Redo SRS export, add command-line options

[benjaminhottell]

Hi team,

Thank you for your work on this project. I've made the following modifications. Please feel free to accept or reject as much as you'd like.

• TSV generation code was rewritten to use Python's built-in CSV module, meaning we do not need to use pandas/numpy
• User can explicitly disable screenshot exports (via --no-export-screenshot). Screenshots are only generated if there is a video stream, and the user has not disabled screenshots.
• User can explicitly disable audio clip exports (via --no-export-audio). Audio clips are only generated if there is an audio stream, and the user has not disabled audio clips.
• User can opt into experimental video exports (via --video-export). Video clips are only generated if there is a video stream, and the user has enabled video exports.
• User can specify a directory to save exported media files. I personally use this to write the media files directly to my Anki collection (this way, I can skip manually copying them over).
• .gitignore updated to include __pycache__
• pandas dependency dropped. numpy dependency also dropped

[aaron-meyers]

This is great - I was hitting issues with srs2cia pypi package (an error about "append" not valid on "DataFrame") which unsurprisingly didn't happen when using @benjaminhottell 's fork that doesn't use pandas anymore. Not sure if there's an incompatibility with a newer version of pandas - I could try running in a python virtualenv to isolate the issue but removing an unneeded pandas dependency seems like a better approach anyway.