How rajeshthappeta/mern-ai-integration-patterns works
This repository documents a progression of MERN AI patterns across separate branches, not a single implementation visible in the supplied tree. The graph captures the README’s ingestion and RAG workflows plus its conversation, tool-calling, and agent capabilities. All repository-node paths point to the README, the only product source provided; the actual branch code and runtime wiring could not be verified.
Main entry points: README.md Validate and save
Components 19
Ingestion
- Article input
README.md - PDF upload
README.md - Validate and save
README.md - Articles collection
README.md - Clean and chunk
README.md - Embedding service
README.md - Chunks collection
README.md
Retrieval and generation
- Question
README.md - Filtered vector search
README.md - Atlas vector index
README.md - Context assembly
README.md - LLM generation
README.md - Answer
README.md
Conversation memory
- Standalone query rewrite
README.md - Chat history
README.md
Tools and agents
- Tool selection
README.md - Tool handlers
README.md - Agent loop
README.md
Other components
- User
Connections 23
- User → Article input: submits article
README.md - User → PDF upload: uploads PDF
README.md - Article input → Validate and save: submits to
README.md - PDF upload → Validate and save: submits to
README.md - Validate and save → Articles collection: saves article
README.md - Validate and save → Clean and chunk: sends text and ID
README.md - Clean and chunk → Embedding service: passes chunks
README.md - Embedding service → Chunks collection: saves vectors
README.md - User → Question: asks question
README.md - Question → Standalone query rewrite: rewrites follow-up
README.md - Chat history → Standalone query rewrite: provides history
README.md - Standalone query rewrite → Embedding service: embeds query
README.md - Embedding service → Filtered vector search: supplies embedding
README.md - Chunks collection → Filtered vector search: searches chunks
README.md - Atlas vector index → Filtered vector search: indexes search
README.md - Filtered vector search → Context assembly: returns top chunks
README.md - Context assembly → LLM generation: supplies context
README.md - LLM generation → Answer: generates answer
README.md - Answer → User: returns answer
README.md - LLM generation → Tool selection: chooses tool
README.md - Tool selection → Tool handlers: dispatches tool
README.md - Agent loop → Tool selection: repeats decisions
README.md - Tool handlers → Agent loop: returns tool result
README.md