rewrite jxl-restore -> jxl-convert

Signed-off-by: Martin Matous <m@matous.dev>

README.md

@@ -90,17 +90,14 @@ Dependencies: openssl
 Usage: ???
 
 
-## jxl-restore.py
-Check whether JPEG version of .jxl exists, remove .jxl if does.
-Attempt at rescuing collection where conversion got messed up.
+## jxl-convert.py
+Recursively convert jpgs to jxls with additional checks.
 
-Will fix one day. Maybe.
+Status: one-off
 
-Status: ancient one-off, broken
+Dependencies: libjxl
 
-Dependencies: None
-
-Usage: None
+Usage: `./jxl-convert.py /path/to/pics`
 
 
 ## invoke-magic.bat

jxl-convert.py

@@ -0,0 +1,70 @@
+#!/usr/bin/env python3
+
+# 3rd-party sw: cjxl (libjxl)
+
+import itertools
+import subprocess
+import sys
+from pathlib import Path
+from typing import Final
+
+source_dir: Final = Path(sys.argv[1])
+if len(sys.argv) > 2:
+    backup_dir: Final = Path(sys.argv[2])
+
+
+def check_source_against_convert(source_dir: Path) -> None:
+    for jxl in source_dir.rglob('*.jxl'):
+        has_original = jxl.with_suffix('.jpg').exists() or jxl.with_suffix('.JPG').exists()
+        if not has_original:
+            print(f'Source image for {jxl} has been lost')
+
+
+def check_source_against_backup(source_dir: Path, backup_dir: Path) -> None:
+    pics_src = { p.relative_to(source_dir) for p in source_dir.rglob('*.JPG') }
+    pics_src |= { p.relative_to(source_dir) for p in source_dir.rglob('*.jpg') }
+
+    pics_bak = { p.relative_to(backup_dir) for p in backup_dir.rglob('*.JPG') }
+    pics_bak |= { p.relative_to(backup_dir) for p in backup_dir.rglob('*.jpg') }
+
+    diff = pics_src - pics_bak
+    if diff:
+        print('Source contains the following additional images:\n', '\n'.join(map(str, diff)))
+    diff = pics_bak - pics_src
+    if diff:
+        print('Backup contains the following additional images:\n', '\n'.join(map(str, diff)))
+
+
+def check_successful_conversion(source_dir: Path) -> None:
+    pics_src = itertools.chain(source_dir.rglob('*.jpg'), source_dir.rglob('*.JPG'))
+
+    for jpg in pics_src:
+        jxl = jpg.with_suffix('.jxl')
+        if not jxl.exists():
+            print(f'Converted counterpart for {jpg} does not exist')
+
+
+def convert(source_dir: Path) -> None:
+    pics_src_it = itertools.chain(source_dir.rglob('*.jpg'), source_dir.rglob('*.JPG'))
+    # set offset to start from image #offset (e.g. in case of a previous error)
+    # e.g. offset = 5 will start with converting 5th image
+    offset = 0
+    pics_src = list(sorted(pics_src_it))[offset:]
+    for i, jpg in enumerate(pics_src):
+        jxl = jpg.with_suffix('.jxl')
+        cmd = ['cjxl', '-v', '--lossless_jpeg=1', '--effort=9', '--brotli_effort=11', jpg, jxl]
+        try:
+            subprocess.run(cmd, check=True, capture_output=True, encoding='utf-8') # type: ignore[arg-type]
+        except subprocess.CalledProcessError as ex:
+            print(f'ERR processing file #{i + offset} {jpg}: {ex}')
+            print(f'cjxl says: {ex.stderr}\n{ex.stdout}')
+            sys.exit(ex.returncode)
+        except Exception as ex:
+            print(f'ERR unexpected: {ex}')
+            sys.exit(1)
+        print(f'Converted #{i + offset} {jpg}')
+
+# check_source_against_convert(source_dir)
+# check_source_against_backup(source_dir, second_dir)
+convert(source_dir)
+# check_successful_conversion(source_dir)

jxl-restore.py

@@ -1,38 +0,0 @@
-#!/usr/bin/env python
-
-import os
-import shutil
-import subprocess
-import sys
-from pathlib import Path
-from typing import Final
-
-source: Final = Path(sys.argv[1])
-dest: Final = Path(sys.argv[2])
-
-def find_source_jpg(dest_name: Path) -> Path:
-    ext_list = ['.jpg', '.jpeg', '.JPG', '.JPEG']
-    for ext in ext_list:
-        res = (Path(source) / relative_path / (dest_name + ext))
-        if os.path.exists(dest_file_path):
-            return res
-    raise RuntimeError('No original found', dest_name)
-
-for dest_root, dirs, files in os.walk(dest):
-    for name in files:
-        dest_name, dest_ext = os.path.splitext(name)
-        if dest_ext.lower() != '.jxl':
-            continue
-        dest_file_path = (Path(dest_root) / (dest_name + '.jpg'))
-        print('scanning for ', dest_file_path, '\n')
-        if os.path.exists(dest_file_path): # both .jxl and jpg exist
-            print(dest_file_path, ' already exists\n')
-            continue
-        relative_path = os.path.relpath(dest_root, dest)
-        print('relapath ', relative_path, '\n')
-        src_file_path = find_source_jpg(dest_name)
-        shutil.copy2(src_file_path, dest_file_path)
-        print('restored ', dest_file_path, ' from ', src_file_path, '\n')
-        #dest_jxl = (Path(dest_root) / (dest_name + '.jxl'))
-        #os.remove(dest_jxl)
-        #print('deleted ', dest_jxl, '\n')