Vitastic

Vitastic is a Javascript x Python based solution to quickly build web interface that serves object detection workloads. By leveraging Microsoft's themable React component library FluentUI powered by Python Flask backend, Vitastic allows you to build your own UI to demonstrate pre-trained image classification, object detection, or semantic segmentation models.

Prerequisites

In order to successfully set up your solution you will need to have provisioned the following:

• Python: A recent Python 3.8 interpreter to run the Flask backend on.
• Yarn: A package and project manager for Node.js applications.
• Object detection model: Deployed either in Azure Custom Vision or Azure ML as a RESTful API. See dstoolkit-objectdetection-accelerator to see how to build an end-to-end training and deployment pipelines in Azure ML.

Getting Started

Backend

The back-end of Vitastic is implemented using python's Flask framework. All source code reside in the Vitastic folder.

It is highly recommended create a virtual environment using e.g. package manager conda before the further package installing:

conda create -n Vitastic python=3.8
conda activate Vitastic

Now you can install all python package dependencies and get the backend running by:

cd Vitastic
pip install -r requirements.txt
flask run

You should be able to see the Flask app serving the app.py script under http://localhost:5000.

Once the backend app starts serving, the app.py script handles the incoming requests and forward the input images to detection jobs written in detection.py. To enable the detection, you need to specify the Azure ML REST endpoint and (optionally) the authentication key in the dotenv file. Complete the information required in the .env_example file and rename the file to .env.

Frontend

The front-end was bootstrapped with Create React App, and all source code can be found under the vitastic-ui directory.

To start the front-end, navigate to the front-end directory, and install react dependencies by running:

cd vitastic-ui
yarn install

After installing, run the react app in development mode by:

yarn start

You can open http://localhost:3000 to view it in your browser.

Frontend Customization

Vitastic also supports you to build a customized version of frontend design by enabling React component configurations. You can navigate to vitastic-ui/src/config.json and modify this configuration file to suit your working scenarios:

cd vitastic-ui/src
nano config.json

Below is a detailed description of each configuration possibilities:

• title: A customizable title appearing in the top-left corner.
• darkTheme: Whether to use fluent UI dark theme, by default set to false.
• defaultConfidence: Model confidence value, by default set to 0.5.
• modelService: Model backbone service, currently supports Azure machine learning aml and Azure custom vision cv.
• modelScope: Backbone model functionality, currently supports classification, object detection or semantic segmentation
• sampleImages: List of image names to be presented as samples, all images should reside in the vitastic-ui/public/img/samples directory. It is optional, and ideally should be of maximal 7 images to for the sake of design.

Repository Details

A high-level structure of the Vitastic repository is as follows:

|-- Vitastic
|   |-- app.py
|   |-- detection.py
|   |-- requirements.txt
|   |-- src
|   |   |-- aml.py
|   |   |-- cv.py
|   |   `-- util.py
|-- docs
`-- vitastic-ui
	|-- package.json
    |-- public
    |   |-- img
    |   `-- index.html
    |-- src
    |   |-- App.css
    |   |-- App.js
    |   |-- AppConfig.json
    |   |-- components
    |   |   |-- ConfigView.js
    |   |   |-- UploadView.js
    |   |   `-- ResultView.js
    |   |-- index.css
    |   `-- index.js
    `-- yarn.lock

The following architecture illustrates how the code is organised into an end to end workflow:

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.

Trademarks

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.