# kallappa2005/docu-rag architecture

> Architecture diagram of the GitHub repository kallappa2005/docu-rag, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-02). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/kallappa2005/docu-rag
- Repository: https://github.com/kallappa2005/docu-rag

## Overview

This is a FastAPI document RAG chatbot: users upload PDFs and ask questions answered using retrieved document context and Groq, with in-memory conversation history. The sampled upload handler only saves files; the README describes a subsequent ingestion pipeline, but its runtime wiring is not shown. The chat path is explicit in source. Unsampled ingestion links below are documented, not verified calls.

Main entry points: [`app/services/rag_chain.py`](https://github.com/kallappa2005/docu-rag/blob/main/app/services/rag_chain.py) (RAG orchestration), [`app/routers/chat.py`](https://github.com/kallappa2005/docu-rag/blob/main/app/routers/chat.py) (Chat endpoint), [`app/main.py`](https://github.com/kallappa2005/docu-rag/blob/main/app/main.py) (FastAPI app), [`app/routers/upload.py`](https://github.com/kallappa2005/docu-rag/blob/main/app/routers/upload.py) (Upload endpoint)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_api["API"]
  node_api_app["FastAPI app<br/>[main.py]"]
  node_upload["Upload endpoint<br/>[upload.py]"]
  node_chat["Chat endpoint<br/>[chat.py]"]
  node_health["Health endpoint<br/>[health.py]"]
  node_upload_response["Upload response<br/>[response_models.py]"]
  node_chat_request["Chat request<br/>[request_models.py]"]
  node_chat_response["Chat response<br/>[response_models.py]"]
end

subgraph group_ingest["Document Ingestion"]
  node_pdf_save["PDF file saving<br/>[pdf_service.py]"]
  node_pdf_loader["PDF extraction<br/>[pdf_loader.py]"]
  node_splitter["Text chunking<br/>[text_splitter.py]"]
  node_embedding["HuggingFace embeddings"]
  node_vector_db[("Chroma vector store<br/>[vector_store.py]")]
end

subgraph group_rag["RAG Answering"]
  node_rag_chain["RAG orchestration<br/>[rag_chain.py]"]
  node_retriever["Semantic retrieval<br/>[retriever.py]"]
  node_prompt["RAG prompt<br/>[prompt.py]"]
  node_groq["Groq LLM client<br/>[groq_service.py]"]
  node_memory[("Chat memory<br/>[chat_memory.py]")]
end

subgraph group_support["Configuration and State"]
  node_settings["Environment settings<br/>[settings.py]"]
end

node_user(("API user"))
node_health_user(("Health-check caller"))

node_user -->|"sends requests"| node_api_app
node_api_app -->|"registers route"| node_upload
node_api_app -->|"registers route"| node_chat
node_api_app -->|"registers route"| node_health
node_user -->|"uploads PDF"| node_upload
node_upload -->|"saves file"| node_pdf_save
node_upload -->|"returns response"| node_upload_response
node_user -->|"asks question"| node_chat
node_chat -->|"accepts request"| node_chat_request
node_chat -->|"asks question"| node_rag_chain
node_chat -->|"returns answer"| node_chat_response
node_rag_chain -->|"retrieves chunks"| node_retriever
node_retriever -->|"searches vectors"| node_vector_db
node_rag_chain -->|"reads and updates"| node_memory
node_rag_chain -->|"formats context"| node_prompt
node_rag_chain -->|"invokes model"| node_groq
node_health_user -->|"checks status"| node_health
node_pdf_save -.->|"loads saved PDF"| node_pdf_loader
node_pdf_loader -.->|"extracts for chunking"| node_splitter
node_splitter -.->|"embeds chunks"| node_embedding
node_embedding -.->|"stores vectors"| node_vector_db
node_vector_db -.->|"serves similarity search"| node_retriever

click node_api_app "https://github.com/kallappa2005/docu-rag/blob/main/app/main.py"
click node_upload "https://github.com/kallappa2005/docu-rag/blob/main/app/routers/upload.py"
click node_chat "https://github.com/kallappa2005/docu-rag/blob/main/app/routers/chat.py"
click node_health "https://github.com/kallappa2005/docu-rag/blob/main/app/routers/health.py"
click node_upload_response "https://github.com/kallappa2005/docu-rag/blob/main/app/models/response_models.py"
click node_chat_request "https://github.com/kallappa2005/docu-rag/blob/main/app/models/request_models.py"
click node_chat_response "https://github.com/kallappa2005/docu-rag/blob/main/app/models/response_models.py"
click node_pdf_save "https://github.com/kallappa2005/docu-rag/blob/main/app/services/pdf_service.py"
click node_pdf_loader "https://github.com/kallappa2005/docu-rag/blob/main/app/services/pdf_loader.py"
click node_splitter "https://github.com/kallappa2005/docu-rag/blob/main/app/services/text_splitter.py"
click node_embedding "https://github.com/kallappa2005/docu-rag/blob/main/app/services/embedding_service.py"
click node_vector_db "https://github.com/kallappa2005/docu-rag/blob/main/app/services/vector_store.py"
click node_rag_chain "https://github.com/kallappa2005/docu-rag/blob/main/app/services/rag_chain.py"
click node_retriever "https://github.com/kallappa2005/docu-rag/blob/main/app/services/retriever.py"
click node_prompt "https://github.com/kallappa2005/docu-rag/blob/main/app/prompts/prompt.py"
click node_groq "https://github.com/kallappa2005/docu-rag/blob/main/app/services/groq_service.py"
click node_memory "https://github.com/kallappa2005/docu-rag/blob/main/app/memory/chat_memory.py"
click node_settings "https://github.com/kallappa2005/docu-rag/blob/main/app/config/settings.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_api_app,node_upload,node_chat,node_health,node_upload_response,node_chat_request,node_chat_response,node_user toneBlue
class node_pdf_save,node_pdf_loader,node_splitter,node_embedding,node_vector_db toneAmber
class node_rag_chain,node_retriever,node_prompt,node_groq,node_memory toneMint
class node_settings toneRose
class node_health_user toneIndigo
```

## Components

### API

- [FastAPI app](https://github.com/kallappa2005/docu-rag/blob/main/app/main.py): `app/main.py`
- [Upload endpoint](https://github.com/kallappa2005/docu-rag/blob/main/app/routers/upload.py): `app/routers/upload.py`
- [Chat endpoint](https://github.com/kallappa2005/docu-rag/blob/main/app/routers/chat.py): `app/routers/chat.py`
- [Health endpoint](https://github.com/kallappa2005/docu-rag/blob/main/app/routers/health.py): `app/routers/health.py`
- [Upload response](https://github.com/kallappa2005/docu-rag/blob/main/app/models/response_models.py): `app/models/response_models.py`
- [Chat request](https://github.com/kallappa2005/docu-rag/blob/main/app/models/request_models.py): `app/models/request_models.py`
- [Chat response](https://github.com/kallappa2005/docu-rag/blob/main/app/models/response_models.py): `app/models/response_models.py`

### Document Ingestion

- [PDF file saving](https://github.com/kallappa2005/docu-rag/blob/main/app/services/pdf_service.py): `app/services/pdf_service.py`
- [PDF extraction](https://github.com/kallappa2005/docu-rag/blob/main/app/services/pdf_loader.py): `app/services/pdf_loader.py`
- [Text chunking](https://github.com/kallappa2005/docu-rag/blob/main/app/services/text_splitter.py): `app/services/text_splitter.py`
- [HuggingFace embeddings](https://github.com/kallappa2005/docu-rag/blob/main/app/services/embedding_service.py): `app/services/embedding_service.py`
- [Chroma vector store](https://github.com/kallappa2005/docu-rag/blob/main/app/services/vector_store.py): `app/services/vector_store.py`

### RAG Answering

- [RAG orchestration](https://github.com/kallappa2005/docu-rag/blob/main/app/services/rag_chain.py): `app/services/rag_chain.py`
- [Semantic retrieval](https://github.com/kallappa2005/docu-rag/blob/main/app/services/retriever.py): `app/services/retriever.py`
- [RAG prompt](https://github.com/kallappa2005/docu-rag/blob/main/app/prompts/prompt.py): `app/prompts/prompt.py`
- [Groq LLM client](https://github.com/kallappa2005/docu-rag/blob/main/app/services/groq_service.py): `app/services/groq_service.py`
- [Chat memory](https://github.com/kallappa2005/docu-rag/blob/main/app/memory/chat_memory.py): `app/memory/chat_memory.py`

### Configuration and State

- [Environment settings](https://github.com/kallappa2005/docu-rag/blob/main/app/config/settings.py): `app/config/settings.py`

### Other components

- **API user**
- **Health-check caller**

## Connections

- API user → FastAPI app: sends requests (evidence: `README.md`)
- FastAPI app → Upload endpoint: registers route (evidence: `app/main.py`)
- FastAPI app → Chat endpoint: registers route (evidence: `app/main.py`)
- FastAPI app → Health endpoint: registers route (evidence: `app/main.py`)
- API user → Upload endpoint: uploads PDF (evidence: `README.md`)
- Upload endpoint → PDF file saving: saves file (evidence: `app/routers/upload.py`)
- Upload endpoint → Upload response: returns response (evidence: `app/routers/upload.py`)
- API user → Chat endpoint: asks question (evidence: `README.md`)
- Chat endpoint → Chat request: accepts request (evidence: `app/routers/chat.py`)
- Chat endpoint → RAG orchestration: asks question (evidence: `app/routers/chat.py`)
- Chat endpoint → Chat response: returns answer (evidence: `app/routers/chat.py`)
- RAG orchestration → Semantic retrieval: retrieves chunks (evidence: `app/services/rag_chain.py`)
- Semantic retrieval → Chroma vector store: searches vectors (evidence: `app/services/retriever.py`)
- RAG orchestration → Chat memory: reads and updates (evidence: `app/services/rag_chain.py`)
- RAG orchestration → RAG prompt: formats context (evidence: `app/services/rag_chain.py`)
- RAG orchestration → Groq LLM client: invokes model (evidence: `app/services/rag_chain.py`)
- Health-check caller → Health endpoint: checks status (evidence: `README.md`)
- PDF file saving → PDF extraction: loads saved PDF (evidence: `README.md`)
- PDF extraction → Text chunking: extracts for chunking (evidence: `README.md`)
- Text chunking → HuggingFace embeddings: embeds chunks (evidence: `README.md`)
- HuggingFace embeddings → Chroma vector store: stores vectors (evidence: `README.md`)
- Chroma vector store → Semantic retrieval: serves similarity search (evidence: `README.md`)
