Update app.py

app.py

@@ -1,180 +1,37 @@
 import streamlit as st
-from PIL import Image
-import numpy as np
-from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer
+import langchain
+st.write('Helldo')
 import os
-
-# Environment variables (set in Space settings)
-hf_token = os.getenv('hf')
-
-# Simplified image preprocessing
-def preprocess_image(image):
-    return image.resize((224, 224)).convert('RGB')
-
-# Smaller VQA model for CPU
-vqa_pipeline = pipeline("visual-question-answering", 
-                       model="dandelin/vilt-b32-finetuned-vqa",
-                       device_map="cpu")
-
-# Optimized LLM setup
-model_id = "Open-Orca/MiniChat-1.5B"  # Smaller alternative to Llama-3
-tokenizer = AutoTokenizer.from_pretrained(model_id, use_fast=True)
-model = AutoModelForCausalLM.from_pretrained(model_id, device_map="cpu")
-
-def generate_response(question, context):
-    prompt = f"Context: {context}\nQuestion: {question}\nAnswer:"
-    inputs = tokenizer(prompt, return_tensors="pt")
-    outputs = model.generate(**inputs, max_new_tokens=50)
-    return tokenizer.decode(outputs[0], skip_special_tokens=True)
-
-# Streamlit App
-st.title("Multimodal Chatbot Lite")
-st.write("Upload an image and ask a question")
-
-uploaded_image = st.file_uploader("Choose an image", type=["jpg", "png", "jpeg"])
-question = st.text_input("Your question")
-
-if uploaded_image and question:
-    try:
-        image = Image.open(uploaded_image)
-        st.image(image, caption="Uploaded Image", use_column_width=True)
-        processed_image = preprocess_image(image)
-        
-        # Get image context
-        context = vqa_pipeline(image=processed_image, question=question, top_k=1)[0]['answer']
-        
-        # Generate response
-        answer = generate_response(question, context)
-        st.write(f"**Answer**: {answer.split('Answer:')[-1].strip()}")
-        
-    except Exception as e:
-        st.error(f"Error: {str(e)}")
-else:
-    st.info("Please upload an image and enter a question")
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-# import streamlit as st
-# import cv2
-# import numpy as np
-# import os
-# from PIL import Image
-# from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer,BitsAndBytesConfig
-# from langchain.chains import LLMChain
-# from langchain.prompts import PromptTemplate
-# from langchain_huggingface import ChatHuggingFace
-# from pydantic import BaseModel, validator
-# from typing import Optional
-
-# hf = os.getenv('hf')
-# os.environ['HUGGINGFACEHUB_API_TOKEN'] = hf
-# os.environ['HF_TOKEN'] = hf
-
-# # Pydantic models for input/output validation
-# class UserInput(BaseModel):
-#     question: str
-
-#     @validator('question')
-#     def check_question(cls, v):
-#         if not v.strip():
-#             raise ValueError('Question cannot be empty')
-#         return v
-
-# class ChatResponse(BaseModel):
-#     answer: str
-#     confidence: Optional[float] = 0.95
-
-#     @validator('answer')
-#     def check_answer(cls, v):
-#         if not v.strip():
-#             raise ValueError('Answer cannot be empty')
-#         return v
-
-# # Image preprocessing with OpenCV
-# def preprocess_image(image):
-#     img = np.array(image)
-#     img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
-#     img = cv2.resize(img, (224, 224))
-#     return img
-
-# # HuggingFace VQA pipeline
-# vqa_pipeline = pipeline("visual-question-answering", model="Salesforce/blip-vqa-base")
-
-# def get_image_context(image, question):
-#     result = vqa_pipeline(image, question, top_k=1)
-#     return result[0]['answer']
-
-# # 🔥 Corrected LangChain setup
-# model_id = "meta-llama/Llama-3.2-1B"
-
-# tokenizer = AutoTokenizer.from_pretrained(model_id, token=hf,use_fast=False)
-# model = AutoModelForCausalLM.from_pretrained(model_id, token=hf,device_map="cpu" )
-
-# llm = ChatHuggingFace(llm=model, tokenizer=tokenizer)
-
-# prompt = PromptTemplate(
-#     input_variables=["image_context", "question"],
-#     template="Based on the image context: {image_context}, answer the question: {question}"
-# )
-# chain = LLMChain(llm=llm, prompt=prompt)
-
-# def generate_response(image_context, question):
-#     return chain.run(image_context=image_context, question=question)
-
-# # Streamlit App
-# st.title("Intelligent Multimodal Chatbot")
-# st.write("Upload an image and ask a question about it.")
-
-# uploaded_image = st.file_uploader("Upload an image", type=["png", "jpg", "jpeg"])
-# question = st.text_input("Ask a question about the image")
-
-# if uploaded_image and question:
-#     try:
-#         user_input = UserInput(question=question)
-#         image = Image.open(uploaded_image)
-#         st.image(image, caption="Uploaded Image", use_column_width=True)
-#         processed_image = preprocess_image(image)
-
-#         image_context = get_image_context(image, question)
-#         response = generate_response(image_context, question)
-
-#         chat_response = ChatResponse(answer=response)
-#         st.write("**Answer**: ", chat_response.answer)
-#         st.write("**Confidence**: ", chat_response.confidence)
-
-#     except Exception as e:
-#         st.error(f"Error: {str(e)}")
-# else:
-#     st.write("Please upload an image and enter a question.")
+import langchain
+import langchain_huggingface
+from pydantic import BaseModel,Field
+from langchain_huggingface import HuggingFaceEndpoint,HuggingFacePipeline,ChatHuggingFace
+from langchain_core.output_parsers import PydanticOutputParser,CommaSeparatedListOutputParser,JsonOutputParser
+from langchain.prompts import PromptTemplate,ChatPromptTemplate,SystemMessagePromptTemplate,HumanMessagePromptTemplate
+from typing import Optional, List
+from langchain_community.document_loaders import UnstructuredPDFLoader
+
+
+
+#  creating the environment
+hk = os.getenv('hf')
+os.environ['HUGGINGFACEHUB_API_TOKEN'] = hk
+os.environ['HF_TOKEN'] = hk
+
+
+# accessing the llm
+# ------ accesssing the llm for geenral prompting -------------------
+llm_skeleton = HuggingFaceEndpoint(repo_id='meta-llama/Llama-3.2-3B-Instruct',
+                    provider = 'novita',
+                    temperature=0.7,
+                    max_new_tokens=150,
+                    task = 'conversational')
+
+
+# ------------- wrapping the llm to be a conversational model ------------------
+llm = ChatHuggingFace(llm=llm_skeleton,
+                    repo_id='meta-llama/Llama-3.2-3B-Instruct',
+                    provider = 'novita',
+                    temperature=0.7,
+                    max_new_tokens=150,
+                    task = 'conversational')