Source for https://measurements.ooni.torproject.org/

To run this locally you need docker, docker-compose and make.

If those are installed you will then be able to run the following:

make serve

This will start the docker containers needed to run the application locally, building also all the required web assets.

Deployment to staging is managed automatically when something is merged into the master branch.

This means that master must always be buildable and working.

Rollover to production is currently handled manually by running the same deploy script used for staging deployments.

It requires docker-machine. Once that is installed to deploy to production you can do:

DEPLOY_HOST="1.1.1.1" scripts/deploy.sh production path/to/id_rsa

You can fill the database with some report files by creating a listing of report files and importing this listing from a text file with:

python -m measurements updatefiles -f listing.txt

An example of listing.txt can be found in dev/fixtures.txt

This application is designed to be run inside of 4 possible environments: development, testing, staging and production.

The meaning of these three environments is:

• development this is the environment used when you are actively developing the application. When in this mode we favour ease of development over similarity of the environment when it's in production or staging.

• testing: this is the environment used for running unit, functional and integration tests automatically on our CI platform. This environment is as close as possible to the real production or staging environment. When running under testing we only expose it locally.

• staging this is the environment used for triggering automatic redeployment of the application when merging into the master branch. When deploying on staging the application will be live on the staging endpoint and should functionally be equivalent to production.

• production this is the environment used when it's ready to be deployed . Production deployments are currently triggerred manually, but some day may be automatic based on tagging a certain git tag.

Configuration management is handled by overriding environment variables and other parameters inside of config/production.yml, config/staging.yml, config/testing.yml. Read the Environments section to learn the meaning of these.

The database for the measurement files needs to be kept up to date as such it's recommended that the following things are done:

On first run you ensure the database is initialized with the correct report files. You can do so in the following way:

python -m measurements updatefiles -t TARGET_DIR

Where TARGET_DIR is generally going to be the same path as REPORTS_DIRECTORY in the configuration file.

Then in the cronjob that runs periodically to sync reports with the measurements server you should generate a listing of all the new reports file (it's sufficient to put their filepaths inside of a text file) and then run:

python -m measurements updatefiles -f update-file.txt

updatefiles will check to see if the files are already in the database, but it's much more performant to not run the update with files that are already in the database.