An alternative approach to parsing PDF tables for RAG, overcoming the limitations of highly normalized formats like CSV and JSON

Why This Approach?

Parsing Challenges: A Case Study with SEC FORM 10-Q

Complexity of Table Structure

Loss of Context

Handling Special Characters and Formats

Addressing the Challenge

The Benefits

Key Code Explanation

1. Extracting Text and Tables from PDFs

2. Cleaning and Sending Tables to Azure OpenAI

3. Writing the Final Output

Conclusion

From PDF Tables to Insights: An Alternative Approach for Parsing PDFs in RAG

Create avatar-assisted voice search experience by integrating speech-to-text, semantic search, RAG and a synthesized avatar for responses.

The evolution of search

The desire for dialogue in search

Beyond question and answer search

Avatar assisted voice search experience

Live demo: creating an avatar assisted voice search experience

Integration details

Speech to search

RAG & cache

Avatar response generation

Summary

Demo assets

AI voice search: Avatar assisted & dialogue driven voice search

Avatar assisted & dialogue driven voice to RAG search

Learn about data chunking in RAG and explore fetch surrounding chunking, a pattern in RAG that uses chunking and Elasticsearch to refine LLM responses.

The motivation: A refined approach to RAG data chunking

The methodology: Intelligent RAG data chunking

Fetch surrounding chunks

Assigning numbers to chunks

Vector database in Elasticsearch

AI search

The RAG pattern

Why intelligent RAG data chunking matters

RAG chunking: Fetch surrounding chunks to refine LLM responses

Intelligent RAG data chunking: Fetch surrounding chunks

Author

Sunile Manjee

Articles

From PDF Tables to Insights: An Alternative Approach for Parsing PDFs in RAG

Avatar assisted & dialogue driven voice to RAG search

Intelligent RAG data chunking: Fetch surrounding chunks