Update app.py

app.py

@@ -1,232 +1,194 @@
-import streamlit as st
-from database import db  
-import html
-import re
-import streamlit.components.v1 as components
-
-def get_text_from_firebase(text_id):
-    """Retrieve text from Firebase Firestore."""
-    try:
-        doc = db.collection("texts").document(text_id).get()
-        if doc.exists:
-            return doc.to_dict()["content"]
-        else:
-            return None
-    except Exception as e:
-        st.error(f"Error retrieving text from Firebase: {e}")
-        return None
-
-def process_text(text):
-    """Clean and structure the extracted text."""
-    # Preserve paragraph breaks
-    text = re.sub(r'\n\s*\n', '\n\n', text)
-    
-    # Convert bullet points to HTML lists
-    text = re.sub(r'^(\s*•\s+)(.*)$', r'<li>\2</li>', text, flags=re.MULTILINE)
-    text = re.sub(r'(<li>.*</li>\n)+', r'<ul>\g<0></ul>\n', text)
-    
-    # Detect headings (lines ending with colons or in all caps)
-    text = re.sub(r'^([A-Z][A-Z\s]+:?)\s*$', r'<h3 class="heading">\1</h3>', text, flags=re.MULTILINE)
-    
-    # Convert URLs to clickable links
-    text = re.sub(r'(https?://\S+)', r'<a href="\1" target="_blank">\1</a>', text)
-    
-    # Preserve original paragraph structure
-    paragraphs = text.split('\n\n')
-    processed = []
-    
-    for p in paragraphs:
-        p = p.strip()
-        if p:
-            if p.startswith('<ul>') or p.startswith('<h3'):
-                processed.append(p)
-            else:
-                processed.append(f'<p>{p}</p>')
-    
-    return '\n'.join(processed)
-
-def main():
-    st.title("Accessibility-Optimized Viewer")
-    
-    # Ask user for Document ID
-    text_id = st.text_input("Enter Document ID to view text:")
-    
-    if text_id:
-        text = get_text_from_firebase(text_id)
-        if text:
-            # Accessibility Controls
-            col1, col2, col3 = st.columns(3)
-            with col1:
-                font_size = st.slider("Font Size", 8, 48, 16)
-                line_height = st.slider("Line Height", 1.0, 2.5, 1.5)
-            with col2:
-                font_color = st.color_picker("Text Color", "#000000")
-                bg_color = st.color_picker("Background Color", "#FFFFFF")
-            with col3:
-                contrast_mode = st.checkbox("High Contrast Mode")
-                dyslexia_font = st.checkbox("Dyslexia-Friendly Font")
-
-            # Apply contrast mode
-            if contrast_mode:
-                font_color = "#FFFFFF"
-                bg_color = "#000000"
-
-            # Create CSS styles
-            font_family = "Arial" if not dyslexia_font else "OpenDyslexic, sans-serif"
-            
-            custom_css = f"""
-            <style>
-                .content {{
-                    font-size: {font_size}px;
-                    color: {font_color};
-                    background-color: {bg_color};
-                    line-height: {line_height};
-                    font-family: {font_family};
-                    padding: 20px;
-                    border-radius: 10px;
-                    margin: 10px 0;
-                    white-space: pre-wrap;
-                    word-wrap: break-word;
-                    hyphens: auto;
-                }}
-                
-                .content p {{
-                    margin: 0.8em 0;
-                }}
-                
-                .content ul {{
-                    margin: 0.8em 20px;
-                    padding-left: 20px;
-                    list-style-type: disc;
-                }}
-                
-                .content h3 {{
-                    font-size: 1.2em;
-                    margin: 1.2em 0 0.5em;
-                    padding-bottom: 3px;
-                    border-bottom: 2px solid {font_color};
-                }}
-                
-                .content a {{
-                    color: {font_color};
-                    text-decoration: underline;
-                    word-break: break-all;
-                }}
-                
-                @keyframes highlight {{ 0% {{background: yellow;}} 100% {{background: transparent;}} }}
-                .highlight {{ animation: highlight 2s; }}
-            </style>
-            """
-            st.markdown(custom_css, unsafe_allow_html=True)
-
-            # Process and display text
-            processed_text = process_text(html.escape(text))
-            
-            # Create a div with an ID that can be targeted by ResponsiveVoice
-            content_html = f'<div id="content-to-speak" class="content">{processed_text}</div>'
-            st.markdown(content_html, unsafe_allow_html=True)
-            
-            # Additional accessibility features
-            with st.expander("More Accessibility Options"):
-                col1, col2 = st.columns(2)
-                with col1:
-                    letter_spacing = st.slider("Letter Spacing (px)", -1, 5, 0)
-                    word_spacing = st.slider("Word Spacing (px)", 0, 10, 0)
-                with col2:
-                    text_align = st.selectbox("Text Alignment", ["left", "justify", "center"])
-                    text_transform = st.selectbox("Text Case", ["none", "uppercase", "lowercase"])
-                
-                # Update CSS
-                st.markdown(f"""
-                <style>
-                    .content {{
-                        letter-spacing: {letter_spacing}px;
-                        word-spacing: {word_spacing}px;
-                        text-align: {text_align};
-                        text-transform: {text_transform};
-                    }}
-                </style>
-                """, unsafe_allow_html=True)
-            
-            # Voice selection for ResponsiveVoice
-            st.subheader("Text-to-Speech Options")
-            voices = ["UK English Female", "UK English Male", "US English Female", "US English Male", 
-                      "Spanish Female", "Spanish Male", "French Female", "French Male", 
-                      "German Female", "German Male", "Italian Female", "Italian Male"]
-            selected_voice = st.selectbox("Select Voice", voices, index=0)
-            
-            # ResponsiveVoice integration
-            # Modified ResponsiveVoice integration
-            responsive_voice_js = f"""
-            <script src="https://code.responsivevoice.org/responsivevoice.js?key=9MY6AzBd"></script>
-            
-            <div id="voice-controls">
-                <button id="speak-button">Read Aloud</button>
-                <button id="pause-button">Pause</button>
-                <button id="resume-button">Resume</button>
-                <button id="stop-button">Stop</button>
-            </div>
-
-            <script>
-            document.addEventListener('DOMContentLoaded', function() {{
-                // Check if ResponsiveVoice is loaded
-                if(typeof responsiveVoice === 'undefined') {{
-                    console.error('ResponsiveVoice not loaded!');
-                    return;
-                }}
-                
-                const getContent = () => {{
-                    const el = document.getElementById('content-to-speak');
-                    return el ? el.innerText : '';
-                }};
-
-                document.getElementById('speak-button').addEventListener('click', () => {{
-                    responsiveVoice.speak(getContent(), "{selected_voice}");
-                }});
-                
-                document.getElementById('pause-button').addEventListener('click', () => {{
-                    responsiveVoice.pause();
-                }});
-                
-                document.getElementById('resume-button').addEventListener('click', () => {{
-                    responsiveVoice.resume();
-                }});
-                
-                document.getElementById('stop-button').addEventListener('click', () => {{
-                    responsiveVoice.cancel();
-                }});
-
-                // Text selection handler
-                document.addEventListener('mouseup', function() {{
-                    const selection = window.getSelection().toString().trim();
-                    if(selection) {{
-                        responsiveVoice.cancel();
-                        responsiveVoice.speak(selection, "{selected_voice}");
-                    }}
-                }});
-            }});
-            </script>
-
-            <style>
-            #voice-controls button {{
-                margin: 2px;
-                padding: 6px 12px;
-                background-color: #4CAF50;
-                color: white;
-                border: none;
-                border-radius: 4px;
-                cursor: pointer;
-            }}
-            </style>
-            """
-
-            # Use Streamlit component to inject the JavaScript
-            components.html(responsive_voice_js, height=70)
-            
-            st.info("You can now click 'Read Aloud' to have the text read to you, or select specific text to have it read.")
-        else:
-            st.error("Text not found. Please check the ID.")
-
-
-if __name__ == "__main__":
+import streamlit as st
+from database import db  
+from firebase_admin import firestore
+from PyPDF2 import PdfReader
+import easyocr
+from PIL import Image
+from io import BytesIO
+import random
+import string
+import base64
+import os
+import numpy as np
+from mistralai import Mistral
+
+def generate_readable_id(length=5):
+    """Generate a random alphanumeric ID."""
+    characters = string.ascii_letters + string.digits  # A-Z, a-z, 0-9
+    return ''.join(random.choice(characters) for _ in range(length))
+
+
+def encode_image(image_file):
+    """Encode image to base64."""
+    return base64.b64encode(image_file.read()).decode('utf-8')
+
+def process_mistral_ocr(file, api_key):
+    """Process OCR using Mistral API."""
+    client = Mistral(api_key=api_key)
+    
+    if file.type == "application/pdf":
+        # Upload the PDF
+        uploaded_pdf = client.files.upload(
+            file={
+                "file_name": file.name,
+                "content": file.getvalue(),
+            },
+            purpose="ocr"
+        )
+        
+        # Get OCR results
+        ocr_response = client.ocr.process(
+            model="mistral-ocr-latest",
+            document={
+                "type": "document_url",
+                "document_url": client.files.get_signed_url(file_id=uploaded_pdf.id).url,
+            }
+        )
+    else:
+        # Process Image
+        base64_image = encode_image(file)
+        ocr_response = client.ocr.process(
+            model="mistral-ocr-latest",
+            document={
+                "type": "image_url",
+                "image_url": f"data:image/jpeg;base64,{base64_image}"
+            }
+        )
+    
+    return ocr_response["text"]
+
+def extract_text_from_pdf(file):
+    """Extract text from a PDF file using PyPDF2."""
+    reader = PdfReader(file)
+    text = ""
+    for page in reader.pages:
+        text += page.extract_text() or ""
+    return text
+
+def extract_text_from_image(image):
+    """Extract text from an image using EasyOCR."""
+    # Convert PIL Image to numpy array for EasyOCR
+    img_array = np.array(image)
+    
+    reader = easyocr.Reader(['en'], gpu=False)
+    results = reader.readtext(img_array)
+    return " ".join([result[1] for result in results])
+
+def save_text_to_firebase(text):
+    """Save text to Firebase Firestore."""
+    try:
+        # Generate a readable ID for the text
+        text_id = generate_readable_id()
+        
+        # Save the text to Firestore
+        db.collection("texts").document(text_id).set({
+            "content": text,
+            "created_at": firestore.SERVER_TIMESTAMP
+        })
+        return text_id
+    except Exception as e:
+        st.error(f"Error saving text to Firebase: {e}")
+        return None
+
+def main():
+    st.title("Text Upload App")
+    
+    # Sidebar for API key configuration
+    with st.sidebar:
+        st.header("OCR Configuration")
+        # Mistral API key input
+        mistral_api_key = st.text_input("Enter Mistral API Key:", type="password")
+        # Save API key to session state
+        if mistral_api_key:
+            st.session_state['mistral_api_key'] = mistral_api_key
+            
+    # Select OCR method
+    ocr_method = st.radio("Choose OCR method:", ("EasyOCR (Default)", "Mistral OCR (Requires API Key)"))
+    
+    # Select input method
+    option = st.radio("Choose input method:", ("Upload Document", "Take Picture"))
+
+    if option == "Upload Document":
+        uploaded_file = st.file_uploader("Upload PDF or Image", type=["pdf", "png", "jpg", "jpeg"])
+        
+        if uploaded_file:
+            with st.spinner("Processing document..."):
+                if ocr_method == "Mistral OCR (Requires API Key)":
+                    # Check if API key is available
+                    if 'mistral_api_key' not in st.session_state or not st.session_state['mistral_api_key']:
+                        st.error("Please provide a Mistral API key in the sidebar to use Mistral OCR")
+                    else:
+                        try:
+                            # Use Mistral OCR
+                            text = process_mistral_ocr(uploaded_file, st.session_state['mistral_api_key'])
+                        except Exception as e:
+                            st.error(f"Error processing with Mistral OCR: {e}")
+                            st.info("Falling back to default OCR method...")
+                            # Fallback to traditional methods
+                            if uploaded_file.type == "application/pdf":
+                                text = extract_text_from_pdf(uploaded_file)
+                            else:
+                                # Reset file pointer to beginning
+                                uploaded_file.seek(0)
+                                image = Image.open(uploaded_file)
+                                text = extract_text_from_image(image)
+                else:
+                    # Use traditional methods
+                    if uploaded_file.type == "application/pdf":
+                        text = extract_text_from_pdf(uploaded_file)
+                    else:
+                        # Reset file pointer to beginning
+                        uploaded_file.seek(0)
+                        image = Image.open(uploaded_file)
+                        text = extract_text_from_image(image)
+                
+                if 'text' in locals() and text and text.strip():
+                    # Show extracted text
+                    with st.expander("View Extracted Text"):
+                        st.write(text)
+                    
+                    # Save text to Firebase
+                    text_id = save_text_to_firebase(text)
+                    if text_id:
+                        st.success(f"Text extracted! Your Document ID: **{text_id}**")
+                else:
+                    st.warning("No text could be extracted")
+
+    elif option == "Take Picture":
+        picture = st.camera_input("Take a picture")
+        
+        if picture:
+            with st.spinner("Processing image..."):
+                if ocr_method == "Mistral OCR (Requires API Key)":
+                    # Check if API key is available
+                    if 'mistral_api_key' not in st.session_state or not st.session_state['mistral_api_key']:
+                        st.error("Please provide a Mistral API key in the sidebar to use Mistral OCR")
+                    else:
+                        try:
+                            # Use Mistral OCR
+                            text = process_mistral_ocr(picture, st.session_state['mistral_api_key'])
+                        except Exception as e:
+                            st.error(f"Error processing with Mistral OCR: {e}")
+                            st.info("Falling back to default OCR method...")
+                            # Fallback to EasyOCR
+                            image = Image.open(BytesIO(picture.getvalue()))
+                            text = extract_text_from_image(image)
+                else:
+                    # Use EasyOCR
+                    image = Image.open(BytesIO(picture.getvalue()))
+                    text = extract_text_from_image(image)
+                
+                if 'text' in locals() and text and text.strip():
+                    # Show extracted text
+                    with st.expander("View Extracted Text"):
+                        st.write(text)
+                    
+                    # Save text to Firebase
+                    text_id = save_text_to_firebase(text)
+                    if text_id:
+                        st.success(f"Text extracted! Your Document ID: **{text_id}**")
+                else:
+                    st.warning("No text could be extracted")
+
+if __name__ == "__main__":
     main()