比較スクリプト
import os
from PIL import Image
import pyheif
import imagehash
from pathlib import Path
import shutil
# === 設定 ===
HEIC_ROOT = Path("＜HEICファイルのパス＞")
JPG_ROOT = Path("＜JPGファイルのパス＞")
OUTPUT_DIR = Path("compare_output")
MATCH_LIST_FILE = Path("matched_files.txt")
HASH_FUNC = imagehash.phash
# === 準備 ===
OUTPUT_DIR.mkdir(exist_ok=True)
heic_hashes = dict()
matched_pairs = []
heic_total = 0
jpg_total = 0
heic_errors = 0
jpg_errors = 0
print("🔍 HEICファイルをスキャン中...")
#print(f"📁 HEICルート: {HEIC_ROOT}")
#if not HEIC_ROOT.exists():
# print("❌ 指定したHEICフォルダが存在しません。パスを確認してください。")
# exit(1)
#else:
# print("📂 HEICフォルダ一覧:")
# for path in HEIC_ROOT.rglob("*"):
# print(" ", path)
#for heic_file in HEIC_ROOT.rglob("*.heic"):
for ext in ["*.heic", "*.HEIC"]:
for heic_file in HEIC_ROOT.rglob(ext):
try:
heif = pyheif.read(heic_file)
image = Image.frombytes(heif.mode, heif.size, heif.data, "raw", heif.mode)
h = str(HASH_FUNC(image))
heic_hashes[h] = heic_file
heic_total += 1
except Exception as e:
print(f"⚠️ HEICエラー: {heic_file} → {e}")
heic_errors += 1
print(f"📸 HEIC読み込み成功: {heic_total} 件 / エラー: {heic_errors} 件")
print("📋 JPGファイルと照合中...")
for ext in ["*.jpg", "*.JPG"]:
for jpg_file in JPG_ROOT.rglob(ext):
try:
with Image.open(jpg_file) as image:
h = str(HASH_FUNC(image))
if h in heic_hashes:
matched_pairs.append((jpg_file, heic_hashes[h]))
jpg_total += 1
except Exception as e:
print(f"⚠️ JPGエラー: {jpg_file} → {e}")
jpg_errors += 1
print(f"📸 JPG読み込み成功: {jpg_total} 件 / エラー: {jpg_errors} 件")
print(f"🔗 一致したペア数: {len(matched_pairs)}")
# === HTMLと画像出力 ===
html_lines = [
"<html><head><meta charset='utf-8'>",
"<style>body{font-family:sans-serif;} img{max-width:300px; margin:5px;} .pair{margin-bottom:40px;}</style>",
"</head><body><h1>HEIC vs JPG 比較結果</h1>"
]
with open(MATCH_LIST_FILE, "w", encoding="utf-8") as f:
for i, (jpg_path, heic_path) in enumerate(matched_pairs):
base = f"pair_{i}"
jpg_out = OUTPUT_DIR / f"{base}_jpg.jpg"
heic_out = OUTPUT_DIR / f"{base}_heic.jpg"
try:
shutil.copy2(jpg_path, jpg_out)
heif = pyheif.read(heic_path)
heic_img = Image.frombytes(heif.mode, heif.size, heif.data, "raw", heif.mode)
heic_img.save(heic_out, "JPEG")
html_lines.append(f"<div class='pair'><h3>{base}</h3>")
html_lines.append(f"<p><b>JPG:</b> {jpg_path}<br><b>HEIC:</b> {heic_path}</p>")
html_lines.append(f"<img src='{jpg_out.name}' alt='JPG'>")
html_lines.append(f"<img src='{heic_out.name}' alt='HEIC'>")
html_lines.append("</div>")
f.write(str(jpg_path) + "\n")
except Exception as e:
print(f"⚠️ 出力失敗: {jpg_path} / {heic_path} → {e}")
html_lines.append("</body></html>")
with open(OUTPUT_DIR / "compare.html", "w", encoding="utf-8") as f:
f.write("\n".join(html_lines))
print(f"✅ 完了！{len(matched_pairs)} ペアを {OUTPUT_DIR}/ に保存しました。")
print("🖼 `compare_output/compare.html` をブラウザで開いて確認してください。")
print(f"📝 削除対象候補リスト: {MATCH_LIST_FILE}")
削除スクリプト
from pathlib import Path
MATCH_LIST_FILE = Path("matched_files.txt")
deleted = 0
if not MATCH_LIST_FILE.exists():
print("❌ matched_files.txt が見つかりません。先に比較スクリプトを実行してください。")
exit(1)
with open(MATCH_LIST_FILE, "r", encoding="utf-8") as f:
jpg_paths = [Path(line.strip()) for line in f if line.strip()]
print(f"🗑 削除対象 JPG 数: {len(jpg_paths)}")
for jpg_path in jpg_paths:
if jpg_path.exists():
try:
jpg_path.unlink()
print(f"✅ 削除: {jpg_path}")
deleted += 1
except Exception as e:
print(f"⚠️ 削除失敗: {jpg_path} → {e}")
else:
print(f"⚠️ ファイル存在しない: {jpg_path}")
print(f"✅ 削除完了: {deleted} 件")
