Create app.py

app.py

@@ -0,0 +1,110 @@
+import gradio as gr
+import cv2
+import numpy as np
+from PIL import Image
+
+def extract_first_line(img_pil):
+    if img_pil is None:
+        return None
+    
+    # Convert to OpenCV format
+    img = np.array(img_pil)
+    if len(img.shape) == 3:
+        img_cv = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
+        gray = cv2.cvtColor(img_cv, cv2.COLOR_BGR2GRAY)
+    else:
+        gray = img
+    
+    # Adaptive thresholding for better handling of uneven lighting
+    # Use lower threshold values for faded manuscripts
+    thresh = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, 
+                                 cv2.THRESH_BINARY_INV, 11, 2)
+    
+    # Alternative: Try multiple threshold values and pick the best
+    # thresh = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)[1]
+    
+    # Smaller morphological kernel to avoid merging separate lines
+    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (20, 3))
+    dilated = cv2.dilate(thresh, kernel, iterations=1)
+    
+    # Find contours (potential lines)
+    contours, _ = cv2.findContours(dilated, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
+    
+    # Filter contours by size to remove noise
+    min_width = img.shape[1] // 10  # Minimum 10% of image width
+    min_height = 5  # Minimum height in pixels
+    
+    valid_contours = []
+    for contour in contours:
+        x, y, w, h = cv2.boundingRect(contour)
+        if w >= min_width and h >= min_height:
+            valid_contours.append((x, y, w, h))
+    
+    # Sort contours by vertical position (top to bottom)
+    if valid_contours:
+        bounding_boxes = sorted(valid_contours, key=lambda b: b[1])
+        
+        # Get topmost box (first visible line)
+        x, y, w, h = bounding_boxes[0]
+        
+        # Add margin but ensure we don't go out of bounds
+        margin = 15
+        y_start = max(0, y - margin)
+        y_end = min(img.shape[0], y + h + margin)
+        x_start = max(0, x - margin)
+        x_end = min(img.shape[1], x + w + margin)
+        
+        crop = img[y_start:y_end, x_start:x_end]
+        
+        if crop.size > 0:
+            return Image.fromarray(crop)
+    
+    # Fallback: return original image if no lines detected
+    return img_pil
+
+def preprocess_voynich_image(img_pil):
+    """Additional preprocessing specifically for Voynich manuscript images"""
+    if img_pil is None:
+        return None
+    
+    img = np.array(img_pil)
+    
+    # Enhance contrast for faded manuscripts
+    lab = cv2.cvtColor(img, cv2.COLOR_RGB2LAB)
+    l, a, b = cv2.split(lab)
+    clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8,8))
+    l = clahe.apply(l)
+    enhanced = cv2.merge([l, a, b])
+    enhanced = cv2.cvtColor(enhanced, cv2.COLOR_LAB2RGB)
+    
+    return Image.fromarray(enhanced)
+
+# Create interface with preprocessing option
+with gr.Blocks(title="Voynich Manuscript Line Extractor") as demo:
+    gr.Markdown("# Voynich Manuscript Line Extractor")
+    gr.Markdown("Upload a scanned folio of the Voynich manuscript. The app will detect and crop the first visible line of text.")
+    
+    with gr.Row():
+        with gr.Column():
+            input_image = gr.Image(type="pil", label="Upload Voynich Folio")
+            enhance_btn = gr.Button("Enhance Image First")
+            extract_btn = gr.Button("Extract First Line")
+            
+        with gr.Column():
+            enhanced_output = gr.Image(label="Enhanced Image")
+            line_output = gr.Image(label="Extracted First Line")
+    
+    enhance_btn.click(
+        fn=preprocess_voynich_image,
+        inputs=input_image,
+        outputs=enhanced_output
+    )
+    
+    extract_btn.click(
+        fn=extract_first_line,
+        inputs=input_image,
+        outputs=line_output
+    )
+
+if __name__ == "__main__":
+    demo.launch()