The Surveys is a demo application based on Kotlin + Vue.js development stack. It's packed with various frequently used features (see the details below) implemented in the basic shape. Mainly I've developed it to have a boilerplate repository for future pet projects 🏗.
This repository will evolve as I have time to add new features and optimize existing functionality.
The idea of application is to provide a platform for conducting surveys and having simple real-time analytics for running polls. There're 2 kinds of actors: coordinators creating surveys, and respondents.
- Surveys list with filters by date or status
- Composing a survey
- Tagging a survey
- Different views for Respondents and Coordinators
- Reviewing real-time survey responses statistics
- Updating user preferences
It showcases the following list of concepts:
I've started using Kotlin on new backend projects 2 years ago, and it's my first repository to have boilerplate code for full-stack apps with Kotlin. Here I'm using Kotlin in pretty much the same Spring setup as I would use Java.
It's arguable whether Spring Data REST is a good fit for a production-ready project. On the other hand, it allows a rapid start having full-blown API implementation with sorting, paging and filtering for given entities.
Any given controller can be extended with additional handlers or reimplemented from scratch as a separate custom REST endpoint.
Hibernate JPA auto schema creation was used during the development. After the first version of the schema had been finalized, I generated the Liquibase changelog from JPA entities and disabled ddl-auto by Hibernate.
To populate schema with some initial data (predefined users), there's sql script on classpath.
JWT-based authentication with role-based access is showcased with tokens issued by the backend. Such a setup must not be used in production, this functionality is only to demonstrate the flow. In a real solution, use IAM solution like Keycloak or 3rd party SaaS.
After reading the official guide on Spring Webflux + RSocket I got an idea to implement some real-time streaming functionality to try Reactive Streams in a demo application setup.
The survey statistics page is implemented using rsocket.js, it reads the survey response stream data from a Kotlin Flow that is populated with data from R2dbc H2 reactive-ready driver. The charts are updated on a survey stats page in real time.
As a sidenote, I've never used Gradle before, always preferring Maven as a tool that makes more sense to me. This was an interesting challenge to create something new with Gradle, especially writing the build script with Kotlin DSL since there're not many ready-to-use examples.
Some features are still in development 👷♂️. Don't consider the current implementation to be polished at this stage. The nearest TODOs:
- User management CRUD table
- User registration
- User profile editing
- Filtering survey list
- Sample integration tests for the rsocket interactions
- Kotlin 1.4
- Spring Boot 2.4.3
- Spring Data REST
- Spring Security (JWT-based authentication)
- Spring WebFlux
- Hibernate JPA
- H2 Database
- Liquibase (database migrations)
- Gradle Build
- JUnit 5
- Vue.js 2
- Vuex
- Vue-router
- Vue-image-upload
- Bootstrap 4
- RSocket
- Apexcharts
Running in a command line
./gradlew bootRunRunning in a Docker container
# build an image of 'surveys-backend'
./gradlew docker
# run a container
docker run -p 8081:8081 -p 8082:8082 surveys-backend:1.0.0-SNAPSHOTRunning in a command line with yarn
cd client
yarn serveIt launches the web application at http://localhost:8080
There're 2 predefined users:
m.lin@surveys.com / !F85BVekCVM3uDKyty (role: Coordinator)
a.mcgill@surveys.com / 12345!abc (role: Respondent)
MIT