Updated

app.py

@@ -3,8 +3,7 @@ import logging
 from fastapi import FastAPI, Request, Header, HTTPException, UploadFile, File
 from fastapi.responses import JSONResponse
 from pydantic import BaseModel
-from langchain_community.chat_models import ChatHF
-from langchain.schema import HumanMessage
+from transformers import pipeline
 from PIL import Image
 import io
 from vector import query_vector
@@ -65,37 +64,40 @@ class VectorRequest(BaseModel):
     query: str
 
 # ==============================
-# HuggingFace Chat Models
+# HuggingFace Pipelines
 # ==============================
-chat_model = ChatHF(model_name="meta-llama/Llama-3.1-8B-Instruct", temperature=0.3)
-disaster_model = ChatHF(model_name="meta-llama/Llama-3.1-8B-Instruct", temperature=0.3)
-market_model = ChatHF(model_name="meta-llama/Llama-3.1-8B-Instruct", temperature=0.3)
+# Conversational models for chat, disaster, marketplace
+chat_pipe = pipeline("conversational", model="meta-llama/Llama-3.1-8B-Instruct")
+disaster_pipe = pipeline("conversational", model="meta-llama/Llama-3.1-8B-Instruct")
+market_pipe = pipeline("conversational", model="meta-llama/Llama-3.1-8B-Instruct")
 
-# Crop Doctor Vision + Language Model
-crop_model = ChatHF(model_name="meta-llama/Llama-3.2-11B-Vision-Instruct", temperature=0.3)
+# Crop Doctor: image + text
+crop_pipe = pipeline("image-to-text", model="meta-llama/Llama-3.2-11B-Vision-Instruct")
 
 # ==============================
 # Helper Functions
 # ==============================
-def run_chat_model(model, prompt: str):
+def run_conversational(pipe, prompt: str):
     try:
-        response = model([HumanMessage(content=prompt)])
-        return response.content
+        output = pipe(prompt)
+        # output is a list of dicts
+        if isinstance(output, list) and len(output) > 0:
+            return output[0].get("generated_text", str(output))
+        return str(output)
     except Exception as e:
-        logger.error(f"Model error: {e}")
+        logger.error(f"Conversational pipeline error: {e}")
         return f"⚠️ Unexpected model error: {str(e)}"
 
-def run_crop_doctor_model(model, image_bytes: bytes, symptoms: str):
-    """Send image + text to vision-language model"""
+def run_crop_doctor(image_bytes: bytes, symptoms: str):
     try:
-        # Convert bytes to image
         image = Image.open(io.BytesIO(image_bytes)).convert("RGB")
         prompt = f"Farmer reports: {symptoms}. Diagnose the crop disease and suggest treatment in simple language."
-        # ChatHF allows messages with image objects as content
-        response = model([HumanMessage(content=prompt, additional_kwargs={"image": image})])
-        return response.content
+        output = crop_pipe(image, prompt=prompt)
+        if isinstance(output, list) and len(output) > 0:
+            return output[0].get("generated_text", str(output))
+        return str(output)
     except Exception as e:
-        logger.error(f"Crop Doctor model error: {e}")
+        logger.error(f"Crop Doctor pipeline error: {e}")
         return f"⚠️ Unexpected model error: {str(e)}"
 
 # ==============================
@@ -103,32 +105,28 @@ def run_crop_doctor_model(model, image_bytes: bytes, symptoms: str):
 # ==============================
 @app.post("/crop-doctor")
 async def crop_doctor(symptoms: str = Header(...), image: UploadFile = File(...), authorization: str | None = Header(None)):
-    """
-    Receives crop image and symptom description.
-    Returns diagnosis and suggested treatment.
-    """
     check_auth(authorization)
     image_bytes = await image.read()
-    result = run_crop_doctor_model(crop_model, image_bytes, symptoms)
-    return {"diagnosis": result}
+    diagnosis = run_crop_doctor(image_bytes, symptoms)
+    return {"diagnosis": diagnosis}
 
 @app.post("/multilingual-chat")
 async def multilingual_chat(req: ChatRequest, authorization: str | None = Header(None)):
     check_auth(authorization)
-    response = run_chat_model(chat_model, req.query)
-    return {"reply": response}
+    reply = run_conversational(chat_pipe, req.query)
+    return {"reply": reply}
 
 @app.post("/disaster-summarizer")
 async def disaster_summarizer(req: DisasterRequest, authorization: str | None = Header(None)):
     check_auth(authorization)
-    response = run_chat_model(disaster_model, req.report)
-    return {"summary": response}
+    summary = run_conversational(disaster_pipe, req.report)
+    return {"summary": summary}
 
 @app.post("/marketplace")
 async def marketplace(req: MarketRequest, authorization: str | None = Header(None)):
     check_auth(authorization)
-    response = run_chat_model(market_model, req.product)
-    return {"recommendation": response}
+    recommendation = run_conversational(market_pipe, req.product)
+    return {"recommendation": recommendation}
 
 @app.post("/vector-search")
 async def vector_search(req: VectorRequest, authorization: str | None = Header(None)):