llama 3

app.py

@@ -1,64 +1,105 @@
-import gradio as gr
-from huggingface_hub import InferenceClient
-
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
-
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
-
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
-
-
-if __name__ == "__main__":
-    demo.launch()
+import os
+import json
+import numpy as np
+import faiss
+import gradio as gr
+from sentence_transformers import SentenceTransformer
+from groq import Groq
+
+# Load FAISS index
+FAISS_INDEX_PATH = "faiss_medical.index"
+index = faiss.read_index(FAISS_INDEX_PATH)
+
+# Load embedding model
+embed_model = SentenceTransformer("all-MiniLM-L6-v2")
+
+# Load FAISS ID → Text Mapping
+with open("id_to_text.json", "r") as f:
+    id_to_text = json.load(f)
+
+# Convert JSON keys to integers (FAISS returns int IDs)
+id_to_text = {int(k): v for k, v in id_to_text.items()}
+
+# Initialize Groq client
+client = Groq(api_key=os.getenv("GROQ_API_KEY"))
+
+def retrieve_medical_summary(query, k=3):
+    """
+    Retrieve the most relevant medical literature from FAISS.
+
+    Args:
+        query (str): The medical question.
+        k (int, optional): Number of closest documents to retrieve. Defaults to 3.
+
+    Returns:
+        str: The most relevant retrieved medical documents.
+    """
+    # Convert query to embedding
+    query_embedding = embed_model.encode([query])
+
+    # Perform FAISS search
+    D, I = index.search(np.array(query_embedding).astype("float32"), k)
+
+    # Retrieve the closest matching text using FAISS index IDs
+    retrieved_docs = [id_to_text.get(int(idx), "No relevant data found.") for idx in I[0]]
+
+    # Ensure all retrieved texts are strings (Flatten lists if needed)
+    retrieved_docs = [doc if isinstance(doc, str) else " ".join(doc) for doc in retrieved_docs]
+
+    # Join multiple retrieved documents into one response
+    return "\n\n---\n\n".join(retrieved_docs) if retrieved_docs else "No relevant data found."
+
+def generate_medical_answer_groq(query, model="llama-3.3-70b-versatile", max_tokens=500, temperature=0.3):
+    """
+    Generates a medical response using Groq's API with LLaMA 3.3-70B, after retrieving relevant literature from FAISS.
+
+    Args:
+        query (str): The patient's medical question.
+        model (str, optional): The model to use. Defaults to "llama-3.3-70b-versatile".
+        max_tokens (int, optional): Max number of tokens to generate. Defaults to 200.
+        temperature (float, optional): Sampling temperature (higher = more creative). Defaults to 0.7.
+
+    Returns:
+        str: The AI-generated medical advice.
+    """
+
+    # Retrieve relevant medical literature from FAISS
+    retrieved_summary = retrieve_medical_summary(query)
+    print("\n🔍 Retrieved Medical Text for Query:", query)
+    print(retrieved_summary, "\n")
+
+    if not retrieved_summary or retrieved_summary == "No relevant data found.":
+        return "No relevant medical data found. Please consult a healthcare professional."
+
+    try:
+        # Send request to Groq API
+        response = client.chat.completions.create(
+            model=model,
+            messages=[
+                {"role": "system", "content": "You are an expert AI specializing in medical knowledge."},
+                {"role": "user", "content": f"Summarize the following medical literature and provide a structured medical answer:\n\n### Medical Literature ###\n{retrieved_summary}\n\n### Patient Question ###\n{query}\n\n### Medical Advice ###"}
+            ],
+            max_tokens=max_tokens,
+            temperature=temperature
+        )
+
+        return response.choices[0].message.content.strip()  # Ensure clean output
+
+    except Exception as e:
+        return f"Error generating response: {str(e)}"
+
+# Gradio Interface
+def ask_medical_question(question):
+    return generate_medical_answer_groq(question)
+
+# Create Gradio Interface
+iface = gr.Interface(
+    fn=ask_medical_question,
+    inputs=gr.Textbox(lines=2, placeholder="Enter your medical question here..."),
+    outputs=gr.Textbox(lines=10, placeholder="AI-generated medical advice will appear here..."),
+    title="Medical Question Answering System",
+    description="Ask any medical question, and the AI will provide an answer based on medical literature."
+)
+
+# Launch the Gradio app
+iface.launch()

requirements.txt

@@ -1 +1,7 @@
-huggingface_hub==0.25.2
+sentence-transformers
+faiss-cpu
+groq
+gradio
+numpy
+nltk
+shutil