next generation of http://api3.geo.admin.ch

Jenkins Build Status:

Checkout the source code:

git clone https://github.com/geoadmin/mf-chsdi3.git

or when you're using ssh key (see https://help.github.com/articles/generating-ssh-keys):

git clone git@github.com:geoadmin/mf-chsdi3.git

Add .pgpass to your environment

cd
touch .pgpass
chmod 600 .pgpass

Open .pgpass and Add

pgcluster.dev.bgdi.ch:5432:*:${username}:${pass}
pgcluster.int.bgdi.ch:5432:*:${username}:${pass}
pgcluster.prod.bgdi.ch:5432:*:${username}:${pass}
pg-sandbox.bgdi.ch:5432:*:${username}:${pass}

Make sure PGUSER and PGPASS is set in your .bashrc (for nosetests, potranslate and sphinx)

export PGUSER=${username} // postgres user (won't be relevant soon)
export PGPASS=${pass}

Add .boto to your environment

cd
touch .boto
chmod 600 .boto

Open .boto and Add (/etc/boto.cfg for main)

[Credentials]
aws_access_key_id = ${keyid}
aws_secret_access_key = ${accesskey}

Nagios Check for Dynamodb Dumps

Create a developer specific build configuration:

cp rc_example rc_user_<username>

Change the port number in the newly created buildout configuration file (In dev mode)

Where "username" is your specific rc configuration. To create the specific build:

make user

If you do this on mf1t, you need to make sure that a correct configuration exists under

/var/www/vhosts/mf-chsdi3/conf

that points to your working directory. If all is well, you can reach your pages at:

http://mf-chsdi3.dev.bgdi.ch/<username>/

Do the following commands inside your working directory. Here's how a standard deploy process is done.

make deploydev SNAPSHOT=true

This updates the source in /var/www... to the latest master branch from github, creates a snapshot and runs nosetests against the test db. The snapshot directory will be shown when the script is done. Note: you can omit the -s parameter if you don't want to create a snapshot e.g. for intermediate releases on dev main.

Once a snapshot has been created, you are able to deploy this snapshot to a desired target. For integration, do

make deployint SNAPSHOT=201512011411

This will run the full nose tests from inside the 201512011411 snapshot directory against the integration db cluster. Only if these tests are successfull, the snapshot is deployed to the integration cluster.

make deployprod SNAPSHOT=201512011411

This will do the corresponding thing for prod (tests will be run against prod backends) The same is valid for demo too:

make deploydemo SNAPSHOT=201512011411

You can disable the running of the nosetests against the target backends by adding notests parameter to the snapshot command. This is handy in an emergency (when deploying an old known-to-work snapshot) or when you have to re-deploy a snapshot that you know has passed the tests for the given backend. To disable the tests, use the following command:

make deployint SNAPSHOT=201512011411 NO_TESTS=notests

Use notests parameter with care, as it removes a level of tests.

Per default the deploy command uses the deploy configuration of the snapshot directory. If you want to use the deploy configuration of directory from which you are executing this command, you can use:

make deployint SNAPSHOT=201512011411 DEPLOYCONFIG=from_current_directory

Call the make deploybranch command in your working directory to deploy your current branch to test (Use make deploybranchint to also deploy it to integration). The code for deployment, however, does not come from your working directory, but does get cloned (first time) or pulled (if done once) directly from github. So you'll likely use this command after you push your branch to github.

The first time you use the command will take some time to execute.

The code of the deployed branch is in a specific directory /var/www/vhosts/mf-geoadmin3/private/branch on both test and integration. The command adds a branch specific configuration to /var/www/vhosts/mf-geoadmin3/conf. This way, the deployed branch behaves exactly the same as any user specific deploy. A deploy to a "demo" instance is possible too (simply use ./deploybranch.sh demo).

Sample path: http://mf-chsdi3.int.bgdi.ch/gjn_deploybranch/ (Don't forget the slash at the end)

To list all the deployed branch: make deletebranch

To delete a given branch: make deletebranch BRANCH_TO_DELETE=my_deployed_branch

Per default the back-link to geoadmin3 points to the main instance. If you want to change that, adapt the geoadminhost variable in the buildout_branch.cfg.in input file and commit it in your branch.

We are able to run our integration tests against different staging environments

For this to work, you need to adapt your personal ~/.pgpass file. It has to include access information for all clusters (add pgcluster0i and pgcluster0)

To run against prod environment: scripts/nose_run.sh -p

To run against int environment: scripts/nose_run.sh -i

To run against dev/test environment: scripts/nose_run.sh

To run against your private environment: make test

To execute all tests, including mapproxy and varnish ones, which are deactivated by default: scripts/nose_run.sh -a

Per default, printing is handled by one single war file on an instances. This war is installed and handled by the 'main' installation.

If you need to work on printing and use your own war, you have to add the following to your personal configuration (rc_user_xxx file):

export PRINT_WAR=ltxxx

In order to create the shared war use:

make printwar APACHE_BASE_PATH=main

We are currently using the FLAKES 8 convention for Python code. You can find more information about our code styling here:

http://www.python.org/dev/peps/pep-0008/
http://pep8.readthedocs.org/en/latest/index.html

You can find additional information about autopep8 here:

https://pypi.python.org/pypi/autopep8/

To check the code styling:

make lint

To autocorrect most linting mistakes

make autolint

Add a pre-commit hook

1. Create a pre-commit file

touch .git/hooks/pre-commit

2. Copy/paste the following script

#!/bin/bash

make lint
if [[ $? != 0 ]];
then
echo "$(tput setaf 1) Nothing has been commited because of styling issues, please fix it according to the comments above $(tput sgr0)"
exit 1
fi

3. Make this it executable

chmod +x .git/hooks/pre-commit

Now commits will be aborted if styling is not respected