Geoviz Application for Searching People And Rescuing - Prototype developed under CHOUCAS Projet)

Geovisualization interface prototype that exploits the Choucas Ontology Alert (Ontologie d'Alerte Choucas - http://purl.org/oac) and whose functionalities aim to support, facilitate and optimize the rescuer's reasoning.

The version whose code is available here is an evolution of the interface prototype described (alongside the ontology) in :
L’Ontologie d’Alerte Choucas : de la modélisation des connaissances à un outil support d’un raisonnement géovisuel - Application à la recherche de victime en haute-montagne. Matthieu Viry & Marlène Villanova-Oliver. SAGEO'19. Clermont-Ferrand, 14 nov. 2019..
Which was extended in :
Ontologie d’Alerte Choucas : de la modélisation des connaissances à un outil support d’un raisonnement géovisuel - Application à la recherche de victime en haute-montagne. Matthieu Viry & Marlène Villanova-Oliver. Geomatica, Association canadienne des sciences géomatiques, 2020, 73 (3), pp.87-103. doi: 10.1139/geomat-2020-0005.

Usage - Option 1, without Docker

In this setup, Ruitor, Neo4j DB and Gaspar are expected to be installed on the same environment.

Installation

Required dependencies: Python 3.7+, git, nodejs & npm.

• Copy this repository:

git clone https://github.com/ANRChoucas/GASPAR/ \
  && cd GASPAR

• Set up a Python virtual environment:

python3 -m virtualenv venv

• Activate the virtual environment:

source venv/bin/activate

• Install Python dependencies:

pip install -r requirements.txt
python -m spacy download fr_core_news_sm

• Install Node dependencies:

npm install

Starting up auxiliary services

Two external services are needed:

1. Neo4j using CHOUCAS.1 database, listening on port 7687 (available on Choucas FTP)
2. Ruitor listening on port 8000 (available on https://github.com/ANRChoucas/Ruitor/)

• Start Neo4j server if not done yet:

neo4j start

• Start Ruitor:

See https://github.com/ANRChoucas/Ruitor/ for more information.

Launching GASPAR

• Activate the virtual environment (if not done yet):

source venv/bin/activate

• Export Neo4j credentials (used by Python code to connect to Neo4j):

export NEO4J_USER="neo4j"
export NEO4J_PASSWORD="secret"

• Start the server application and watch for changes in JS code:

python server_app.py & npm run watch

This command starts the Python server application while watching for changes in Javascript code: the application is now accessible in a web-browser on http://localhost:8008/.

Usage - Option 2, with Docker

In this setup, Gaspar will be running in a Docker container and Neo4j DB and Ruitor will be running on the host, listening respectively on port 7687 and 8000.

Required dependencies: Git, Docker 18.03+ (Windows / MacOS) or 20.10+ (GNU/Linux) (these minimum versions are required because of the use of "host.docker.internal" networking feature).

Building the container

It may take some time because it has to fetch and install some dependencies but it's a one-time operation.

• Copy this repository:

git clone https://github.com/ANRChoucas/GASPAR/ \
  && cd GASPAR

• Actually build the container:

docker build -t "gaspar_choucas" .

Starting up auxiliary services

Be sure that Neo4j accepts non-local connections, by uncommenting the following line in neo4j.conf:

dbms.connectors.default_listen_address=0.0.0.0

Otherwise, it is up to you to launch Neo4j and Ruitor.

Launching GASPAR

docker run \
  --publish "8008:8008" \
  --add-host host.docker.internal:host-gateway \ # only on GNU/Linux
  --env NODE_ENV=developement \
  --env HOST_IP=host.docker.internal \
  --env NEO4J_USER=neo4j \ # replace by your neo4j username
  --env NEO4J_PASSWORD=secret-neo4j-password \ # replace by your neo4j password
  -it "gaspar_choucas:latest"

This command starts the Python server application inside Docker container: the application is now accessible in a web-browser on http://localhost:8008/.

Video demonstrations

• Guided tour of the interface (no sound) (october 2019)
• Commented use-case (comments in french) (october 2019)
• Use-case after integrating Neo4j DB and Ruitor (no sound) (mars 2021)
• Use-case with the new choices of graphical portrayal for the location zones (no sound) (mars 2022)

Previous versions

Other developments have been done on the visualisation prototype beforehand:

• the version from Matthew Sreeves internship is on branch matthews-jul2018
• the version from Jacques Gautier work is on branch jacquesg-jul2019

Other details

Technical overview

Server code uses aiohttp.

Client code notably uses webpack (to bundle the code), lumino (providing a flexible tab-based layout), openlayers, itowns and sweetalert.

Browser compatibility

Intended browser compatibility : last 2 (desktop) versions of Chrome and Firefox + last version of Opera and Safari.

Code style

Python source code follows PEP8 conventions (use flake8 to ensure it) and Javascript code style follows AirBNB guidelines (with some minor amendments) and is enforced with ESLint (see .eslintrc.js file for details).

Repository structure

GASPAR                   <--- GASPAR interface "root" folder
│
├── data
│   └── osm               <--- Some data extracted from OSM (GeoJSON format)
│
├── dist                  <--- Static assets after compilation
│
├── src                   <--- Source files for client code
│   ├── css
│   ├── img
│   ├── js
│   └── json              <--- JSONschemas derived from Choucas Ontology Alert
│
└── server_app.py         <--- Source file for server code

License

The code is provided under BSD 3-clause licence.

Geospatial features in use (in data/osm) are coming from OpenStreetMap (credits: OpenStreetMap Contributors), the licence is the ODbL.