app.py

import pandas as pd
import PIL
from PIL import Image
from PIL import ImageDraw
import gradio as gr
import torch
import easyocr
import fitz  # PyMuPDF

# Function to extract images from PDF
def pdf_to_images(pdf_path):
    doc = fitz.open(pdf_path)
    images = []
    for page_num in range(len(doc)):
        page = doc.load_page(page_num)
        pix = page.get_pixmap()
        img = Image.frombytes("RGB", [pix.width, pix.height], pix.samples)
        images.append(img)
    return images

def draw_boxes(image, bounds, color='yellow', width=2):
    draw = ImageDraw.Draw(image)
    for bound in bounds:
        p0, p1, p2, p3 = bound[0]
        draw.line([*p0, *p1, *p2, *p3, *p0], fill=color, width=width)
    return image

def inference(pdf, lang):
    reader = easyocr.Reader(lang)
    images = pdf_to_images(pdf.name)
    results = []
    for i, img in enumerate(images):
        img_path = f'page_{i + 1}.jpg'
        img.save(img_path)
        bounds = reader.readtext(img_path)
        draw_boxes(img, bounds)
        result_img_path = f'result_{i + 1}.jpg'
        img.save(result_img_path)
        results.append((result_img_path, pd.DataFrame(bounds).iloc[:, 1:]))
    return results

title = 'EasyOCR'
description = 'Realtime EasyOCR.'
article = "<p style='text-align: center'><a href='https://www.jaided.ai/easyocr/'>OCR for written scripts.</a> | <a href='https://github.com/JaidedAI/EasyOCR'>Github Repo</a></p>"
examples = [['example.pdf',['en']]]
css = ".output_image, .input_image {height: 40rem !important; width: 100% !important;}"
choices = [
    "en",
    "hi",
   ]
gr.Interface(
    inference,
    [gr.inputs.File(type='file', label='Input PDF'), gr.inputs.CheckboxGroup(choices, type="value", default=['en'], label='language')],
    [gr.outputs.File(type='file', label='Output Images'), gr.outputs.Dataframe(headers=['text', 'confidence'])],
    title=title,
    description=description,
    article=article,
    examples=examples,
    css=css,
    enable_queue=True
    ).launch(debug=True)