feat: Improve Easy RAG module with complete implementation

langchain4j-easy-rag-spring-boot-starter/README.md

@@ -0,0 +1,224 @@
+# LangChain4j Easy RAG Spring Boot Starter
+
+This Spring Boot starter provides easy integration for Retrieval-Augmented Generation (RAG) functionality using LangChain4j.
+
+## Features
+
+- 🚀 **Auto-configuration** - Automatic setup of RAG components
+- 📄 **Flexible Document Loading** - Support for single files, directories, and glob patterns
+- 🔄 **Recursive Traversal** - Load documents from nested directory structures
+- 🎯 **Glob Pattern Matching** - Filter documents by file patterns (e.g., `*.txt`, `*.md`)
+- 🧩 **Customizable Components** - Override default beans for custom behavior
+- 📦 **Embedded Models** - Includes BGE Small quantized embedding model by default
+
+## Quick Start
+
+### 1. Add Dependency
+
+```xml
+<dependency>
+    <groupId>dev.langchain4j</groupId>
+    <artifactId>langchain4j-easy-rag-spring-boot-starter</artifactId>
+    <version>1.10.0-beta18-SNAPSHOT</version>
+</dependency>
+```
+
+### 2. Configure Application Properties
+
+```yaml
+langchain4j:
+  easy-rag:
+    ingestion:
+      documents:
+        path: /path/to/your/documents
+```
+
+### 3. Provide an Embedding Store Bean
+
+```java
+@Configuration
+public class RagConfiguration {
+
+    @Bean
+    public EmbeddingStore<TextSegment> embeddingStore() {
+        return new InMemoryEmbeddingStore<>();
+    }
+}
+```
+
+That's it! Documents will be automatically loaded and ingested on application startup.
+
+## Configuration Options
+
+### Basic Configuration
+
+```yaml
+langchain4j:
+  easy-rag:
+    ingestion:
+      documents:
+        # Path to documents (file or directory)
+        path: /path/to/documents
+
+        # Enable recursive directory traversal (default: false)
+        recursion: true
+
+        # Glob pattern for file filtering (e.g., *.txt, *.md)
+        glob: "*.txt"
+```
+
+## Usage Examples
+
+### Example 1: Load Single File
+
+```yaml
+langchain4j:
+  easy-rag:
+    ingestion:
+      documents:
+        path: /data/document.txt
+```
+
+### Example 2: Load All Files in Directory
+
+```yaml
+langchain4j:
+  easy-rag:
+    ingestion:
+      documents:
+        path: /data/documents
+```
+
+### Example 3: Load Files Recursively
+
+```yaml
+langchain4j:
+  easy-rag:
+    ingestion:
+      documents:
+        path: /data/documents
+        recursion: true
+```
+
+### Example 4: Load Only Text Files
+
+```yaml
+langchain4j:
+  easy-rag:
+    ingestion:
+      documents:
+        path: /data/documents
+        glob: "*.txt"
+```
+
+### Example 5: Load Markdown Files Recursively
+
+```yaml
+langchain4j:
+  easy-rag:
+    ingestion:
+      documents:
+        path: /data/docs
+        glob: "*.md"
+        recursion: true
+```
+
+## Customization
+
+### Custom Embedding Model
+
+Override the default embedding model by providing your own bean:
+
+```java
+@Configuration
+public class CustomRagConfiguration {
+
+    @Bean
+    public EmbeddingModel easyRagEmbeddingModel() {
+        return OpenAiEmbeddingModel.builder()
+                .apiKey(System.getenv("OPENAI_API_KEY"))
+                .modelName("text-embedding-ada-002")
+                .build();
+    }
+}
+```
+
+### Custom Document Splitter
+
+Configure custom splitting behavior:
+
+```java
+@Configuration
+public class CustomRagConfiguration {
+
+    @Bean
+    public DocumentSplitter easyRagDocumentSplitter(TokenCountEstimator estimator) {
+        return DocumentSplitters.recursive(
+                500,  // Max chunk size
+                50,   // Overlap
+                estimator
+        );
+    }
+}
+```
+
+### Custom Embedding Store
+
+Use a persistent embedding store:
+
+```java
+@Configuration
+public class CustomRagConfiguration {
+
+    @Bean
+    public EmbeddingStore<TextSegment> embeddingStore() {
+        return ElasticsearchEmbeddingStore.builder()
+                .serverUrl("http://localhost:9200")
+                .indexName("documents")
+                .build();
+    }
+}
+```
+
+## Default Components
+
+The starter provides the following default beans (can be overridden):
+
+| Bean Name | Type | Description |
+|-----------|------|-------------|
+| `easyRagEmbeddingModel` | `EmbeddingModel` | BGE Small quantized model |
+| `easyRagTokenCountEstimator` | `TokenCountEstimator` | HuggingFace token estimator |
+| `easyRagDocumentSplitter` | `DocumentSplitter` | Recursive splitter (300 tokens, 30 overlap) |
+
+## Supported File Types
+
+The document loader supports various file formats through LangChain4j's `FileSystemDocumentLoader`:
+
+- Plain text (`.txt`)
+- Markdown (`.md`)
+- PDF (`.pdf`)
+- Microsoft Word (`.doc`, `.docx`)
+- And more...
+
+## How It Works
+
+1. **Startup**: On application startup, the `easyRagDocumentIngestor` runner is executed
+2. **Loading**: Documents are loaded based on the configured path, glob pattern, and recursion settings
+3. **Splitting**: Documents are split into chunks using the configured `DocumentSplitter`
+4. **Embedding**: Text chunks are converted to embeddings using the configured `EmbeddingModel`
+5. **Storage**: Embeddings are stored in the configured `EmbeddingStore`
+
+## Requirements
+
+- Java 17 or higher
+- Spring Boot 3.3.8 or higher
+
+## Related Documentation
+
+- [LangChain4j Documentation](https://docs.langchain4j.dev/)
+- [Spring Boot Integration Guide](https://docs.langchain4j.dev/tutorials/spring-boot-integration)
+
+## License
+
+This project is licensed under the Apache License 2.0 - see the [LICENSE](../../LICENSE) file for details.
+

langchain4j-easy-rag-spring-boot-starter/pom.xml

@@ -7,7 +7,7 @@
     <parent>
         <groupId>dev.langchain4j</groupId>
         <artifactId>langchain4j-spring</artifactId>
-        <version>1.8.0-beta15-SNAPSHOT</version>
+        <version>1.10.0-beta18-SNAPSHOT</version>
         <relativePath>../pom.xml</relativePath>
     </parent>