Compare commits

..

5 Commits

Author SHA1 Message Date
Nathan 5ab5571cf4 ignore specst/vscode 2026-05-21 21:29:25 -06:00
Nathan fd5c8a9ab4 use over 61 workers 2025-12-23 17:01:00 -07:00
Nathan f1cf096069 specst 2025-12-17 15:36:54 -07:00
Nathan 1885b1d55d include jpgs skipped in skipped count 2025-12-11 15:32:18 -07:00
Nathan 4416949d7c detect and quarantine corrupt pngs 2025-12-11 15:21:36 -07:00
4 changed files with 143 additions and 3032 deletions
+5
View File
@@ -1,3 +1,8 @@
input/
output/
corrupted/
.vscode/
.specstory/
-4
View File
@@ -1,4 +0,0 @@
# SpecStory project identity file
/.project.json
# SpecStory explanation file
/.what-is-this.md
+135 -11
View File
@@ -7,12 +7,21 @@ Compresses all PNG files in subdirectories with maximum parallelism.
import os
import sys
import argparse
import platform
from pathlib import Path
from concurrent.futures import ProcessPoolExecutor, as_completed
from PIL import Image
import multiprocessing
import time
# Try to unlock ProcessPoolExecutor on Windows to bypass 61-worker limit
try:
import unlock_processpool
unlock_processpool.please()
UNLOCKED = True
except ImportError:
UNLOCKED = False
def compress_png(input_path, output_path, force_bitdepth=None):
"""Compress a single PNG file.
@@ -31,13 +40,55 @@ def compress_png(input_path, output_path, force_bitdepth=None):
new_size = os.path.getsize(output_path)
savings = original_size - new_size
savings_pct = (savings / original_size * 100) if original_size > 0 else 0
return (str(input_path), True, None, original_size, new_size, savings_pct, True)
return (str(input_path), True, None, original_size, new_size, savings_pct, True, False)
# Ensure output directory exists
output_path.parent.mkdir(parents=True, exist_ok=True)
original_size = os.path.getsize(input_path)
# Check for corrupted/empty files
if original_size == 0:
return (str(input_path), False, "CORRUPTED: File is 0 bytes (empty/placeholder)", 0, 0, 0, False, True)
# Try to open and validate the image
try:
img = Image.open(input_path)
# Force load to detect corruption early
img.load()
except Exception as e:
return (str(input_path), False, f"CORRUPTED: Cannot open/load image - {str(e)}", original_size, 0, 0, False, True)
# Validate image dimensions
if img.width == 0 or img.height == 0:
return (str(input_path), False, f"CORRUPTED: Invalid dimensions ({img.width}x{img.height})", original_size, 0, 0, False, True)
# Check if image appears completely black (potential corruption indicator)
# But be careful - some images might legitimately be black
# We'll only flag this if the file is suspiciously small for its dimensions
try:
# Sample a few pixels to check if all are black
sample_size = min(100, img.width * img.height)
pixels = list(img.getdata()[:sample_size])
if pixels:
# Check if all sampled pixels are black (0 or (0,0,0) or (0,0,0,0))
all_black = True
for pixel in pixels:
if isinstance(pixel, (tuple, list)):
if any(p > 0 for p in pixel):
all_black = False
break
else:
if pixel > 0:
all_black = False
break
# If all black AND file is suspiciously small, flag as potentially corrupted
if all_black and original_size < (img.width * img.height * 0.1): # Less than 0.1 bytes per pixel
return (str(input_path), False, f"CORRUPTED: Image appears all black with suspiciously small file size", original_size, 0, 0, False, True)
except:
# If we can't check pixels, continue anyway
pass
# Determine target bit depth
if force_bitdepth == '8':
@@ -131,9 +182,11 @@ def compress_png(input_path, output_path, force_bitdepth=None):
new_size = os.path.getsize(output_path)
savings = original_size - new_size
savings_pct = (savings / original_size * 100) if original_size > 0 else 0
return (str(input_path), True, None, original_size, new_size, savings_pct, False)
return (str(input_path), True, None, original_size, new_size, savings_pct, False, False)
except Exception as e:
return (str(input_path), False, str(e), 0, 0, 0, False)
# Check if error might indicate corruption
is_corrupted = "truncated" in str(e).lower() or "cannot identify" in str(e).lower() or "corrupt" in str(e).lower()
return (str(input_path), False, str(e), 0, 0, 0, False, is_corrupted)
def find_image_files(input_dir):
"""Find all PNG and JPG files in subdirectories."""
@@ -163,6 +216,31 @@ def get_output_path(input_path, input_dir, output_dir):
# Create output path
return output_dir / relative_path
def move_to_corrupted(input_path, input_dir, corrupted_dir):
"""Move a corrupted file to the corrupted folder, preserving directory structure."""
try:
input_path = Path(input_path)
input_dir = Path(input_dir)
corrupted_dir = Path(corrupted_dir)
# Get relative path from input directory
try:
relative_path = input_path.relative_to(input_dir)
except ValueError:
relative_path = input_path.name
# Create destination path
dest_path = corrupted_dir / relative_path
# Ensure destination directory exists
dest_path.parent.mkdir(parents=True, exist_ok=True)
# Move the file
input_path.rename(dest_path)
return True, None
except Exception as e:
return False, str(e)
def format_size(size_bytes):
"""Format file size in human readable format."""
for unit in ['B', 'KB', 'MB', 'GB']:
@@ -235,17 +313,32 @@ def main():
print(f"Found {len(png_files)} PNG files to process.")
# Create output directory
# Create output and corrupted directories
output_dir.mkdir(exist_ok=True)
corrupted_dir = Path('corrupted')
corrupted_dir.mkdir(exist_ok=True)
print(f"Corrupted files will be moved to: {corrupted_dir}")
# Use all available CPU cores
max_workers = multiprocessing.cpu_count()
cpu_count = multiprocessing.cpu_count()
if platform.system() == 'Windows' and not UNLOCKED:
# Windows ProcessPoolExecutor has a maximum of 61 workers (unless unlocked)
max_workers = min(cpu_count, 61)
if cpu_count > 61:
print(f"Detected {cpu_count} CPU threads, but Windows limits ProcessPoolExecutor to 61 workers.")
print("Install 'unlock-processpool-win' package to use all cores: pip install unlock-processpool-win")
else:
max_workers = cpu_count
if UNLOCKED:
print(f"Using unlock-processpool-win to bypass Windows 61-worker limit")
print(f"Using {max_workers} worker processes for compression...")
print("-" * 80)
compressed = 0
skipped = 0
skipped = len(jpg_files) # Start with JPG files counted as skipped
failed = 0
corrupted = 0
corrupted_files = []
total_original_size = 0
total_new_size = 0
start_time = time.time()
@@ -261,7 +354,7 @@ def main():
# Process results as they complete
for future in as_completed(future_to_file):
result = future.result()
file_path, success, error, orig_size, new_size, savings_pct, was_skipped = result
file_path, success, error, orig_size, new_size, savings_pct, was_skipped, is_corrupted = result
if success:
if was_skipped:
@@ -280,7 +373,7 @@ def main():
# Update every file or every 0.5 seconds, whichever comes first
if processed == 1 or time_since_update >= 0.5:
rate = processed / elapsed if elapsed > 0 else 0
remaining = len(png_files) - processed
remaining = len(png_files) - processed - failed
eta_seconds = remaining / rate if rate > 0 and remaining > 0 else 0
total_savings = total_original_size - total_new_size
@@ -290,11 +383,22 @@ def main():
elapsed_str = format_time(elapsed)
print(f"[{processed:5d}/{len(png_files)}] "
f"Compressed: {compressed} | Skipped: {skipped} | "
f"Compressed: {compressed} | Skipped: {skipped} | Corrupted: {corrupted} | "
f"Speed: {rate:.1f} files/sec | "
f"Elapsed: {elapsed_str} | ETA: {eta_str} | "
f"Saved: {format_size(total_savings)} ({total_savings_pct:.1f}%)", end='\r')
last_update_time = current_time
else:
if is_corrupted:
corrupted += 1
# Move corrupted file to corrupted folder
move_success, move_error = move_to_corrupted(file_path, input_dir, corrupted_dir)
if move_success:
corrupted_files.append((file_path, error, f"Moved to {corrupted_dir / Path(file_path).relative_to(input_dir)}"))
print(f"\n[CORRUPTED] {file_path}: {error} -> Moved to corrupted folder")
else:
corrupted_files.append((file_path, error, f"Failed to move: {move_error}"))
print(f"\n[CORRUPTED] {file_path}: {error} (Failed to move: {move_error})")
else:
failed += 1
print(f"\n[ERROR] Failed to compress {file_path}: {error}")
@@ -308,9 +412,13 @@ def main():
print("\n" + "=" * 80)
print(f"Compression complete!")
print(f"Successfully compressed: {compressed} files")
print(f"Skipped (already exist): {skipped} files")
skipped_existing = skipped - len(jpg_files) if skipped >= len(jpg_files) else 0
if jpg_files:
print(f"Ignored (JPG/JPEG): {len(jpg_files)} files")
print(f"Skipped: {skipped} files ({skipped_existing} already exist, {len(jpg_files)} JPG/JPEG)")
else:
print(f"Skipped (already exist): {skipped} files")
if corrupted > 0:
print(f"Corrupted (bad PNGs): {corrupted} files")
print(f"Failed: {failed} files")
print(f"Total time: {format_time(total_time)}")
print(f"Average speed: {avg_rate:.2f} files/second")
@@ -318,6 +426,22 @@ def main():
print(f"Compressed size: {format_size(total_new_size)}")
print(f"Total savings: {format_size(total_savings)} ({total_savings_pct:.1f}%)")
# Print list of corrupted files if any
if corrupted_files:
print("\n" + "=" * 80)
print("CORRUPTED FILES LIST:")
print("=" * 80)
for item in corrupted_files:
if len(item) == 3:
file_path, error, move_status = item
print(f" {file_path}")
print(f" Reason: {error}")
print(f" Status: {move_status}")
else:
file_path, error = item
print(f" {file_path}")
print(f" Reason: {error}")
if __name__ == '__main__':
main()