MBARI Media Management - Demonstrate how to register a video in the `vampire-squid` video asset manager

register_video.py

import datetime
import json
import sys
import requests


host = "http://localhost:8084"
client_secret = "foo"


# Useful vampire-squid endpoints
vampire_squid = "{}/vam/v1".format(host)
auth_url = "{}/auth".format(vampire_squid)
media_url = "{}/media".format(vampire_squid)


def main(url, timeString):

    # -- Authenticate so that we can make a post request
    jwt_auth_header = {"Authorization": "APIKEY " + client_secret}
    jwt_response = requests.post(auth_url, headers=jwt_auth_header).json()
    access_token = jwt_response['access_token']

    # -- Use POST to create media
    # This is totally mock data. The only real data is the URL and
    # time stamp. In production the media_data would be parse from the
    # file system, file names and or file metadata
    auth_header = {"Authorization": "Bearer " + access_token}
    media_data = {"video_sequence_name": "Ventana 0952",
                  "camera_id": "Ventana",
                  "video_name": "Ventana 0952 {}".format(timeString),
                  "uri": url,
                  "start_timestamp": timeString,
                  "duration_millis": 123456}

    return requests.post(media_url, media_data, headers=auth_header).json()


if __name__ == "__main__":
    """
      Args:
         url -> A URL to a file. Should start with http but file URLS will work to
         time_string -> Start time of the video. Formatted as yyyy-MM-ddTHH:mm:ssZ 
                        Just use UTC times for now.

    """
    url = sys.argv[1]
    time_string = sys.argv[2]
    response_body = main(url, time_string)
    print(response_body)

This is a super simple mock up using some mock values. You can use it as:

python register_video.py <movie url> <time stamp as iso8601>

Example:

python register_video.py \
    http://m3.shore.mbari.org/videos/M3/mezzanine/DocRicketts/2018/07/1047/D1047_20180730T162551Z_h264.mp4 \  
    2018-07-30T16:25:51Z

A media_data can have the following params:

Required parameters

• video_sequence_name: String name that used to group videos from a single camera deployment
• camera_id: String name of that identifies the camera used to collect the video
• video_name: String name of the video segment from the deployment
• uri: The URL that points to your video. Other URI types are allowed.
• start_timestamp: UTC time formatted as yyyy-MM-ddTHH:mm:ss[.ssss]Z for example 2018-07-30T16:25:51Z or 2018-07-30T16:25:51.123Z

Optional parameters

• duration_millis: The length of the video in milliseconds
• container: Identifier for the media container. No real constraints but I usually use the mimetype of the video file here
• video_codec: String to ID the codec used to encode the video track. There's no standard so use what makes sense to you
• audio_codec: String to ID the coded used to encode the audio track.
• width: The width of the video in pixels (e.g. 1920)
• height: The height of the video in pixels (e.g. 1080)
• frame_rate: Decimal frame rate. Since it's a decimal this will only be an approximation of the actual framerate.
• size_bytes: Size of the video file in bytes
• video_description: Free form text to describe the video
• sha512: sha512 check sum encoded as a HEX string