Add ExaDocumentLoader module to langchain4j-community

[harikrishna553]

Add ExaDocumentLoader module to langchain4j-community

Summary

This PR adds a new community module langchain4j-community-document-loader-exa to the LangChain4j ecosystem.
It provides a ready-to-use connector for the Exa Search API, allowing retrieval of web documents as Document objects enriched with metadata.

---

Features

• ExaDocumentLoader: Loads documents from Exa Search API based on a query.
• Search types via ExaSearchType:
  • AUTO – Balanced default mode combining quality and latency.
  • FAST – Optimized for low-latency, factual queries.
  • DEEP – Comprehensive multi-step and agentic retrieval.
  • NEURAL – Semantic/embedding-based exploratory search.
• Flexible content resolution:
  • Includes full text if available and enabled.
  • Falls back to highlights or title when full text is unavailable.
• Metadata extraction:
  • Title, URL, Exa result ID, publication date, author, relevance score.
• Builder pattern:
  • Configure API key, number of results, search type, text inclusion, custom HTTP client, and custom ObjectMapper.
• Consistent error handling:
  • All API and parsing errors are wrapped in ExaDocumentLoaderException.

---

Example Usage

ExaDocumentLoader loader = ExaDocumentLoader.builder()
    .apiKey("your-api-key")
    .numResults(5)
    .searchType(ExaSearchType.AUTO)
    .includeText(true)
    .build();

List<Document> documents = loader.loadDocuments("latest AI research");