app.py

import gradio as gr
import io
import sys

try:
    from doctr.io import DocumentFile
    from doctr.models import ocr_predictor
except ImportError:
    print("Error: Failed to import doctr. Please ensure it's installed correctly.")
    print("Python version:", sys.version)
    print("Python path:", sys.path)
    raise

# Initialize the OCR model
try:
    model = ocr_predictor(det_arch='db_resnet50', reco_arch='crnn_vgg16_bn', pretrained=True)
except Exception as e:
    print(f"Error initializing OCR model: {e}")
    raise

def ocr_process(file):
    try:
        # Read the uploaded file
        if file.name.lower().endswith('.pdf'):
            doc = DocumentFile.from_pdf(file.name)
        else:
            # Assume it's an image if not PDF
            image_stream = io.BytesIO(file.read())
            doc = DocumentFile.from_images(image_stream)
        
        # Perform OCR
        result = model(doc)
        
        # Extract text from the result
        extracted_text = ""
        for page in result.pages:
            for block in page.blocks:
                for line in block.lines:
                    for word in line.words:
                        extracted_text += word.value + " "
                    extracted_text += "\n"
                extracted_text += "\n"
        
        return extracted_text.strip()
    except Exception as e:
        return f"Error processing file: {str(e)}"

# Create Gradio interface
iface = gr.Interface(
    fn=ocr_process,
    inputs=gr.File(label="Upload PDF or Image"),
    outputs=gr.Textbox(label="Extracted Text"),
    title="OCR with doctr",
    description="Upload a PDF or image file to extract text using OCR."
)

# Launch the interface
iface.launch()