README.md

This tool was developed to automate the process of downloading videos and transcripts, so as to facilitate the generation of large amounts of data for machine learning.

• pytube
  • pip install git+https://github.com/rmerzouki/pytube
• xmltodict
  • pip install xmltodict
• Python 3.6 and above

To use VidGrab within your own Python script, you would first import the library and call the get function.

import vidgrab
dict_out = vidgrab.get(url, video_save_path, caption_save_path, write_caption, debug, write_if_both)

Note that video_save_path and caption_save_path take in a file path to the saved video or caption file, e.g. "./videos/video1.mp4", "./videos/caption1.json".

The last 3 arguments are optional. Their meanings and default values are detailed below in the "Argument Descriptions and Flags" section.

To use this tool with the command line interface (CLI), you would use the following command:

python vidgrab.py -u <video url> -v <video save path> -c <caption save path>

Here is an example of using the tool with the CLI with optional flags:

python get.py -u <video url> -v <video save path> -c <caption save path> -w -d -b

This gets a video at <video url>, saves the video to <video save path> and saves relevant information to <caption save path> as a .json file if captions exist for the video. It also prints out useful debugging information that can help identify the sources of any errors.

The output will be a .json file containing the relevant information. If using the tool within Python script, the function also returns the relevant information as a Python dictionary.

The output follows the following format:

{
    'video_id': <string>,
    'url': <string>,
    'video_path': <string>,
    'chunks': <list>
    'intervals': <2D list>
    'transcript_type': <string: "human" / "asr_youtube"
}

The output in the failure cases will contain information about the error.

No English captions are available.

{
    'video_id': <string>,
    'url': <string>,
    'video_path': <string>,
    'chunks': None,
    'intervals': None,
    'transcript_type': None
}

Video is not found.

{
   "error": "video_not_found"
}

Unable to download the video.

{
   "error": "download_failed"
}