Skip to content

FHIR/sql-on-fhir-v2

Repository files navigation

SQL on FHIR® (v2.0)

Introduction

This project provides the source for the SQL on FHIR v2.0 Implementation Guide.

SQL on FHIR is a specification that defines a standard way to define portable, tabular projections of FHIR data.

The FHIR® standard is a great fit for RESTful and JSON-based systems, helping make healthcare data liquidity real. This spec aims to take FHIR usage a step further, making FHIR work well with familiar and efficient SQL engines and surrounding ecosystems.

We do this by creating simple, tabular views of the underlying FHIR data that are tailored to specific needs. Views are defined with FHIRPath expressions in a logical structure to specify things like column names and unnested items.

Read the specification →

Check the existing implementations page or register your own.

Check out the interactive playground.

FHIR Foundation Project Statement

  • Maintainers: Ryan Brush, Dan Gottlieb, John Grimes, Josh Mandel, Nikolai Ryzhikov, Arjun Sanyal
  • Issues / Discussion: analytics on FHIR @chat.fhir.org
  • License: MIT
  • Contribution Policy: CONTRIBUTING.md
  • Security Information: This repository consists of code to build a static implementation guide, and an implementation for reference only.
  • Compliance Information: This specification and its reference implementation are tested using the test suite in the tests directory. See Tests Overview for more information.

Content

Content as markdown is now found in input/pagecontent. Also see sushi-config.yaml for additional settings, including configuration for the menu.

Local Build

This is a Sushi project and can use HL7 IG Publisher to build locally:

  1. Clone this respository

  2. Run ./scripts/_updatePublisher.sh to get the latest IG publisher

  3. Install sushi if you don't have it already with: npm i fsh-sushi

  4. Run ./scripts/_genonce.sh to generate the IG

  5. Run open output/index.html to view the IG website

    Instructions for viewing the IG in a local http-server...
    npm i http-server
    cd output
    http-server  # Will launch the content in a new browser tab.

Building tests, see test README

Testing Implementation

This specification contains a set of tests in the /tests directory, which are set of test case files, each covering one aspect of implementation. A test case is represented as JSON document with title and description attributes, a set of fixtures (FHIR resources) as the resources attribute, and an array of test objects.

A test object has a unique title, a ViewDefinition as the view attribute, and and expected set of resulting rows in the expect attribute.

Tests Overview

Test cases are organized as individual JSON documents within the /tests directory. Each test case file is structured to include a combination of attributes that define the scope and expectations of the test. The main components of a test case file are:

  • Title (title attribute): A brief, descriptive title that summarizes the aspect of the implementation being tested.
  • Description (description attribute): A detailed explanation of what the test case aims to validate, including any relevant context or specifications that the test is based on.
  • FHIR version: A list of FHIR version strings like ['4.0.1', '5.0.0']. This applies to all FHIR resources in the test suite. The version numbers come from this ValueSet and can only include "Release" versions.
  • Fixtures (resources attribute): A set of FHIR resources that serve as input data for the test. These fixtures are essential for setting up the test environment and conditions.
  • Test Objects (tests attribute): An array of objects, each representing a unique test scenario within the case. Every test object includes:
    • Title (title attribute): A unique, descriptive title for the test object, differentiating it from others in the same test case.
    • Tags (tags attribute): A list of strings that categorize the test case into relevant groups. This attribute helps in organizing and filtering test cases based on their scope or focus. See Reserved Tags.
    • ViewDefinition (view attribute): Specifies the ViewDefinition being tested. This attribute outlines the expected data view or transformation applied to the input fixtures.
    • Expected Result (expect attribute): An array of rows that represent the expected outcome of the test. This attribute is crucial for validating the correctness of the implementation against the defined expectations.

Below is an abstract representation of what a test case file might look like:

{
  // unique name of test
  'title': 'title',
  'description': '...',
  'fhirVersion': ['5.0.0', '4.0.1'],
  // fixtures
  'resources': [
    {'resourceType': 'Patient', 'id': 'pt-1'},
    {'resourceType': 'Patient', 'id': 'pt-2'}
  ]
  'tests': [
    ...
    {
      'title': 'title of test case',
      'tags': ['shareable'],
      // ViewDefintion
      'view': {
        'select': [
          {'column': [{'name': 'id', 'path': 'id'}]}
        ]},
      // expected result
      'expect': [
        {'id': 'pt-1'},
        {'id': 'pt-2'}
      ]
    }
    ...
  ]
 }

Reserved Tags

The following tags are reserved for categorizing test cases based on their applicability to profiles within the core specification:

  • shareable: Test cases that validate conformance with the Shareable View Definition profile.
  • tabular: Test cases that validate conformance with the Tabular View Definition profile.
  • experimental: Test cases that cover experimental aspects of the specification.

Implement Test Runner

To ensure comprehensive validation and interoperability, it is recommended for implementers to integrate the test suite contained in this repository directly into their projects. This can be achieved efficiently by adding this repository as a git submodule to your project.

Furthermore, implementers are advised to develop a test runner based on the following guidelines to execute the test cases and generate a test report. This process is essential for verifying the implementation against the specified test cases.

The test runner should be designed to automate the execution of test cases and generate a comprehensive test report. Follow these steps to implement your test runner:

Test Execution Workflow

  • Iterate through each file in the test directory: Treat each file as a distinct testcase. These files are JSON documents containing the test specifications.
  • Read and parse each testcase file: Load the content of the testcase to access its structure, including the title, description, fixtures (resources), and the tests array.
  • Load fixtures into your implementation (if required): Before running the tests, ensure that the input data (FHIR resources specified in the resources attribute) is correctly loaded into your implementation's environment.
  • Execute each test:
    • For every test object within the tests array of a testcase, evaluate the view against the loaded fixtures by calling a function like evaluate(test.view, testcase.resources).
    • Compare the result of the evaluation with the expected results specified in the expect attribute of the test object.

Generating the Test Report

The test runner should produce a test_report.json file containing the results of the test executions. The structure of the test report is a map where:

  • each key is the name of a test file,
  • each value is a map with a single tests list,
  • each element of the tests list has a name and a result field, reporting whether the name test passed or not. Beside passed, the result map may also have a reason text field describing why the test did not pass. Here is an example:
//example test_report.json
{
  "logic.json": {
    "tests": [
      {
        "name": "filtering with 'and'",
        "result": {
          "passed": true
        }
      },
      {
        "name": "filtering with 'or'",
        "result": {
          "passed": false,
          "reason": "skipped"
        }
      },
      ...
    ]
  },
  ...
}

Reporting Your Test Results

After running the test suite and generating a test_report.json file with the outcomes of your implementations test runs, the next step is to make these results accessible for review and validation. Publishing your test report to a publicly accessible HTTP server enables broader visibility and verification of your implementations compliance with the specifications. This guide outlines the process of publishing your test report and registering your implementation.

You can validate the structure of your test report file using the test report JSON schema.

Publishing the Test Report

  1. Choose a Hosting Service: Select an HTTP server or a cloud storage service (such as AWS S3, Google Cloud Storage, or Microsoft Azure Blob Storage) that supports setting CORS (Cross-Origin Resource Sharing) policies. This is crucial for enabling the test report to be accessed from different origins.

  2. Upload Your Test Report:

    • Ensure your test_report.json is ready for publication.
    • Upload the file to your chosen service. If you're using cloud storage, you might need to create a bucket or container if you haven't already.
  3. Enable CORS:

    • Configure CORS settings on your HTTP server or cloud storage bucket to allow requests from https://fhir.github.io. This typically involves setting a CORS policy that includes this origin.
    • An example CORS policy for a cloud storage service might look like:
      [
        {
          "AllowedOrigins": ["https://fhir.github.io"],
          "AllowedMethods": ["GET"],
          "AllowedHeaders": ["*"],
          "MaxAgeSeconds": 3000
        }
      ]
  4. Verify Access:

    • After configuring CORS, verify that the test_report.json can be accessed from a browser without encountering CORS errors. You can do this by attempting to fetch the report from a webpage hosted on https://fhir.github.io or using developer tools in your browser.

Registering Your Implementation

Once your test report is published and accessible, the final step is to register your implementation in the test_report/public/implementations.json file. This file serves as a registry of available implementations and their test results, facilitating discovery and comparison.

  1. Format of implementations.json:

    • The implementations.json file is a JSON document that lists implementations along with URLs to their test reports.
    • Each entry should include the name of your implementation and the URL to the published test_report.json.
  2. Add Your Implementation:

    • Clone or fork the repository containing the implementations.json if necessary.
    • Add an entry for your implementation in the format:
      {
          "name": "YourImplName",
          "description": "<description>",
          "url": "<link-to-the-site>",
          "testResultsUrl": "<link-to-test-results>"
      },
    • Ensure that the URL is directly accessible and points to the latest version of your test report.
  3. Submit Your Changes:

    • Commit your changes to the implementations.json file.
    • If you're working on a fork or a branch, submit a pull request to the main repository to merge your changes.
    • Your implementation will appear on the implementations page.

By following these steps, you'll not only make your test results publicly available, you'll also contribute to a collective resource that benefits the entire FHIR implementation community. Your participation helps in demonstrating interoperability and compliance with the specifications, fostering trust and collaboration among developers and organizations.