Datahub Job Testing

This is a simple test runner for datahub jobs. It is designed to be used in a CI/CD pipeline to verify that jobs are working as expected.

Usage

This tool expects a manifest file to be present inside your datahub config project. The manifest file should contain a list of tests to run. For each test, the runner will spin up a new datahub, upload the defined required datasets, and run the job. It will then compare the output with the expected output.

With CLI

djt path/to/manifest.json

Import as a module

package tests

import (
    ...
    djt "github.com/mimiro-io/datahub-job-testing"
)

func TestMyJob(t *testing.T) {
    ...
    manifest := "path/to/manifest.json"
    tr := djt.NewTestRunner(manifest)
    if ! tr.RunAllTests() {
		// tests didn't pass
    }
    ...
}

Test configuration

Each test case defined has the following properties:

{
  "name": "Case1: Unique test name",
  "description": "Description of the test case",
  "includeCommon": true, # Include common configuration in this test run. Default is false
  "jobPath": "relative/filepath/to/my/job.json",
  "requiredDatasets": [
    {
      "name": "sdb.Animal",
      "path": "tests/testdata/case1/sdb-Animal.json"
    }
  ],
  "expectedOutput": "tests/expected/myJob/case1.json"
}

Note: All file paths in the manifest file are relative to the repo root of the datahub config project

Common configuration

Some configuration is common to all tests. To add datasets for all test cases, use the top-level property common.requiredDatasets. (See example manifest for details.)

Log output from transforms

If you're having trouble debugging a failing test, you can enable logging from the transforms by setting the loglevel to 'error' in the Log() call. This will print the output from the transforms to the console. Example:

Log("my log line", "error")