ussenuk/Congolese-Swahili-TODS-ontology-pipeline

# Humanitarian Dialogue System with RDF Ontologies A task-oriented dialogue system in Congolese Swahili for humanitarian assistance in displacement camps. This system integrates Rasa NLU with RDF ontologies to provide semantic query capabilities. ## Project Overview This system helps displaced people and humanitarian workers access information about: - Health facilities and services - Water sources - Food distribution - Camp locations and facilities - Security information The system uses RDF (Resource Description Framework) to create a structured knowledge base that can be queried using SPARQL. ## Key Components 1. **RDF Ontology**: Defines humanitarian domain concepts using OWL/RDF 2. **Triple Store**: Stores RDF data and provides SPARQL query endpoint 3. **Rasa NLU**: Natural language understanding for Congolese Swahili 4. **Custom Actions**: Connect Rasa intents to SPARQL queries ## Setup Instructions ### 1. Install Dependencies ```bash # Install Python requirements pip install -r requirements.txt # Additional packages for RDF support pip install rdflib SPARQLWrapper ``` ### 2. Set Up Triple Store We recommend Apache Jena Fuseki as the SPARQL server: > **Important**: Apache Jena Fuseki 5.4.0 requires Java 17 or later. Please ensure you have the correct Java version installed before proceeding. ```bash # Check your Java version java -version # Download and extract Fuseki wget dlcdn.apache.org unzip apache-jena-fuseki-5.4.0.zip cd apache-jena-fuseki-5.4.0 # Start Fuseki server (development mode) ./fuseki-server --mem /humanitarian ``` If you're using an older Java version and cannot upgrade, you can use an older version of Fuseki that is compatible with your Java installation: ```bash # For Java 11, use Fuseki 4.x wget archive.apache.org unzip apache-jena-fuseki-4.8.0.zip cd apache-jena-fuseki-4.8.0 ./fuseki-server --mem /humanitarian ``` Or …