feat: code syntax highlighting

pyproject.toml

@@ -14,9 +14,12 @@ 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,<12",
+    "pillow-avif-plugin~=1.5",
     "pillow-heif~=0.22",
     "pillow-jxl-plugin~=1.3",
     "py7zr==1.0.0",
@@ -29,6 +32,7 @@ dependencies = [
     "SQLAlchemy~=2.0",
     "srctools~=2.6",
     "structlog~=25.3",
+    "superqt>=0.7.6",
     "toml~=0.10",
     "typing_extensions~=4.13",
     "ujson~=5.10",

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",
@@ -187,13 +188,15 @@ class MediaCategories:
         ".dhtml",
         ".fgd",
         ".fish",
+        ".gd",
         ".gitignore",
         ".h",
         ".hpp",
         ".htm",
         ".html",
         ".inf",
         ".ini",
+        ".java",
         ".js",
         ".json",
         ".json5",
@@ -296,16 +299,32 @@ class MediaCategories:
         ".cr2",
         ".cr3",
         ".crw",
+        ".dcs",
+        ".dcr",
         ".dng",
+        ".drf",
+        ".erf",
+        ".k25",
+        ".kdc",
+        ".mdc",
+        ".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",
@@ -316,24 +335,34 @@ class MediaCategories:
         ".heic",
         ".heif",
         ".icns",
-        ".j2k",
+        ".jpeg",
+        ".jpg",
+        ".jpe",
+        ".jif",
         ".jfif",
-        ".jp2",
+        ".jfi",
         ".jpeg_large",
-        ".jpeg",
         ".jpg_large",
-        ".jpg",
+        ".jp2",
+        ".j2k",
+        ".jpf",
+        ".jpm",
         ".jpg2",
+        ".j2c",
+        ".jpc",
+        ".jpx",
+        ".mj2",
         ".jxl",
         ".png",
         ".psb",
         ".psd",
-        ".tif",
         ".tiff",
+        ".tif",
         ".webp",
     }
     _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 +404,11 @@ 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"}
@@ -410,6 +439,7 @@ class MediaCategories:
         ".m4v",
         ".mkv",
         ".mov",
+        ".movie",
         ".mp4",
         ".webm",
         ".wmv",
@@ -500,6 +530,9 @@ 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 +599,15 @@ 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_panel/thumbnail/text_display_controller.py

@@ -0,0 +1,42 @@
+from pathlib import Path
+from typing import TYPE_CHECKING
+
+import structlog
+from superqt.utils import CodeSyntaxHighlight
+
+from tagstudio.qt.views.preview_panel.thumbnail.text_display_view import TextDisplayView
+
+if TYPE_CHECKING:
+    from tagstudio.qt.ts_qt import QtDriver
+
+logger = structlog.get_logger(__name__)
+
+
+class TextDisplayController(TextDisplayView):
+    """A widget for displaying a plaintext file."""
+
+    def __init__(self, driver: "QtDriver") -> None:
+        super().__init__()
+        self.driver = driver
+
+    def set_file(self, path: Path) -> None:
+        """Sets the text file that the text display displays.
+
+        Args:
+            path (Path): The path of the text file.
+
+        """
+        language: str = path.suffix[1:]
+
+        try:
+            with open(path) as text_file:
+                content: str = text_file.read()
+
+            CodeSyntaxHighlight(
+                self.document(), language, self.driver.settings.syntax_highlighting_style
+            )
+            self.setText(content)
+        except ValueError:
+            logger.warn(f"[TextDisplayController] Couldn't find lexer for `{language}`")
+        except FileNotFoundError:
+            logger.error(f"[TextDisplayController] Couldn't find file {path}")

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
@@ -49,11 +50,19 @@ def __get_image_stats(self, filepath: Path) -> FileAttributeData:
                     stats.width = image.width
                     stats.height = image.height
             except (
-                rawpy._rawpy._rawpy.LibRawIOError,  # pyright: ignore[reportAttributeAccessIssue]
+                rawpy._rawpy.LibRawIOError,  # pyright: ignore[reportAttributeAccessIssue]
                 rawpy._rawpy.LibRawFileUnsupportedError,  # pyright: ignore[reportAttributeAccessIssue]
                 FileNotFoundError,
             ):
                 pass
+        elif MediaCategories.IMAGE_EXR_TYPES.contains(ext, mime_fallback=True):
+            try:
+                exr_file = OpenEXR.File(str(filepath))
+                part = exr_file.parts[0]
+                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))
@@ -138,6 +147,10 @@ def display_file(self, filepath: Path) -> FileAttributeData:
             else:
                 self._display_image(filepath)
                 return self.__get_image_stats(filepath)
+        # Text/Code
+        elif MediaCategories.PLAINTEXT_TYPES.contains(ext, mime_fallback=True):
+            self._display_text(filepath)
+            return self.__get_image_stats(filepath)
         # Other Types (Including Images)
         else:
             self._display_image(filepath)

src/tagstudio/qt/global_settings.py

@@ -70,6 +70,7 @@ class GlobalSettings(BaseModel):
     show_filepath: ShowFilepathOption = Field(default=ShowFilepathOption.DEFAULT)
     tag_click_action: TagClickActionOption = Field(default=TagClickActionOption.DEFAULT)
     theme: Theme = Field(default=Theme.SYSTEM)
+    syntax_highlighting_style: str = Field(default="github-dark")
     splash: Splash = Field(default=Splash.DEFAULT)
     windows_start_command: bool = Field(default=False)
 

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