Refactored components to work asynchronously and updated openai sdk version

[mateoperezrivera]

This pull request focuses on refactoring the codebase to use asynchronous operations, which should improve the performance and responsiveness of the application. The most important changes include converting synchronous Azure SDK calls to their asynchronous counterparts, updating the document analysis and chunking functions to be asynchronous, and refactoring the TextEmbedder and ChunkEmbeddingHelper classes to support asynchronous operations.

Asynchronous Azure SDK Calls:

• chunker/chunk_documents_docint.py: Replaced synchronous imports with asynchronous ones (DefaultAzureCredential, BlobServiceClient, SecretClient) and added aiohttp and asyncio imports.
• embedder/text_embedder.py: Replaced synchronous SecretClient and DefaultAzureCredential with their asynchronous versions and updated the TextEmbedder class to use AsyncAzureOpenAI.

Asynchronous Document Analysis:

• chunker/chunk_documents_docint.py: Converted get_secret, analyze_document_rest, get_chunk, and chunk_document functions to be asynchronous. [1] [2] [3] [4]

Asynchronous Chunking:

• chunker/chunk_documents_raw.py: Updated chunk_document function to be asynchronous and to use asynchronous ChunkEmbeddingHelper. [1] [2]

Refactoring Helper Classes:

• chunker/chunk_metadata_helper.py: Refactored ChunkEmbeddingHelper to support asynchronous operations, including an asynchronous create method and generate_chunks_with_embedding method.
• embedder/text_embedder.py: Refactored TextEmbedder to include asynchronous methods for embedding content and retrieving secrets. [1] [2]

These changes aim to improve the efficiency of document processing by leveraging asynchronous I/O operations