snappergps-app

This repository contains the front-end and a (small) part of the back-end of the SnapperGPS web application.

It is the companion app for your SnapperGPS receiver. Use it to configure your SnapperGPS receiver for your next deployment and to process the collected data after a completed deployment.

Find the remainder of the back-end in the snappergps-backend repository.

Table of contents

• Technologies
• Repository structure
• Setting up a new SnapperGPS Heroku app
• Setting up a new SnapperGPS app without Heroku
• Database
• Files
• WebUSB messages
• Offline mode
• Further notes
• Acknowledgement

Technologies

The front-end is designed as a Progressive Web App and hence aims to combine the advantages of a web app and a native app.

The part of the back-end in this repository targets the Node.js runtime and is intended to be hosted on the Heroku cloud platform, although, the latter is not strictly necessary. It stores data in a PosgreSQL database.

Both parts are written in JavaScript.

The front-end communicates with your SnapperGPS receiver via the WebUSB API. This allows for secure communication without the need to install a driver. However, it requires a web browser and an operating system that support the WebUSB API. Examples of browsers that currently (2022) support the WebUSB API are Microsoft Edge and Google Chrome. Mozilla Firefox and Safari currently do not support the WebUSB API. Examples of operating systems that currently support the WebUSB API are macOS, Microsoft Windows, and Linux operating systems like Android, Ubuntu, and Chrome OS. iOS and iPadOS currently do not support the WebUSB API.

A service worker is present to enable the app to run offline, to keep the it up-to-date, and to serve push notifications.

Repository structure

The JavaScript (and related content) for the back-end is split between the following files:

• The Procfile that defines the command to start the app,
• The dependencies in package.json,
• The entry point server.js that also connects to the database,
• Very few Heroku settings in settings.js,
• The routes to fulfil the user requests in routes.js, and
• Database functions in js/dbFunctions.js.

The directory views contains Embedded JavaScript templates (EJS) that generate the individual views. The most important ones are:

• views/index.ejs for the home page,
• views/configure.ejs for configuration of a connected SnapperGPS receiver,
• views/upload.ejs for uploading data from a connected SnapperGPS receiver to the cloud, and
• views/search.ejs and views/view.ejs for viewing processed data.

The sub-directory views/partials contains elements that are used across different views.

Resources available to the user are stored in public:

• Cascading style sheets in public/css,
• Icons and photos in public/images,
• JavaScript for the individual views in public/js,
• Important functions that keep the app automatically up-to-date, enable it to run offline, and to make push notifications work in public/service-worker.js, and
• The web app manifest public/snappergps.webmanifest that comprises all definitions to turn the SnapperGPS website into a PWA.

Setting up a new SnapperGPS Heroku app

• Make sure that you have git installed and configured.
• Open a terminal in a directory of your choice.
• Clone the SnapperGPS app repository via SSH ot HTTPS:

git clone [email protected]:SnapperGPS/snappergps-app.git
# or
git clone https://github.com/SnapperGPS/snappergps-app.git

• Change to the new directory:

cd snappergps-app

• Sign up for a Heroku account.
• Go to your Heroku app dashboard.
• Select New and Create new app.
• Choose an app name, e.g., my-snappergps-app, and click Create app.
• Choose a deployment method. In the following, we will assume that you chose Heroku Git.
• Download and install the Heroku command line interface (CLI).
• Go back to your terminal and login to your Heroku account:

heroku login

• Add the Heroku app as a remote for the repository:

heroku git:remote -a my-snappergps-app

• To make the interactive maps on the upload view and the download view display properly, obtain a Mapbox access token.
• Edit public/js/upload/uploadUI.js, public/js/view/viewUI.js, and public/js/animate/animateUI.js and set the variable mapboxAccessToken to your token.
• To make push notifications work, generate a VAPID key pair for you email address.
• Edit public/js/upload/uploadUI.js and set the variable vapidPublicKey to your public key.
• Fill in your privacy policy in views/privacy.ejs.
• Apply further desired changes to the source code of the SnapperGPS app.
• Stage, commit, and push (deploy) your changes:

git add public/js/upload/uploadUI.js public/js/view/viewUI.js my/changed.file
git commit -m "Change something"
git push heroku main

• If you want, visit your own SnapperGPS app on https://my-snappergps-app.herokuapp.com.
• Create a Heroku Postgres SQL database using a Heroku plan, e.g., the free hobby-dev plan:

heroku addons:create heroku-postgresql:hobby-dev

• Obtain and note down the URL, the name, the user, and the password for the database from the following command, which returns a URL with the structure postgres://database_user:database_password@database_url:port/database_name:

heroku config

• Alternatively, you can connect an external PosgreSQL database instead of using one provisioned by Heroku:

heroku config:add DATABASE_URL=postgres://database_user:database_password@database_url:port/database_name

• In both cases, set up the database SQL schema using the provided file:

heroku pg:psql --file snappergps_db_schema.sql

• If you would now upload data via the app, it should appear in the database, which you can either verify by creating an SQL dataclip in the Heroku data center or with an SQL query from a local PosgreSQL client.
• Set up the Python back-end.
• Enter the credentials of your database into the config.py file of the back-end.
• The app should now be fully functional.

Setting up a new SnapperGPS app without Heroku

The SnapperGPS app can be set up on a server of your choice without using Heroku. If you want to do so, pay attention to the following points:

• The server needs the Node.js runtime and the NPM package manager to serve the app.
• For a git-based deployment, it also needs git.
• You need to set up a PosgreSQL database and manually connect it to the app by setting DATABASE_URL to a URL with the structure postgres://database_user:database_password@database_url:port/database_name.
• The SnapperGPS app requires HTTPS.
• In general, follow the same steps as for a Heroku-hosted app.

Database

The SQL file snappergps_db_schema.sql describes the schema of the PosgreSQL database. It consists of four tables:

• uploads has a single row for every user upload, either directly from a SnapperGPS receiver or from a file. It contains meta data like an upload timestamp datetime, the ID of the SnapperGPS receiver device_id, and the unique upload_id that identifies this upload. In addition, it optionally holds information to communicate the processing progress to the user, including an email address, a push notification subscription, and a Telegram chat_id. Finally, there is also data set by the processing back-end, e.g., the status of the processing.
• snapshots has a single row for each uploaded GNSS signal snapshot. Each row is identified by a unique snapshot_id and linked to an upload via an upload_id. Besides that each row holds a measurement timestamp datetime in UTC, the raw signal snapshot data where each bit represents an amplitude measurement and the bit-order is little, a temperature measurement from the MCU in degrees Celsius, and a battery voltage measurement in volts.
• reference_points has a single row for each user-provided starting point, described by latitude lat and longitude lng and linked to an upload via an upload_id.
• positions is populated by the processing back-end with position estimates consisting of estimated_lat and estimated_lng in decimal degrees. In addition, each row contains an estimated_time_correction in seconds and an uncertainty estimate estimated_horizontal_error. Each row is linked to a signal snapshot via a snapshot_id.

Files

Instead of uploading data directly to the database, the users can choose to transfer the data from their SnapperGPS receiver to their host computer and store it in a local file. This is done via the Upload view, too. Two file formats are available. The legacy file format consisting of a CSV file with meta data and an individual binary file for each snapshot in a ZIP-compressed directory is described in snapshot-gnss-data. The alternative is a single (mostly) human-readable JSON file for a whole recording that contains a little bit of meta data (deviceID, firmwareDescription, firmwareVersion) and an array of individual GNSS signal snapshots. Each snapshot is defined by a measurement timestamp in UTC, a temperature measurement in degrees Celsius, a batteryVoltage measurement in volts, and the actual data. The latter are the raw signal amplitudes with one bit per value. The values are stored as byte stream where the bit order is little and which is encoded using Base64.

WebUSB messages

The SnapperGPS web app uses the WebUSB API to securely communicate with a SnapperGPS receiver. The custom USB messages are defined in the readme of snappergps-firmware.

Offline mode

Three views of the web app run offline: Home, Configure, and Upload. This is made possible by a service worker, which is defined in public/service-worker.js. When the user visits the homepage for the first time, then the service worker is registered. (You can check this in your browser's developer tools at Application -> Service Workers.) Once the worker is installed, it gets an install event and caches all the files that are listed in the respective function. Currently, these are all public files that are required to run the three views mentioned above offline. If you add or remove files used by these views, remember to change the file list of the service worker, too. If a user visits the page again, then the service worker intercepts any fetch request and provides the data from the cache if no network is available to enable an offline experience. If a network is available, it will attempt to update the cache. The development options allow you to simulate an offline experience, too, using the Network tab and the throttling options.

Further notes

• If you want to run the app locally before you deploy it on a server (probably a good idea), then you can find information how to do it here. For this, you need Node.js and npm on your machine.
• If you add new resources (files) that shall be part of the offline version of the app, then make sure that the service worker caches them (see above).
• Note that the Python back-end is not part of this repository. To process snapshots that you have uploaded to the database, you need to run the script process_queue.py. This requires maintaining a local navigation database, which you can achieve with maintain_navigation_data.py. For more information, see the readme in the respective repository.
• If you want to release the app in an app store such as Google Play or the Microsoft Store, then you can use the PWA builder to package it. Afterwards, follow these instructions to publish it on Google Play or these instructions for the Microsoft Store.

Acknowledgements

Jonas Beuchert and Alex Rogers are based in the Department of Computer Science of the University of Oxford.

Jonas Beuchert is funded by the EPSRC Centre for Doctoral Training in Autonomous Intelligent Machines and Systems (University of Oxford Project Code: DFT00350-DF03.01, UKRI/EPSRC Grant Reference: EP/S024050/1) and works on SnapperGPS as part of his doctoral studies. The implementation of SnapperGPS was co-funded by an EPSRC IAA Technology Fund (D4D00010-BL14).

Parts of the SnapperGPS web app are based on work by Peter Prince.

This documentation is licensed under a Creative Commons Attribution 4.0 International License.