J-Node Abstract System

A web application to publish conference information and to collect, review and publish conference abstracts. It is run by the INCF Japan Node and succeeds GCA-Web, the conference application of the German Neuroinformatics Node (G-Node), which is no longer maintained. It is a rewrite in Spring Boot and React that follows the design of GCA-Web and took over its data.

• Public pages show the conferences with their schedules, locations, floor plans and accepted abstracts.
• Authors write abstracts with authors, affiliations, figures and references, and submit them for review.
• Conference owners review the submitted abstracts, assign them to presentation groups and numbers, and manage the conference settings.
• Site admins create conferences and accounts. There is no public sign up.

Technology

• Backend: Spring Boot 4 (Java 17 or later), Spring Security, Spring Data JPA, MariaDB/MySQL (H2 for tests)
• Frontend: React 19, TypeScript, Redux Toolkit, React Bootstrap, built with Vite
• The frontend is built into the backend, so the application is a single executable jar.

Requirements

• JDK 17 or later
• Node.js 24 and npm for frontend development. The Maven build downloads its own Node.js, so building the jar needs only the JDK.
• MariaDB or MySQL for production and for development with real data

Development

Local configuration

Settings in application.properties in the working directory override the defaults in src/main/resources/application.properties, which also lists the available settings as comments. The file in the working directory is ignored by git, so keep local and production settings there. A minimal local configuration:

spring.datasource.url=jdbc:mariadb://localhost/abstracts
spring.datasource.username=abstracts
spring.datasource.password=secret
spring.jpa.hibernate.ddl-auto=update

app.url=http://localhost:5173
app.admins=admin@example.com
app.path.banners=./banners
app.path.figures=./figures
app.mail.mock=true

Without a database setting the application uses an in-memory H2 database, which starts empty. As accounts are created only by site admins, use a database with at least one account whose mail is listed in app.admins.

Running

Start the backend on port 8080, skipping the frontend build:

./mvnw spring-boot:run -Dskip.npm -Dskip.installnodenpm

Start the frontend development server on port 5173, which forwards /api to the backend:

npm install
npm run dev

Then open http://localhost:5173. With app.mail.mock=true, mails such as password reset links are written to the backend log instead of being sent.

Tests, linting and formatting

Command	Purpose
./mvnw test	Run the backend tests against an in-memory database. They ignore application.properties in the working directory.
./mvnw spotless:apply	Format the Java sources (profile in eclipse-formatter.xml).
./mvnw spotless:check	Check the format of the Java sources.
npm run lint	Lint and check the format of the frontend with Biome.
npm run format	Fix the lint findings and format the frontend with Biome.

VS Code formats Java files on save with the same Eclipse profile and the frontend with Biome (see .vscode).

Building

./mvnw clean package

This builds the frontend into src/main/resources/static and packages everything into target/abstracts-<version>.jar. Run it with an application.properties in the working directory:

java -jar abstracts-1.0.0.jar

Configuration

Application settings:

Setting	Default	Description
app.url	http://localhost:8080	Base URL of the site, used for links in mails.
app.admins	(none)	Comma separated mail addresses of the site admins.
app.read-only	false	Only site admins can log in, everyone else just reads the published information.
app.path.banners	./banners	Directory of the conference logos and thumbnails.
app.path.figures	./figures	Directory of the abstract figures.
app.mail.mock	true	Write mails to the log instead of sending them.
app.mail.from.address	abstracts@example.com	Sender address of mails.
app.mail.from.name	The J-Node Abstract System	Sender name of mails.

Spring settings that matter in production:

Setting	Description
spring.datasource.url, .username, .password	Database connection.
spring.jpa.hibernate.ddl-auto=update	Keeps the database schema in step with the entities, see below.
spring.mail.host, .port, .username, .password and spring.mail.properties.*	SMTP server.
server.servlet.session.cookie.secure=true	Sends the session cookie over HTTPS only.

Deployment

The database schema is updated by Hibernate on startup (spring.jpa.hibernate.ddl-auto=update), which adds missing tables and columns but never drops or changes existing ones. Back up the database before deploying a new version.

Check the production application.properties before starting a new version:

• app.admins is set, and every listed address has an account. Without it the site has no admins.
• app.url is the public URL of the site.
• app.mail.mock=false and the spring.mail.* settings point to the SMTP server. Otherwise mails are only logged.
• server.servlet.session.cookie.secure=true if the site is served over HTTPS.
• app.read-only=true while the site should only show published information.
• The directories of app.path.banners and app.path.figures are writable by the application.

Times are stored without time zone. Conference dates and schedules are local times of the venue and are shown as they are. Timestamps such as modification times are recorded in UTC, as the application runs in UTC, and are shown in the time zone of the browser.

Project structure

src/main/java/jp/neuroinf/abstracts
  controller/   REST endpoints under /api, and forwarding of application paths to the frontend
  service/      business logic and permission checks
  entity/       JPA entities
  dto/, form/   response and request bodies
  core/         security and application configuration
  component/    mail, file storage, tokens and login attempt limiting
frontend/src
  features/     pages and Redux slices by feature (abstract, account, conference, user)
  api/          API clients
  entities/     types of the API data
scripts/        build helpers, e.g. copying MathJax so that the site serves it itself

Acknowledgements

This system is based on the design of GCA-Web by the German Neuroinformatics Node (G-Node), including its data model and review workflow. We thank the G-Node for developing GCA-Web and making it available under an open source license.

License

This project is licensed under the MIT License, see LICENSE.