Centralise file type detection and add support for additional RAW for…

README.md

@@ -6,8 +6,11 @@ Command-line tools for organising photos by date, managing RAW/JPG pairs, and op
 
 ## Supported formats
 
-- RAW: .raf
-- JPG: .jpg, .jpeg
+- **RAW**: `.cr2`, `.cr3`, `.nef`, `.arw`, `.raf`, `.orf`, `.rw2`, `.dng`, `.pef`, `.srw`, `.x3f`
+- JPG: `.jpg`, `.jpeg`
+
+
+> ⚠️ **Note:** Detection is based on file extensions only (case-insensitive). Files with incorrect or missing extensions may not be handled correctly.
 
 ## Prerequisites - System Tools
 

src/photo_tools/commands/clean_unpaired_raws.py

@@ -6,9 +6,6 @@
 
 logger = logging.getLogger(__name__)
 
-RAW_EXTENSIONS = {".raf"}
-JPG_EXTENSIONS = {".jpg", ".jpeg"}
-
 Reporter = Callable[[str, str], None]
 RawMatcher = Callable[[Path, list[Path]], bool]
 

src/photo_tools/commands/keep_five_star_raws.py

@@ -7,8 +7,6 @@
 
 logger = logging.getLogger(__name__)
 
-RAW_EXTENSIONS = {".raf"}
-JPG_EXTENSIONS = {".jpg", ".jpeg"}
 
 Reporter = Callable[[str, str], None]
 RawMatcher = Callable[[Path, list[Path]], bool]

src/photo_tools/commands/optimise.py

@@ -5,11 +5,11 @@
 from PIL import Image
 
 from photo_tools.core.validation import validate_input_dir
+from photo_tools.image.file_types import is_jpg
 from photo_tools.image.optimisation import optimise_jpeg, resize_to_max_width
 
 logger = logging.getLogger(__name__)
 
-IMAGE_EXTENSIONS = {".jpg", ".jpeg"}
 
 MAX_WIDTH = 2500
 MAX_FILE_SIZE_BYTES = 500 * 1024
@@ -34,10 +34,7 @@ def optimise(
     failed_count = 0
 
     for file_path in input_path.iterdir():
-        if not file_path.is_file():
-            continue
-
-        if file_path.suffix.lower() not in IMAGE_EXTENSIONS:
+        if not is_jpg(file_path):
             logger.debug("Skipping (not a supported image): %s", file_path.name)
             continue
 

src/photo_tools/commands/organise_by_date.py

@@ -4,16 +4,11 @@
 from pathlib import Path
 
 from photo_tools.core.validation import validate_input_dir
+from photo_tools.image.file_types import is_jpg, is_raw
 from photo_tools.image.metadata import get_image_date
 
 logger = logging.getLogger(__name__)
 
-IMAGE_EXTENSIONS = {
-    ".jpg",
-    ".jpeg",
-    ".raf",
-}
-
 Reporter = Callable[[str, str], None]
 
 
@@ -37,10 +32,7 @@ def organise_by_date(
     cleaned_suffix = suffix.strip() if suffix and suffix.strip() else None
 
     for file_path in input_path.iterdir():
-        if not file_path.is_file():
-            continue
-
-        if file_path.suffix.lower() not in IMAGE_EXTENSIONS:
+        if not (is_jpg(file_path) or is_raw(file_path)):
             logger.debug("Skipping unsupported file: %s", file_path.name)
             continue
 
@@ -85,7 +77,6 @@ def organise_by_date(
         report("info", f"Moved {file_path.name} -> {target_dir}")
 
     # Summary
-
     if dry_run:
         report("summary", f"Dry run complete: would move {dry_run_count} file(s)")
     else:

src/photo_tools/commands/separate_raws.py

@@ -4,10 +4,10 @@
 from pathlib import Path
 
 from photo_tools.core.validation import validate_input_dir
+from photo_tools.image.file_types import is_raw
 
 logger = logging.getLogger(__name__)
 
-RAW_EXTENSIONS = {".raf"}
 OUTPUT_DIR = "raws"
 
 Reporter = Callable[[str, str], None]
@@ -29,10 +29,7 @@ def separate_raws(
     skipped_existing_count = 0
 
     for file_path in input_path.iterdir():
-        if not file_path.is_file():
-            continue
-
-        if file_path.suffix.lower() not in RAW_EXTENSIONS:
+        if not is_raw(file_path):
             logger.debug("Skipping (not RAW): %s", file_path.name)
             continue
 

src/photo_tools/image/file_types.py

@@ -0,0 +1,25 @@
+from pathlib import Path
+
+RAW_EXTENSIONS = {
+    ".cr2",
+    ".cr3",
+    ".nef",
+    ".arw",
+    ".raf",
+    ".orf",
+    ".rw2",
+    ".dng",
+    ".pef",
+    ".srw",
+    ".x3f",
+}
+
+JPG_EXTENSIONS = {".jpg", ".jpeg"}
+
+
+def is_raw(file_path: Path) -> bool:
+    return file_path.is_file() and file_path.suffix.lower() in RAW_EXTENSIONS
+
+
+def is_jpg(file_path: Path) -> bool:
+    return file_path.is_file() and file_path.suffix.lower() in JPG_EXTENSIONS

src/photo_tools/image/raw_utils.py

@@ -4,11 +4,10 @@
 from pathlib import Path
 
 from photo_tools.core.validation import validate_input_dir
+from photo_tools.image.file_types import is_jpg, is_raw
 
 logger = logging.getLogger(__name__)
 
-RAW_EXTENSIONS = {".raf"}
-JPG_EXTENSIONS = {".jpg", ".jpeg"}
 
 Reporter = Callable[[str, str], None]
 RawMatcher = Callable[[Path, list[Path]], bool]
@@ -34,17 +33,10 @@ def move_raws_by_rule(
     dry_run_count = 0
     skipped_existing_count = 0
 
-    jpg_files = [
-        f
-        for f in jpg_path.iterdir()
-        if f.is_file() and f.suffix.lower() in JPG_EXTENSIONS
-    ]
+    jpg_files = [f for f in jpg_path.iterdir() if is_jpg(f)]
 
     for raw_file in raw_path.iterdir():
-        if not raw_file.is_file():
-            continue
-
-        if raw_file.suffix.lower() not in RAW_EXTENSIONS:
+        if not is_raw(raw_file):
             continue
 
         if not should_move(raw_file, jpg_files):

tests/image/test_file_types.py

@@ -0,0 +1,118 @@
+from pathlib import Path
+
+import pytest
+
+from photo_tools.image.raw_utils import is_jpg, is_raw
+
+
+@pytest.mark.parametrize(
+    ("filename"),
+    [
+        "image.cr2",
+        "image.CR2",
+        "image.cr3",
+        "image.nef",
+        "image.NeF",
+        "image.arw",
+        "image.raf",
+        "image.orf",
+        "image.rw2",
+        "image.dng",
+        "image.pef",
+        "image.srw",
+        "image.x3f",
+    ],
+)
+def test_is_raw_returns_true_for_supported_raw_files(
+    tmp_path: Path,
+    filename: str,
+) -> None:
+    file_path = tmp_path / filename
+    file_path.touch()
+
+    assert is_raw(file_path) is True
+
+
+@pytest.mark.parametrize(
+    ("filename"),
+    [
+        "image.jpg",
+        "image.jpeg",
+        "image.png",
+        "image.txt",
+        "image",
+    ],
+)
+def test_is_raw_returns_false_for_non_raw_files(
+    tmp_path: Path,
+    filename: str,
+) -> None:
+    file_path = tmp_path / filename
+    file_path.touch()
+
+    assert is_raw(file_path) is False
+
+
+def test_is_raw_returns_false_for_directory(tmp_path: Path) -> None:
+    dir_path = tmp_path / "image.cr2"
+    dir_path.mkdir()
+
+    assert is_raw(dir_path) is False
+
+
+def test_is_raw_returns_false_for_non_existing_path(tmp_path: Path) -> None:
+    file_path = tmp_path / "missing.cr2"
+
+    assert is_raw(file_path) is False
+
+
+@pytest.mark.parametrize(
+    ("filename"),
+    [
+        "image.jpg",
+        "image.jpeg",
+        "image.JPG",
+        "image.JPEG",
+    ],
+)
+def test_is_jpg_returns_true_for_supported_jpg_files(
+    tmp_path: Path,
+    filename: str,
+) -> None:
+    file_path = tmp_path / filename
+    file_path.touch()
+
+    assert is_jpg(file_path) is True
+
+
+@pytest.mark.parametrize(
+    ("filename"),
+    [
+        "image.cr2",
+        "image.nef",
+        "image.png",
+        "image.txt",
+        "image",
+    ],
+)
+def test_is_jpg_returns_false_for_non_jpg_files(
+    tmp_path: Path,
+    filename: str,
+) -> None:
+    file_path = tmp_path / filename
+    file_path.touch()
+
+    assert is_jpg(file_path) is False
+
+
+def test_is_jpg_returns_false_for_directory(tmp_path: Path) -> None:
+    dir_path = tmp_path / "image.jpg"
+    dir_path.mkdir()
+
+    assert is_jpg(dir_path) is False
+
+
+def test_is_jpg_returns_false_for_non_existing_path(tmp_path: Path) -> None:
+    file_path = tmp_path / "missing.jpg"
+
+    assert is_jpg(file_path) is False