video", "-v", metavar="FILE[:STREAM]", action=InputStreamAction,
        help="File name for the default video input stream (can be the "
            "same as for other input streams). You can optionally specify "
            "the default video stream number to use if the file contains "
            "more than one (this can be overidden in a configuration "
            "file). If you don't specify a stream number, it defaults "
            "to 0 (i.e., the first video stream in the file).")
    
    parser.add_argument(
        "--configuration", "--config", "-c", dest="config", metavar="FILE",
        help="File name for the podcast segment configuration (plain text). "
            "See config_help.md details on the file "
            "format.".format(p=globals.PROGRAM))
    
    parser.add_argument(
        "--input-prefix", "-i", dest="prefix", metavar="PATH", default=".",
        help="Path to be prefixed to all INPUT files. This includes the "
            "configuration file, if applicable, and any files specified "
            "within the configuration file. Input files that already "
            "include the input prefix will not have it added again.")
    
    parser.add_argument(
        "--preview", "-p", metavar="RATE", nargs="?", const="1",
        help="Generate a preview of the podcast by only rendering a "
            "subset of the video frames. RATE is the number of frames "
            "per second to render (default 1 fps). You can specify "
            "fractions (e.g., 1/10 for one frame every 10 seconds). "
            'NOTE: if you get a "too few arguments" error when using '
            "this option, you've probably not provided an fps value and "
            "placed the option just before the output filename. "
            "Either move --preview earlier in the option list, add "
            'a "--" between it and the output filename, or provide a '
            "fps value.")
    
    parser.add_argument(
        "--debug", "-d", action="store_true",
        help="Print debugging output (overrides --quiet).")
    
    parser.add_argument(
        "--keep", "-k", action="store_true",
        help="Don't delete any generated temporary files.")
    
    parser.add_argument(
        "--quiet", "-q", action="store_true",
        help="Mute all console output (overridden by --debug).")

    args = parser.parse_args()
    
    return args


def prefix_path(prefix, path):
    """Prepend prefix to path, unless path already starts with prefix"""
    assert(prefix is not None)
    if not path:
        return None
    elif path == "." or path.startswith(prefix):
        return path
    else:
        return os.path.join(prefix, path)


def check_arguments(args):
    """Sanity check the command line arguments."""
    fn = "check_arguments"
    # Prepend input files with --input-prefix where applicable.
    args.audio, args.video, args.config = map(
        prefix_path, [args.prefix] * 3, [args.audio, args.video, args.config])

    if args.quiet:
        globals.log.setLevel(logging.WARNING)
        
    # --debug overrides --quiet.
    if args.debug:
        globals.log.setLevel(logging.DEBUG)
        globals.log.debug("{fn}(): args = {a}".format(fn=fn, a=args))
    
    # Must specify at least one of --audio, --video, --config.
    if not any([args.audio, args.video, args.config]):
        globals.log.error("must specify at least one of --audio, --video, "
                          "or --config")
        sys.exit(1)
    
    if not os.path.exists(args.prefix):
        globals.log.error('input prefix "{p}" does not '
                          "exist".format(p=args.prefix))
        sys.exit(1)
    

def get_configuration(args):
    """Load podcast configuration."""
    # Fill in missing file names for default input streams.
    fn = "get_configuration"
    # These types can have default input files and streams.
    type_mapping = {
        "audio": {"file": args.audio, "stream": args.audio_stream_number},
        "video": {"file": args.video, "stream": args.video_stream_number}}
    globals.log.info("Processing configuration...")
    if (args.config):
        config = parse_configuration_file(args.config)
        
        # Check that applicable default input streams have been specified.
        for i, c in enumerate(config):
            type = c["type"]
            
            # Add prefix to filename, if applicable.
            if c["filename"] and (c["filename"] != "^"):
                config[i]["filename"] = prefix_path(args.prefix,
                                                    config[i]["filename"])
            
            if (type in type_mapping):
                default_file = type_mapping[type]["file"]
                default_stream = type_mapping[type]["stream"]
                error_string = ("attempting to use default {s} input file, "
                    "but --{s} hasn't been specified".format(s=type))
            else:
                default_file = None
                default_stream = 0
                error_string = ("attempting to use a default input file, "
                    "but the {s} type doesn't support this".format(s=type))

            # No filename in configuration.
            if (not c["filename"]):
                if (default_file):
                    config[i]["filename"] = default_file
                # No filename on command line either.
                else:
                    globals.log.error(error_string)
                    sys.exit(1)
            
            # No stream number in configuration. Note: 0 is a valid
            # stream number, so explicitly check for None.
            if (c["num"] is None):
                # Assume 0 if no stream on command line either.
                if (default_stream is None):
                    config[i]["num"] = 0
                else:
                    config[i]["num"] = default_stream
    
    # No configuration file.
    else:
        conf_list = []
        for m in type_mapping:
            default_file = type_mapping[m]["file"]
            default_stream = type_mapping[m]["stream"]
            if (default_file and default_stream is not None):
                conf_list += ["[{type}:{file}:{stream}]".format(
                    type=m, file=default_file, stream=default_stream)]
        globals.log.debug("{fn}(): default config = "
                          "{c}".format(fn=fn, c=conf_list))
        config = parse_configuration_string("\n".join(conf_list))
    
    return config


def get_file_duration(file):
    """Calculate the duration a media file as a timedelta object."""
    fn = "get_file_duration"
    command = FFprobeCommand([file])
    globals.log.debug("{fn}(): {cmd}".format(fn=fn, cmd=command))
    # Only consider the first stream. If it's the only stream in the
    # file, great; otherwise it seems reasonable to assume that all
    # streams in the same file will have the same duration.
    ss, ms = command.get_entries(
        section="format", find_list=["duration"])[0].split(".")
    ms = ms[:3].ljust(3, "0")
    globals.log.debug("{fn}(): ss = {ss}, ms = {ms}".format(fn=fn, ss=ss,
                                                            ms=ms))
    return datetime.timedelta(seconds=int(ss), milliseconds=int(ms))


def make_new_segment(type, filename, punch_in, punch_out, num):
    """Make a new segment instance of the correct class."""
    fn = "make_new_segment"
    globals.log.debug("{fn}(): type = {t}".format(fn=fn, t=type))
    globals.log.debug("{fn}(): filename = {f}".format(fn=fn, f=filename))
    globals.log.debug("{fn}(): punch in = {i}".format(fn=fn, i=punch_in))
    globals.log.debug("{fn}(): punch out = {o}".format(fn=fn, o=punch_out))
    globals.log.debug("{fn}(): num = {n}".format(fn=fn, n=num))
    
    if (type == "audio"):
        return AudioSegment(file=filename, punch_in=punch_in,
                            punch_out=punch_out, input_stream=num)
    elif (type == "video"):
        return VideoSegment(file=filename, punch_in=punch_in,
                            punch_out=punch_out, input_stream=num)
    elif (type == "frame"):
        return FrameSegment(file=filename, punch_in=punch_in,
                            punch_out=punch_out, frame_number=num)
    else:
        return None


def process_timestamp_pair(args, times):
    """Constructs timedelta instances from a pair of config timestamps."""
    fn = "process_timestamp_pair"
    globals.log.debug("{fn}(): times[0] = {t}".format(fn=fn, t=times[0]))
    globals.log.debug("{fn}(): times[1] = {t}".format(fn=fn, t=times[1]))
    
    # If the first item in the timestamp list in the configuration file
    # is a filename, the parser inserts a zero timestamp before it. We
    # can therefore guarantee that the first item of the pair will
    # always be a timestamp.
    t0 = datetime.timedelta(
        hours=times[0]["hh"], minutes=times[0]["mm"],
        seconds=times[0]["ss"], milliseconds=times[0]["ms"])
    if (times[1]):
        if (len(times[1]) == 1): # filename
            t1 = t0 + get_file_duration(
                prefix_path(args.prefix, times[1]["filename"]))
        elif (len(times[1]) == 4): # normal timestamp
            t1 = datetime.timedelta(
                hours=times[1]["hh"], minutes=times[1]["mm"],
                seconds=times[1]["ss"], milliseconds=times[1]["ms"])
        else:
            globals.log.error("{fn}():unreadable timestamp {t}".format(
                fn=fn, t=times[1]))
            t1 = None
    else:
        t1 = None
    
    globals.log.debug("{fn}(): t0 = {t}".format(fn=fn, t=t0))
    globals.log.debug("{fn}(): t1 = {t}".format(fn=fn, t=t1))
    return t0, t1


def process_time_list(args, type, filename, num, time_list):
    """Process an audio or video stream and build a list of segments."""
    fn = "process_time_list"
    if (os.path.exists(filename) and type in ["audio", "video"]):
        stream_duration = get_file_duration(filename)
    else:
        stream_duration = 0
    segments = []
    globals.log.debug("{fn}(): stream duration = {d}".format(
        fn=fn, d=stream_duration))
    
    # No timestamps: punch in at 0, out at stream duration.
    if (len(time_list) == 0):
        punch_in = datetime.timedelta()
        punch_out = stream_duration
        segments.append(make_new_segment(type, filename, punch_in,
                                         punch_out, num))
    else:
        # Process each pair of timestamps as punch in, out. If there's
        # an odd number of items, the last one is processed separately.
        for t in zip(time_list[::2], time_list[1::2]):
            punch_in, punch_out = process_timestamp_pair(args, t)
            if (punch_in == punch_out):
                globals.log.warning(
                    "punch in ({i}s) and punch out ({o}s) times are "
                    "equal; no segment will be "
                    "generated".format(i=punch_in.total_seconds(),
                                       o=punch_out.total_seconds()))
                continue
            elif (punch_out < punch_in):
                globals.log.error(
                    "punch out time ({i}s) falls before punch in time "
                    "({o}s); can't generate a valid "
                    "segment".format(i=punch_in.total_seconds(),
                                     o=punch_out.total_seconds()))
                sys.exit(1)
            segments.append(make_new_segment(type, filename, punch_in,
                                             punch_out, num))
    
    # Odd number of timestamps: punch in at last timestamp,
    # out at stream duration.
    if (len(time_list) % 2 != 0):
        punch_in, _ = process_timestamp_pair(args, [time_list[-1], None])
        punch_out = stream_duration - punch_in
        segments.append(make_new_segment(type, filename, punch_in,
                                         punch_out, num))
    return segments


def process_input_streams(args, config):
    """Process a list of stream specifications and build a list of segments."""
    fn = "process_input_streams"
    globals.log.info("Processing input streams...")
    segments = []
    for cnf in config:
        globals.log.debug("{fn}(): type = {t}".format(fn=fn, t=cnf["type"]))
        globals.log.debug(
            "{fn}(): filename = {f}".format(fn=fn, f=cnf["filename"]))
        globals.log.debug("{fn}(): num = {n}".format(fn=fn, n=cnf["num"]))
        globals.log.debug("{fn}(): times = {t}".format(fn=fn, t=cnf["times"]))
    
        segments += process_time_list(args, cnf["type"], cnf["filename"],
                                      cnf["num"], cnf["times"])  
    return segments


def smallest_video_dimensions(args, segments):
    """Compute the smallest frame dimensions across all video inputs."""
    fn = "smallest_video_dimensions"
    width = 2048
    height = 1536
    for s in segments:
        if isinstance(s, FrameSegment):
            continue
        command = FFprobeCommand([s.input_file])
        globals.log.debug("{fn}(): {cmd}".format(fn=fn, cmd=command))
        w, h = command.get_entries(find_list=["width", "height"])
        if (w * h) < (width * height):
            width, height = w, h
    return width, height


def process_frame_segments(args, segments, width, height):
    """Post-process frame segments to set frame images, etc."""
    fn = "process_frame_segments"
    globals.log.info("Processing frames...")
    frame_segments = [s for s in segments if isinstance(s, FrameSegment)]
    n = len(frame_segments)
    globals.log.debug("{fn}(): num frames = {n}".format(fn=fn, n=n))
    progress = ProgressBar(max_value=n,
                           quiet=args.quiet or args.debug or n == 0)
    progress.update(0)
    for i, f in enumerate(frame_segments):
        try:
            globals.log.debug(
                "{fn}(): frame (before) = {b}".format(fn=fn, b=f))
            # Frame segments that use a frame from the previous segment.
            if (f.input_file == "^"):
                if (f.segment_number > 0):
                    prev = segments[f.segment_number - 1]
                    globals.log.debug(
                        "{fn}(): prev = {p}".format(fn=fn, p=prev))
                    prev.generate_temp_file(args.output, width=width,
                                            height=height)
                    f.use_frame(
                        prev.generate_frame(f.frame_number, args.output,
                                            width=width, height=height))
                else:
                    globals.log.error(
                        "frame segment {s} is attempting to use the last "
                        "frame of a non-existent previous "
                        "segment".format(s=f.segment_number))
                    sys.exit(1)
            # Frame segments whose frame comes from a PDF file.
            else:
                _, suffix = os.path.splitext(f.input_file)
                if (suffix.lower() == ".pdf"):
                    f.use_frame(f.generate_temp_file(args.output, width=width,
                                            height=height))
                else:
                    globals.log.error(
                        'unexpected input file type "{s}" for frame segment '
                        "{f}".format(s=suffix, f=f.segment_number))
                    sys.exit(1)
            progress.update(i)
            globals.log.debug("{fn}(): frame (after) = ""{a}".format(fn=fn, a=f))
        except SegmentException as e:
            progress.finish()
            globals.log.error(e.message)
            sys.exit(1)
    else:
        progress.finish()


def render_podcast(args, audio_segments, video_segments, output, duration):
    """Stitch together the various input components into the final podcast."""
    fn = "render_podcast"
    globals.log.info("Rendering final podcast...")
    command = FFmpegConcatCommand(has_audio=len(audio_segments) > 0,
                                  has_video=len(video_segments) > 0,
                                  max_progress=duration,
                                  quiet=args.quiet and not args.debug)
    input_files = Segment.input_files()
    for f in input_files:
        if (input_files[f]):
            command.append_input_options(input_files[f])
        command.append_input_options(["-i", f])
    for s in (audio_segments + video_segments):
        command.append_filter(s.trim_filter())
    command.append_concat_filter("a", [s for s in audio_segments])
    command.append_normalisation_filter()
    command.append_concat_filter("v", [s for s in video_segments])
    if args.preview:
        globals.log.info("PREVIEW MODE: {fps} fps".format(fps=args.preview))
        command.append_output_options(["-r", args.preview])
    command.append_output_options([output])
    globals.log.debug("{fn}(): {c}".format(fn=fn, c=command))
    if (command.run() != 0):
        globals.log.error("Failed to render final podcast")


def cleanup(segments):
    """Clean up generated temporary files."""
    globals.log.info("Cleaning up...")
    for s in segments:
        s.delete_temp_files()


def main():
    fn = "main"
    logging.basicConfig(
        level=logging.INFO,
        format="%(levelname)s: {p}: %(message)s".format(p=globals.PROGRAM))
    segments = None
    
    try:
        args = parse_command_line()
        check_arguments(args)
    
        config = get_configuration(args)
    
        segments = process_input_streams(args, config)
        globals.log.debug("{fn}(): audio segments = {a}".format(
            fn=fn, a=[s for s in segments if isinstance(s, AudioSegment)]))
        globals.log.debug("{fn}(): video segments = {v}".format(
            fn=fn, v=[s for s in segments if isinstance(s, VideoSegment)]))
    
        audio_segments = [s for s in segments if isinstance(s, AudioSegment)]
        video_segments = [s for s in segments if isinstance(s, VideoSegment)]
    
        audio_duration = sum([s.get_duration() for s in audio_segments])
        video_duration = sum([s.get_duration() for s in video_segments])
        globals.log.debug("{fn}(): audio duration = "
                          "{a}".format(fn=fn, a=audio_duration))
        globals.log.debug("{fn}(): video duration = "
                          "{v}".format(fn=fn, v=video_duration))
    
        if (len(audio_segments) and len(video_segments)):
            if (audio_duration != video_duration):
                globals.log.warning("total video duration ({v}s) doesn't match "
                            "total audio duration "
                            "({a}s)".format(v=video_duration, a=audio_duration))
        
        width, height = smallest_video_dimensions(args, video_segments)
        globals.log.debug("{fn}(): width = {w}, height = "
                          "{h}".format(fn=fn, w=width, h=height))
        
        process_frame_segments(args, segments, width, height)
    
        globals.log.debug("{fn}(): input files = "
                          "{i}".format(fn=fn, i=Segment.input_files()))
    
        render_podcast(args, audio_segments, video_segments, args.output,
                       max(audio_duration, video_duration))

    except (KeyboardInterrupt):
        pass
    finally:
        if segments and not args.keep:
            cleanup(segments)


if (__name__ == "__main__"):
    main()
             "See config_help.md details on the file "
             "format.".format(p=globals.PROGRAM))
     parser.add_argument(
+        "--copy-audio", dest="process_audio", action="store_false",
+        default=True,
+        help="Disable additional processing of the source audio. The audio "
+            "will still be re-encoded using the specified audio codec "
+            "because the concatenation filter requires it, but extra "
+            "processing such as reduction of channels or normalisation "
+            "will not be carried out. (Implies --no-normalise.)")
+    parser.add_argument(
+        "--copy-video", dest="process_video", action="store_false",
+        default=True,
+        help="Disable additional processing of the source video. The video "
+            "will still be re-encoded using the specified video codec "
+            "because the concatenation filter requires it, but extra "
+            "processing such as remapping of colours will not be "
+            "carried out.")
+    parser.add_argument(
+        "--no-normalise", dest="normalise", action="store_false",
+        default=True,
+        help="Disable normalisation of the source audio level (implied by "
+            "--copy-audio).")
+    parser.add_argument(
+        "--audio-codec", dest="audio_codec", metavar="CODEC",
+        default="pcm_s16le",
+        help="Specify ffmpeg audio codec for output (default pcm_s16le). "
+            "See the output of ffmpeg -codecs for possible codecs.")
+    parser.add_argument(
+        "--video-codec", dest="video_codec", metavar="CODEC", default="h264",
+        help="Specify ffmpeg video codec for output (default h264). "
+            "See the output of ffmpeg -codecs for possible codecs.")
+    parser.add_argument(
         "--input-prefix", "-i", dest="prefix", metavar="PATH", default=".",
         help="Path to be prefixed to all INPUT files. This includes the "
             "configuration file, if applicable, and any files specified "
             "within the configuration file. Input files that already "
         prefix_path, [args.prefix] * 3, [args.audio, args.video, args.config])
     if args.quiet:
         globals.log.setLevel(logging.WARNING)
+    # --copy-audio implies --no-normalise.
+    if not args.process_audio:
+        args.normalise = False
     # --debug overrides --quiet.
     if args.debug:
         globals.log.setLevel(logging.DEBUG)
     # Odd number of timestamps: punch in at last timestamp,
     # out at stream duration.
     if (len(time_list) % 2 != 0):
+        globals.log.debug("{fn}(): odd number of timestamps".format(fn=fn))
         punch_in, _ = process_timestamp_pair(args, [time_list[-1], None])
+        punch_out = stream_duration - punch_in
+        punch_out = stream_duration
         segments.append(make_new_segment(type, filename, punch_in,
                                          punch_out, num))
     return segments
     globals.log.info("Rendering final podcast...")
     command = FFmpegConcatCommand(has_audio=len(audio_segments) > 0,
                                   has_video=len(video_segments) > 0,
                                   max_progress=duration,
+                                  quiet=args.quiet and not args.debug)
+                                  quiet=args.quiet and not args.debug,
+                                  process_audio=args.process_audio,
+                                  process_video=args.process_video,
+                                  audio_codec=args.audio_codec,
+                                  video_codec=args.video_codec)
     input_files = Segment.input_files()
     for f in input_files:
         if (input_files[f]):
             command.append_input_options(input_files[f])
         command.append_input_options(["-i", f])
     for s in (audio_segments + video_segments):
         command.append_filter(s.trim_filter())
     command.append_concat_filter("a", [s for s in audio_segments])
+    command.append_normalisation_filter()
+    if (args.normalise):
+        command.append_normalisation_filter()
     command.append_concat_filter("v", [s for s in video_segments])
     if args.preview:
         globals.log.info("PREVIEW MODE: {fps} fps".format(fps=args.preview))
         command.append_output_options(["-r", args.preview])

                Ignore Space
               Show notes
              View
            
          
          30 ■■■■■
          shell_command.py
     """An ffmpeg shell command with a complex concat filter."""
    _expect_patterns = [r"time=(\d\d):(\d\d):(\d\d\.\d\d)"]
    
    def __init__(self, input_options=[], output_options=[], quiet=False,
                 max_progress=100, has_audio=False, has_video=False):
                 max_progress=100, has_audio=False, has_video=False,
                 process_audio=True, process_video=True,
                 audio_codec="pcm_s16le", video_codec="h264"):
        super(FFmpegConcatCommand, self).__init__(
            input_options, output_options)
        self.progress = ProgressBar(max_value=max_progress, quiet=quiet)
        self.has_video = has_video
        self.process_video = process_video
        self.video_codec = video_codec
        if (self.has_video):
            self.prepend_output_options(["-codec:v", "h264",
                                         "-pix_fmt", "yuv420p",
                                         "-map", "[vconc]"])
            self.prepend_output_options(["-map", "[vconc]"])
            if (self.process_video):
                self.prepend_output_options(["-pix_fmt", "yuv420p"])
            self.prepend_output_options(["-codec:v", self.video_codec])
        self.has_audio = has_audio
        self.process_audio = process_audio
        self.audio_codec = audio_codec
        if (self.has_audio):
            self.prepend_output_options(["-codec:a", "pcm_s16le",
                                         "-ac", "1",
                                         "-map", "[anorm]"])
            if (self.process_audio):
                self.prepend_output_options(["-ac", "1",
                                             "-map", "[anorm]"])
            else:
                self.prepend_output_options(["-map", "[aconc]"])
            self.prepend_output_options(["-codec:a", self.audio_codec])
        self.filters = []
    
    def append_filter(self, filter):
        """Append a filter to the filters list."""

            #!/usr/bin/env python

import datetime
import distutils.spawn
import json
import tempfile
import os.path
import re

import pexpect

from progress_bar import (ProgressBar)


class ShellCommand(object):
    """A shell command.
    
    _executable contains the full path to the relevant executable.
    
    _base_options is a list of standard general options for this
    command.
    """
    _executable = ""
    _base_options = []
    _expect_patterns = []
    
    @staticmethod
    def shellquote(s):
        """Quote a string so it can be safely pasted into the shell."""
        # Note: pipes/shlex.quote() only wraps '' around things,
        # it doesn't do things like \( \), which we also need.
        # We double-quote everything because that makes it easier
        # to deal with single quotes.
        if s:
            regexes = [
                # grouping parentheses for convert
                re.compile(r"^(\(|\))$"),
                # double quote (in middle of string)
                re.compile(r'^(.+".*)$'),
                # quotes around entire string
                re.compile(r"""^((?P<quote>['"]).*(?P=quote))$"""),
                # command line switch (starts with one or two "-")
                re.compile(r"^(--?.+)$"),
                # command path (starts with "/")
                re.compile(r"^(/[\w/]+)$")]
            substitutions = [r"\\\1", r"'\1'", r"\1", r"\1", r"\1"]
            num_matches = 0
            for subst in zip(regexes, substitutions):
                (s, n) = subst[0].subn(subst[1], s)
                num_matches += n
            if num_matches == 0:
                # catch-all for non-word characters when none of the other
                # substitutions apply
                s = re.compile(r"^(.*\W+.*)$").sub(r'"\1"', s)
        return s
    
    def __init__(self, input_options=[], output_options=[], quiet=False):
        self.input_options = input_options
        self.output_options = output_options
        self.progress = None
        self.process = None
    
    def __repr__(self):
        return "<{cls}: {cmd}>".format(
            cls=self.__class__.__name__,
            cmd=self.command_string(quote=True))

    def append_input_options(self, items=[]):
        """Add a list of items to the end of the input options."""
        self.input_options += items
    
    def prepend_input_options(self, items=[]):
        """Add a list of items at the front of the input options."""
        self.input_options = items + self.input_options
    
    def append_output_options(self, items=[]):
        """Add a list of items to the end of the output options."""
        self.output_options += items
    
    def prepend_output_options(self, items=[]):
        """Add a list of items at the front of the output options."""
        self.output_options = items + self.output_options
    
    def executable_string(self, quote=False):
        """Return the executable as a string."""
        if quote:
            return ShellCommand.shellquote(self._executable)
        else:
            return self._executable
    
    def argument_string(self, quote=False):
        """Return the list of arguments as a string."""
        args = self._base_options + self.input_options + self.output_options
        if quote:
            return " ".join([ShellCommand.shellquote(a) for a in args])
        else:
            return " ".join(args)
    
    def argument_list(self):
        """Return a combined list of all arguments."""
        return self._base_options + self.input_options + self.output_options
    
    def command_string(self, quote=False):
        """Return the entire command as a string."""
        return "{exe} {arg}".format(exe=self.executable_string(quote),
                                    arg=self.argument_string(quote))
    
    def process_pattern(self, pat):
        """Respond to a pexpect pattern. Return True on EOF."""
        return (pat == 0)
    
    def run(self):
        """Execute the command in a subprocess."""
        self.process = pexpect.spawn(self.executable_string(),
                                     self.argument_list())
        # EOF is *always* the first pattern.
        patterns = self.process.compile_pattern_list(
            [pexpect.EOF] + self._expect_patterns)
        try:
            while True:
                i = self.process.expect_list(patterns, timeout=None)
                if self.process_pattern(i):
                    break
        except (KeyboardInterrupt):
            pass
        finally:
            if self.progress:
                self.progress.finish()
            self.process.close()
            return self.process.exitstatus
    
    def get_output(self):
        """Execute the command in a subprocess and return the output."""
        return pexpect.run(self.command_string(quote=True))


class ConvertCommand(ShellCommand):
    """An ImageMagick convert command."""
    _executable = distutils.spawn.find_executable("convert")
    _base_options = ["-density", "600",
                     "xc:dimgrey", "null:", # dark grey background
                     "("]
    
    def __init__(self, input_options=[], output_options=[], width=2048, height=1536):
        super(ConvertCommand, self).__init__(input_options, output_options)
        self._base_options = (["-size", "{w}x{h}".format(w=width, h=height)] +
                              self._base_options)
        self.append_input_options(
            ["-resize", "{w}x{h}".format(w=width, h=height),
             "-background", "white", 
             "-alpha", "remove",
             "-type", "truecolor", # force RGB (this and next line)
             "-define", "colorspace:auto-grayscale=off"])
        self.prepend_output_options([")",
                                     "-gravity", "center",
                                     "-layers", "composite",
                                     "-flatten"])
    

class FFprobeCommand(ShellCommand):
    """An ffprobe shell command."""
    _executable = distutils.spawn.find_executable("ffprobe")
    _base_options = ["-loglevel", "error",
                     "-show_entries", "format:stream",
                     "-print_format", "json"]
    
    def __init__(self, input_options=[], output_options=[]):
        super(FFprobeCommand, self).__init__(input_options, output_options)
        self.entries = None
        # The input file should be the last input option.
        assert(os.path.exists(self.input_options[-1]))
        self.last_modified = os.path.getmtime(self.input_options[-1])
    
    def get_entries(self, section="stream", find_list=[]):
        """Fetch specified attributes from the input file."""
        # Re-fetch if the file's changed since we last looked.
        modified = os.path.getmtime(self.input_options[-1])
        if (not self.entries) or (modified > self.last_modified):
            js = json.loads(self.get_output())
            self.entries = {"format": js["format"], "stream": js["streams"][0]}
        return [self.entries[section][f] for f in find_list]

    
class FFmpegCommand(ShellCommand):
    """A "simple" ffmpeg shell command."""
    _executable = distutils.spawn.find_executable("ffmpeg")
    _base_options = ["-y", "-nostdin"]
        

class FFmpegConcatCommand(FFmpegCommand):
    """An ffmpeg shell command with a complex concat filter."""
    _expect_patterns = [r"time=(\d\d):(\d\d):(\d\d\.\d\d)"]
    
    def __init__(self, input_options=[], output_options=[], quiet=False,
                 max_progress=100, has_audio=False, has_video=False,
                 process_audio=True, process_video=True,
                 audio_codec="pcm_s16le", video_codec="h264"):
        super(FFmpegConcatCommand, self).__init__(
            input_options, output_options)
        self.progress = ProgressBar(max_value=max_progress, quiet=quiet)
        self.has_video = has_video
        self.process_video = process_video
        self.video_codec = video_codec
        if (self.has_video):
            self.prepend_output_options(["-map", "[vconc]"])
            if (self.process_video):
                self.prepend_output_options(["-pix_fmt", "yuv420p"])
            self.prepend_output_options(["-codec:v", self.video_codec])
        self.has_audio = has_audio
        self.process_audio = process_audio
        self.audio_codec = audio_codec
        if (self.has_audio):
            if (self.process_audio):
                self.prepend_output_options(["-ac", "1",
                                             "-map", "[anorm]"])
            else:
                self.prepend_output_options(["-map", "[aconc]"])
            self.prepend_output_options(["-codec:a", self.audio_codec])
        self.filters = []
    
    def append_filter(self, filter):
        """Append a filter to the filters list."""
        if (filter):
            self.filters.append(filter)
        
    def append_normalisation_filter(self):
        """Append a normalisation audio filter to the complex filter."""
        if (self.has_audio):
            self.append_filter("[aconc] dynaudnorm=r=0.25:f=10:b=y [anorm]")
    
    def append_concat_filter(self, type, segments=[]):
        """Append a concat filter to the filters list"""
        if (len(segments) > 1):
            self.append_filter(
                "{inspecs} concat=n={n}:v={v}:a={a} [{t}conc]".format(
                    inspecs=" ".join([s.output_stream_specifier()
                                      for s in segments]),
                    n=len(segments), v=int(type == "v"),
                    a=int(type == "a"), t=type))
        elif (len(segments) == 1):
            self.append_filter(
                "{inspec} {a}null [{t}conc]".format(
                    inspec=segments[0].output_stream_specifier(),
                    a=type if type == "a" else "", t=type))
        
    def build_complex_filter(self):
        """Build the complete complex filter.
        
        Filters in the filtergraph are separated by ";".
        """
        return "{f}".format(f=";".join(self.filters))
    
    def argument_string(self, quote=False):
        """Return the list of arguments as a string."""
        args = (self._base_options + self.input_options +
                ["-filter_complex", self.build_complex_filter()] +
                self.output_options)
        if quote:
            return " ".join([ShellCommand.shellquote(a) for a in args])
        else:
            return " ".join(args)
    
    def argument_list(self):
        """Return a combined list of all arguments."""
        return (self._base_options + self.input_options +
                ["-filter_complex", self.build_complex_filter()] +
                self.output_options)
    
    def process_pattern(self, pat):
        """Respond to a pexpect pattern. Return True on EOF."""
        if (pat == 1):
            elapsed = datetime.timedelta(
                hours=int(self.process.match.group(1)),
                minutes=int(self.process.match.group(2)),
                seconds=float(self.process.match.group(3)))
            self.progress.update(elapsed.total_seconds())
        return (pat == 0)
    

if (__name__ == "__main__"):
    print ShellCommand()
    print ConvertCommand(input_options=["in.pdf[12]"],
                         output_options=["out.png"])
    
    # FFprobeCommand expects the input file to exist.
    f = tempfile.NamedTemporaryFile()
    print FFprobeCommand(input_options=["-i", f.name])
    f.close()
        
    print FFmpegCommand(input_options=["-i", "in.mov"],
                        output_options=["out.mov"])
    concat = FFmpegConcatCommand(input_options=["-i", "in.mov"],
                                 output_options=["out.mov"], has_audio=True)
    concat.append_normalisation_filter()
    print concat
    
    print "Quoting:"
    for s in ["foobar", "foo bar baz", "foo(bar)", "[foobar]", "foo[bar",
              "foo 'bar'", '"foobar"', "'foobar'", 'foo"bar', "foo.bar",
              "(", ")", "'", "/foo/bar/baz/quux", "-f", "--foobar"]:
        print "  {s} => echo {sub}".format(s=s, sub=ShellCommand.shellquote(s))

            #!/usr/bin/env python

import datetime
import distutils.spawn
import json
import tempfile
import os.path
import re

import pexpect

from progress_bar import (ProgressBar)


class ShellCommand(object):
    """A shell command.
    
    _executable contains the full path to the relevant executable.
    
    _base_options is a list of standard general options for this
    command.
    """
    _executable = ""
    _base_options = []
    _expect_patterns = []
    
    @staticmethod
    def shellquote(s):
        """Quote a string so it can be safely pasted into the shell."""
        # Note: pipes/shlex.quote() only wraps '' around things,
        # it doesn't do things like \( \), which we also need.
        # We double-quote everything because that makes it easier
        # to deal with single quotes.
        if s:
            regexes = [
                # grouping parentheses for convert
                re.compile(r"^(\(|\))$"),
                # double quote (in middle of string)
                re.compile(r'^(.+".*)$'),
                # quotes around entire string
                re.compile(r"""^((?P<quote>['"]).*(?P=quote))$"""),
                # command line switch (starts with one or two "-")
                re.compile(r"^(--?.+)$"),
                # command path (starts with "/")
                re.compile(r"^(/[\w/]+)$")]
            substitutions = [r"\\\1", r"'\1'", r"\1", r"\1", r"\1"]
            num_matches = 0
            for subst in zip(regexes, substitutions):
                (s, n) = subst[0].subn(subst[1], s)
                num_matches += n
            if num_matches == 0:
                # catch-all for non-word characters when none of the other
                # substitutions apply
                s = re.compile(r"^(.*\W+.*)$").sub(r'"\1"', s)
        return s
    
    def __init__(self, input_options=[], output_options=[], quiet=False):
        self.input_options = input_options
        self.output_options = output_options
        self.progress = None
        self.process = None
    
    def __repr__(self):
        return "<{cls}: {cmd}>".format(
            cls=self.__class__.__name__,
            cmd=self.command_string(quote=True))

    def append_input_options(self, items=[]):
        """Add a list of items to the end of the input options."""
        self.input_options += items
    
    def prepend_input_options(self, items=[]):
        """Add a list of items at the front of the input options."""
        self.input_options = items + self.input_options
    
    def append_output_options(self, items=[]):
        """Add a list of items to the end of the output options."""
        self.output_options += items
    
    def prepend_output_options(self, items=[]):
        """Add a list of items at the front of the output options."""
        self.output_options = items + self.output_options
    
    def executable_string(self, quote=False):
        """Return the executable as a string."""
        if quote:
            return ShellCommand.shellquote(self._executable)
        else:
            return self._executable
    
    def argument_string(self, quote=False):
        """Return the list of arguments as a string."""
        args = self._base_options + self.input_options + self.output_options
        if quote:
            return " ".join([ShellCommand.shellquote(a) for a in args])
        else:
            return " ".join(args)
    
    def argument_list(self):
        """Return a combined list of all arguments."""
        return self._base_options + self.input_options + self.output_options
    
    def command_string(self, quote=False):
        """Return the entire command as a string."""
        return "{exe} {arg}".format(exe=self.executable_string(quote),
                                    arg=self.argument_string(quote))
    
    def process_pattern(self, pat):
        """Respond to a pexpect pattern. Return True on EOF."""
        return (pat == 0)
    
    def run(self):
        """Execute the command in a subprocess."""
        self.process = pexpect.spawn(self.executable_string(),
                                     self.argument_list())
        # EOF is *always* the first pattern.
        patterns = self.process.compile_pattern_list(
            [pexpect.EOF] + self._expect_patterns)
        try:
            while True:
                i = self.process.expect_list(patterns, timeout=None)
                if self.process_pattern(i):
                    break
        except (KeyboardInterrupt):
            pass
        finally:
            if self.progress:
                self.progress.finish()
            self.process.close()
            return self.process.exitstatus
    
    def get_output(self):
        """Execute the command in a subprocess and return the output."""
        return pexpect.run(self.command_string(quote=True))


class ConvertCommand(ShellCommand):
    """An ImageMagick convert command."""
    _executable = distutils.spawn.find_executable("convert")
    _base_options = ["-density", "600",
                     "xc:dimgrey", "null:", # dark grey background
                     "("]
    
    def __init__(self, input_options=[], output_options=[], width=2048, height=1536):
        super(ConvertCommand, self).__init__(input_options, output_options)
        self._base_options = (["-size", "{w}x{h}".format(w=width, h=height)] +
                              self._base_options)
        self.append_input_options(
            ["-resize", "{w}x{h}".format(w=width, h=height),
             "-background", "white", 
             "-alpha", "remove",
             "-type", "truecolor", # force RGB (this and next line)
             "-define", "colorspace:auto-grayscale=off"])
        self.prepend_output_options([")",
                                     "-gravity", "center",
                                     "-layers", "composite",
                                     "-flatten"])
    

class FFprobeCommand(ShellCommand):
    """An ffprobe shell command."""
    _executable = distutils.spawn.find_executable("ffprobe")
    _base_options = ["-loglevel", "error",
                     "-show_entries", "format:stream",
                     "-print_format", "json"]
    
    def __init__(self, input_options=[], output_options=[]):
        super(FFprobeCommand, self).__init__(input_options, output_options)
        self.entries = None
        # The input file should be the last input option.
        assert(os.path.exists(self.input_options[-1]))
        self.last_modified = os.path.getmtime(self.input_options[-1])
    
    def get_entries(self, section="stream", find_list=[]):
        """Fetch specified attributes from the input file."""
        # Re-fetch if the file's changed since we last looked.
        modified = os.path.getmtime(self.input_options[-1])
        if (not self.entries) or (modified > self.last_modified):
            js = json.loads(self.get_output())
            self.entries = {"format": js["format"], "stream": js["streams"][0]}
        return [self.entries[section][f] for f in find_list]

    
class FFmpegCommand(ShellCommand):
    """A "simple" ffmpeg shell command."""
    _executable = distutils.spawn.find_executable("ffmpeg")
    _base_options = ["-y", "-nostdin"]
        

class FFmpegConcatCommand(FFmpegCommand):
    """An ffmpeg shell command with a complex concat filter."""
    _expect_patterns = [r"time=(\d\d):(\d\d):(\d\d\.\d\d)"]
    
    def __init__(self, input_options=[], output_options=[], quiet=False,
                 max_progress=100, has_audio=False, has_video=False):
        super(FFmpegConcatCommand, self).__init__(
            input_options, output_options)
        self.progress = ProgressBar(max_value=max_progress, quiet=quiet)
        self.has_video = has_video
        if (self.has_video):
            self.prepend_output_options(["-codec:v", "h264",
                                         "-pix_fmt", "yuv420p",
                                         "-map", "[vconc]"])
        self.has_audio = has_audio
        if (self.has_audio):
            self.prepend_output_options(["-codec:a", "pcm_s16le",
                                         "-ac", "1",
                                         "-map", "[anorm]"])
        self.filters = []
    
    def append_filter(self, filter):
        """Append a filter to the filters list."""
        if (filter):
            self.filters.append(filter)
        
    def append_normalisation_filter(self):
        """Append a normalisation audio filter to the complex filter."""
        if (self.has_audio):
            self.append_filter("[aconc] dynaudnorm=r=0.25:f=10:b=y [anorm]")
    
    def append_concat_filter(self, type, segments=[]):
        """Append a concat filter to the filters list"""
        if (len(segments) > 1):
            self.append_filter(
                "{inspecs} concat=n={n}:v={v}:a={a} [{t}conc]".format(
                    inspecs=" ".join([s.output_stream_specifier()
                                      for s in segments]),
                    n=len(segments), v=int(type == "v"),
                    a=int(type == "a"), t=type))
        elif (len(segments) == 1):
            self.append_filter(
                "{inspec} {a}null [{t}conc]".format(
                    inspec=segments[0].output_stream_specifier(),
                    a=type if type == "a" else "", t=type))
        
    def build_complex_filter(self):
        """Build the complete complex filter.
        
        Filters in the filtergraph are separated by ";".
        """
        return "{f}".format(f=";".join(self.filters))
    
    def argument_string(self, quote=False):
        """Return the list of arguments as a string."""
        args = (self._base_options + self.input_options +
                ["-filter_complex", self.build_complex_filter()] +
                self.output_options)
        if quote:
            return " ".join([ShellCommand.shellquote(a) for a in args])
        else:
            return " ".join(args)
    
    def argument_list(self):
        """Return a combined list of all arguments."""
        return (self._base_options + self.input_options +
                ["-filter_complex", self.build_complex_filter()] +
                self.output_options)
    
    def process_pattern(self, pat):
        """Respond to a pexpect pattern. Return True on EOF."""
        if (pat == 1):
            elapsed = datetime.timedelta(
                hours=int(self.process.match.group(1)),
                minutes=int(self.process.match.group(2)),
                seconds=float(self.process.match.group(3)))
            self.progress.update(elapsed.total_seconds())
        return (pat == 0)
    

if (__name__ == "__main__"):
    print ShellCommand()
    print ConvertCommand(input_options=["in.pdf[12]"],
                         output_options=["out.png"])
    
    # FFprobeCommand expects the input file to exist.
    f = tempfile.NamedTemporaryFile()
    print FFprobeCommand(input_options=["-i", f.name])
    f.close()
        
    print FFmpegCommand(input_options=["-i", "in.mov"],
                        output_options=["out.mov"])
    concat = FFmpegConcatCommand(input_options=["-i", "in.mov"],
                                 output_options=["out.mov"], has_audio=True)
    concat.append_normalisation_filter()
    print concat
    
    print "Quoting:"
    for s in ["foobar", "foo bar baz", "foo(bar)", "[foobar]", "foo[bar",
              "foo 'bar'", '"foobar"', "'foobar'", 'foo"bar', "foo.bar",
              "(", ")", "'", "/foo/bar/baz/quux", "-f", "--foobar"]:
        print "  {s} => echo {sub}".format(s=s, sub=ShellCommand.shellquote(s))
     """An ffmpeg shell command with a complex concat filter."""
     _expect_patterns = [r"time=(\d\d):(\d\d):(\d\d\.\d\d)"]
     def __init__(self, input_options=[], output_options=[], quiet=False,
+                 max_progress=100, has_audio=False, has_video=False):
+                 max_progress=100, has_audio=False, has_video=False,
+                 process_audio=True, process_video=True,
+                 audio_codec="pcm_s16le", video_codec="h264"):
         super(FFmpegConcatCommand, self).__init__(
             input_options, output_options)
         self.progress = ProgressBar(max_value=max_progress, quiet=quiet)
         self.has_video = has_video
+        self.process_video = process_video
+        self.video_codec = video_codec
         if (self.has_video):
+            self.prepend_output_options(["-codec:v", "h264",
+                                         "-pix_fmt", "yuv420p",
+                                         "-map", "[vconc]"])
+            self.prepend_output_options(["-map", "[vconc]"])
+            if (self.process_video):
+                self.prepend_output_options(["-pix_fmt", "yuv420p"])
+            self.prepend_output_options(["-codec:v", self.video_codec])
         self.has_audio = has_audio
+        self.process_audio = process_audio
+        self.audio_codec = audio_codec
         if (self.has_audio):
+            self.prepend_output_options(["-codec:a", "pcm_s16le",
+                                         "-ac", "1",
+                                         "-map", "[anorm]"])
+            if (self.process_audio):
+                self.prepend_output_options(["-ac", "1",
+                                             "-map", "[anorm]"])
+            else:
+                self.prepend_output_options(["-map", "[aconc]"])
+            self.prepend_output_options(["-codec:a", self.audio_codec])
         self.filters = []
     def append_filter(self, filter):
         """Append a filter to the filters list."""

Show line notes below