Pystrukts!

Some classical algorithms and data structures in pure Python.

Table of Contents

• Pystrukts
• Running Tests with Sonar

Pystrukts

Pystrukts is a collection of classical algorithms and data structures implementations written in Python. It's main objective is to contain easy to understand algorithms so that anyone can use its implementations for studying.

Running Tests with Sonar

This project supports running sonarqube analysis by creating docker local containers via docker-compose commands which execute the instructions on the docker-compose.yml file.

• sonar: brings up the sonarqube server which is used for the code and code coverage analysis
• sonar scanner: container which contains the sonar-scanner cli
• tests: container which runs the projec tests

The first step is to bring up the sonar server:

docker-compose up sonar  # authenticate on localhost:9000 and create an auth token there

If it's the first time running the sonar, go to http://localhost:9000 and use the credentials admin:admin (user:password) to login. Also, a token will be generated for authenticating with this local sonar server on later steps such as: c11361e5ce0719a8be5249dcc329f31 (example). Write this token down for later steps.

After that, we should run the project tests or the sonar analysis will be incomplete. For that, run:

docker-compose up tests  # will generate, via volume mounts, coverage.xml and test-results.xml files

This will generate coverage.xml and test-results.xml files required by sonar. One very important thing is that the coverage.xml will contain the test paths using absolute paths like /app/algorithms. As a consequence, we must put all of our source code in an /app folder inside the sonar_scanner container. This is crucial for the sonar_scanner to be able to understand the coverage paths on the analysis. The following docker-compose.yml snippet shows this mechanism:

sonar_scanner:
  container_name: algorithms_sonar_scanner
  image: sonarsource/sonar-scanner-cli:latest
  environment:
    - SONAR_HOST_URL=http://sonar:9000
  working_dir: /app # this is the same /app folder in which the tests are run!
  volumes:
    - .:/app # coverage.xml results filepath must begin with '/app'

PS: if pytest runs with --cov=. option, the absolute path consideration with /app is not really required as sonar_scanner will just raise a WARN but will understand the coverage.xml in anyway. However, if the project runs pytest with the option --cov=algorithm (coverage of the src code folder only), which changes the coverage.xml file's absolute path, then if the /app convention is not followed, sonar_scanner will raise ERROR instead of warnings and will NOT perform any coverage analysis. In order to avoid such problems, we run pytest with --cov=. so that no errors are observed even if we locally run start-tests.sh to run tests.

As a final step, we run the sonar analysis by running the sonar_scanner service which will capture our source code via volumes and send to the sonar server. Now, we execute a docker-compose run command by passing the auth token that we noted in the previous steps, and that's it:

docker-compose run -e SONAR_LOGIN=c11361e5ce0719a8be5249dcc329f31 sonar_scanner  # change with your local auth token

Now, just navigate to http://localhost:9000 (sonar server page) and check out your code quality!