Spaces:
Running
Running
| import os | |
| import fitz # PyMuPDF for PDF handling | |
| import pytesseract # OCR for text extraction | |
| from PIL import Image | |
| import tempfile | |
| import streamlit as st | |
| def extract_text_with_tesseract(pdf_path): | |
| """ | |
| Extract text with bounding box positions using Tesseract OCR. | |
| :param pdf_path: Path to the input PDF file. | |
| :return: List of dictionaries containing text and positions for each page. | |
| """ | |
| extracted_data = [] | |
| doc = fitz.open(pdf_path) | |
| for page_num in range(len(doc)): | |
| page = doc.load_page(page_num) | |
| pix = page.get_pixmap(dpi=300) # Convert PDF page to high-resolution image | |
| image_path = f"temp_page_{page_num}.png" | |
| pix.save(image_path) | |
| # Perform OCR using Tesseract | |
| img = Image.open(image_path) | |
| ocr_result = pytesseract.image_to_data(img, output_type=pytesseract.Output.DICT) | |
| page_data = [] | |
| for i in range(len(ocr_result["text"])): | |
| if ocr_result["text"][i].strip(): # Ignore empty text | |
| page_data.append({ | |
| "text": ocr_result["text"][i], | |
| "x0": ocr_result["left"][i], | |
| "y0": ocr_result["top"][i], | |
| "x1": ocr_result["left"][i] + ocr_result["width"][i], | |
| "y1": ocr_result["top"][i] + ocr_result["height"][i], | |
| "font_size": ocr_result["height"][i] | |
| }) | |
| extracted_data.append(page_data) | |
| # Cleanup temporary image | |
| os.remove(image_path) | |
| return extracted_data | |
| def overlay_text_with_fonts(pdf_path, extracted_data, output_pdf_path): | |
| """ | |
| Overlay extracted text onto the original PDF using PyMuPDF. | |
| :param pdf_path: Path to the input PDF file. | |
| :param extracted_data: Extracted text and positions. | |
| :param output_pdf_path: Path to save the output PDF file. | |
| """ | |
| doc = fitz.open(pdf_path) | |
| default_font = "Helvetica" | |
| for page_num, page_data in enumerate(extracted_data): | |
| page = doc[page_num] | |
| for item in page_data: | |
| page.insert_text( | |
| (item["x0"], item["y0"]), | |
| item["text"], | |
| fontsize=item["font_size"] / 2, # Adjust font size for better scaling | |
| fontname=default_font, | |
| color=(0, 0, 0) # Black text | |
| ) | |
| doc.save(output_pdf_path) | |
| print(f"PDF saved to: {output_pdf_path}") | |
| def process_pdf(uploaded_pdf, output_pdf_path): | |
| """ | |
| Process the uploaded PDF to extract text using Tesseract and overlay it. | |
| :param uploaded_pdf: Uploaded PDF file. | |
| :param output_pdf_path: Path to save the output PDF file. | |
| """ | |
| with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as temp_pdf: | |
| temp_pdf.write(uploaded_pdf.read()) | |
| temp_pdf_path = temp_pdf.name | |
| extracted_data = extract_text_with_tesseract(temp_pdf_path) | |
| overlay_text_with_fonts(temp_pdf_path, extracted_data, output_pdf_path) | |
| os.remove(temp_pdf_path) | |
| # Streamlit App | |
| def main(): | |
| st.title("PDF OCR and Text Conversion Tool") | |
| st.write("Upload a PDF to extract and overlay text as editable layers.") | |
| uploaded_file = st.file_uploader("Upload PDF", type=["pdf"]) | |
| if uploaded_file: | |
| output_pdf_path = "converted_output.pdf" | |
| with st.spinner("Processing your PDF..."): | |
| process_pdf(uploaded_file, output_pdf_path) | |
| st.success("PDF processing complete!") | |
| # Provide a download button for the processed PDF | |
| with open(output_pdf_path, "rb") as f: | |
| st.download_button( | |
| label="Download Converted PDF", | |
| data=f, | |
| file_name="converted_output.pdf", | |
| mime="application/pdf" | |
| ) | |
| os.remove(output_pdf_path) | |
| if __name__ == "__main__": | |
| main() | |