Create app.py

app.py

@@ -0,0 +1,236 @@
+import streamlit as st
+import PyPDF2
+import openai
+from io import BytesIO
+import io
+from reportlab.pdfgen import canvas
+from reportlab.lib.pagesizes import letter, A4
+from reportlab.pdfbase import pdfmetrics
+from reportlab.pdfbase.ttfonts import TTFont
+from reportlab.lib.utils import simpleSplit
+from reportlab.lib.colors import black
+import arabic_reshaper
+from bidi.algorithm import get_display
+import os
+
+
+# Get API key from Hugging Face secrets
+api_key = os.environ.get('OPENAI_API_KEY')
+
+def register_fonts():
+    """Register fonts for different languages"""
+    try:
+        # Using Jameel Noori Nastaleeq for Urdu
+        pdfmetrics.registerFont(TTFont('Jameel', 'fonts/Jameel Noori Nastaleeq.ttf'))
+        # Using Noto Naskh Arabic for Arabic
+        pdfmetrics.registerFont(TTFont('NotoNaskhArabic', 'fonts/NotoNaskhArabic-Regular.ttf'))
+        # Using Noto Sans for other languages
+        pdfmetrics.registerFont(TTFont('NotoSans', 'fonts/NotoSans-Regular.ttf'))
+    except Exception as e:
+        st.warning(f"Font files not found. Default fonts will be used. Error: {str(e)}")
+
+def extract_text_from_pdf(pdf_file):
+    """Extract text from uploaded PDF file"""
+    pdf_reader = PyPDF2.PdfReader(pdf_file)
+    text = ""
+    for page in pdf_reader.pages:
+        text += page.extract_text()
+    return text
+
+def create_pdf(text, target_language):
+    """Create a PDF file from text with proper language support"""
+    buffer = BytesIO()
+    c = canvas.Canvas(buffer, pagesize=A4)
+    width, height = A4
+    
+    # Set initial Y position from top
+    y = height - 50
+    margin = 50
+    
+    # Special handling for Urdu
+    if target_language == "Urdu":
+        try:
+            # Use Jameel Noori Nastaleeq font for Urdu
+            c.setFont('Jameel', 16)
+            # For Urdu text, don't reshape but apply BIDI
+            text = text.strip()
+            lines = text.split('\n')
+        except:
+            c.setFont('Helvetica', 12)
+            st.warning("Urdu font not loaded properly. Using default font.")
+    elif target_language == "Arabic":
+        try:
+            c.setFont('NotoNaskhArabic', 14)
+            text = arabic_reshaper.reshape(text)
+            text = get_display(text)
+            lines = text.split('\n')
+        except:
+            c.setFont('Helvetica', 12)
+            lines = text.split('\n')
+    else:
+        try:
+            c.setFont('NotoSans', 12)
+            lines = text.split('\n')
+        except:
+            c.setFont('Helvetica', 12)
+            lines = text.split('\n')
+
+    # Draw text with proper spacing
+    line_height = c._fontsize * 2.5 if target_language == "Urdu" else c._fontsize * 1.5
+    
+    for line in lines:
+        if y < 50:  # If near bottom of page
+            c.showPage()
+            y = height - 50
+            # Reset font for new page
+            if target_language == "Urdu":
+                c.setFont('Jameel', 16)
+            elif target_language == "Arabic":
+                c.setFont('NotoNaskhArabic', 14)
+            else:
+                c.setFont('NotoSans', 12)
+
+        # Calculate line width for RTL positioning
+        line_width = c.stringWidth(line, c._fontname, c._fontsize)
+        
+        # Position text based on language
+        if target_language in ['Arabic', 'Urdu']:
+            x = width - margin - line_width  # Right-aligned
+            # For Urdu, we'll write the original text without reshaping
+            if target_language == "Urdu":
+                c.drawRightString(width - margin, y, line)
+            else:
+                c.drawString(x, y, line)
+        else:
+            x = margin  # Left-aligned
+            c.drawString(x, y, line)
+            
+        y -= line_height
+
+    c.save()
+    buffer.seek(0)
+    return buffer
+
+def translate_text(text, target_language,api_key):
+    """Translate text using OpenAI API with improved prompting"""
+    try:
+        client = openai.OpenAI(api_key=api_key)
+        
+        # Enhanced prompt for better translation
+        system_prompt = f"""You are a professional translator specializing in {target_language}. 
+        Translate the following text to {target_language}, ensuring:
+        1. Technical terms are accurately translated
+        2. Maintain formal language and proper grammar
+        3. Preserve formatting and structure
+        4. Keep proper nouns and technical terms like 'AI', 'LLMs', 'Python' in English where appropriate
+        5. Use culturally appropriate expressions
+        6. For Urdu/Arabic, ensure proper character connections and diacritics
+        7. Maintain professional and accurate technical translations
+        """
+        
+        response = client.chat.completions.create(
+            model="gpt-3.5-turbo",
+            messages=[
+                {"role": "system", "content": system_prompt},
+                {"role": "user", "content": text}
+            ],
+            temperature=0.3
+        )
+        return response.choices[0].message.content
+    except Exception as e:
+        return f"Translation error: {str(e)}"
+
+# Set page config
+st.set_page_config(page_title="PDF Translator", layout="wide")
+
+# Try to register fonts at startup
+register_fonts()
+
+# Main app interface
+st.title("PDF Document Translator")
+
+# Add custom CSS for better text display
+st.markdown("""
+<style>
+    .stTextArea textarea {
+        font-size: 16px !important;
+    }
+</style>
+""", unsafe_allow_html=True)
+
+# # API Key input with better security
+# api_key = st.text_input("Enter your OpenAI API Key", type="password")
+
+# Language selection
+languages = {
+    "English": "English",
+    "Urdu": "Urdu",
+    "Arabic": "Arabic",
+    "Roman English": "Roman English",
+    "Roman Urdu": "Roman Urdu",
+    "Hindi": "Hindi",
+    "Spanish": "Spanish",
+    "French": "French"
+}
+
+# File uploader
+uploaded_file = st.file_uploader("Upload your PDF file", type="pdf")
+
+# Language selector
+target_language = st.selectbox(
+    "Select target language",
+    options=list(languages.keys())
+)
+
+# Create two columns for original and translated text
+col1, col2 = st.columns(2)
+
+if uploaded_file is not None and api_key:
+    # Extract text from PDF
+    with st.spinner("Extracting text from PDF..."):
+        text = extract_text_from_pdf(uploaded_file)
+    
+    # Show original text
+    with col1:
+        st.subheader("Original Text")
+        st.text_area("", value=text, height=400, key="original_text")
+    
+    # Initialize session state for translated text
+    if 'translated_text' not in st.session_state:
+        st.session_state.translated_text = None
+    
+    # Translate button
+    if st.button("Translate"):
+        with st.spinner("Translating..."):
+            translated_text = translate_text(text, languages[target_language], api_key)
+            st.session_state.translated_text = translated_text
+        
+        # Show translated text
+        with col2:
+            st.subheader(f"Translated Text ({target_language})")
+            st.text_area("", value=translated_text, height=400, key="translated_text")
+    
+    # Show download button if translation exists
+    if st.session_state.translated_text:
+        # Create PDF button
+        if st.download_button(
+            label="Download Translated PDF",
+            data=create_pdf(st.session_state.translated_text, target_language),
+            file_name=f"translated_{target_language}.pdf",
+            mime="application/pdf"
+        ):
+            st.success("PDF downloaded successfully!")
+
+elif not api_key:
+    st.warning("Please enter your OpenAI API key to proceed.")
+
+# Add instructions and notes
+st.markdown("""
+### Instructions:
+1. Upload your PDF file
+2. Select your target language
+3. Click 'Translate' to get your translation
+4. Review the translation
+5. Click 'Download Translated PDF' to save as PDF
+
+""")