The objective of this project is to develop a full stack web platform for the AI in Architecture studio, whereby the students can display both their ML models and results, and deploy them onto whichever cloud technology IAAC so chooses.
The frontend will take the form of a website, with a static landing page containing a description of the studio, student-provided “about” pages that will describe the work of each group, and interactive Mapbox maps that will visualise the student’s data layers. Furthermore, on the Mapbox interface, users will have access to a range of student-developed liv models (running on the developed backend server described below), where live ML models can be called upon to return inferences, also visualised on the map.
Built using React, and styled with Tailwind, the frontend will house the interactive Mapbox interface, as well as static pages for a description of the studio and of each group. Other common frontend web technology, such as HTML and CSS, will also be used.
The standalone backend server will act as a REST API to handle fetch requests from the aforementioned frontend. This backend server will host all data and models provided by the students, including static data layers in the form of GeoJSONs and Mapbox Style JSONs, as well as student-trained ML models, to respond to live inference requests.
Containerised using Docker and managed with Docker-compose, the Python-based Flask web server will use Gunicorn to serve the student’s static files (*.geojson and style.json files), ML models, and pre-/post-processing scripts (Python functions) over a REST API. An NGINX load balancer will be placed in front of the server. The backend is where the student groups will be uploading their data layers, Mapbox styles, and ML model pipelines, so a specification for how this will be used will also be developed and provided.
To start the backend server locally, you'll need Python3 (this was built and tested using Python3.8).
Then, clone the repo, create a new virtual environment, activate it, and install all the required packages. (Note, I'm using Linux and the Fish Shell, so some commands may look different).
python3.8 -m virtualenv venv
source venv/bin/activate.fish
pip install -r requirements.txtTo spin up the backend, set up the required environment variable and run Flask.
set -x FLASK_APP main
flask run