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

• Added --preview option (closes #17).

• Now checks whether a filename already starts with the input prefix before adding it (closes #25).

• Deleted leftover development debugging output. ...

• Added code to resize frame inputs to the size of the smallest video segment where applicable (closes #23). Still need to add resizing for differently sized video inputs.

• Redesigned FFprobeCommand to retrieve media file entries in JSON format for easier extraction and processing. ...

• Changed default of --input-prefix to “.”.
• Added check that the input prefix actually exists.
• Ensured that the segments list variable actually exists when cleanup happens.

• Ensured that prefix isn’t added if input filename is “^”. ...

• Rearranged display of frame processing progress bar.
• Better handling of frame segment processing errors:
    – Added SegmentException.
    – Checked all subprocess exit status values.
• Moved cleanup() call into finally clause.
• Added Segment._temp_files_list to make deletion of temporary files more predictable, and eliminate inadvertent deletion of original source files for PDF frame inputs.
• Refactored segment temporary file generation (closes #22).

Merge branch 'master' of https://github.com/Otago-InfoSci-Database/Tools

• Ensured progress bar is silent when maximum value is zero.

• Added debug code to output number of frame segments.

• Rewrote get_configuration() to fix a crasher when processing frame inputs. Also refactored to eliminate duplicate code. ...

• Defaulted args.prefix to "".
• Renamed some variables.
• Added some documentation and debugging calls.

• Added --input-prefix option (closes #13). ...

• Fixed an undiscovered bug in process_timestamp_pair() that would have caused a crash when there was an odd number of timestamps (failed to check whether second timestamp is defined).

• Added progress bars for all long-running operations (closes #18). ...

• Added new ProgressBar class.
• Switched from subprocess to pexpect for running shell commands.
• Added general shell quoting for debugging output of commands.
• Rewrote the way command strings and argument lists are generated.
• Removed “-loglevel error” from ffmpeg commands to enable capturing progress information (pexpect doesn’t echo the output anyway).

• Updated print_progress(): ...

– Added optional width argument.
    – Moved newline character to the start of the line rather than the end to stop the cursor overprinting the start of the progress bar.

• Fixed parsing bug that produced incorrectly structured timestamp lists when using a duration file (introduced by restructuring of grammar).

• Added print_progress() function and implemented progress bar for frame processing (partial progress—hah!—towards #18).

• Ensured that audio/video output options and filters are only included when there actually is an audio/video stream (closes #19). ...

• Rearranged exception handling on main() so that a KeyboardInterrupt is caught no matter where you are in the process.

• Removed some leftovers from the deleted --config-help option.

• Removed the --config-help option as there appears to be no easy way to create additional standalone help-style options with argparse (it complains that there aren’t enough arguments because the positional argument is missing). The help text has been put into a separate Markdown file instead. (closes #12)

• Moved frame segment post-processing into a function. ...

• Ensured warnings about unequal stream durations are only output if there is both an audio and a video stream (closes #20).
• Fixed obsolete leftover occurrence of “options” in VideoSegment.get_last_frame_number().

• Normalised terminology around “streams” (closes #21).

• Only print warning about differing audio and video durations when there’s actually some audio (may need further work?).

• Ensured that audio_stream_number and video_stream_number always default to None.

• Finished support for generating frames directly from PDF slides (closes #8). ...

• Specified defaults for audio and video stream numbers.
• Fixed some typos on comments and debug output.
• Added "/" as a valid filename character for config files.
• Added some comments explaining the more obscure class variables in the Segment class.
• Generalised Segment.generate_last_frame() into generate_frame().
• Ensured that generate_temp_file() returns the generated filename.

• Enabled --config-help option (no content yet).

• Updated help messages for --audio and --video to reflect ability to specify default stream number.

• Added support for default stream numbers on the command line (in the form filename:0). Includes a custom argparse action handler to split the stream number from the filename. NOTE: not fully tested! (closes #11)

• Removed --frame option (closes #10).

• Moved segment classes into separate module (closes #7). ...

• Moved global configuration into separate module.
• Made format of debug logging more consistent.
• Added a few more info logging messages.
• Renamed configuration parser module.

• Moved shell command classes into separate module.

• Added -pix_fmt output video option to ensure that concatenated segments have the same pixel format (especially for motion JPEG segments generated from still images). ...

• Also added “file not found” exception handling to FrameSegment.delete_temp_files().
• Deleted obsolete read_segments_from_file().