|
|
import os |
|
|
import requests |
|
|
from datasets import load_dataset |
|
|
from sentence_transformers import SentenceTransformer |
|
|
import numpy as np |
|
|
from fastapi import FastAPI |
|
|
from fastapi.middleware.cors import CORSMiddleware |
|
|
from pydantic import BaseModel |
|
|
from dotenv import load_dotenv |
|
|
|
|
|
load_dotenv() |
|
|
GROQ_API_KEY = os.getenv("GROQ_API_KEY") |
|
|
|
|
|
app = FastAPI() |
|
|
|
|
|
app.add_middleware( |
|
|
CORSMiddleware, |
|
|
allow_origins=["*"], |
|
|
allow_credentials=True, |
|
|
allow_methods=["*"], |
|
|
allow_headers=["*"], |
|
|
) |
|
|
|
|
|
dataset = load_dataset("altaidevorg/women-health-mini") |
|
|
|
|
|
conversation_data = [ |
|
|
turn["content"] |
|
|
for conv in dataset["train"] |
|
|
for turn in conv["conversations"] |
|
|
] |
|
|
|
|
|
model = SentenceTransformer("all-MiniLM-L6-v2", device="cpu") |
|
|
|
|
|
conversation_embeddings = model.encode(conversation_data) |
|
|
|
|
|
GROQ_API_URL = "https://api.groq.com/openai/v1/chat/completions" |
|
|
|
|
|
|
|
|
def get_more_relevant_rsponse(query): |
|
|
query_embedding = model.encode([query]) |
|
|
similarities = np.dot(conversation_embeddings, query_embedding.T).flatten() |
|
|
best_match_idx = np.argmax(similarities) |
|
|
return conversation_data[best_match_idx] |
|
|
|
|
|
@app.post("/chat/") |
|
|
async def chat_with_bot(user_query:dict): |
|
|
prompt = user_query.get("message", "") |
|
|
|
|
|
if not prompt : |
|
|
return {"response": "Prompt is required!"} |
|
|
|
|
|
history_response = get_more_relevant_rsponse(prompt) |
|
|
|
|
|
context_prompt = f""" |
|
|
You are a chatbot named fille AI specialized in women's health. Provide **clear, factual, and supportive** responses. |
|
|
If the user's question involves medical advice, remind them to consult a healthcare professional. |
|
|
|
|
|
User Question: {prompt} |
|
|
|
|
|
Below is a similar question/response from the knowledge base that may contain helpful information: |
|
|
{history_response} |
|
|
|
|
|
Please provide your own professional, friendly, and informative response that addresses the user's specific question. |
|
|
""" |
|
|
|
|
|
headers = { |
|
|
"Authorization": f"Bearer {GROQ_API_KEY}", |
|
|
"Content-Type": "application/json" |
|
|
} |
|
|
|
|
|
payload = { |
|
|
"model": "llama3-70b-8192", |
|
|
"messages": [{"role": "user", "content": context_prompt}] |
|
|
} |
|
|
|
|
|
response = requests.post(GROQ_API_URL, json=payload, headers=headers) |
|
|
|
|
|
if response.status_code == 200: |
|
|
bot_reply = response.json()["choices"][0]["message"]["content"] |
|
|
return { |
|
|
"response": bot_reply |
|
|
} |
|
|
|
|
|
else: |
|
|
return {"response": "Error in fetching the data from groq AI"} |
|
|
|
|
|
if __name__ == "__main__" : |
|
|
import uvicorn |
|
|
uvicorn.run(app, host="0.0.0.0", port=8000) |
