Skip to content

Latest commit

 

History

History

double-agent-stacks

README.md

DoubleAgent Stacks

This project has reference implementations of testing a scraper stack using the [DoubleAgent][../double-agent] suite of tests.

Setup Test Suite:

  1. Modify external/userAgentConfig.json to include browser ids you wish to test (<browser.toLowercase()>-<major>-<minor ?? 0>).
  2. Run yarn 2 to create a file called userAgentsToTest.json at .data/external/2-user-agents-to-test. This file contains all the user agents that DoubleAgent should test. Grow or shrink this list based on how long you are willing to spend testing ;)
  3. Setup your DoubleAgent SSL Certificates and Host File entries (follow directions in the code here).
    • Why? DoubleAgent uses cross domain requests and https domains, so your local setup needs to be able to process these.

Testing your Scraper:

This project leverages yarn workspaces. To get started, first run yarn build from the root directory.

NOTE: You should start with the `Setup Test Suite` step.

You'll test your scraper from the "Stacks" directory. A server called "collect-controller" creates all the "assignments" that your scraper should run. Each assignment includes a UserAgent (OS + Browser) and a series of pages to click, navigate through.

To start a collect-controller:

  1. Run yarn start from the DoubleAgent repo.
  • The API will return assignments one at a time until all tests have been run. Include a scraper engine you're testing with a query string or header called "scraper".
  1. Run a "stack" to automatically click through all of the assignments. You can mimic the Puppeteer or Unblocked classes and corresponding Runner classes with your own scraper.

yarn 3 runs Unblocked Agent. yarn 3-puppeteer runs Puppeteer.

  • If you are operating in a different language, you can see the flow of assignments in the AssignmentRunner and AssignmentsClient classes.
  • Behind the scenes of the runner class, when you create a new Test suite, you tell the server where to point to your userAgentsToTest.json file that you generated during the Setup step above (eg, ./data/external/2-user-agents-to-test/userAgentsToTest.json).
  1. As tests run, you will see "profiles" of your raw data downloaded into ./data/external/3-assignments/.
  2. After all tests are run, you can run yarn 4 to generate probe-ids from all your profiles and compare them to real browser probe-ids. Under the ./data/external/4-assignment-results, any entries in the -signature files are signatures that differed from the probeId. The corresponding file contains the test details of the failed tests.
  • NOTE: currently, these results are optimized for display on version 2 of the Scraper Report website. If you would like to submit a pull request to display differences in a human readable format, it would be tremendously appreciated!!