Community Screencaps

Twitter Bot that posts screenshots from all 6 seasons of Community every 2 hours: https://twitter.com/Community_Caps

Running and Requirements

We developed this with Python 3.7. You should have the ffmpeg library installed on your system for processing frames.

You can install python package dependencies with the included requirements.txt file:

python -m pip install -r requirements.txt

Scripts

process.py will process videos pulled from the directory specified by --input into a set of frames and place them into the directory specified with the --output argument with a URL-safe filename. Additionally, it will create the file frames.yaml that specifies metadata about the frames created. Additionally, one must specify a URL that the images will be published and avaialble at. Run python process.py --help for more details.

Example usage:

python process.py --input=videos/ --output=images/ --frames-per-second=14 --remote_url=https://jacobysuh.gitlab.io/community/images

Details:

• Requires a directory that contains the videos to be processed passed as --input. This directory can have an directory sub-structure, but all videos must have the extension ".mkv".
• Requires a directory to place images into, specified with --output
• Produces frames at a specified interval using the ffmpeg command and places them in a directory called staging/
• Moves files from staging/ to the output directory with url-safe file names
• Produces frames.yaml, which is a YAML text files specifying the content of the output directory with metadata about each image. An example entry in the file is:

0001-Community-S01E01-Pilot.jpg: # the file name of the frame in the images/ directory
  season: 01
  episode: 01
  name: Pilot
  include: true
  url: https://jacobysuh.gitlab.io/community/images/0001-Community_S01E01_Pilot.jpg

One entry is included for each file in the ouput directory.

publish.py uses the tweepy Python package to publish individual frames to our Twitter page. It uses the contents of frames.yaml to select an image at random from images/. It then builds a tweet using the image selected and its associated metadata from frames.yaml and pushes it to Twitter.