ClassCapture_AVProcessor

The audio video processor for interacting with the ClassCapture back end API. This contains the process AVProcessor.py, which will run every night on the secondary VM.

Why Python

Since our audio/video processing algorithms were already being developed in Python, the most clear choice seemed to be for our processor to be written in Python as well.

Contents of this repository

• AVProcessor.py
  • This file is the main process responsible for performing periodic processing of videos in the ClassCapture backend API.
• Logs
  • This file contains timestamped log files, with the various actions performed by the AV processor.
  • If the process runs once per night, we should get 1 file per day.
• ProcessedVideos / UnprocessedVideos
  • Both of these directories will remain empty 90% of the time, as they are temporary holding locations for files in the middle of being processed.
  • It may be worth looking into just deleting and recreating these directories each time, but they don't hurt.
• config.cfg
  • This file is written using Python's ConfigParser to keep track of which files have been processed, and with which algorithm.

Command line arguments

• Required arguments:
  • CCuser: username to login to classcapture.
  • CCpass: password to login to classcapture.
  • CCurl: url for classcapture
  • sftpuser: username to login to the VM hosting the API server.
  • sftppass: password to login to the VM hosting the API server.
  • sftpurl: address of VM hosting API.
• Optional flags:
  • -i, --ignoreConfig: Ignores the config.cfg when checking which files need to be updates. Will update every file instead.
  • -n, --noUpdate: Will run the AVProcessor without updating the config.cfg.

A/V Algorithms

• These are the algorithms that will be run on the videos stored in the API server.

• Stored under Algs/ where each algorithm has it's own directory.

• To add a new algorithm:

  1. Create a subdirectory under Algs/ with the necessary files for your algorithm.
  2. Add an import to import from the Algs/<subdirectory>/, importing the necessary python function. The python function should take 2 parameters, the input file, and the resulting output file from the algorithm.
  3. Add an empty __init__.py file into the Algs/<subdirectory>/
  4. Add to the algs dictionary, mapping the algorithm name to the given 2 parameter function from above. For example, if I had an a file AlvinsAlg.py, which contained a function myStabilize(inname, outname). The resulting changes to AVProcessor.py would look like this:

   # Added this import line
   from Alg/<subdirectory>/ import AlvinsAlg
   ...
   algs = { 
   	...
   	# Added this line to the algs dict
   	"AlvinsAlgorithmThingy" : AlvinsAlg.myStabilize,
   	...
   	}
  

• Algorithms included:

  • TimsVstab: <Tim put your description of the algorithm here>
  • DirectCopy: Creates an exact copy of the video, more for testing purposes
  • DirectCopy2: Same as DirectCopy

How to deploy using CronTab

A machine's CronTab is just a simple text file that has a list of commands meant to run at a certain time, which is performed by the Cron Daemon. In order to schedule a VM to run this processor, you would do something like:

First edit the CronTab file.

crontab -e

Add a line like this to the file:

00 4 * * * cd ~/ClassCapture_AVProcessor/ && python AVProcessor.py --CCuser=<username here> --CCpass=<password here> --CCurl=<ClassCapture URL> --sftpuser=<login for API VM> --sftppass=<password for API> --sftpurl=<API Url>