Merge pull request #671 from aarhusstadsarkiv/dev-matca

Version 1.0.4

digiarch/__version__.py

@@ -1 +1 @@
-__version__ = "1.0.3"
+__version__ = "1.0.4"

digiarch/cli.py

@@ -37,12 +37,16 @@
 from click import pass_context
 from click import Path as ClickPath
 from click import version_option
+from PIL import Image
 from PIL import UnidentifiedImageError
+from PIL.Image import DecompressionBombError
 from pydantic import TypeAdapter
 
 from .__version__ import __version__
 from .database import FileDB
 
+Image.MAX_IMAGE_PIXELS = int(50e3**2)
+
 
 def handle_rename(file: File, action: RenameAction) -> Union[tuple[Path, Path], tuple[None, None]]:
     old_path: Path = file.get_absolute_path()
@@ -71,7 +75,7 @@ def handle_end(ctx: Context, database: FileDB, exception: ExceptionManager, *log
     program_end: HistoryEntry = HistoryEntry.command_history(
         ctx,
         "end",
-        data=1 if exception.exception else 0,
+        data=repr(exception.exception) if exception.exception else None,
         reason="".join(format_tb(exception.traceback)) if exception.traceback else None,
     )
 
@@ -196,15 +200,20 @@ def app_identify(
 
                 file_history: list[HistoryEntry] = []
 
-                with ExceptionManager(UnidentifiedImageError) as image_exception:
+                with ExceptionManager(
+                    Exception,
+                    UnidentifiedImageError,
+                    DecompressionBombError,
+                    allow=[OSError, IOError],
+                ) as identify_error:
                     file = File.from_file(path, root, siegfried, actions, custom_signatures)
 
-                if image_exception.exception:
+                if identify_error.exception:
                     file = File.from_file(path, root, siegfried)
                     file.action = "manual"
                     file.action_data = ActionData(
                         manual=ManualAction(
-                            reason=image_exception.exception.__class__.__name__,
+                            reason=identify_error.exception.__class__.__name__,
                             process="Identify and fix error.",
                         ),
                     )
@@ -213,7 +222,7 @@ def app_identify(
                             ctx,
                             "file:identify:error",
                             file.uuid,
-                            repr(image_exception.exception),
+                            repr(identify_error.exception),
                             "".join(format_tb(exception.traceback)) if exception.traceback else None,
                         ),
                     )

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "digiarch"
-version = "1.0.3"
+version = "1.0.4"
 description = "Tools for the Digital Archive Project at Aarhus Stadsarkiv"
 authors = ["Aryan Muhammadi Landi <[email protected]>", "Nina Jensen <[email protected]>", "Aarhus Stadsarkiv <[email protected]>"]
 license = "GPL-3.0"
@@ -12,7 +12,7 @@ include = ["pyproject.toml"]
 
 [tool.poetry.dependencies]
 python = "^3.9"
-acacore = {git = "https://github.com/aarhusstadsarkiv/acacore.git", tag = "v1.1.1"}
+acacore = {git = "https://github.com/aarhusstadsarkiv/acacore.git", tag = "v1.1.2"}
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.0"

tests/test_cli.py

@@ -98,7 +98,8 @@ def test_identify(tests_folder: Path, files_folder: Path, files_folder_copy: Pat
 
     with FileDB(files_folder_copy / "_metadata" / "files.db") as database:
         last_history: HistoryEntry = sorted(database.history, key=lambda h: h.time).pop()
-        assert last_history.data == 1
+        assert isinstance(last_history.data, str)
+        assert last_history.data.startswith("FileNotFoundError")
         assert last_history.reason is not None