Added an async client, table extraction from markdown to list of list with headers

[AswanthManoj]

PR: Add AsyncOmniParse Client, Table Extraction, and Pydantic Models

This PR introduces several enhancements to our OmniParse project:

1. AsyncOmniParse Client:

   • Implemented a new asynchronous client class AsyncOmniParse for interacting with the OmniParse server.
   • Added methods for parsing various types of documents (PDF, PPT, DOCX), images, videos, audio, and websites.
   • Included error handling for unsupported file types and API request failures.
   • Added comprehensive docstrings with usage examples for better developer experience.

2. Table Extraction from Markdown:

   • Introduced functions to extract tables from markdown content:
     • extract_markdown_tables: Extracts raw markdown tables from a string.
     • markdown_to_tables: Converts markdown tables to a list of TableObj instances.
   • Tables are now represented as list of lists with separate headers, improving data structure for further processing.

3. Pydantic Models:

   • Added several Pydantic models to ensure type safety and data validation:
     • ImageObj: Represents extracted images with name, binary data, and MIME type.
     • TableObj: Represents extracted tables with headers and data.
     • MetaData: Stores metadata about parsed documents.
     • ParsedDocument: Represents the complete parsed document with markdown content, images, tables, and metadata.

These additions provide a robust, type-safe, and asynchronous interface for interacting with OmniParse.

[adithya-s-k]

@AswanthManoj Thank you very much ⭐for contributing

will review the PR ,merge it and use it as a base for the the SDK