Update app.py

app.py

@@ -1,83 +1,54 @@
 import streamlit as st
-from PIL import Image
-import torch
 import easyocr
+import openai
+from openai import OpenAI
+from PIL import Image
 import numpy as np
-import openai  # Using OpenAI GPT (or replace with GROQ API)
 import io
-from transformers import CLIPModel, CLIPImageProcessor
 
-# ✅ Fix: set_page_config() must be the first Streamlit command
+# ✅ Fix: Ensure set_page_config is the first Streamlit command
 st.set_page_config(page_title="Multimodal AI Assistant", layout="wide")
 
-# ---- Load CLIP Model (Vision Only) ---- #
-@st.cache_resource
-def load_clip_model():
-    model = CLIPModel.from_pretrained(
-        "fxmarty/clip-vision-model-tiny", 
-        ignore_mismatched_sizes=True  # ✅ Fix size mismatch
-    )
-    processor = CLIPImageProcessor.from_pretrained("fxmarty/clip-vision-model-tiny")
-    return model, processor
-
-model, processor = load_clip_model()
-
-# ---- Load OCR (EasyOCR) ---- #
-@st.cache_resource
-def load_ocr():
-    return easyocr.Reader(['en'])
-
-reader = load_ocr()
+# ✅ OpenAI API Key Setup (Replace with your actual key)
+client = OpenAI(api_key="your_openai_api_key")  # Set your OpenAI API key
 
-# ---- Streamlit UI ---- #
-st.title("🖼️ Multimodal AI Assistant")
-st.write("Upload an image, extract text, and ask questions!")
+# ✅ Initialize OCR Reader
+reader = easyocr.Reader(['en'])
 
-# ---- Upload Image ---- #
-uploaded_file = st.file_uploader("📤 Upload an image", type=["jpg", "png", "jpeg"])
+# ✅ Streamlit App Layout
+st.title("📸 Multimodal AI Assistant")
+st.write("Upload an image and ask questions based on the extracted text.")
 
-extracted_text = None  # Variable to store extracted text
+# ✅ File Uploader
+uploaded_file = st.file_uploader("Upload an image", type=["png", "jpg", "jpeg"])
 
-if uploaded_file is not None:
-    # Convert file to image format
-    image = Image.open(uploaded_file).convert("RGB")
-
-    # ✅ Fix: use `use_container_width` instead of `use_column_width`
+if uploaded_file:
+    # ✅ Display uploaded image
+    image = Image.open(uploaded_file)
     st.image(image, caption="Uploaded Image", use_container_width=True)
 
-    # ✅ Convert PIL image to NumPy array for EasyOCR
-    image_np = np.array(image)
-
-    # ✅ Fix: Pass the correct format to EasyOCR
-    with st.spinner("🔍 Extracting text from image..."):
-        extracted_text_list = reader.readtext(image_np, detail=0)
-
-    extracted_text = " ".join(extracted_text_list)  # Combine extracted text
-
-    st.write("### 📝 Extracted Text:")
-    if extracted_text:
-        st.success(extracted_text)
-    else:
-        st.warning("No readable text found in the image.")
+    # ✅ Convert image to a format that EasyOCR supports
+    image_bytes = io.BytesIO(uploaded_file.getvalue()).read()
+    extracted_text = reader.readtext(image_bytes, detail=0)
 
-# ---- Question Answering Section ---- #
-if extracted_text:
-    user_question = st.text_input("💡 Ask a question about the extracted text:")
+    # ✅ Show extracted text
+    extracted_text_str = " ".join(extracted_text) if extracted_text else "No text found"
+    st.subheader("📝 Extracted Text:")
+    st.write(extracted_text_str)
 
-    if user_question:
-        with st.spinner("🤖 Thinking..."):
-            # Using OpenAI GPT API (replace with GROQ or Hugging Face LLM if needed)
-            openai.api_key = "YOUR_OPENAI_API_KEY"  # Store securely in a .env file
+    # ✅ Ask a question about the extracted text
+    user_query = st.text_input("Ask a question about the extracted text:")
 
-            response = openai.ChatCompletion.create(
+    if user_query:
+        with st.spinner("Thinking... 💭"):
+            response = client.chat.completions.create(
                 model="gpt-3.5-turbo",
                 messages=[
-                    {"role": "system", "content": "You are an AI assistant helping answer questions based on extracted text from an image."},
-                    {"role": "user", "content": f"Extracted text: {extracted_text}\n\nQuestion: {user_question}"}
+                    {"role": "system", "content": "You are an AI assistant analyzing extracted text from images."},
+                    {"role": "user", "content": f"Extracted text: {extracted_text_str}\n\nUser question: {user_query}"}
                 ]
             )
 
-            answer = response["choices"][0]["message"]["content"]
-        
-        st.write("### 🤖 AI Answer:")
-        st.success(answer)
+            answer = response.choices[0].message.content
+            st.subheader("🤖 AI Answer:")
+            st.write(answer)