New to our project? Be sure to review the OpenMRS 3 Frontend Developer Documentation. You may find the Introduction especially helpful.
Also see the API documentation for @openmrs/esm-framework, which is contained in this repository.
Below is the documentation for this repository.
O3 Frontend Core
This is a monorepo containing the core packages for the O3 frontend. These packages handle cross-cutting concerns such as the configuration and extension systems, the core framework, global state management, the styleguide, and more.
Available Packages
Application
This contains tooling and the app shell.
Framework
The following common libraries have been developed. They may also be used independently of the app shell.
- @openmrs/esm-api: helps make API calls to the backend
- @openmrs/esm-config: validation and storage of frontend configuration
- @openmrs/esm-context: provides the AppContext for sharing contextual state across the app
- @openmrs/esm-dynamic-loading: provides functionality for dynamically loading frontend modules using Webpack Module Federation dynamic remotes
- @openmrs/esm-error-handling: handles errors
- @openmrs/esm-expression-evaluator: provides functions that allow evaluation of user-defined expressions in a safer way than eval()
- @openmrs/esm-extensions: implementation of a frontend component extension system
- @openmrs/esm-feature-flags: hide features that are in progress
- @openmrs/esm-globals: useful global variables and types
- @openmrs/esm-navigation: navigation utilities, breadcrumbs, and history
- @openmrs/esm-offline: provides offline functionality
- @openmrs/esm-react-utils: utilities for React components
- @openmrs/esm-routes: provides helper functions for working with
routes.jsonfiles in O3 - @openmrs/esm-state: brings in state management
- @openmrs/esm-styleguide: styling and UI capabilities
- @openmrs/esm-translations: common translations and utilities
- @openmrs/esm-utils: general utility and helper functions
All libraries are aggregated in the @openmrs/esm-framework package:
Frontend modules
A set of frontend modules provides the core technical functionality of the application.
- @openmrs/esm-devtools-app
- @openmrs/esm-help-menu-app
- @openmrs/esm-implementer-tools-app
- @openmrs/esm-login-app
- @openmrs/esm-primary-navigation-app
- @openmrs/esm-offline-tools-app
Development
Getting Started
To set up the repository for development, run the following commands:
yarn
yarn setup
Note: If
yarn setupfails or causes system resource issues, use this alternative instead:yarn build --concurrency 1Both commands build all the packages — once either completes successfully, you can proceed to running the app shell and the framework.
Building
To build all packages in the repository, run the following command:
yarn build
You can also verify all packages in one step using yarn:
yarn verify
Running the app shell and the framework
yarn run:shell
run:shell runs only the shell and the framework (not the frontend modules).
Running the frontend modules in apps
yarn start --sources packages/apps/<app folder>
For example, to run the Login app, run:
yarn start --sources packages/apps/esm-login-app
This will spin up a development server with hot module reloading so any changes you make to the code in the app will be reflected automatically in the browser.
Running the tooling
cd packages/tooling/openmrs
yarn build
./dist/cli.js
Running tests
Unit tests
To run tests for all packages, run:
yarn turbo run test
To run tests in watch mode, run:
yarn turbo run test:watch
To run tests for a specific package, pass the package name to the --filter flag. For example:
yarn turbo run test --filter="@openmrs/esm-styleguide"
To run a specific test file, run:
yarn turbo run test -- login
This command runs tests only in files whose names match the provided string.
You can also run those matching tests in watch mode by running:
yarn turbo run test:watch -- login.test
To generate a coverage report, run:
yarn turbo run coverage
By default, turbo will cache test runs. This means that re-running tests without changing any of the related files will return the cached logs from the last run. To bypass the cache, run tests with the force flag, as follows:
yarn turbo run test --force
Running End-to-End (E2E) tests
E2E tests run against a complete OpenMRS stack in Docker containers. The test runner handles everything automatically: building the frontend, starting containers, running tests, and cleanup.
Prerequisites:
Install Playwright browsers before your first run:
npx playwright install
Running E2E tests:
yarn test-e2e
This single command will:
- Build the frontend with your local changes
- Start Docker containers (backend, database, gateway, frontend)
- Wait for the backend to be ready
- Run all E2E tests in headless mode
- Automatically stop and remove containers when done
Common options:
# Run tests with browser visible
yarn test-e2e -- --headed
# Run tests in Playwright's interactive UI mode
yarn test-e2e -- --ui
# Run a specific test file
yarn test-e2e -- login.spec.ts
# Keep containers running on test failure (for debugging)
yarn test-e2e --keep-on-failure
# CLI-friendly output (no browser popup, useful for LLM tools)
yarn test-e2e --list
Running tests on multiple branches:
You can run E2E tests on different branches simultaneously in separate terminal windows. Each run automatically uses a unique port and container names based on your git branch, so there's no interference between concurrent test runs.
Testing against a remote instance:
To test against a remote instance instead of Docker containers, set the E2E_BASE_URL environment variable:
E2E_BASE_URL=https://dev3.openmrs.org/openmrs yarn playwright test
Note: When testing against a remote instance, use yarn playwright test directly (not yarn test-e2e) since you don't need the Docker environment.
Read the e2e testing guide to learn more about End-to-End tests in this project.
Linking the framework
You probably want to try out your changes to a framework library in a frontend module. Unfortunately, getting a working development environment for this is very finicky; no one technique works for all frontend modules all the time.
Note that even though frontend modules import from @openmrs/esm-framework, the package you need to link is the sub-library; for example, if you are trying to test changes in packages/framework/esm-api, you will need to link that sub-library.
If you're unsure whether your version of a core package is running, add a console.log at the top level of a file you're working on.
Here are the tools at your disposal for trying to get this to work:
Yarn link
This should be the first thing you try. To link the styleguide, for example, you would use
yarn link ../path/to/openmrs-esm-core/packages/framework/esm-styleguide
This will add a line to the "resolutions" section of the package.json file which uses the portal: protocol. The other protocol is link:. If you need to make changes to the esm-framework package, you will need to link it in as well. However, linking @openmrs/esm-framework as a portal created by yarn link will not work; instead manually add the line to the resolutions field in the package.json file:
"resolutions": {
"@openmrs/esm-framework": "link:../path/to/openmrs-esm-core/packages/framework/esm-framework"
}
Yalc
Sometimes, the build tooling will simply not work with yarn link. In this case, you will need to use yalc.
Install yalc on your computer with:
npm install -g yalc
Then, link the repository you are working on. For esm-api, for example, run:
# In this repository
cd packages/framework/esm-api
yalc publish
cd ../../../openmrs-esm-patient-chart # for example
yalc link @openmrs/esm-api
In order for Patient Chart to receive further updates you make to esm-api, you will need to run yalc push in the esm-api directory and yalc update in the Patient Chart directory.
Running with a local version of the core packages
After linking the packages (using yarn link or yalc), the build tooling is satisfied, but you must do one more step to get the frontend to load these dependencies at runtime.
Here, there are two options:
Method 1: Using the frontend dev server
In order to get your local version of the core packages to be served in your local dev server, you will need to link the tooling as well.
yarn link /path/to/esm-core/packages/tooling/openmrs
You can try using yalc for this as well, if yarn link doesn't work. Or manually create a link: resolution in package.json.
In packages/shell/esm-app-shell, run:
yarn build:development --watch
to ensure the built app shell is updated with your changes and available to Patient Chart. Then run your Patient Chart dev server as usual, with yarn start.
Method 2: Using import map overrides
In this repository, start the app shell with yarn run:shell. Then, in the Patient Chart repository, cd into whatever package(s) you are working on and run yarn serve from there. Then use the import map override tool in the browser to tell the frontend to load your local Patient Chart packages.
Once it's working
Please note that any of these techniques will modify the package.json file. These changes must be undone before creating your PR. If you used yarn link, you can undo these changes by running:
yarn unlink --all
in the Patient Chart repo.
Version and release
We use Yarn workspaces to handle versioning in this monorepo.
To increment the version, run the following command:
yarn release [version]
Where version corresponds to:
patchfor bug fixes e.g.,3.2.0→3.2.1minorfor new features that are backwards-compatible e.g.,3.2.0→3.3.0majorfor breaking changes e.g.,3.2.0→4.0.0
Note that this command will not create a new tag, nor publish the packages. After running it, make a PR or merge to main with the resulting changeset. Note that the release commit message must resemble (chore) Release vx.x.x where x.x.x is the new version number prefixed with v. This is to avoid triggering a pre-release build when effecting a version bump.
Once the version bump commit is merged, go to GitHub and draft a new release.
The tag should be prefixed with v (e.g., v3.2.1), while the release title should just be the version number (e.g., 3.2.1). The creation of the GitHub release will cause GitHub Actions to publish the packages, completing the release process.
Don't run
npm publish,yarn publish, orlerna publish. Use the above process.
Important Notes About Version Updates
When releasing a new major version (e.g., moving from v6 to v7), you must:
- Update all peerDependencies that reference
@openmrs/packages in every package that depends on them. - Change the version notation from the current major version to the new one (e.g., from
6.xto7.x).
Example:
// Before (during v6)
"peerDependencies": {
"@openmrs/esm-config": "6.x",
"@openmrs/esm-utils": "6.x"
}
// After (for v7)
"peerDependencies": {
"@openmrs/esm-config": "7.x",
"@openmrs/esm-utils": "7.x"
}
This ensures that all packages use compatible versions and breaking changes are properly tracked.
Design Patterns
For documentation about our design patterns, please visit our design system documentation website.
Bumping Playwright
Be sure to update the Playwright version in the Bamboo Playwright Docker image whenever making version changes.
Also, ensure you specify fixed (pinned) versions of Playwright in the package.json file to maintain consistency between the Playwright version used in the Docker image for Bamboo test execution and the version used in the codebase.
Documentation
The API documentation for @openmrs/esm-framework (under packages/framework/esm-framework/docs/) is generated by TypeDoc from the framework source. It's regenerated automatically by CI once a PR with framework changes has been approved — the bot pushes a (chore) Add docs commit back to the PR branch, so you don't need to run TypeDoc before submitting.
To preview locally:
yarn document
If a PR shouldn't update the docs (rare — typically only for reverts or release-version bumps), a maintainer can apply the skip-docs label to bypass regeneration.