This app implements a gRPC + flask server that takes detection handler requests (the result of video object detection) and streams them via HTTP to web clients.

• One of the answers to How to implement server push in Flask framework on stackoverflow.com
• HTTP server events - https://www.html5rocks.com/en/tutorials/eventsource/basics/
• Redis publish/subscribe - https://github.com/andymccurdy/redis-py/#publish--subscribe

When retrieving a message from the pubsub when many have been pushed, there was a redis error (subscribe scheduled to be closed ASAP for overcoming of output buffer limits) that caused the connection to be closed. This was resolved by using this configuration in redis.conf - client-output-buffer-limit pubsub 0 0 0

A conda environment is created to install required packages.

• Run conda env create -f env.yaml to setup a conda environment

• Run the basic startup script bash start.sh <redis-channel-name>

• set an environmental variable for the flask app settings

  export SETTINGS=settings.cfg

• if running on a network (not localhost), set host name where the react app is running. This will help to avoid CORS errors.

export STREAMING_NETWORK_HOST=192.168.100.55

• startup the server with a command such as the one below

  python app.py 50002 <redis-channel-name>

Individual tests can be run like this:

bash run_with_env.sh python -m pytest web_handler_test.py --disable-warnings --log-cli-level=DEBUG

Eric Njogu

MIT