Create app.py

app.py

@@ -0,0 +1,47 @@
+import chromadb
+import pandas as pd
+import gradio as gr
+from llama_cpp import Llama
+
+# Load ChromaDB for symptom retrieval
+chroma_client = chromadb.PersistentClient(path="./chroma_db")
+collection = chroma_client.get_or_create_collection(name="symptom_tree")
+
+# Load and insert CSV data into ChromaDB
+csv_path = "/mnt/data/enhanced_symptom_tree_with_measures.csv"
+df = pd.read_csv(csv_path)
+for _, row in df.iterrows():
+    collection.add(
+        documents=[row["Primary Symptom"]],
+        metadatas={
+            "follow_up_questions": row["Follow-up Questions"],
+            "possible_diseases": row["Possible Diseases"],
+            "recommended_measures": row["Recommended Measures"]
+        },
+        ids=[row["Primary Symptom"]]
+    )
+
+# Load Mistral 7B model (GGUF format for low-resource deployment)
+llm = Llama(model_path="mistral-7b.Q4_K_M.gguf", n_ctx=2048)
+
+def chatbot(symptom: str):
+    # Retrieve symptom details from ChromaDB
+    results = collection.query(query_texts=[symptom], n_results=1)
+    if not results["documents"]:
+        return "I'm sorry, I couldn't find relevant information."
+    
+    metadata = results["metadatas"][0]
+    follow_up_qs = metadata["follow_up_questions"].split(";")
+    possible_diseases = metadata["possible_diseases"].split(";")
+    measures = metadata["recommended_measures"].split(";")
+    
+    # Generate chatbot response using Mistral 7B
+    prompt = f"You are a doctor. A patient reports {symptom}. Ask 2 follow-up questions: {follow_up_qs}"
+    response = llm(prompt)["choices"][0]["text"]
+    
+    return f"\n**Follow-up Questions:**\n {' '.join(follow_up_qs)}\n\n**Possible Diseases:**\n {' '.join(possible_diseases)}\n\n**Recommended Measures:**\n {' '.join(measures)}\n\n**Chatbot Response:**\n {response}"
+
+# Create Gradio UI
+demo = gr.Interface(fn=chatbot, inputs="text", outputs="markdown", title="Chest X-Ray Chatbot")
+
+demo.launch(share=True)