Full Summary:

A multilingual AI-powered semantic search engine for Japanese literature, featuring intelligent context-aware search results and natural language understanding.

• Semantic Search: AI-powered understanding of queries in both Japanese and English
• Context-Aware Results: Each result includes intelligent summaries explaining relevance and context
• Multilingual Support: Search Japanese literature with queries in multiple languages
• Intuitive UI: Clean, scholarly interface inspired by traditional Japanese design
• Vector Database: Uses embeddings for sophisticated content matching beyond keyword search
• Scalable Architecture: Designed to grow from 5 to 100+ books seamlessly

• Backend: Java Spring Boot
• Frontend: Thymeleaf templates with responsive CSS/JavaScript
• Database: PostgreSQL (metadata & book information)
• Vector Search: Pinecone vector database
• AI Integration: OpenAI embeddings and ChatGPT for intelligent summaries
• Cloud: AWS EC2

Frontend (Thymeleaf) → Spring Boot API → Vector Search Service → PostgreSQL
                                      ↘ Pinecone Vector DB
                                      ↘ OpenAI API (embeddings + summaries)

• Java 17+
• Maven 3.6+
• PostgreSQL 12+
• API keys for:
• OpenAI (for embeddings and AI summaries)
• Pinecone (for vector search)

1. Clone the repository

   git clone https://github.com/yourusername/bungaku-kensaku.git
   cd bungaku-kensaku

    

2. Set up configuration

   # Copy example configuration
   cp src/main/resources/application-example.properties src/main/resources/application-local.properties
   
   # Edit application-local.properties with your actual credentials

    

3. Configure environment variables

   export OPENAI_API_KEY="your-openai-api-key"
   export PINECONE_API_KEY="your-pinecone-api-key"
   export DB_PASSWORD="your-database-password"

    

4. Set up PostgreSQL database

   CREATE DATABASE bungaku_kensaku_db;
   CREATE USER your_username WITH PASSWORD 'your_password';
   GRANT ALL PRIVILEGES ON DATABASE bungaku_kensaku_db TO your_username;

    

5. Run the application

   mvn spring-boot:run

    

6. Access the application

   • Navigate to http://localhost:8080
   • Default demo credentials:
     • Username: demo
     • Password: changeme
   • Or configure your own via environment variables:

     export DEMO_USERNAME=your-username
     export DEMO_PASSWORD=your-password

      

src/main/java/com/senseisearch/
├── controller/        # REST API endpoints
├── service/           # Business logic (search, AI, document processing)
├── repository/        # JPA data access layer
├── model/             # Entity classes
├── config/            # Configuration classes
└── util/              # Utility classes

src/main/resources/
├── templates/               # Thymeleaf HTML templates
├── static/                  # CSS, JavaScript, images
└── application*.properties  # Configuration files

1. PDF Upload: Documents are uploaded and stored in PostgreSQL
2. Text Extraction: Content is extracted and chunked (500 tokens with overlap)
3. Embedding Generation: Each chunk is converted to vector embeddings using OpenAI
4. Vector Storage: Embeddings are indexed in Pinecone for fast similarity search
5. Metadata Storage: Book information and chunk metadata stored in PostgreSQL

1. Query Processing: User query is converted to embeddings
2. Similarity Search: Pinecone finds most relevant text chunks
3. Context Assembly: Retrieved chunks are enriched with book/chapter context
4. AI Summary: OpenAI generates intelligent explanations for each result
5. Result Presentation: Clean, organized results with context and relevance explanations

• Academic Research: Deep exploration of philosophical and literary texts
• Study Groups: Finding relevant passages for discussion topics
• Cross-Reference Search: Discovering connections between different works
• Multilingual Access: Non-Japanese speakers accessing Japanese literature
• Contextual Learning: Understanding passages within their broader narrative context