Trails Visualization

A web-app to display model predictions of trail use.

Build Status

Development

The project is split into two separate apps for front end and back end:

trails-viz/
  |- trails-viz-app
  |- trails-viz-api

trails-viz-app is the front end written in JavaScript using the Vue framework. trails-viz-api is the backend written in Python using Flask

Development requires nodejs >= 10.x.x and Python == 3.7.x

While opening the project in IDE, it's recommended to open the front end and backend as two separate projects. That way the IDE would be able to recognize the former as a JavaScript project and the later as a Python project.

Backend Setup

Install the dependencies globally or inside a virtual environment (recommended). Navigate to the root of trails-viz-api, setup flask environment variables and start the development server. One app specific environment variable DATA_FILES_ROOT is required to set the location where all the csv and shp files are:

$ cd trails-viz-api
$ pip install -r requirements.txt
$ export FLASK_APP=trailsvizapi
$ export FLASK_ENV=development
$ export DATA_FILES_ROOT=<path/to/trails-viz-data/>
$ flask run

This will start a localhost server on port 5000

Frontend Setup

Navigate to root of trails-viz-app, install dependencies and start development server:

$ cd trails-viz-app
$ npm install
$ npm run serve

This will start the server on port 8080. Navigate to http://localhost:8080.

Note: If you encounter CORS policy issues when viewing the dashboard on the development server, try different browsers. Firefox works, but Chrome and Safari may result in bugs.

Build

Front end and back end has must be built separately. Run the following commands form the root of the project - trails-viz

# build front end
$ cd trails-viz-app
$ npm run build
$ cd ..
# build back end
$ cd trails-viz-api
$ python setup.py bdist_wheel

This will create a dist folder in trails-viz-app with all the static html contents and trailsvizapi-x.x.x-py3-none-any.whl file in trails-viz-api/dist folder.

Now build the docker image from project root - trails-viz

$ docker build -t trails-viz:latest .

Running on Production

A conainer can be started from the docker image, which will run both the app and web inside it. To start the container, two arguments are required to run the app - the directory on the host which contains the csv and shape files must be mounted onto the container and environment variable DATA_FILES_ROOT must be set. Also host port must be mapped if not using default.

docker run -v <host-path-to-data>:/app/data-files \
-p 9000:80 \
-e DATA_FILES_ROOT=/app/data-files/ \
vivekkr12/trails-viz:latest

This will start the app on port 9000 on the host machine. If running behind a web server such as apache, the virtualhost can be configured to forward to the port. This may require proxies for both the gui and the api. On our production machine, we do this with:

ProxyPass /trails_site/ http://localhost:9000/
ProxyPassReverse /trails_site/ http://localhost:9000/
ProxyPass /api/ http://localhost:9000/api/
ProxyPassReverse /api/ http://localhost:9000/api/

Adding a New Project

The instructions for adding a new project to the dashboard are documented here.