Updated

app.py

@@ -69,8 +69,6 @@ class VectorRequest(BaseModel):
 # ==============================
 # MODELS PER ENDPOINT (Meta Models, Conversational)
 # ==============================
-
-# 1. Crop Doctor
 crop_template = PromptTemplate(
     input_variables=["symptoms"],
     template="You are AgriCopilot, a multilingual AI assistant created to support farmers. Farmer reports: {symptoms}. Diagnose the most likely disease and suggest treatments in simple farmer-friendly language."
@@ -85,7 +83,6 @@ crop_llm = HuggingFaceEndpoint(
     max_new_tokens=1024
 )
 
-# 2. Multilingual Chat
 chat_template = PromptTemplate(
     input_variables=["query"],
     template="You are AgriCopilot, a supportive multilingual AI guide built for farmers. Farmer says: {query}"
@@ -100,7 +97,6 @@ chat_llm = HuggingFaceEndpoint(
     max_new_tokens=1024
 )
 
-# 3. Disaster Summarizer
 disaster_template = PromptTemplate(
     input_variables=["report"],
     template="You are AgriCopilot, an AI disaster-response assistant. Summarize in simple steps: {report}"
@@ -115,7 +111,6 @@ disaster_llm = HuggingFaceEndpoint(
     max_new_tokens=1024
 )
 
-# 4. Marketplace Recommendation
 market_template = PromptTemplate(
     input_variables=["product"],
     template="You are AgriCopilot, an AI agricultural marketplace advisor. Farmer wants to sell or buy: {product}. Suggest best options and advice."
@@ -134,8 +129,8 @@ market_llm = HuggingFaceEndpoint(
 # ENDPOINT HELPERS
 # ==============================
 def run_conversational_model(model, prompt: str):
-    """Wraps prompt into HF conversational format"""
-    return model.invoke({
+    """Wraps prompt into HF conversational format and extracts text"""
+    result = model.invoke({
         "inputs": {
             "past_user_inputs": [],
             "generated_responses": [],
@@ -143,6 +138,14 @@ def run_conversational_model(model, prompt: str):
         }
     })
 
+    if isinstance(result, dict) and "generated_text" in result:
+        return result["generated_text"]
+
+    if isinstance(result, list) and len(result) > 0 and "generated_text" in result[0]:
+        return result[0]["generated_text"]
+
+    return str(result)  # fallback
+
 # ==============================
 # ENDPOINTS
 # ==============================
@@ -152,7 +155,7 @@ async def crop_doctor(req: CropRequest, authorization: str | None = Header(None)
     prompt = crop_template.format(symptoms=req.symptoms)
     try:
         response = run_conversational_model(crop_llm, prompt)
-        return {"diagnosis": str(response)}
+        return {"diagnosis": response}
     except HfHubHTTPError as e:
         return {"error": f"HuggingFace error: {str(e)}"}
 
@@ -162,7 +165,7 @@ async def multilingual_chat(req: ChatRequest, authorization: str | None = Header
     prompt = chat_template.format(query=req.query)
     try:
         response = run_conversational_model(chat_llm, prompt)
-        return {"reply": str(response)}
+        return {"reply": response}
     except HfHubHTTPError as e:
         return {"error": f"HuggingFace error: {str(e)}"}
 
@@ -172,7 +175,7 @@ async def disaster_summarizer(req: DisasterRequest, authorization: str | None =
     prompt = disaster_template.format(report=req.report)
     try:
         response = run_conversational_model(disaster_llm, prompt)
-        return {"summary": str(response)}
+        return {"summary": response}
     except HfHubHTTPError as e:
         return {"error": f"HuggingFace error: {str(e)}"}
 
@@ -182,7 +185,7 @@ async def marketplace(req: MarketRequest, authorization: str | None = Header(Non
     prompt = market_template.format(product=req.product)
     try:
         response = run_conversational_model(market_llm, prompt)
-        return {"recommendation": str(response)}
+        return {"recommendation": response}
     except HfHubHTTPError as e:
         return {"error": f"HuggingFace error: {str(e)}"}