Welcome to the Open City Tookit!
The Open City Toolkit (OCT) was developed in cooperation between the CityScienceLab of the HafenCity University Hamburg (HCU) and Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH (GIZ) in India and Ecuador. It is an open source tool and the software for this project is based entirely on open source components. The Open City Toolkit is a web-based geographic information system (GIS) for multi-touch tables that is optimised for the use by non-GIS-experts. It supports integrated and participatory urban planning processes, fostering dialogue between governments and citizens and exchange of knowledge and data between government departments. The main functionality of the Open City Toolkit is to visualise and analyse complex urban data, jointly among local practitioners and with citizens.
OCT.GIF-.15.seconds-.with.text.mp4
You can also look under the Wiki section to read the administrator's manual or the user manual.
The Open City Toolkit is a framework connecting several external tools in order to implement a flexible and easy-to-use web GIS solution. A Linux system equipped with several applications is required as a base system:
- GeoServer
- GRASS GIS
- enscript + ghostscript
- Node.js
The following instructions provide guidance for the installation of all required components.
You can quickly set up a running system via Docker.
To build and start the stack with Docker Compose, first open docker-compose.yaml in your root directory and change the volume directories to your local absolute paths:
...
volumes:
geoserver_data:
driver: local
driver_opts:
type: none
o: bind
device: ./geoserver_data_dir # change this to your local absolute path
grass:
driver: local
driver_opts:
type: none
o: bind
device: ./grass # change this to your local absolute path
output:
driver: local
driver_opts:
type: none
o: bind
device: ./output # change this to your local absolute paththen run this command in your root directory:
docker-compose up --build
For the webapp it may be required to change some environment variables. These are configured in the environment section of the webapp service in docker-compose.yml.
GEOSERVER_URL: The base URL of the local GeoServer instance. This should normally be the public IP or domain of the server, port 8080.INITIAL_LAT,INITIAL_LON: Initial center coordinates for the map view.USE_LANG: Language of the user interface (default: English). A matching translations file must exist inwebapp/i18n.
If Docker Compose is not available, you can also build the GeoServer and webapp containers separately:
docker build -t oct-geoserver geoserver
docker build -t oct-webapp webapp
Run the containers using the newly created images:
docker run -dti \
-v `pwd`/geoserver_data_dir:/usr/share/geoserver/data_dir \
-p 8080:8080 oct-geoserver
docker run -dti \
-e DATA_FROM_BROWSER_DIR=/oct/data_from_browser \
-e GRASS_DIR=/oct/grass \
-e OUTPUT_DIR=/oct/output \
-e GEOSERVER_URL=http://localhost:8080/ \
-v `pwd`/geoserver_data_dir:/usr/share/geoserver/data_dir \
-v `pwd`/grass:/oct/grass \
-v `pwd`/output:/oct/output \
-p 3000:3000 oct-webapp
The geoserver_data, grass and output directories are mounted into the container as volumes in order to make their contents persistent and accessible from the host system.
While the stack is running, the app is served at http://localhost:3000 and GeoServer is available at http://localhost:8080/geoserver/.
A Linux system is required. Any modern and up-to-date Linux environment can be used, however, these instructions are valid for a Debian Stable system.
It is recommended to install the app into a home directory of a dedicated user created for this purpose (e.g., oct_user). Here it is assumed that the dedicated user’s home directory is /home/oct_user.
Download the contents of this repository to /home/oct_user/oct.
GeoServer requires a Java Runtime Environment. Install the JRE:
apt-get install openjdk-11-jre-headless
Use a current stable version of GeoServer. Download a platform-independent binary of GeoServer (e.g., geoserver-2.17.2-bin.zip) from geoserver.org.
While you can install GeoServer in an arbitrary location, it is recommended to install it in /usr/share/geoserver. Create the directory:
mkdir /usr/share/geoserver
Move the downloaded GeoServer archive to the new directory and unzip it:
mv geoserver-2.17.2-bin.zip /usr/share/geoserver
cd /usr/share/geoserver
unzip ./geoserver-2.17.2-bin.zip
Because stylesheets are in CSS format, it is required to install a CSS extension for GeoServer. Download the extension that matches your GeoServer version, e.g.: https://build.geoserver.org/geoserver/2.17.x/ext-latest/geoserver-2.17-SNAPSHOT-css-plugin.zip
Move the downloaded file and unzip it:
mv geoserver-2.17-SNAPSHOT-css-plugin.zip /usr/share/geoserver/webapps/geoserver/WEB-INF/lib/
cd /usr/share/geoserver/webapps/geoserver/WEB-INF/lib/
unzip geoserver-2.17-SNAPSHOT-css-plugin.zip
After installing GeoServer, it is required to create a symbolic link pointing to /home/oct_user/oct/geoserver_data_dir, because GRASS GIS will export the results to that directory.
rm -r /usr/share/geoserver/data_dir
ln -s /home/oct_user/oct/geoserver_data_dir /usr/share/geoserver/data_dir
GeoServer is now ready to run. To start GeoServer, execute:
GEOSERVER_HOME=/usr/share/geoserver /usr/share/geoserver/bin/startup.sh
To stop it:
GEOSERVER_HOME=/usr/share/geoserver /usr/share/geoserver/bin/shutdown.sh
GRASS GIS is the core component of the backend: a highly developed, generic-purpose, cross-platform GIS system. It is required to install GRASS GIS version 7.1 or newer. Download GRASS GIS from: https://grass.osgeo.org/ or install it using a package manager:
apt-get install grass-core
Enscript is a command line tool used to create statistical output. Ghostscript is a package containing tools to manage PostScript files, including a PS to PDF converter. Install both packages:
apt-get install enscript ghostscript
The recommended way of installing Node.js in Linux is via NodeSource (follow the instructions for your distribution). On a Debian system:
curl -sL https://deb.nodesource.com/setup_14.x | bash -
apt-get install -y nodejs
Now change to the webapp directory and install the dependencies:
cd /home/oct_user/oct/webapp
npm install
Before starting the app, you need to set a few environment variables in the .env file:
DATA_FROM_BROWSER_DIR: directory to store temporary messages from the clientGRASS_DIR: directory to store map filesOUTPUT_DIR: directory to store human-readable analysis outputsGEOSERVER_URL: URL of the GeoServerINITIAL_LAT: latitude of map view center on startupINITIAL_LON: longitude of map view center on startup
Now to start the server, run:
node app.js
Open a browser at http://localhost:3000 – you should now see the user interface of the Open City Toolkit.
If you use the server in production, it is recommended to use a process manager, such as pm2. To install pm2, run:
npm install -g pm2
To start a process for OCT:
pm2 start /home/oct_user/oct/webapp/app.js
For better user experience, the OCT allows the user to add metadata to map layers.
Once added correctly, the metadata will be queried in various places within in the app, e.g. in the attribute table shown after the user clicks 'show attribute' in the query module.
The metadata should be added in a metadata.json file under the grass/metadata folder.
The metadata should conform to the following format:
[
{
"table": "<map layer name - find out the layer names under the grass/global folder>",
"description": "<description of the layer>",
"columns": [
{
"column": "<column name - find out the column names by running db.columns command in GRASS>",
"description": "<description of the column>"
},
{
"column": "",
"description": ""
},
]
},
{
"table": "",
"description": "",
"columns": [
{
"column": "",
"description": ""
}
]
}
]