# imprintlab/medical-graph-rag architecture

> Architecture diagram of the GitHub repository imprintlab/medical-graph-rag, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-09-30). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/imprintlab/medical-graph-rag
- Repository: https://github.com/imprintlab/medical-graph-rag

## Overview

Medical-Graph-RAG turns medical documents into a searchable graph and uses retrieval to answer a clinical question; the README also describes hierarchical graph linking and a simpler vector-RAG baseline. The graph highlights the documented build and inference stages, plus nano_graphrag’s internal graph, storage, and LLM layers. The README establishes the product workflow; source excerpts and the SOURCE INDEX support the internal nano_graphrag and selected CAMEL relationships. Many listed CAMEL modules are framework capabilities, and their runtime wiring into this application is not shown.

Main entry points: [`run.py`](https://github.com/imprintlab/medical-graph-rag/blob/main/run.py) (Run Workflow), [`nano_graphrag/graphrag.py`](https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/graphrag.py) (GraphRAG Runtime), [`retrieve.py`](https://github.com/imprintlab/medical-graph-rag/blob/main/retrieve.py) (Medical Retrieval), [`camel/workforce/manager_node.py`](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/workforce/manager_node.py) (Workforce Manager)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_app["Medical RAG"]
  node_run["Run Workflow<br/>[run.py]"]
  node_load["Medical Data Load<br/>[dataloader.py]"]
  node_chunk["Document Chunking<br/>[data_chunk.py]"]
  node_agentchunk["Agentic Chunking<br/>[agentic_chunker.py]"]
  node_construct["Graph Construction<br/>[creat_graph.py]"]
  node_merge["Graph Cleanup<br/>[cleangraph.py]"]
  node_hierarchy["Graph Linking"]
  node_retrieval["Medical Retrieval<br/>[retrieve.py]"]
  node_answer["Answer Synthesis<br/>[run.py]"]
end

subgraph group_nano["Graph RAG Engine"]
  node_nanoengine["GraphRAG Runtime<br/>[graphrag.py]"]
  node_nanoops["Graph Operations<br/>[_op.py]"]
  node_nanostore[("Graph Storage<br/>[_storage.py]")]
  node_nanollm["LLM Adapter<br/>[_llm.py]"]
  node_nanobase["Engine Base<br/>[base.py]"]
end

subgraph group_camel["CAMEL Framework"]
  node_camelagent["Chat Agent<br/>[chat_agent.py]"]
  node_cameltask["Task Model<br/>[task.py]"]
  node_workmanager["Workforce Manager<br/>[manager_node.py]"]
  node_worker["Workforce Worker<br/>[worker_node.py]"]
  node_taskchannel["Task Channel<br/>[task_channel.py]"]
  node_retriever["Vector Retriever"]
  node_embed["Embeddings<br/>[base.py]"]
  node_loader["Document Loader<br/>[unstructured_io.py]"]
end

subgraph group_data["Data And Storage"]
  node_neo4j[("Neo4j Graph Store<br/>[neo4j_graph.py]")]
  node_jsonstore[("JSON Key Store<br/>[json.py]")]
end

node_user(("Questioner"))
node_external{{"LLM Service"}}
node_pubmed{{"PubMed Search"}}

node_user -->|"starts"| node_run
node_run -.->|"reads data"| node_load
node_run -->|"chunks data"| node_chunk
node_run -->|"builds graph"| node_construct
node_run -->|"merges graph"| node_merge
node_run -->|"links levels"| node_hierarchy
node_run -->|"retrieves evidence"| node_retrieval
node_retrieval -.->|"supplies context"| node_answer
node_run -->|"runs inference"| node_answer
node_run -.->|"uses runtime"| node_nanoengine
node_nanoengine -->|"runs operations"| node_nanoops
node_nanoengine -->|"uses storage"| node_nanostore
node_nanoengine -->|"calls LLM"| node_nanollm
node_nanoengine -->|"uses base"| node_nanobase
node_nanostore -->|"uses key store"| node_jsonstore
node_nanollm -.->|"requests completion"| node_external
node_run -.->|"searches papers"| node_pubmed
node_retriever -->|"embeds query"| node_embed
node_retriever -->|"processes documents"| node_loader
node_workmanager -->|"creates coordinator"| node_camelagent
node_workmanager -->|"posts tasks"| node_taskchannel
node_worker -->|"gets assigned tasks"| node_taskchannel
node_worker -->|"processes tasks"| node_cameltask
node_retrieval -.->|"queries graph"| node_neo4j

click node_run "https://github.com/imprintlab/medical-graph-rag/blob/main/run.py"
click node_load "https://github.com/imprintlab/medical-graph-rag/blob/main/dataloader.py"
click node_chunk "https://github.com/imprintlab/medical-graph-rag/blob/main/data_chunk.py"
click node_agentchunk "https://github.com/imprintlab/medical-graph-rag/blob/main/agentic_chunker.py"
click node_construct "https://github.com/imprintlab/medical-graph-rag/blob/main/creat_graph.py"
click node_merge "https://github.com/imprintlab/medical-graph-rag/blob/main/cleangraph.py"
click node_hierarchy "https://github.com/imprintlab/medical-graph-rag/blob/main/three_layer_import.py"
click node_retrieval "https://github.com/imprintlab/medical-graph-rag/blob/main/retrieve.py"
click node_answer "https://github.com/imprintlab/medical-graph-rag/blob/main/run.py"
click node_nanoengine "https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/graphrag.py"
click node_nanoops "https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/_op.py"
click node_nanostore "https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/_storage.py"
click node_nanollm "https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/_llm.py"
click node_nanobase "https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/base.py"
click node_camelagent "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/agents/chat_agent.py"
click node_cameltask "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/tasks/task.py"
click node_workmanager "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/workforce/manager_node.py"
click node_worker "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/workforce/worker_node.py"
click node_taskchannel "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/workforce/task_channel.py"
click node_retriever "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/retrievers/vector_retriever.py"
click node_embed "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/embeddings/base.py"
click node_loader "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/loaders/unstructured_io.py"
click node_neo4j "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/storages/graph_storages/neo4j_graph.py"
click node_jsonstore "https://github.com/imprintlab/medical-graph-rag/blob/main/camel/storages/key_value_storages/json.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_run,node_load,node_chunk,node_agentchunk,node_construct,node_merge,node_hierarchy,node_retrieval,node_answer toneBlue
class node_nanoengine,node_nanoops,node_nanostore,node_nanollm,node_nanobase toneAmber
class node_camelagent,node_cameltask,node_workmanager,node_worker,node_taskchannel,node_retriever,node_embed,node_loader toneMint
class node_neo4j,node_jsonstore toneRose
class node_user,node_external,node_pubmed toneIndigo
```

## Components

### Medical RAG

- [Run Workflow](https://github.com/imprintlab/medical-graph-rag/blob/main/run.py): `run.py`
- [Medical Data Load](https://github.com/imprintlab/medical-graph-rag/blob/main/dataloader.py): `dataloader.py`
- [Document Chunking](https://github.com/imprintlab/medical-graph-rag/blob/main/data_chunk.py): `data_chunk.py`
- [Agentic Chunking](https://github.com/imprintlab/medical-graph-rag/blob/main/agentic_chunker.py): `agentic_chunker.py`
- [Graph Construction](https://github.com/imprintlab/medical-graph-rag/blob/main/creat_graph.py): `creat_graph.py`
- [Graph Cleanup](https://github.com/imprintlab/medical-graph-rag/blob/main/cleangraph.py): `cleangraph.py`
- [Graph Linking](https://github.com/imprintlab/medical-graph-rag/blob/main/three_layer_import.py): `three_layer_import.py`
- [Medical Retrieval](https://github.com/imprintlab/medical-graph-rag/blob/main/retrieve.py): `retrieve.py`
- [Answer Synthesis](https://github.com/imprintlab/medical-graph-rag/blob/main/run.py): `run.py`

### Graph RAG Engine

- [GraphRAG Runtime](https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/graphrag.py): `nano_graphrag/graphrag.py`
- [Graph Operations](https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/_op.py): `nano_graphrag/_op.py`
- [Graph Storage](https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/_storage.py): `nano_graphrag/_storage.py`
- [LLM Adapter](https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/_llm.py): `nano_graphrag/_llm.py`
- [Engine Base](https://github.com/imprintlab/medical-graph-rag/blob/main/nano_graphrag/base.py): `nano_graphrag/base.py`

### CAMEL Framework

- [Chat Agent](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/agents/chat_agent.py): `camel/agents/chat_agent.py`
- [Task Model](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/tasks/task.py): `camel/tasks/task.py`
- [Workforce Manager](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/workforce/manager_node.py): `camel/workforce/manager_node.py`
- [Workforce Worker](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/workforce/worker_node.py): `camel/workforce/worker_node.py`
- [Task Channel](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/workforce/task_channel.py): `camel/workforce/task_channel.py`
- [Vector Retriever](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/retrievers/vector_retriever.py): `camel/retrievers/vector_retriever.py`
- [Embeddings](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/embeddings/base.py): `camel/embeddings/base.py`
- [Document Loader](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/loaders/unstructured_io.py): `camel/loaders/unstructured_io.py`

### Data And Storage

- [Neo4j Graph Store](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/storages/graph_storages/neo4j_graph.py): `camel/storages/graph_storages/neo4j_graph.py`
- [JSON Key Store](https://github.com/imprintlab/medical-graph-rag/blob/main/camel/storages/key_value_storages/json.py): `camel/storages/key_value_storages/json.py`

### Other components

- **Questioner**
- **LLM Service**
- **PubMed Search**

## Connections

- Questioner → Run Workflow: starts
- Run Workflow → Medical Data Load: reads data
- Run Workflow → Document Chunking: chunks data
- Run Workflow → Graph Construction: builds graph
- Run Workflow → Graph Cleanup: merges graph
- Run Workflow → Graph Linking: links levels
- Run Workflow → Medical Retrieval: retrieves evidence
- Medical Retrieval → Answer Synthesis: supplies context
- Run Workflow → Answer Synthesis: runs inference
- Run Workflow → GraphRAG Runtime: uses runtime
- GraphRAG Runtime → Graph Operations: runs operations (evidence: `nano_graphrag/graphrag.py`)
- GraphRAG Runtime → Graph Storage: uses storage (evidence: `nano_graphrag/graphrag.py`)
- GraphRAG Runtime → LLM Adapter: calls LLM (evidence: `nano_graphrag/graphrag.py`)
- GraphRAG Runtime → Engine Base: uses base (evidence: `nano_graphrag/graphrag.py`)
- Graph Storage → JSON Key Store: uses key store (evidence: `nano_graphrag/_storage.py`)
- LLM Adapter → LLM Service: requests completion
- Run Workflow → PubMed Search: searches papers
- Vector Retriever → Embeddings: embeds query (evidence: `camel/retrievers/vector_retriever.py`)
- Vector Retriever → Document Loader: processes documents (evidence: `camel/retrievers/vector_retriever.py`)
- Workforce Manager → Chat Agent: creates coordinator (evidence: `camel/workforce/manager_node.py`)
- Workforce Manager → Task Channel: posts tasks (evidence: `camel/workforce/manager_node.py`)
- Workforce Worker → Task Channel: gets assigned tasks (evidence: `camel/workforce/worker_node.py`)
- Workforce Worker → Task Model: processes tasks (evidence: `camel/workforce/worker_node.py`)
- Medical Retrieval → Neo4j Graph Store: queries graph
