Update app.py

app.py

@@ -7,42 +7,28 @@ from langchain_core.runnables.history import RunnableWithMessageHistory
 from langchain_groq import ChatGroq
 import os
 
-# Load BLIP model and processor
-processor = AutoProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
-model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")
-
-
-
-# Define the image description function
-def image_description(image):
-    inputs = processor(images=image, return_tensors="pt")
-    out = model.generate(**inputs)
-    return processor.decode(out[0], skip_special_tokens=True)
-
-# Define the problem categorization function
-def problem_categorization(problem_description):
-    department_keywords = {
-        "Cleaning": [
-            "trash", "garbage", "dirt", "dirty", "litter", "unclean", "filthy", "sanitize",
-            "disinfect", "hygiene", "cleanliness", "stain", "spill", "clutter", "pest",
-            "bedbugs", "roaches", "dust", "graffiti"
-        ],
-        "Electrical": [
-            "power", "electric", "light", "lamp", "flicker", "outage", "wiring", "circuit",
-            "switch", "socket", "fan", "air conditioner", "generator", "fuse", "overload",
-            "short circuit", "malfunction", "burnt", "spark", "battery"
-        ],
-        "Plumbing": [
-            "leak", "pipe", "flood", "drain", "clog", "sewage", "overflow", "toilet", "sink",
-            "basin", "shower", "tap", "faucet", "plumber", "gasket", "water pressure", "burst",
-            "plumbing", "seepage", "moisture", "damp", "fixture", "water supply", "clogging"
-        ]
-    }
-    for department, terms in department_keywords.items():
-        if any(term in problem_description for term in terms):
-            return department
-    return "NA"
+def category(image):
+  model = genai.GenerativeModel(model_name="gemini-1.5-pro")
 
+  prompt = '''
+  "Analyze the given image and identify three distinct problems. For each problem, provide a brief description highlighting the issue. Additionally, offer a general description of the image, including key elements, objects, or the overall scene depicted. Follow the format below:
+
+**General Description:**
+[Provide a brief summary of the image, mentioning the main elements or scene.]
+
+**Problem 1:**
+- **Description:** [Briefly describe the first issue identified in the image.]
+
+**Problem 2:**
+- **Description:** [Briefly describe the second issue identified in the image.]
+
+**Problem 3:**
+- **Description:** [Briefly describe the third issue identified in the image.]"
+'''
+
+  response = model.generate_content([prompt, image])
+
+  print(response.text)
 # Define the chatbot model
 # os.environ["GROQ_API_KEY"] = "GROQQQ"
 chat_model = ChatGroq(model="llama3-8b-8192")
@@ -139,14 +125,8 @@ chain = prompt | chat_model
 def handle_input(image, text, session_id):
     if image:
         # Handle image input
-        image_desc = image_description(image)
-        category = problem_categorization(image_desc)
-        if category=='NA':
-            response="Please upload a better image"
-        else:
-            response=f"Your request has been forwarded to the {category} department"
-            if text:
-                response+=f"With the description: {text}"
+        image_desc = category(image)
+        return image_desc
     else:
         # Handle text input
         if text=='':