therobotacademy / loopback.io

LoopBack project site - now with documentation!

loopback.io

LoopBack community site, http://loopback.io. This website now includes the LoopBack documentation.

NOTE: The website is served from the gh-pages branch.

This repository is provided under the MIT License.

Jekyll installation

https://kristofclaes.github.io/2016/06/19/running-jekyll-locally-with-docker/

This post uses docker-compose, but as we are running a single container we really do not need it. We will use the image: jekyll/jekyll:pages

$ mkdir ~/jekyll
$ cd ~/jekyll
$ docker pull jekyll/jekyll:pages
$ docker run -it --restart unless-stopped -p 4000:4000 -v $PWD:/srv/jekyll --name jekyll jekyll/jekyll:pages jekyll serve --watch --incremental

This opens a bare jekyll site at port 4000. Visit: http://localhost:4000

Building and running loopback.io

We need to use another image variation: jekyll/jekyll:builder

Clone loopback.io inside the container ('_config.yml file modified to run loopback.io in port 4000):

$ docker run -it -p 4000:4000 --name loopback jekyll/jekyll:builder bash

$ git clone https://github.com/therobotacademy/loopback.io
$ chown -R jekyll:jekyll loopback.io
$ cd loopback.io
$ bundle install
$ npm start

npm start = bundle exec jekyll serve --incremental

You can add the option --port <PORT> to run the server in another port. Visit http://localhost:4000 (default configured as 4000)

NOTE: The docs part will be at http://localhost:4000/doc. It's not yet linked from the main "overview" part of the site, but will be once we launch (RSN).

To run in another port:

$ bundle exec jekyll serve --incremental --port 4000 # This is script 'start'

!!! NOT RESPONDING

For running if the container is stopped:

$ docker start loopback
$ docker exec -it loopback bash
> cd loopback.io
> npm start

Official Loopback.io instructions from here

Setup

To preview the website locally:

1. Install Ruby and Bundler if you don't have them already.

2. Clone this repo (you might use the SSH URL instead of HTTPS).:

git clone https://github.com/strongloop/loopback.io.git

3. cd to the repository directory and run the following command:

$ cd loopback.io
$ bundle install

Bundler will look in the Gemfile for which gems to install. The github-pages gem includes the same version of Jekyll and other dependencies as used by GitHub Pages, so that your local setup mirrors GitHub Pages as closely as possible.

Run and view site locally

Run Jekyll using the following command:

$ npm start

Then, load http://localhost:4001/ on your browser.

NOTE: The docs part will be at http://localhost:4001/doc. It's not yet linked from the main "overview" part of the site, but will be once we launch (RSN).

Formatting

Jekyll uses a variant of Markdown known as Kramdown.

Jekyll uses the Liquid template engine for templating.

Incorporating external READMEs

The documentation incorporates README files from a number of LoopBack example repositories. We use the get-readmes utility to fetch the README files directly from GitHub. Here is how to update the READMEs

1. npm install (first time/setup only)
2. npm run fetch-readmes

From there, the README markdown files are incorporated into documentation articles using the standard Jekyll "include" syntax as follows (for example):

---
title: "Angular example app"
lang: en
layout: readme
source: loopback-example-angular
keywords: LoopBack
tags:
sidebar: lb2_sidebar
permalink: /doc/en/lb2/Angular-example-app.html
summary: A brief tutorial on creating an Angular client app using the Loopback AngularJS SDK.
---

{% include readmes/loopback-example-angular.md %}

Incorporating updates from loopback-next

We use a node script update-lb4-docs to copy over contents from loopback-next repository using @loopback/docs package. The script is responsible for copying over the markdown documentation and related tables, as well as the sidebar used for LoopBack 4 content. The changes are then picked up as part of Travis CI's builds along with README update scripts and deployed to GitHub Pages once successful. The upgrade-swagger-ui.js script is also run as part of the builds. If you'd like to make documentation changes for LoopBack 4, please do so on its own repository.

Linting Readmes

There is an additional npm script that "lints" the readmes for markdown formatting problems. It is currently "experimental", see this issue for more info.

You can run this script thus:

$ npm run lint-readmes

LoopBack.io Docs Search

Docs Search is powered by IBM Watson Discovery.

How it works

LoopBack Docs are uploaded to Discovery by Travis CI (before deploying a new site -- after changes have been merged into gh-pages branch) automatically. The script creates a new collection and uploads each doc to that. The content of the docs along with certain metadata can be found in _site/posts.json. Travis uploads each Object in that file as a separate document (only uploads documentation pages).

Travis CI has encrypted variables which sets the environment variables for Discovery username / password. Core team can reach out to Taranveer Virk / Diana Lau for these credentials if needed.

• Once docs have been uploaded, Travis will delete the previous collection as the Search URL uses the oldest collection available to serve results.

If you want to try this out locally, you can create an account for Watson Discovery (Use the Lite plan which is free). Set the following environment variables locally:

export DISCOVERY_USERNAME="username"
export DISCOVERY_PASSWORD="password"

You can now upload docs to your Watson Discovery instance by running

npm run discovery-upload

Front End

Each documentation page has a search bar the top. The search will redirect the user to a /search page to show the results. The results are retrieved from a IBM Cloud Function that acts as a proxy to protect Watson Discovery credentials. The search bar also contains a hidden form input that sets the sidebar value and Watson Discovery filters content based on this value to return context aware results. Ex: Searching for the word controller from a LoopBack 4 documentation page will return LoopBack 4 results. To search all documents you can search from the /search page. Community Docs and Contribution docs are included in all results.

Cloud Function

The code for the Cloud function can be found here. It is deployed to the same account and must be edited directly in IBM Cloud (formerly BlueMix). The repository exists to document code changes and have peer reviews. Credentials for accessing the Cloud Function can be obtained from Taranveer Virk / Diana Lau OR you can ask them to re-deploy changes.

Your own version

You can deploy the code on your own BlueMix account and upload the URL for Discovery in search.html. For the function to work, you must set the following parameters in the function settings:

discovery_username: username
discovery_password: password