! We will deprecate this project in the next few weeks.

Gulliver

Gulliver is a directory of Progressive Web Apps.

Contents

In Gulliver's landing page you can browse the set of currently registered PWAs as depicted in the following landing page snapshot:

If you click on a particular PWA, Gulliver takes you to a detail page showing the results of an evaluation done on that specific PWA using the Lighthouse PWA Analyzer tool (Details page #1), and a view of the associated web app manifest file for the application (Details Page #2):

Details Page #1	Details Page #2

Gulliver itself has been implemented as a PWA; therefore it is designed to work well on any kind of device, including desktop web browsers (see landing page), and on mobile devices (see details page).

FAQ

Visit our FAQ Page

Requirements

Gulliver was built using the ExpressJS web framework for Node.js, and uses the Google Cloud Platform (GCP) for computing and storage services.

The following components are required to run the project (tested on macOS):

1. NodeJS (LTS version ~6.11.0). A JavaScript runtime built on Chrome's V8 JavaScript engine. (How to verify? Run

   node --version

   .) If you have a later version, install the LTS version with

   nvm

   .

2. Google Cloud SDK. A set of tools for the Google Cloud Platform (GCP) that you can use to access the Google Compute Engine and the Google Cloud Storage, which are two components of GCP used by Gulliver. (How to verify? Run

   gcloud --version

   .)

3. Memcached. A distributed memory object caching system. (How to verify? Run

   memcached

   (the command should appear to hang), and then

   telnet localhost 11211

   in a separate terminal. In the

   telnet

   window, typing

   version

   it should report the

   memcached

   version. If you don't have it, see these instructions to install memcached.)

In addition, you will need to set up a GCP project, and configure OAuth:

1. Create a Google Cloud Platform project. A GCP project forms the basis of accessing the GCP. Then, run

   gcloud init

   to configure

   gcloud

   locally, if you get the error "Could not load the default credentials" run

   gcloud auth login

   .

2. Get the OAuth client id and client secret associated with this project. (How to verify? There's no automatic way, but see Creating a Google API Console project and client ID for how to create one. Make sure you list

   http://localhost:8080

   as one of the

   Authorized JavaScript origins

   .)

Finally (and optionally), you need a Firebase project, and the Firebase Cloud Messaging "Server key" and "Sender ID":

1. Create a Firebase project.

2. Get Firebase Cloud Messaging "Server key" and "Sender ID" associated with this project. Select "Project settings" and then "Cloud Messaging". The URL should be of the form https://console.firebase.google.com/project/$FIREBASE_PROJECT/settings/cloudmessaging. (How to verify? There's no automatic way, but the "Server key" should be a long string of >100 characters, and the "Sender ID" a >10 digit number.)

Running Gulliver

1. Clone the GitHub repository:

   git clone https://github.com/GoogleChrome/gulliver.git

2. Switch into the project directory:

   cd gulliver

3. Create indexes for the Google Cloud Datastore:

   gcloud datastore create-indexes index.yaml

4. (Optional) Deploy cron jobs for scheduled PWA updates:

   gcloud app deploy cron.yaml

5. Install Memcached and run it on

   localhost:11211

   . Check these installation instructions for guidance.

6. Run

   npm install

   to install dependencies.

7. Configure your project either via a config file or environment variables (which override the corresponding keys in the config file). To create a config file, copy the sample config and adjust the values accordingly:

$ cp config/config.example.json config/config.json
$ vim config/config.json

1. Start Gulliver via

   npm start

   .

2. Gulliver should now be running at

   http://localhost:8080

   .

Running Tests

To verify that everything is working properly you can run the project's tests:

1. npm test

   to run lint + tests + coverage report.

2. npm run mocha

   to run all the tests only.

3. npm run coverage

   to run tests + coverage report.

Lighthouse PWA Analyzer

Gulliver reports an evaluation of the "progressiveness" of each registered PWA. This evaluation is done by Lighthouse, which is a tool that runs a set of checks validating the existence of the features, capabilities, and performance that should characterize a PWA. You can learn more about Lighthouse in the GitHub repository, or in this video.

References

To find out more about what PWAs are and how to go about incorporating the principles of PWAs into the development of your applications, check the following references which provide introductory information and references:

• Progressive Web Apps: Documentation entry point. Here you will find several resources to get started developing PWAs

• Progressive Web Apps: Escaping Tabs without Losing our Soul: Introductory article with historical perspective

• Getting Started with Progressive Web Apps: Sound introduction on the fundamental elements behind the development of PWAs

• The Building Blocks of PWAs: Interesting overall view of PWAs.

License

See LICENSE for more.

Disclaimer

This is not a Google product.