This is the front end repository for VA.gov. It contains application code and templates used across the site.
There are several repositories that contain the code and content used to build VA.gov. If you're looking to get started running VA.gov locally, you should read the Getting Started documentation.
Once you have the site set up locally, these are some common commands you might find useful:
I want to... | Then you should... |
---|---|
fetch all dependencies | yarn install ; run this any time package.json changes |
build both static HTML pages and applications | yarn build |
run the webpack dev server | yarn watch |
The vets-website
build has two main functions:
- Build the application assets (JS, CSS)
- Create the static HTML pages
vets-website
uses Webpack to bundle application
assets.
To build all applications, run the following:
yarn build:webpack
To recompile your application when you make changes, run:
yarn watch
You can also limit the applications Webpack builds with --env.entry
:
yarn watch --env.entry static-pages,auth
The entryname
for your application can be found in its manifest.json
file.
If you're developing a feature that requires the API, but can't or don't want to
run it locally, you can specify --env.api
:
yarn watch --env.api https://dev-api.va.gov
Note: If you try to log on, ID.me will redirect you to the environment that the API is set up for. So in the above example, you'd be redirected back to dev.va.gov.
VA.gov contains many pages that include content generated from a Drupal-based content model. When testing changes to static pages, or to see what your application looks like on VA.gov, you'll need to build these static pages using the following commands:
yarn build
(add —pull-drupal
to get the latest content from production - needs active socks proxy connection)
- run once to pull the latest Drupal content and build the static HTML files
- need to run this again when adding new templates based on new Drupal entities
yarn watch:content
- watches for changes to liquid templates or CSS
- separated from
yarn watch
because of JS memory issues
yarn preview
- You can run this concurrently with
yarn watch
. It adds local routes needed to preview Drupal nodes (e.g./preview?nodeId=XX
).
If you do not have access to the SOCKS proxy, you can fetch the latest cached version of the content with the following:
yarn fetch-drupal-cache
"Templating" refers to the process of data representing website content being compiled into an HTML file.
After setup, the vagov-content
repo should be a sibling repo to vets-website
. vagov-content
contains static content files in the form of Markdown (.md
) in the vagov-content/pages
directory. At the top of each Markdown file, there are declarations of key-value pairs referred to as "frontmatter", which are often used to define metadata about the page or to describe how the page should be processed during the HTML compilation.
The most common key in frontmatter is a layout
property, which refers to a file in vets-website/src/site/layouts
. For example, this layout
property refers to this layout file in vets-website
. The frontmatter of the .md
file are passed to the layout file as the global scope.
Example of a Markdown processed throughout a layout file
For example, pretend this is a .md
in vagov-content.
---
layout: favorite-movies.html
movies:
- Ace Ventura
- Disney Cars
- Toy Story
---
I love movies! Here are some of my recommendations.
The layout
property would refer to a layout file vets-website/src/site/layouts/favorite-movies.html
, which could look like -
<p>{{ contents }}</p>
<ul>
{% for movie in movies %}
<li>{{ movie }}</li>
{% endfor %}
</ul>
This templating language is called Liquid. The body of the Markdown file (which resides below the frontmatter of the page and may contain HTML) will be available to the template in a contents
property. This would result in the following HTML snippet.
<p>I love movies! Here are some of my recommendations.</p>
<ul>
<li>Ace Ventura</li>
<li>Disney Cars</li>
<li>Toy Story</li>
</ul>
After setup, your project should have issued a network request to fetch data from the VA CMS, Drupal. To verify, the response JSON should be located at vets-website/.cache/localhost/drupal/pages.json
. To refresh the content, run yarn build:content --pull-drupal
. The vets-website/.cache
directory will be emptied and a new network request will be issued.
The JSON file returned by the CMS will be parsed into separate pages during the vets-website
compilation process. Consider opening the vets-website/.cache/localhost/drupal/pages.json
file to explore the data structures.
Example of a Drupal entity processed throughout a layout file
{
"entityBundle": "landing_page",
"entityId": "79",
"entityPublished": true,
"title": "VA records",
"entityUrl": {
"breadcrumb": [
{
"url": {
"path": "/",
"routed": true
},
"text": "Home"
},
{
"url": {
"path": "",
"routed": true
},
"text": "Records"
}
],
"path": "/records"
},
"fieldIntroText": "Access your VA records and documents online to more easily manage your benefits."
}
Whereas Markdown files contain a clear layout
declaration in frontmatter, the entityBundle
is used to form the relationship between a CMS data structure to a layout. The entityBundle
can be considered a type of CMS "node" represented by this data structure. Layouts used solely by CMS data structures have a unique file extension - .drupal.liquid
. In this example, the layout file would be vets-website/src/site/layouts/landing_page.drupal.liquid
.
<h1>{{ title }}</h1>
<p>{{ fieldIntroText }}</p>
This layout would result in this HTML snippet -
<title>VA records</title>
<p>Access your VA records and documents online to more easily manage your benefits.</p>
The entityUrl.path
property is used to determine the page's path on the website. This example would be visible at /records/
.
Console output
While running your website locally, open the console panel of your browser's developer tools. Visit /health-care
and observe the following output.
This data structure represents the data structure that was processed throughout the layout file during the website compilation.
DOM attributes
To further demystify the dynamic nature of the templating process, a pattern of the project is to use `data-` attributes to describe various templates. For example, observe the various `data-template` properties in the following DOM.There is also an HTML comment at the top of the file to help describe the current page.
The implementation of the Liquid templating language used by our project is TinyLiquid. We highly recommend using TinyLiquid's README as your resource while writing a template, as there are small differences from Shopify's language spec.
CI will build both applications and content with the following:
yarn build
To run all unit tests, use:
yarn test:unit
If you want to run only one test file, you can provide the path to it:
yarn test:unit src/applications/path/to/test-file.unit.spec.js
To run all tests for a folder in src/applications, you can use app-folder:
yarn test:unit --app-folder hca
To run all tests in a directory, you can use a glob pattern:
yarn test:unit src/applications/path/to/tests/**/*.unit.spec.js*
To run tests with some extra debugging info, you can pass a log-level:
yarn test:unit --log-level debug
To run tests with coverage output, you can pass the coverage option:
yarn test:unit --coverage
For help with test runner usage, you can run:
yarn test:unit --help
- E2E or browser tests primarily run in Cypress.
- Some older, existing tests run in Nightwatch, but those are deprecated.
Before running Cypress tests, first make sure that:
vets-website
is served locally on port 3001- You can do this with
yarn watch
- You can do this with
vets-api
is NOT running- Any required APIs will be mocked by the Cypress test that needs them.
To open the Cypress test runner UI and run any tests within it:
yarn cy:open
To run Cypress tests from the command line:
yarn cy:run
To run specific Cypress tests from the command line:
# Running one specific test.
yarn cy:run --spec "path/to/test-file.cypress.spec.js"
# Running multiple specific tests.
yarn cy:run --spec "path/to/test-a.cypress.spec.js,path/to/test-b.cypress.spec.js"
# Running tests that match a glob pattern.
yarn cy:run --spec "src/applications/my-app/tests/*"
yarn cy:run --spec "src/applications/my-app/tests/**/*"
# Running tests that match multiple glob patterns.
yarn cy:run --spec "src/applications/a/tests/**/*,src/applications/b/tests/**/*"
To run Cypress tests from the command line on a specific browser:
yarn cy:run --headless --browser chrome
yarn cy:run --headless --browser firefox
# Without --headless, the test runner will open and run the test.
yarn cy:run --browser chrome
yarn cy:run --browser firefox
For other options with yarn cy:run
, the same options for cypress run
are applicable.
To run Nightwatch tests, you first need three things:
- Install the Java JDK on MacOS (if needed):
brew update brew tap adoptopenjdk/openjdk brew cask install adoptopenjdk8
vets-website
served locally on port 3001- You can do this with
yarn watch
- You can do this with
vets-api
to NOT be running- The browser tests will use a simple mock api on port 3000, but only if nothing is already attached to that port
yarn test:e2e
Just like with unit tests, you can also specify the path to the test file
yarn test:e2e src/applications/path/to/test-file.e2e.spec.js
In separate terminal from your local dev server, run
yarn mock-api --responses path/to/responses.js
See the mocker-api usage
documentation for how to use
the responses.js
.
If you need to log in, go to your browser dev tools console and enter
localStorage.setItem('hasSession', true)
and refresh the page. This will then
trigger a /v0/user
call, which will then get the mocked response of a logged-in
user. (Assuming you've mocked that response, of course.)
Responses to common API requests, such as /v0/user
and
/v0/maintenance_windows
, you can use
src/platform/testing/local-dev-mock-api/common.js
const commonResponses = require('src/platform/testing/local-dev-mock-api/common');
module.exports = {
...commonResponses,
'GET path/to/endpoint': { foo: 'bar' },
};
After a while, you may run into a less common task. We have a lot of commands for doing very specific things.
I want to... | Then you should... |
---|---|
build the production site (dev features disabled). | NODE_ENV=production yarn build --buildtype vagovprod |
fetch the latest content cache from S3 | yarn fetch-drupal-cache (does not require SOCKS proxy access) |
reset local environment (clean out node modules, Babel cache, and runs npm install ) |
yarn reset:env |
run only the app pages on the site for local development without building content. | yarn watch --env.scaffold |
run the site for local development with automatic rebuilding of Javascript and sass with css sourcemaps | yarn watch:css-sourcemaps then visit http://localhost:3001/ . You may also set --env.buildtype and NODE_ENV though setting NODE_ENV to production will make incremental builds slow. |
run the site for local development with automatic rebuilding of code and styles for specific apps | yarn watch --env.entry disability-benefits,static-pages . Valid application names are in each app's manifest.json under entryName |
run the site for local development with automatic rebuilding of code and styles for static content | yarn watch:static |
run the site so that devices on your local network can access it | yarn watch --host 0.0.0.0 --public 198.162.x.x:3001 Note that we use CORS to limit what hosts can access different APIs, so accessing with a 192.168.x.x address may run into problems |
run all unit tests and watch | yarn test:watch |
run only e2e tests | Make sure the site is running locally (yarn watch ) and run the tests with yarn test:e2e |
run e2e tests in headless mode | yarn test:e2e:headless |
run all linters | yarn lint |
run only javascript linter | yarn lint:js |
run only sass linter | yarn lint:sass |
run lint on JS and fix anything that changed | yarn lint:js:changed:fix |
run automated accessibility tests | yarn build && yarn test:accessibility |
run visual regression testing | Start the site. Generate your baseline image set using yarn test:visual:baseline . Make your changes. Then run yarn test:visual . |
test for broken links | Build the site. Broken Link Checking is done via a Metalsmith plugin during build. Note that it only runs on build not watch. |
add new npm modules | yarn add my-module . Use the --dev flag for modules that are build or test related. |
get the latest json schema | yarn update:schema . This updates our vets-json-schema vets-json-schema https://github.com/department-of-veterans-affairs/ to the most recent commit. |
check test coverage | yarn test:coverage |
run bundle analyzer on our production JS bundles | yarn build-analyze |
generate a stats file for analysis by bundle analyzer | NODE_ENV=production yarn build:webpack --env.buildtype=vagovprod --env.analyzer . Note that if you get an error like FetchError: request to http://prod.cms.va.gov/graphql failed you need to be on the SOCKS proxy |
load the analyzer tool on a stats file | yarn analyze |
add a new React app | yarn new:app (make sure you have vagov-content sibling to vets-website ) |
Browser | Minimum version | Note |
---|---|---|
Internet Explorer | 11 | |
Microsoft Edge | 13 | |
Safari / iOS Safari | 9 | |
Chrome / Android Web view | 44 | Latest version with >0.5% of traffic |
Firefox | 52 | Latest version with >0.5% of traffic |
- VA.gov Knowledge Hub
- Docs Directory
- Manual and Automated 508 Testing