/pull-request-microservice

Primary LanguageJavaScript

Pull Request Microservice

Gain insights on GitHub pull requests!

Libraries & Tools Used

  • Express
  • Nodemon: restart server with each code change
  • Dotenv: so that environment variables can be read
  • Axios: to make API calls to GitHub
  • Jest: testing
  • Supertest
  • Cors: allow for future security configuration (not configured)
  • Babel: allow for ES6 import/export syntax

Setup

Ensure that node and npm are installed and up to date. Also use a GitHub account to create a new personal access token to add to a .env file so that successful GitHub API requests can be made. No need to define a particular scope for the token.

  1. Run git clone git@github.com:Stearnzy/pull-request-microservice.git in a new directory
  2. cd into pull-request-microservice
  3. touch a .env file at the project's root and add PORT=<desiredPortNumber> and GITHUB_ACCESS_TOKEN=<yourGHAccessToken>
  4. Run npm install
  5. Run npm test -- --coverage to run tests with code coverage reporting
  6. Start server with npm start and start hitting endpoints! The endpoint is structured as follows:
http://localhost:5000/api/github/:organization/:repository/open-pull-requests

Decisions

  1. ES6 imports syntax for a more modern application.
  2. Chose a funcitonal path instead of a class-oriented path due to my previous experience working in Hapi.js, with our API structured in more of a functional way.
  3. Though not necessary to call the GitHub endpoints for the necessary data for this task, I decided to add Authorization headers to the GitHub service methods to prevent rate limiting.

Wins

  • This was my first time testing with Jest.
    • Learned about using mock functions and using them in place of functions nested within the current function to test (see the test for collectOpenPullRequestGeneralData utilizing jest.mock('../services/getOpenPullRequests.js'); to use the mock getOpenPullRequests function nested within the collectOpenPullRequestGeneralData function).
    • Was able to test solely on mocked data, including integration tests
    • Project has high code coverage percentage
  • Scalability: Utilized a recursive function to iterate through pages of pull requests (in the event that there were more than 100) to ensure all pull requests for a repository are returned.

Places to Improve

  1. Being my first time using Express since my brief interaction with it at Turing, my file naming and structure could be improved. I saw there are many different approaches to file naming conventions, so I went with a template that made sense to me.
  2. If this was a bigger project, I would have a more effective error logging/handling system in place. One example would be having a logger for each of the GitHub service methods so that I could provide a more detailed error message to the user, such as if the organization or repository name that was provided does not exist.
  3. Reflecting on this project, perhaps a class-oriented, OOP approach may have had benefits over the functional style I chose. Perhaps naming conventions such as for route and controllers could have been improved by this.
  4. A particular style guide could be added to this project.
  5. In terms of testing, I did not know how to test app.listen within the index.js file, so I would add that in the future once I figured out how. As in most cases, more edge case testing is always better, but I wanted to focus on demonstrating mocking with Jest and accounting for multiple-page responses from the GitHub API, while providing happy-path unit and integration tests.

Thank you so much for seeing my work!