NgAssessment

Prerequisites

To start the app required are:

• Node@10.15.3
• npm@6.4.1

Start app

Run npm start to start the application. This will start:

• fake rest api server at http://localhost:3000
• angular app at http://localhost:4200

Guidelines for monorepo

Enterprise angular monorepo patterns

Fake REST API

Mocked backed is implemented with json-server.

Creates rest endpoints from a json-config at fake-rest-api/endpoints.json

Run npm run fake:rest:api to create a server.

Server portal with resource description is available at http://localhost:3000

Code linting

To lint all apps & libs run npm run lint

To lint apps & libs separately run either of following:

• ng lint shipping
• ng lint shipping-e2e
• ng lint shipping-backend
• ng lint data
• ng lint ui-components

Linting of all apps & libs is done automatically on pre-commit and pre-push

Unit tests

To unit test all apps & libs via Jest run npm run test. By default if no tests are specified for app / lib it will pass.

• ng test shipping
• ng test shipping-e2e
• ng test shipping-backend
• ng test data
• ng test ui-components

Unit testing of all apps & libs is done automatically on pre-push.

Run npm run affected:test to execute the unit tests affected by a change.

Coverage

In order to test coverage of all apps & libs run npm run test:coverage Results are available in coverage/ dir in root.

Coverage is checked while unit-testing the app in pre-push (will fail if unit tests don't meet coverage thresholds).

Coverage thresholds are setup in jest.config.js :

• branches: 100
• functions: 100
• lines: 100
• statements: -10

E2E tests

Run ng e2e my-app to execute the end-to-end tests via Cypress.

Run npm run affected:e2e to execute the end-to-end tests affected by a change.

By default e2e tests are automatically run on pre-push.

Consistent code style

Consistent code style is enforced with Prettier.

Settings are available in .prettierignore & .prettierrc

To prettify the codebase manually, run npm run prettify

Executed automatically on:

• pre-commit
• pre-push
• in ci

Application shell

Run npm run build:shell to build the shipping angular app with app shell. To verify that app shell was built successfully check the contents of dist/apps/shipping/index.html to contain rendered contents of apps/shipping/src/app/app-shell/.

For profiling open the shipping app in chrome and in Performance tab hit Shift + Cmd + E.

HMR

In order to start the shipping front-end app with hot module replacement enabled, run npm run shipping:hmr

The web-page will reflect occurring changes in the code without live-reload (page refresh).

Build

Run ng build <project> to build the project. The build artifacts will be stored in the dist/ directory. Use the --prod flag for a production build.

SSR

To verify that SSR works, run:

• npm run ssr to build client and server and serve with server-side rendering
  • UI at http://localhost:4000/
  • API at http://localhost:4000/api
• curl http://localhost:4000/ > shipping.html that will put all the contents of the retrieved html file via http request to curl.html in the root.

All contents of the app component will be rendered to retrieved html file. If you do the same with npm run start the retrieved html file will only contain <app-root></app-root> element but not its content.

Shared UI-components library

Multiple front-end apps might be consuming reusable shared UI-components. Reusing shared UI-components with well-defined APIs will make development and delivery of domain related value easier & faster. Thus shared UI-components should be placed into ui-components library, located at libs/ui-components.

• to lint data library run ng lint ui-components
• to test data library run ng test ui-components

Shared data library

Front-end and back-end apps should use as much shared code, as possible, in order to prevent diverging of the code for same entities in FE and BE, which will make the entire app more error-prone eventually. Thus for keeping shared code between FE and BE apps there is a lintable & testable data library, located at libs/data.

• to lint data library run ng lint data
• to test data library run ng test data

Browser compatibility

Configuration for browser compatibility are set in package.json in "browserslist". Current options are to support last 2 versions of browsers with > 1% share.

Budgets

Custom budgets for the entire shipping app size are set to the following options in angular.json:

• baseline = 750kb
• maximumWarning = 200kb
• maximumError = 400kb

Deploy to gh-pages

Deployed app will be available at https://nongrata081.github.io/ng-assessment/

In order to deploy to gh-pages make sure that in dist/ directory root built bundles with index.html are present.

Install angular-cli-ghpages npm i -g angular-cli-ghpages

Build the app ng build prod

Publish to gh ngh

Dependency graph

Run npm run dep-graph to generate & open the graph of dependencies in the workspace.

Run npm run affected:dep-graph to see the graph of affected entities in the app (since retesting & rebuilding everything always will not be scalable and will take too much time).

CI/CD

.travis.yml has next steps for travis ci:

• commit lint
• lint
• unit test
• e2e

Bundle analyze

npm run build:stats to build the shipping app with stats.json files for es5 & es2015

npm run analyze:bundle:es5 to use webpack-bundle-analyzer for es5 bundle stats

npm run analyze:bundle:es2015 to use webpack-bundle-analyzer for es2015 bundle stats

Styleguide / UI components library app

Is implemented with Storybook.

Config files available at .storybook/

npm run storybook to run the app npm run build-storybook to build static files, output dir by default storybook-static/

App's documentation

Generation of static documentation for apps is implemented with compodoc.

npm run compodoc:shipping to generate docs for shipping ng app

npm run compodoc:shipping-backend to generate docs for shipping-backend nest app

Generated docs are stored in compodoc/

Performance

Auditing for web performance and best practice metrics is implemented via Lighthouse.

npm run lighthouse:mobile for mobile report

npm run lighthouse:desktop for desktop report

Generated reports will open after running scripts & are stored in lighthouse/

Visual regression testing

Is implemented with backstopjs

npm run backstop:reference to create reference images (stored in backstop_data/bitmaps_reference)

npm run backstop:test to test current app against reference images (stored in backstop_data/bitmaps_test)

npm run backstop:test:ci to test in CI without opening browser report (stored in backstop_data/ci_test)

npm run backstop:approve to approve tested images

npm run backstop:open:latest to open latest test report (stored in backstop_data/html_report)

Accessibility testing

Is implemented with pa11y.

npm run a11y:test to test http://localhost:4000/ and generate html report in a11y/

Exit codes:

0 - Pa11y ran successfully, and there are no errors

1 - Pa11y failed run due to a technical fault

2 - Pa11y ran successfully but there are errors in the page

Make sure to have the app served on :4000 before running a11y test

---

Enforce gitflow branches

Enforce best practice of gitflow branch naming patterns by failing push if branch name doesn't correspond to the gitflow pattern. Implemented with enforce-gitflow-branches

Commit linting

Checks if your commit messages meet the conventional commit format. Note: Please check rules / rules before using it.

In general the pattern mostly looks like this:

type(scope?): subject (scope is optional)

Real world examples can look like this:

chore: run tests on travis ci

fix(server): send cors headers

feat(blog): add comment section

---

Common types according to commitlint-config-conventional (based on the the Angular convention) can be:

• build
• ci
• chore
• docs
• feat
• fix
• perf
• refactor
• revert
• style
• test

These can be modified by your own configuration.

You still can commit with git commit, however the recommended way to do it is by using cli-prompt for commits.

Implemented with commitlint

---

CLI prompt for commits

You still can commit with git commit, however the recommended way to do it is by using cli-prompt for commits - commitizen-cli. Run yarn commit to open a prompt with a step-by-step wizard with hints to properly fill in the commit data.

Implemented with commitizen

Following conventional commit format will allow us to automatically generate changelog from commit history, that would have chapters (features, bugfixes, etc.) and would look like this

Recommended version bump

Run npm run recommended:bump to get a hit on a version for bump for release from conventional commit history.

Implemented with conventional-recommended-bump

Changelog generation

Run npm run generate:changelog to generate a changelog. Generates a changelog based on commits since the last semver tag that match the pattern of a "Feature", "Fix", "Performance Improvement" or "Breaking Changes". Conventional commits are a necessary prerequisite, enabling automatic changelog generation from commit history.

Implemented with conventional-changelog.

---

Recommended flow:

1. Run npm run recommended:bump to find out what is the recommended version for bump
2. Depending on which version you want to bump, run one of the following (this will also create git tag):

   yarn version --major
   yarn version --minor
   yarn version --patch
   

3. Generate changelog with npm run generate:changelog
4. Commit package.json and CHANGELOG.md files
5. Push including tags with git push --follow-tags
6. Manually create a release draft in Github and copypaste changelog’s release markdown to it
7. Publish release

---

According to gitflow, you first branch out to a new branch (e.g. feature/some-feature) to work on some functionality. While developing on that branch, you might make multiple commits. Since changelog is generated based on existing conventional commits (of types: feat, fix, perf, revert, revert), one has to shape the commit-history in such a way, that the resulting generated changelog on one-hand is clean (has no duplicate data as would be in case of having multiple conventional commits for one feature in a single branch), and on the other hand contains all the necessary data (a separate bullet in the changelog list which you want to be present there).

Your tools to achieve that are:

• Squash commits into a single commit, thus have a single bullet in the changelog generated from it.
• Keep separate commits for the functionality you want to be present as bullets in the changelog

---

Why squash commits?

Squashing related commits is considered a good practice, since it helps to keep the git history clean (especially valuable when there are a lot of people developing and pushing the code) and prevent its pollution with commits (e.g. an enterprise project with tens of developers making and pushing commits daily, without squashing very soon it will become hard to reason about the contents of git log due to enormous amount of commits).

Let's say you branched out to a new feature branch and developed a feature, and made several commits:

If you would then generate a changelog without squashing related commits, you would get a bloated changelog like this:

Now imagine, if you have 10 or more commits per feature? And then imagine, if you have 10 features? And then imagine, if you have 10 developers, each pushing 10 features? At the end of the day you have 1000 commits for a release contributed to by 10 developers, polluted git history and changelog and difficult times with reasoning about git log if you have to.

At this step, squashing commits comes to rescue.

---

How to squash commits?

In order to squash commits run git rebase -i HEAD~N, where N stands for amount of last commits, which you want to be able to perform actions on in rebase editor (squash, edit, etc). In this case we would like to squash last 5 commits, so run git rebase -i HEAD~5, which opens git rebase editor :

We would like to squash all feature functionality-specific commits into a single generic commit, so we leave pick for commit we want to keep, and replace word pick with squash for commits we want to squash:

Then write:quit and we get combination of squashed commits (where we can either edit or comment certain commit messages, if needed):

Then write:quit again and we get:

At this step our last 5 commits were squashed into a single commit, what we can see from git log:

Now, if we generate a changelog, it will have just one generic feature commit, into which we squashed other feature-specific commits:

So as a result you get a clean changelog without duplicate data, yet at the same time you have more details about which commits does the generic one contain in git-log

---

This project was generated using Nx.

Quick Start & Documentation

Nx Documentation

30-minute video showing all Nx features

Interactive Tutorial

Generate an application

Run ng g @nrwl/angular:app my-app to generate an application.

You can use any of the plugins above to generate applications as well.

When using Nx, you can create multiple applications and libraries in the same workspace.

Generate a library

Run ng g @nrwl/angular:lib my-lib to generate a library.

You can also use any of the plugins above to generate libraries as well.

Libraries are sharable across libraries and applications. They can be imported from @ng-assessment/mylib.

Development server

Run ng serve my-app for a dev server. Navigate to http://localhost:4200/. The app will automatically reload if you change any of the source files.

Code scaffolding

Run ng g component my-component --project=my-app to generate a new component.

Further help

Visit the Nx Documentation to learn more.