ManualTrigger Automate

This workflow is ideal for:

• Data Scientists looking to automate the process of document analysis and citation extraction.
• Researchers who need to interact with large documents and retrieve specific information efficiently.
• Developers seeking to integrate AI chat capabilities into their applications using Google Drive and vector databases.
• Educators wanting to create interactive learning tools that leverage document-based queries.

This workflow addresses the challenge of efficiently extracting relevant information from large documents. By integrating with Google Drive and utilizing LangChain's capabilities, it allows users to:

• Download documents from Google Drive.
• Split documents into manageable chunks for easier processing.
• Store and retrieve relevant chunks using a vector database (Pinecone).
• Answer queries based on the content of the documents, providing citations for references.

1. Trigger the Workflow: The workflow begins when the user clicks the "Execute Workflow" button.
2. Set File URL: The workflow sets the URL of the file to be downloaded from Google Drive.
3. Download File: The specified file is downloaded from Google Drive.
4. Load Document: The downloaded file is loaded as a binary document for processing.
5. Chunking: The document is split into smaller, manageable chunks using a recursive character text splitter.
6. Embedding Creation: Each chunk is converted into embeddings using OpenAI's API for further processing.
7. Store in Vector Database: The embeddings are stored in a Pinecone vector database for efficient retrieval.
8. Chat Trigger: The workflow listens for incoming chat messages, which will trigger the query process.
9. Query Processing: When a message is received, the workflow determines how many chunks to send for processing.
10. Retrieve Relevant Chunks: The workflow fetches the top chunks that match the user's query from the vector database.
11. Prepare Context: The relevant chunks are formatted for the AI model to understand.
12. Answer Generation: The AI model generates an answer based on the context provided, including citations from the document.
13. Compose Final Response: The final response is composed, including the answer and the citations for reference.