Skip to content

Latest commit

 

History

History
90 lines (64 loc) · 3.99 KB

file.md

File metadata and controls

90 lines (64 loc) · 3.99 KB

AIStreamer Ingestion - File/Data Streaming

Google Cloud Video Intelligence Streaming API allows customers to stream a video file to Google Streaming API server and receives analytics results in real time. It also allows customers to divide a video file into multiple data chunks, and stream each data chunk individually.

Supported video codecs

Not all video files can be supported in Google Cloud Video Intellgence Streaming API. Please refer to Google Cloud Documentation for list of supported video codecs during file streaming.

You can determine if your video file supports streaming by using a named pipe and the FFMPEG tool. The following sample commands decode a video into PNG images at a rate of 0.2 fps. If your file can be successfully decoded using these commands, then you can stream it using the Video Intelligence streaming API.

$ export FILE_NAME=/path_to_file/file_name
$ mkfifo /tmp/inpipe
$ ffmpeg -i /tmp/inpipe -r 0.2 output_%04d.png
$ cat $FILE_NAME > /tmp/inpipe

Python client library

Python examples are available to our customers. We provide one Python file for each feature.

Command line to run our example Python code:

$ export GOOGLE_APPLICATION_CREDENTIALS=/path_to_credential/credential_json
$ ./streaming_label_detection.py $FILE_NAME
$ ./streaming_shot_detectin.py $FILE_NAME
$ ./streaming_explicit_content_detection.py $FILE_NAME
$ ./streaming_object_tracking.py $FILE_NAME

Here, $GOOGLE_APPLICATION_CREDENTIALS specifies where GCP credential json file is located.

Make sure to set correct timeout value when initiating the gRPC connection. If you need to stream 1 hour of video, timeout value should be at least 3600 (unit: seconds).

# streaming_annotate_video returns a generator.
# timeout argument specifies the maximum allowable time duration between
# the time that the last packet is sent to Google video intelligence API
# and the time that an annotation result is returned from the API.
# timeout argument is represented in number of seconds.
responses = client.streaming_annotate_video(requests, timeout=3600)

C++ client library

C++ examples are available to our customers. A single binary example streaming_client_main is provided to support all features. Please follow the build instruction.

Command line to run our example C++ code:

$ export GOOGLE_APPLICATION_CREDENTIALS=/path_to_credential/credential_json
$ export CONFIG=/path_to_config/config_json
$ export FILE_NAME=/path_to_file/file_name
$ export TIMEOUT=3600
$ ./streaming_client_main --video_path=$FILE_NAME --config=$CONFIG --timeout=$TIMEOUT

Here, $GOOGLE_APPLICATION_CREDENTIALS specifies where GCP credential json file is located.

Example config files $CONFIG for each feature can be found here.

Make sure to set correct timeout flag in the command line. If you need to stream 1 hour of video, timeout value should be at least 3600 (unit: seconds).

Other languages (Java, NodeJS)

Support for Java and NodeJS is available in Google Cloud Documentation.

Build instruction

We provide a Docker example, and build/deployment dependency rules have been properly configured in env/Dockerfile.

Please refer to Docker documentation for detailed instructions.

Flow control

Google Cloud Video Intelligence Streaming API server has inherent flow control. A StreamingAnnotateVideoRequest message will be rejected, and gRPC streaming connection will be stopped immediately in the following two cases:

  1. when AIStreamer ingestion client is sending requests to Google servers too frequently
  2. when AIStreamer ingestion client is sending too much data to Google servers (beyond 20Mbytes per second).