refactor: thumbnail renderers

pyproject.toml

@@ -14,8 +14,10 @@ dependencies = [
     "ffmpeg-python~=0.2",
     "humanfriendly==10.*",
     "mutagen~=1.47",
+    "numexpr~=2.14.1",
     "numpy~=2.2",
     "opencv_python~=4.11",
+    "openexr~=3.4.3",
     "Pillow>=10.2,<=11",
     "pillow-avif-plugin~=1.5",
     "pillow-heif~=0.22",

src/tagstudio/core/media_types.py

@@ -41,6 +41,7 @@ class MediaType(str, Enum):
     FONT = "font"
     IMAGE_ANIMATED = "image_animated"
     IMAGE_RAW = "image_raw"
+    IMAGE_EXR = "image_exr"
     IMAGE_VECTOR = "image_vector"
     IMAGE = "image"
     INSTALLER = "installer"
@@ -51,6 +52,7 @@ class MediaType(str, Enum):
     PACKAGE = "package"
     PDF = "pdf"
     PLAINTEXT = "plaintext"
+    POWERPOINT = "powerpoint"
     PRESENTATION = "presentation"
     PROGRAM = "program"
     SHADER = "shader"
@@ -109,7 +111,6 @@ class MediaCategories:
         ".psd",
     }
     _AFFINITY_PHOTO_SET: set[str] = {".afphoto"}
-    _KRITA_SET: set[str] = {".kra", ".krz"}
     _ARCHIVE_SET: set[str] = {
         ".7z",
         ".gz",
@@ -297,15 +298,25 @@ class MediaCategories:
         ".cr3",
         ".crw",
         ".dng",
+        ".erf",
+        ".mef",
+        ".mos",
+        ".mrw",
         ".nef",
         ".nrw",
         ".orf",
+        ".pef",
         ".raf",
         ".raw",
         ".rw2",
         ".srf",
         ".srf2",
+        ".sr2",
+        ".srw",
+        ".x3f",
+        ".3fr",
     }
+    _IMAGE_EXR_SET: set[str] = {".exr"}
     _IMAGE_VECTOR_SET: set[str] = {".eps", ".epsf", ".epsi", ".svg", ".svgz"}
     _IMAGE_RASTER_SET: set[str] = {
         ".apng",
@@ -334,6 +345,7 @@ class MediaCategories:
     }
     _INSTALLER_SET: set[str] = {".appx", ".msi", ".msix"}
     _IWORK_SET: set[str] = {".key", ".pages", ".numbers"}
+    _KRITA_SET: set[str] = {".kra", ".krz"}
     _MATERIAL_SET: set[str] = {".mtl"}
     _MODEL_SET: set[str] = {".3ds", ".fbx", ".obj", ".stl"}
     _OPEN_DOCUMENT_SET: set[str] = {
@@ -375,11 +387,11 @@ class MediaCategories:
         "license",
         "readme",
     }
+    _POWERPOINT_SET: set[str] = {".pptx"}
     _PRESENTATION_SET: set[str] = {
         ".key",
         ".odp",
         ".ppt",
-        ".pptx",
     }
     _PROGRAM_SET: set[str] = {".app", ".bin", ".exe"}
     _SOURCE_ENGINE_SET: set[str] = {".vtf"}
@@ -500,6 +512,9 @@ class MediaCategories:
         is_iana=False,
         name="raw image",
     )
+    IMAGE_EXR_TYPES = MediaCategory(
+        media_type=MediaType.IMAGE_EXR, extensions=_IMAGE_EXR_SET, is_iana=False, name="exr image"
+    )
     IMAGE_VECTOR_TYPES = MediaCategory(
         media_type=MediaType.IMAGE_VECTOR,
         extensions=_IMAGE_VECTOR_SET,
@@ -566,9 +581,15 @@ class MediaCategories:
         is_iana=False,
         name="plaintext",
     )
+    POWERPOINT_TYPES = MediaCategory(
+        media_type=MediaType.POWERPOINT,
+        extensions=_POWERPOINT_SET,
+        is_iana=False,
+        name="powerpoint",
+    )
     PRESENTATION_TYPES = MediaCategory(
         media_type=MediaType.PRESENTATION,
-        extensions=_PRESENTATION_SET,
+        extensions=_PRESENTATION_SET | _POWERPOINT_SET,
         is_iana=False,
         name="presentation",
     )

src/tagstudio/qt/controllers/preview_thumb_controller.py

@@ -6,6 +6,7 @@
 from typing import TYPE_CHECKING
 
 import cv2
+import OpenEXR
 import rawpy
 import structlog
 from PIL import Image, UnidentifiedImageError
@@ -54,6 +55,15 @@ def __get_image_stats(self, filepath: Path) -> FileAttributeData:
                 FileNotFoundError,
             ):
                 pass
+        elif MediaCategories.IMAGE_EXR_TYPES.contains(ext, mime_fallback=True):
+            try:
+                exr_file = OpenEXR.File(str(filepath))
+                part = exr_file.parts[0]
+                logger.debug("[PreviewThumb]", part=part)
+                stats.width = part.width()
+                stats.height = part.height()
+            except Exception:
+                pass
         elif MediaCategories.IMAGE_RASTER_TYPES.contains(ext, mime_fallback=True):
             try:
                 image = Image.open(str(filepath))
@@ -69,6 +79,7 @@ def __get_image_stats(self, filepath: Path) -> FileAttributeData:
         elif MediaCategories.IMAGE_VECTOR_TYPES.contains(ext, mime_fallback=True):
             pass  # TODO
 
+        logger.debug("[PreviewThumb]", stats=stats)
         return stats
 
     def __get_gif_data(self, filepath: Path) -> tuple[bytes, tuple[int, int]] | None:

src/tagstudio/qt/helpers/file_wrappers/archive/archive_file.py

@@ -0,0 +1,21 @@
+from abc import ABC, abstractmethod
+from pathlib import Path
+from typing import Literal
+
+
+class ArchiveFile(ABC):
+    @abstractmethod
+    def __init__(self, path: Path, mode: Literal["r"]) -> None:
+        pass
+
+    @abstractmethod
+    def get_name_list(self) -> list[str]:
+        raise NotImplementedError
+
+    @abstractmethod
+    def has_file_name(self, file_name: str) -> bool:
+        raise NotImplementedError
+
+    @abstractmethod
+    def read(self, file_name: str) -> bytes:
+        raise NotImplementedError

src/tagstudio/qt/helpers/file_wrappers/archive/rar_file.py

@@ -0,0 +1,52 @@
+from pathlib import Path
+from types import TracebackType
+from typing import Literal, Self
+
+import rarfile
+
+from tagstudio.qt.helpers.file_wrappers.archive.archive_file import ArchiveFile
+
+
+class RarFile(ArchiveFile):
+    """Wrapper around rarfile.RarFile."""
+
+    def __init__(self, path: Path, mode: Literal["r"]) -> None:
+        super().__init__(path, mode)
+        self.path = path
+        self.__rar_file: rarfile.RarFile = rarfile.RarFile(path, mode)
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exception_type: type[BaseException] | None,
+        exception_value: BaseException | None,
+        exception_traceback: TracebackType | None,
+    ) -> None:
+        self.__rar_file.close()
+
+    def get_name_list(self) -> list[str]:
+        without_own_file_name: map = map(
+            lambda file_name: file_name.replace(f"{self.path.name}/", ""),
+            self.__rar_file.namelist(),
+        )
+        without_empty_items: filter = filter(None, without_own_file_name)
+
+        return list(without_empty_items)
+
+    def has_file_name(self, file_name: str) -> bool:
+        return file_name in self.get_name_list()
+
+    def read(self, file_name: str) -> bytes | None:
+        search_paths: list[Path] = [Path(file_name), Path(self.path.name, file_name)]
+        try:
+            for file_path in search_paths:
+                try:
+                    return self.__rar_file.read(file_path)
+                except KeyError:
+                    continue
+
+            return None
+        except KeyError as e:
+            raise e

src/tagstudio/qt/helpers/file_wrappers/archive/seven_zip_file.py

@@ -0,0 +1,59 @@
+from pathlib import Path
+from types import TracebackType
+from typing import Literal, Self
+
+import py7zr
+
+from tagstudio.qt.helpers.file_wrappers.archive.archive_file import ArchiveFile
+
+
+class SevenZipFile(ArchiveFile):
+    """Wrapper around py7zr.SevenZipFile."""
+
+    def __init__(self, path: Path, mode: Literal["r"]) -> None:
+        super().__init__(path, mode)
+        self.path = path
+        self.__seven_zip_file: py7zr.SevenZipFile = py7zr.SevenZipFile(path, mode)
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exception_type: type[BaseException] | None,
+        exception_value: BaseException | None,
+        exception_traceback: TracebackType | None,
+    ) -> None:
+        self.__seven_zip_file.close()
+
+    def get_name_list(self) -> list[str]:
+        without_own_file_name: map = map(
+            lambda file_name: file_name.replace(f"{self.path.name}/", ""),
+            self.__seven_zip_file.namelist(),
+        )
+        without_empty_items: filter = filter(None, without_own_file_name)
+
+        return list(without_empty_items)
+
+    def has_file_name(self, file_name: str) -> bool:
+        return file_name in self.get_name_list()
+
+    def read(self, file_name: str) -> bytes | None:
+        # py7zr.SevenZipFile must be reset after every extraction
+        # See https://py7zr.readthedocs.io/en/stable/api.html#py7zr.SevenZipFile.extract
+        self.__seven_zip_file.reset()
+
+        factory = py7zr.io.BytesIOFactory(limit=10485760)  # 10 MiB
+
+        search_paths: list[Path] = [Path(file_name), Path(self.path.name, file_name)]
+        try:
+            for file_path in search_paths:
+                try:
+                    self.__seven_zip_file.extract(targets=[str(file_path)], factory=factory)
+                    return factory.get(file_path).read()
+                except KeyError:
+                    continue
+
+            return None
+        except KeyError as e:
+            raise e

src/tagstudio/qt/helpers/file_wrappers/archive/tar_file.py

@@ -0,0 +1,52 @@
+import tarfile
+from pathlib import Path
+from types import TracebackType
+from typing import Literal, Self
+
+from tagstudio.core.utils.types import unwrap
+from tagstudio.qt.helpers.file_wrappers.archive.archive_file import ArchiveFile
+
+
+class TarFile(ArchiveFile):
+    """Wrapper around tarfile.TarFile."""
+
+    def __init__(self, path: Path, mode: Literal["r"]) -> None:
+        super().__init__(path, mode)
+        self.path = path
+        self.__tar_file: tarfile.TarFile = tarfile.TarFile(path, mode)
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exception_type: type[BaseException] | None,
+        exception_value: BaseException | None,
+        exception_traceback: TracebackType | None,
+    ) -> None:
+        self.__tar_file.close()
+
+    def get_name_list(self) -> list[str]:
+        without_own_file_name: map = map(
+            lambda file_name: file_name.replace(f"{self.path.name}/", ""),
+            self.__tar_file.getnames(),
+        )
+        without_empty_items: filter = filter(None, without_own_file_name)
+
+        return list(without_empty_items)
+
+    def has_file_name(self, file_name: str) -> bool:
+        return file_name in self.get_name_list()
+
+    def read(self, file_name: str) -> bytes | None:
+        search_paths: list[Path] = [Path(file_name), Path(self.path.name, file_name)]
+        try:
+            for file_path in search_paths:
+                try:
+                    return unwrap(self.__tar_file.extractfile(str(file_path))).read()
+                except KeyError:
+                    continue
+
+            return None
+        except KeyError as e:
+            raise e

src/tagstudio/qt/helpers/file_wrappers/archive/zip_file.py

@@ -0,0 +1,51 @@
+import zipfile
+from pathlib import Path
+from types import TracebackType
+from typing import Literal, Self
+
+from tagstudio.qt.helpers.file_wrappers.archive.archive_file import ArchiveFile
+
+
+class ZipFile(ArchiveFile):
+    """Wrapper around zipfile.ZipFile."""
+
+    def __init__(self, path: Path, mode: Literal["r"]) -> None:
+        super().__init__(path, mode)
+        self.path = path
+        self.__zip_file: zipfile.ZipFile = zipfile.ZipFile(path, mode)
+
+    def __enter__(self) -> Self:
+        return self
+
+    def __exit__(
+        self,
+        exception_type: type[BaseException] | None,
+        exception_value: BaseException | None,
+        exception_traceback: TracebackType | None,
+    ) -> None:
+        self.__zip_file.close()
+
+    def get_name_list(self) -> list[str]:
+        without_own_file_name: map = map(
+            lambda file_name: file_name.replace(f"{self.path.name}/", ""),
+            self.__zip_file.namelist(),
+        )
+        without_empty_items: filter = filter(None, without_own_file_name)
+
+        return list(without_empty_items)
+
+    def has_file_name(self, file_name: str) -> bool:
+        return file_name in self.get_name_list()
+
+    def read(self, file_name: str) -> bytes | None:
+        search_paths: list[Path] = [Path(file_name), Path(self.path.name, file_name)]
+        try:
+            for file_path in search_paths:
+                try:
+                    return self.__zip_file.read(str(file_path))
+                except KeyError:
+                    continue
+
+            return None
+        except KeyError as e:
+            raise e