Routines for processing MMIF files to create visual indexes ("visaids") and other useful output.
These routines require existing MMIF files containing annotations of videos. MMIF files with relevant annotations can be created with the CLAMS scenes-with-text detection app.
The use_swt module is a stand-alone (not presupposing any particular workflow or media source) application for creating visaids from an existing MMIF file and the corresponding media file.
The process_swt module includes functions for processing MMIF produced by the CLAMS swt-detection app.
The post_proc_item module includes functions called by run_job from clams-kitchen and calls functions in process_swt to perform postprocessing on MMIF produced by swt-detection.
To install the necessary dependencies, navigate to the project's root directory and run:
pip install .For guidance on usage of the stand-alone CLI, run visswt -h.
To see a list of the TimeFrame annotations in a MMIF file from SWT-detection, run:
visswt -d my_swt_output.mmifTo create a visaid using the sample MMIF file in this repository, download the corresponding media file to the sample_files directory. Then run:
visswt -d -v sample_files/cpb-aacip-4071f72dd46_swt_v72.mmif sample_files/cpb-aacip-4071f72dd46.mp4The easiest way to integrate visaid creation into another Python project is by importing proc_visaid directly from the visaid_builder package and calling it. For an example, see the visaid_builder/integration_example.py file.