Update app.py

app.py

@@ -1,107 +1,123 @@
-import gradio as gr
-import easyocr
+import PIL
 from PIL import Image
-import pdf2image
-import tempfile
-import os
-import cv2
-import numpy as np
+from PIL import ImageDraw
+import gradio as gr
 import torch
+import easyocr
+import re
 
-# Initialize the OCR reader (this will download models on first run)
-reader = easyocr.Reader(['en'], gpu=torch.cuda.is_available())
+# Download example images (same as before)
+# Download example images
+torch.hub.download_url_to_file('https://github.com/JaidedAI/EasyOCR/raw/master/examples/english.png', 'english.png')
+torch.hub.download_url_to_file('https://i.imgur.com/mwQFd7G.jpeg', 'Hindi.jpeg')
+torch.hub.download_url_to_file('https://github.com/JaidedAI/EasyOCR/raw/master/examples/thai.jpg', 'thai.jpg')
+torch.hub.download_url_to_file('https://github.com/JaidedAI/EasyOCR/raw/master/examples/french.jpg', 'french.jpg')
+torch.hub.download_url_to_file('https://github.com/JaidedAI/EasyOCR/raw/master/examples/chinese.jpg', 'chinese.jpg')
+torch.hub.download_url_to_file('https://github.com/JaidedAI/EasyOCR/raw/master/examples/japanese.jpg', 'japanese.jpg')
+torch.hub.download_url_to_file('https://github.com/JaidedAI/EasyOCR/raw/master/examples/korean.png', 'korean.png')
 
-def preprocess_image(img):
-    """Preprocess image to improve OCR accuracy for handwritten text"""
-    # Convert PIL Image to numpy array
-    img_array = np.array(img)
-    
-    # Check if the image is already grayscale
-    if len(img_array.shape) == 2:
-        gray = img_array
-    else:
-        gray = cv2.cvtColor(img_array, cv2.COLOR_RGB2GRAY)
-    
-    # Apply adaptive thresholding for better handling of different lighting conditions
-    binary = cv2.adaptiveThreshold(
-        gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2
-    )
+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 format_extracted_text(bounds):
+    return " ".join([text for _, text, _ in bounds])
+
+def highlight_search_results(text, search_query):
+    if not search_query:
+        return text, []
+    pattern = re.compile(re.escape(search_query), re.IGNORECASE)
+    matches = list(pattern.finditer(text))
+    highlighted_text = pattern.sub(lambda m: f"**{m.group()}**", text)
+    return highlighted_text, matches
+
+def inference(img, lang):
+    reader = easyocr.Reader(lang)
+    bounds = reader.readtext(img)
+    im = PIL.Image.open(img)
+    draw_boxes(im, bounds)
+    im.save('result.jpg')
     
-    # Noise removal
-    kernel = np.ones((1, 1), np.uint8)
-    binary = cv2.morphologyEx(binary, cv2.MORPH_CLOSE, kernel)
-    binary = cv2.morphologyEx(binary, cv2.MORPH_OPEN, kernel)
+    extracted_text = format_extracted_text(bounds)
     
-    return binary
+    return ['result.jpg', extracted_text]
 
-def extract_text_from_image(img):
-    """Extract text from an image using EasyOCR"""
-    # Preprocess for better handwriting recognition
-    processed_img = preprocess_image(img)
+def search_text(text, search_query):
+    highlighted_text, matches = highlight_search_results(text, search_query)
     
-    # Use EasyOCR to extract text
-    results = reader.readtext(processed_img)
-    
-    # Combine all detected text
-    text = '\n'.join([result[1] for result in results])
+    if matches:
+        result = f"Found {len(matches)} occurrence(s) of \"{search_query}\":\n"
+        for i, match in enumerate(matches, 1):
+            context_start = max(0, match.start() - 20)
+            context_end = min(len(text), match.end() + 20)
+            context = text[context_start:context_end]
+            result += f"{i}. ...{context}...\n"
+    else:
+        result = f"No occurrences of \"{search_query}\" found."
     
-    return text.strip()
+    return highlighted_text, result
 
-def extract_text_from_pdf(pdf_path):
-    """Extract text from all pages of a PDF file"""
-    # Convert PDF to images
-    with tempfile.TemporaryDirectory() as path:
-        images = pdf2image.convert_from_path(pdf_path, output_folder=path)
-        
-        # Extract text from each page
-        full_text = []
-        for img in images:
-            text = extract_text_from_image(img)
-            full_text.append(text)
-        
-        return "\n\n--- Page Break ---\n\n".join(full_text)
+title = 'Image To Text OCR Converter'
+subtitle = 'Extract Hindi/English or both or any Text From Image'
+description = 'This application is being built on the request of IIT R Internship Assignment. It allows users to upload a single image, processes the image to extract text using OCR, and provides a basic search feature.'
 
-def process_file(file):
-    """Process the uploaded file (PDF or image)"""
-    if file is None:
-        return "No file uploaded. Please upload an image or PDF file."
-    
-    try:
-        file_extension = os.path.splitext(file.name)[1].lower()
-        
-        if file_extension == ".pdf":
-            # Process PDF
-            return extract_text_from_pdf(file.name)
-        elif file_extension in [".jpg", ".jpeg", ".png", ".bmp", ".tiff", ".tif"]:
-            # Process Image
-            img = Image.open(file.name)
-            return extract_text_from_image(img)
-        else:
-            return "Unsupported file format. Please upload a PDF or image file (JPG, PNG, BMP, TIFF)."
-    except Exception as e:
-        return f"Error processing file: {str(e)}"
-
-# Create Gradio interface
-with gr.Blocks(title="Handwritten Text OCR Extractor") as app:
-    gr.Markdown("# Handwritten Text OCR Extraction Tool")
-    gr.Markdown("Upload an image or PDF containing handwritten text to extract the content.")
-    
+note = 'Please keep patience while processing the OCR, as it may take a few seconds to complete'
+
+
+alternative_link = "[Alternative: Ready-to-use OCR using Vercel](https://iitr-haq-nawaz-maliks-projects.vercel.app/)"
+
+examples = [
+    ['english.png', ['en']],
+    ['Hindi.jpeg', ['hi', 'en']],
+    ['thai.jpg', ['th', 'en']],
+    ['french.jpg', ['fr', 'en']],
+    ['chinese.jpg', ['ch_sim', 'en']],
+    ['japanese.jpg', ['ja', 'en' ]],
+    ['korean.png', ['ko', 'en' ]]
+]
+
+css = """
+.output_image, .input_image {height: 40rem !important; width: 100% !important;}
+.search_results {margin-top: 1rem; padding: 1rem; background-color: #f0f0f0; border-radius: 4px;}
+.centered-title {text-align: center; font-size: 2.5em; font-weight: bold; margin-bottom: 0.5em;}
+.centered-subtitle {text-align: center; font-size: 1.5em; margin-bottom: 1em;}
+.alternative-link {text-align: center; margin-top: 1em; font-style: italic;}
+"""
+
+
+choices = [
+    "abq", "ady", "af", "ang", "ar", "as", "ava", "az", "be", "bg", "bh", "bho", "bn", "bs", "ch_sim", "ch_tra", 
+    "che", "cs", "cy", "da", "dar", "de", "en", "es", "et", "fa", "fr", "ga", "gom", "hi", "hr", "hu", "id", 
+    "inh", "is", "it", "ja", "kbd", "kn", "ko", "ku", "la", "lbe", "lez", "lt", "lv", "mah", "mai", "mi", "mn", 
+    "mr", "ms", "mt", "ne", "new", "nl", "no", "oc", "pi", "pl", "pt", "ro", "ru", "rs_cyrillic", "rs_latin", 
+    "sck", "sk", "sl", "sq", "sv", "sw", "ta", "tab", "te", "th", "tjk", "tl", "tr", "ug", "uk", "ur", "uz", "vi"
+]
+
+with gr.Blocks(css=css) as iface:
+    gr.Markdown(f"# {title}")
+    gr.Markdown(f"## {subtitle}")
+    gr.Markdown(description)
+    gr.Markdown(note)
+    gr.Markdown(alternative_link)
     with gr.Row():
-        with gr.Column():
-            file_input = gr.File(label="Upload Image or PDF")
-            extract_button = gr.Button("Extract Text")
+        with gr.Column(scale=2):
+            input_image = gr.Image(type="filepath", label="Upload Image")
+            lang_select = gr.CheckboxGroup(choices=choices, label="Select Languages", value=['hi', 'en'])
+            ocr_button = gr.Button("Perform OCR")
         
-        with gr.Column():
-            text_output = gr.Textbox(label="Extracted Text", lines=10, placeholder="Extracted text will appear here...")
+        with gr.Column(scale=3):
+            output_image = gr.Image(type="filepath", label="OCR Result")
+            extracted_text = gr.Markdown(label="Extracted Text")
+            search_box = gr.Textbox(label="Search in extracted text")
+            search_button = gr.Button("Search")
+            search_results = gr.Markdown(label="Search Results")
     
-    extract_button.click(fn=process_file, inputs=[file_input], outputs=[text_output])
+    ocr_button.click(inference, inputs=[input_image, lang_select], outputs=[output_image, extracted_text])
+    search_button.click(search_text, inputs=[extracted_text, search_box], outputs=[extracted_text, search_results])
     
-    gr.Markdown("### Notes:")
-    gr.Markdown("- For best results, ensure the handwriting is clear and the image is well-lit")
-    gr.Markdown("- The system works best with dark text on light background")
-    gr.Markdown("- The first run may take longer as it downloads the OCR models")
-    gr.Markdown("- Multiple page PDFs will show page breaks in the output")
-
-# Launch the app
-if __name__ == "__main__":
-    app.launch()
+    gr.Examples(examples, inputs=[input_image, lang_select])
+
+iface.launch()