Pledari Grond - Java API

This repository contains the code for the Java API of Pledari Grond, a Romansh language dictionary and grammar resource.

Technology Stack

• Java
• Spring Boot - Web framework and dependency injection
• MongoDB - Document database for dictionary data
• Apache Lucene - Full-text search indexing
• Maven - Build and dependency management
• Docker - Containerization for local development

Development Setup

Prerequisites

• Java 21
• Maven 3.6+
• Docker (for local MongoDB instance)

Getting Started

1. Clone the repository:

   git clone <repository-url>
   cd pg

2. Start MongoDB with Docker:

   docker-compose up -d

   This will start:

   • MongoDB on port 27000 (accessible with rootuser/rootpass)
   • Mongo Express (web UI) on port 8081 (login with admin/pass)

3. Configure the application: Edit api/src/main/resources/application.properties to adjust settings as needed. The default configuration connects to the Docker MongoDB instance.

4. Build the project:

   mvn clean install

5. Run the API:

   cd api
   mvn spring-boot:run

   The API will be available at http://localhost:8080

Configuration

Project configuration is located in api/src/main/resources/application.properties. Key settings include:

• Database connection: MongoDB host, port, credentials
• Admin user: Default admin user credentials (configure on first startup)
• Data directories: Locations for Lucene indexes, backups, exports, and corpus data
• JWT secret: For authentication tokens

Data Directory

The data/ directory contains various runtime and persistent data:

• backup/ - Automated backups of dictionary databases
• corpus/ - Text corpus files from La Quotidiana
• db_dump/ - Database dump files for export/import
• export/ - Generated export files (dictionary exports)
• hunspell/ - Spellchecker files and Git repository
• lucene_index/ - Apache Lucene search indexes

Project Structure

The repository consists of several modules:

• api: Spring Boot application handling web requests to the API.
• common: Data definitions used across different modules.
• corpus: A simple corpus of La Quotidiana texts used to allow the editors to search for examples.
• dictionary: Handles dictionary data persistence in MongoDB, including data models and repository layer.
• inflection: Logic for inflecting words (conjugation, pluralization) in various Romansh idioms. Conjugation generation partially relies on a copy of the Maalr's maalr.conjugator module.
• lucene: Creates Apache Lucene indexes for dictionary data. This includes one main index and two suggestion indexes (German and Romansh) for cases where no exact search results are found.
• names: Manages storage of names (typical Romansh first/last names, geographical names, etc.) used by spellcheckers.
• parent: Parent module for all project modules, defining common pom entries.
• pronunciation: Handles audio pronunciation generation and conversion using JAVE (Java Audio Video Encoder).
• spellchecker: Handles generation of spellchecker files.
• users: Manages user data and authentication in MongoDB.

License

This project is licensed under the Apache License 2.0. See the LICENSE file for details.