Fabric8 Recommender

Fabric8 Recommender is a stack analysis feature.

Running the app

Set NODE_ENV

If you’re just trying to test the application, please use inmemory mode which will load the app with mock data for you. If you, however, want to contribute to the codebase, unset it back or to "development" (default) mode and rebuild.

$ export NODE_ENV=inmemory # (1)
$ export NODE_ENV=development # (2)
$ export NODE_ENV=production # (3)

<1> In-memory mode for trying-out the app
<2> Development mode for contributing to the source
<2> Production mode for deploying the application

Once you’re done setting the environment, you can proceed with the next step(s)

Note	If you’re directly trying to run the app in dev mode, you can skip this step, as `NODE_ENV` is treated as `"development"` by default.

First run

If you’re trying to run the app for the first time:

$ npm install

Then, start the app with:

$ npm start

Fresh run

If you trying to refresh your installation, you need to run:

$ npm run reinstall

Then, start the app with:

$ npm start

Testcase run

To run the linter, unit tests, and functional test use:

$ npm test

Other useful scripts

The package.json file’s scripts: section lists all the tasks we run.

Here are some of the most useful/frequently used scripts you may need to run:

Scipt	Command	Description
Lint	`$ npm run lint`	Runs the TypeScript and Angular 2 linter
Validation	`$ npm run validate`	Validates the webpack build
Unit Tests	`$ npm run test:unit`	Runs the unit tests
Functional Tests	`$ npm run test:func`	Runs the functional tests
Continuous Tests	`$ npm run watch:test`	Looks for changes in source code and runs unit tests

Building the app

Production build

To generate production build, set API URL and run build script as follows:

$ npm run build:prod

The build output will be under dist directory.

To create a docker image, run this command immediately after the production build completion:

Library Build

For production:

To build the fabric8-stack-analysis-ui as an npm library, use:

$ npm run build

The created library will be placed in dist.

Important

You shouldn’t ever publish the build manually, instead you should let the CD pipeline do a semantic release.

For development:

To build fabric8-stack-analysis-ui as an npm library and embed it into a webapp such as fabric8-ui, you should:

Step 1: Run npm run watch:library in the source directory: This will build fabric8-stack-analysis-ui as a library and then set up a watch task to rebuild any ts, html and scss files you change.
Step 2: Run npm link <path to fabric8-stack-analysis-ui>/dist-watch --production: In the webapp into which you are embedding. This will create a symlink from node_modules/fabric8-stack-analysis-ui to the dist-watch directory and install that symlinked node module into your webapp.
Step 3: Run your webapp in development mode: Make sure you have a watch on node_modules/fabric8-stack-analysis-ui enabled. You will have access to both JS and SASS sourcemaps if your webapp is properly setup.

Note	`fabric8-ui` is setup to do reloading and sourcemaps automatically when you run `npm start`.

To hit stack analysis api in standalone mode** Put a token in the environment variable with key as 'STACK_API_TOKEN'.

CSS and LESS

fabric8-stack-analysis-ui uses LESS for it’s stylesheets. It also uses the Angular emulation of the shadow dom, so you will normally want to place your styles in the .component.LESS file next to the html and the typescript.

We use mixins to avoid polluting components with uncessary style classes, and to avoid an explosion of shared files.

The src/assets/stylesheets/ directory includes a shared directory. These are shared global styles that we will refactor out in to a shared library at some point. Only update these styles if you are making a truly global style, and are going to synchronise your changes across all the various UI projects.

Contributing to the app

The development guide is part of the contributors' instructions. Please check it out in order to contribute to this project.

c-koehler / fabric8-recommender Goto Github PK