ml-cv-trader-server

ml-cv-trader - short for 'machine learning-computer vision-trader' - is a a demo/experiment that runs gesture recognition from your webcam's video stream in order to determine whether to buy or sell stocks while consuming a simulated market data feed over Solace PubSub+ and sending the end result back to Solace for processing on the server side.

Installing

This project uses an express server & typescript. Install it and the associated dependencies by running the following command:

npm install

Configuring the server with a Solace PubSub+ Broker

Spin up a Solace PubSub+ broker container or in the cloud.

Modify ./config/solace-config.ts with the appropriate username and credentials. If using Solace Cloud, you can retrieve the credentials from the Web Messaging section of the Connect Tab of a Solace cloud instance as shown below:

Developing

Once you've installed dependencies with npm install, start a development server using the following command:

npm serve

Building

To create a production version of this app:

npm run tsc
npm run start:prod

Running the server side component

A svelte js ui component is required for this demo - ml-cv-trader-ui. Visit the repo to setup and run.