Update app.py

app.py

@@ -1,17 +1,22 @@
-from fastapi import FastAPI
+from fastapi import FastAPI, HTTPException
 from pydantic import BaseModel
 from sentence_transformers import SentenceTransformer
 import faiss
 import pandas as pd
 import os
+import logging
 from groq import Groq
 from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
 
+# ✅ Initialize FastAPI
 app = FastAPI()
 
-# ✅ Set up Groq API key
-os.environ["GROQ_API_KEY"] = "your-groq-api-key"  # Replace with your actual API key
-client = Groq(api_key=os.environ["GROQ_API_KEY"])
+# ✅ Securely Fetch API Key
+GROQ_API_KEY = os.getenv("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")  # Use environment variable for security
+if not GROQ_API_KEY:
+    raise ValueError("GROQ_API_KEY is missing. Set it as an environment variable.")
+
+client = Groq(api_key=GROQ_API_KEY)
 
 # ✅ Load AI Models
 similarity_model = SentenceTransformer("sentence-transformers/all-mpnet-base-v2")
@@ -20,8 +25,12 @@ summarization_model = AutoModelForSeq2SeqLM.from_pretrained("google/long-t5-tglo
 summarization_tokenizer = AutoTokenizer.from_pretrained("google/long-t5-tglobal-base")
 
 # ✅ Load datasets
-recommendations_df = pd.read_csv("treatment_recommendations.csv")
-questions_df = pd.read_csv("symptom_questions.csv")
+try:
+    recommendations_df = pd.read_csv("treatment_recommendations.csv")
+    questions_df = pd.read_csv("symptom_questions.csv")
+except FileNotFoundError as e:
+    logging.error(f"Missing dataset file: {e}")
+    raise HTTPException(status_code=500, detail="Dataset files are missing.")
 
 # ✅ FAISS Index for disorder detection
 treatment_embeddings = similarity_model.encode(recommendations_df["Disorder"].tolist(), convert_to_numpy=True)
@@ -46,10 +55,13 @@ def retrieve_questions(user_input):
     input_embedding = embedding_model.encode([user_input], convert_to_numpy=True)
     _, indices = question_index.search(input_embedding, 1)  # ✅ Retrieve only 1 question
 
+    if indices[0][0] == -1:
+        return "I'm sorry, I couldn't find a relevant question."
+
     # ✅ Extract only the first meaningful question
     question_block = questions_df["Questions"].iloc[indices[0][0]]
     split_questions = question_block.split(", ")  
-    best_question = split_questions[0]  # ✅ Select the first clear question
+    best_question = split_questions[0] if split_questions else question_block  # ✅ Select the first clear question
 
     return best_question  # ✅ Return a single question as a string
 
@@ -73,17 +85,20 @@ def generate_empathetic_response(user_input, retrieved_question):
     Generate only one empathetic response.
     """
 
-    chat_completion = client.chat.completions.create(
-        messages=[
-            {"role": "system", "content": "You are a helpful, empathetic AI psychiatrist."},
-            {"role": "user", "content": prompt}
-        ],
-        model="llama3-8b",  # ✅ Use Groq's LLaMA-3 Model
-        temperature=0.8,  # Adjust for natural variation
-        top_p=0.9
-    )
-
-    return chat_completion.choices[0].message.content  # ✅ Return only one response
+    try:
+        chat_completion = client.chat.completions.create(
+            messages=[
+                {"role": "system", "content": "You are a helpful, empathetic AI psychiatrist."},
+                {"role": "user", "content": prompt}
+            ],
+            model="llama3-8b",  # ✅ Use Groq's LLaMA-3 Model
+            temperature=0.8,
+            top_p=0.9
+        )
+        return chat_completion.choices[0].message.content  # ✅ Return only one response
+    except Exception as e:
+        logging.error(f"Groq API error: {e}")
+        return "I'm sorry, I couldn't process your request."
 
 # ✅ API Endpoint: Get Empathetic Questions (Hybrid RAG)
 @app.post("/get_questions")
@@ -111,6 +126,10 @@ def detect_disorders(request: SummaryRequest):
     full_chat_text = " ".join(request.chat_history)
     text_embedding = similarity_model.encode([full_chat_text], convert_to_numpy=True)
     distances, indices = index.search(text_embedding, 3)
+
+    if indices[0][0] == -1:
+        return {"disorders": "No matching disorder found."}
+
     disorders = [recommendations_df["Disorder"].iloc[i] for i in indices[0]]
     return {"disorders": disorders}
 
@@ -121,6 +140,6 @@ def get_treatment(request: SummaryRequest):
     detected_disorders = detect_disorders(request)["disorders"]
     treatments = {
         disorder: recommendations_df[recommendations_df["Disorder"] == disorder]["Treatment Recommendation"].values[0]
-        for disorder in detected_disorders
+        for disorder in detected_disorders if disorder in recommendations_df["Disorder"].values
     }
     return {"treatments": treatments}