Table of Contents
Background Waterpoints Prequisites Linux Windows Installation Installation using a virtual machine Usage Deployment to Heroku Contribute
Taarifa is an open source platform for the crowd sourced reporting and triaging of infrastructure related issues. Think of it as a bug tracker for the real world which helps to engage citizens with their local government.
The Taarifa platform is built around the Taarifa API, a RESTful service offering that clients can interact with to create and triage 'bugreports' relating to public infrastructure (e.g., the public toilet is broken).
For information on how to get inovoled, scroll to the Contributing section at the bottom of the page.
This repository contains an example application around Waterpoint Management built on top of the core API. It contains scripts to import Waterpoint data (resources) which then become available through the API to file requests against (e.g., waterpoint 2345 is broken and needs fixing).
There is also an angularjs web application that illustrates how a user can interact with the API and data through a browser.
You may choose to install into a virtual machine as described further down.
pip >= 1.5 is required. if you have an older version you can upgrade (may
pip install -U pip
If you are running Ubuntu you can simply execute the install.sh script
Some of the commands are executed with sudo and require sudo permission for the currently logged in user.
On other distributions, use the package manager to install the packages corresponding to those listed in install.sh.
Windows can be used as the platform to run Taarifa - the main caveat here however is that as you install the dependencies, these are not added to the $PATH variable - this needs to be added manually.
Details for the following steps are the same as for MacOS/Linux (except for the actual package installation):
Install and run MongoDB, which does not automagically come with a service, so it needs to be started manually. Open a command prompt and run:
"c:\Program Files\MongoDB 2.6 Standard\bin\mongod.exe" --dbpath c:\mongo_databases\taarifa\
The following steps are part of the bootstrap.sh script, so you may choose to execute that instead.
Requires Python, pip and the Taarifa API to be installed and MongoDB to be running.
pip install virtualenv virtualenvwrapper
For windows only, install virtualenvwrapper-win using pip:
pip install virtualenvwrapper-win
Set up virtualenvwrapper according to your shell and create a virtualenv:
If you already created the virtualenv for the Taarifa API, activate it:
Clone the repository
git clone https://github.com/taarifa/TaarifaWaterpoints
Change into directory and install the requirements
cd TaarifaWaterpoints pip install -r requirements/dev.txt
Ensure you have node.js and npm installed. Then, from the
TaarifaWaterpoints directory, install the npm dependencies:
Install the Grunt and Bower command line interface You may need to have administrator (root) rights to run some of the commands below. On Windows this typically means running the command prompt window (CMD) as Administrator..:
npm install -g grunt-cli npm install -g bower
Finally, install the frontend dependencies using Bower:
Continue with the usage section.
Instead of following the installation instructions above you may choose to set up a virtual machine with all dependencies installed. This process is fully automated using Vagrant and the provided Vagrantfile. Note that the Vagrantfile is included in the repository and needs not be downloaded.
Clone the repositories into the same root folder. This is required since these local folders are mounted in the VM such that you can edit files either on the host or in the VM.
git clone https://github.com/taarifa/TaarifaAPI git clone https://github.com/taarifa/TaarifaWaterpoints cd TaarifaWaterpoints
Start the VM. This may take quite a while the very first time as the VM image needs to be downloaded (~360MB) and the VM provisioned with all dependencies. On every subsequent use these steps are skipped.
In case provisioning fails due to e.g. loss of network connection, run the provisioning scripts again until successful:
Connect to the virtual machine and change into the TaarifaWaterpoints folder:
vagrant ssh cd TaarifaWaterpoints
You can then continue with the usage section below. The ports are automatically forwarded so you can access the API and frontend from your host browser. Note that both the TaarifaAPI and the TaarifaWaterpoints folders in the VM are mounted from the host i.e. changes made on the host are immediately reflected in the VM and vice versa. This allows you to work on the code either on the host or in the VM according to your preference.
When using a virtual machine, run the following commands in the VM.
Make sure the virtualenv is active:
From the TaarifaWaterpoints directory run the following commands to create the waterpoint schemas:
python manage.py create_facility python manage.py create_service
Then upload the waterpoint data:
python manage.py upload_waterpoints <path/to/waterpoints/file.csv>
Start the application from the TaarifaWaterpoints directory by running:
python manage.py runserver -r -d
By default the API server is only accessible from the local machine. If access from the outside is required (e.g. when running from inside a VM), run:
python manage.py runserver -h 0.0.0.0 -r -d
-d cause the server to run in debug mode and reload
automatically when files are changed.
To verify things are working, open a browser (on the host when using the VM) and navigate to:
This should show a list of all the waterpoint resources currently in the database.
To work on the frontend web application start the grunt server (with the API server running on port 5000) using:
grunt serve --watch
Then navigate to (on the host when using the VM):
Grunt watches the app folder for changes and automatically reloads the frontend in your browser as soon as you make changes.
To build the frontend (which is automatically done on deployment), use:
This creates a distribution in the dist folder, which is served via the Flask development server running on port 5000. The build step needs to be run again whenever the frontend in the app folder changes. Running grunt serve is not required in this case.
Taarifa uses pip to install and manage python dependencies.
Conventionally this uses requirements.txt, but Heroku automatically installs from there. Therefore a requirements folder is used as following:
- Dev and deploy requirements in requirements/base.txt
- Development only in requirements/dev.txt
- Deployment only in requirements/deploy.txt
heroku app:create <name>
This will add a new Git remote heroku, which is used to deploy the app. Run git remote -v to check. To add the remote manually, do:
git remote add heroku firstname.lastname@example.org:<name>.git
Since Taarifa uses Python for the API and Node.js to build the frontend, Heroku build packs for both stacks are required. heroku-buildpack-multi enables the use of multiple build packs, configured via the .buildpacks file. Before deploying for the first time, the app needs to be configured to use it:
heroku config:set BUILDPACK_URL=https://github.com/ddollar/heroku-buildpack-multi.git
Add the MongoLab Sandbox to provide the MongoDB database
heroku addons:add mongolab
To be able to import the data into the MongoLab database, copy down the heroku configuration to a .env file you can use with foreman:
Make sure the virtualenv is active:
Create the waterpoint schemas and upload the waterpoint data, which may take several hours:
foreman run python manage.py create_facility foreman run python manage.py create_service foreman run python manage.py upload_waterpoints <path/to/waterpoints/file.csv>
Alternatively, you can import a dump of your local database and import it. If mongod is not running, create a dump directly from the database files in a dump folder in your current directory:
sudo -u mongodb mongodump --journal --db TaarifaAPI --dbpath /var/lib/mongodb
This assumes you have followed the MongoDB installation instructions on Ubuntu. Otherwise you might not need to run the command as the mongodb user and your database directory might be /data/db.
Import the dump into your MongoLab database, running the following command:
mongorestore -h <host> -d <database> -u <user> -p <password> /path/to/dump/TaarifaAPI/
Extract host, database, user and password from the MONGOLAB_URI Heroku configuration variable:
heroku config:get MONGOLAB_URI
Once finished you are ready to deploy:
git push heroku master
To set up a custom domain for the deployed app, register with heroku:
heroku domains:add <domain>
and add a DNS record for it:
<domain>. 10800 IN CNAME <appname>.herokuapp.com.
There is still much left do do and Taarifa is currently undergoing rapid development. We aspire to be a very friendly and welcoming community to all skill levels.
If you are comfortable you can also take a look at the github issues and comment/fix to you heart's content.
We use the github pull request model for all contributions. Refer to the contributing guidelines for further details.