import os
#!/usr/bin/env -S uv run
import click
import cv2
import numpy as np
from PIL import Image
from pdf2image import convert_from_path
import img2pdf

def ensure_temp_dir():
    """Ensure temporary directory exists and return its path."""
    temp_dir = "temp_processed_images"
    os.makedirs(temp_dir, exist_ok=True)
    return temp_dir

def get_temp_files():
    """Get list of temporary PNG files in order."""
    temp_dir = ensure_temp_dir()
    files = [f for f in os.listdir(temp_dir) if f.endswith('.png')]
    files.sort()  # Ensure correct page order
    return [os.path.join(temp_dir, f) for f in files]

@click.group()
def cli():
    """PDF cleaning toolbox for musical scores."""
    pass

@cli.command()
@click.argument('input_pdf', type=click.Path(exists=True))
def extract(input_pdf):
    """Extract pages from PDF to temporary directory."""
    temp_dir = ensure_temp_dir()
    
    # Convert PDF to images
    print(f"Extracting pages from {input_pdf}...")
    pages = convert_from_path(input_pdf, dpi=400)
    
    # Save each page
    for i, page in enumerate(pages):
        output_path = os.path.join(temp_dir, f"page_{i:03d}.png")
        page.save(output_path, "PNG", optimize=False)
        print(f"Saved page {i+1}/{len(pages)}")
    
    print(f"Extracted {len(pages)} pages to {temp_dir}/")

@cli.command()
def deskew():
    """Deskew all pages in temporary directory."""
    temp_files = get_temp_files()
    if not temp_files:
        print("No pages found in temporary directory. Run 'extract' first.")
        return
    
    for file_path in temp_files:
        print(f"Deskewing {os.path.basename(file_path)}...")
        
        # Read image
        image = cv2.imread(file_path)
        
        # Convert to grayscale
        gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
        
        # Apply threshold to get binary image
        _, binary = cv2.threshold(gray, 127, 255, cv2.THRESH_BINARY)
        
        # Create a rectangular kernel that's wider than it is tall
        # This helps detect horizontal lines
        kernel_length = np.array(binary).shape[1]//80
        horizontal_kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (kernel_length, 1))
        
        # Detect horizontal lines
        horizontal_lines = cv2.erode(binary, horizontal_kernel, iterations=3)
        horizontal_lines = cv2.dilate(horizontal_lines, horizontal_kernel, iterations=3)
        
        # Use probabilistic Hough transform to detect line segments
        line_segments = cv2.HoughLinesP(
            cv2.bitwise_not(horizontal_lines),
            rho=1,
            theta=np.pi/180,
            threshold=100,
            minLineLength=binary.shape[1]//4,  # Lines must be at least 1/4 of image width
            maxLineGap=20
        )
        
        if line_segments is not None and len(line_segments) > 0:
            # Calculate angles of detected line segments
            angles = []
            for line in line_segments:
                x1, y1, x2, y2 = line[0]
                if x2 - x1 == 0:  # Avoid division by zero
                    continue
                angle = np.degrees(np.arctan2(y2 - y1, x2 - x1))
                # Only consider angles that are close to horizontal
                if abs(angle) < 20:
                    angles.append(angle)
            
            if angles:
                # Use median angle to avoid outliers
                median_angle = np.median(angles)
                
                # Only rotate if the angle is significant but not too large
                if 0.5 < abs(median_angle) < 20:
                    height, width = image.shape[:2]
                    center = (width/2, height/2)
                    rotation_matrix = cv2.getRotationMatrix2D(center, median_angle, 1.0)
                    rotated = cv2.warpAffine(image, rotation_matrix, (width, height),
                                           flags=cv2.INTER_CUBIC,
                                           borderMode=cv2.BORDER_REPLICATE)
                    
                    # Save rotated image
                    cv2.imwrite(file_path, rotated)
                    print(f"  Rotated by {median_angle:.2f} degrees")
                else:
                    print("  No significant rotation needed")
            else:
                print("  No valid horizontal lines found")
        else:
            print("  No line segments detected")

@cli.command()
@click.argument('output_pdf', type=click.Path())
def finalize(output_pdf):
    """Combine processed pages into final PDF and clean up."""
    temp_dir = ensure_temp_dir()
    temp_files = get_temp_files()
    if not temp_files:
        print("No pages found in temporary directory. Run 'extract' first.")
        return
    
    print(f"Combining {len(temp_files)} pages into {output_pdf}...")
    
    # Convert to PDF
    with open(output_pdf, "wb") as f:
        f.write(img2pdf.convert(temp_files))
    
    # Clean up temporary files
    print("Cleaning up temporary files...")
    for file_path in temp_files:
        os.remove(file_path)
    os.rmdir(temp_dir)
    
    print("PDF created successfully and temporary files removed!")

if __name__ == '__main__':
    cli()