ci: Add test for OCR languages

Test that the languages that we provide to users for OCR match the
languages that are installed in the container image

Fixes #417

tests/test_ocr.py

@@ -0,0 +1,41 @@
+import platform
+import subprocess
+
+import pytest
+
+from dangerzone.isolation_provider.container import Container
+from dangerzone.logic import DangerzoneCore
+
+
+# TODO: Perform an equivalent test on Qubes.
+# NOTE: We skip running this test on Windows/MacOS, because our current CI cannot run
+# Docker in these platforms. It's not a problem anyways, because the result should be
+# the same in all container-based platforms.
+@pytest.mark.skipif(platform.system() != "Linux", reason="Container-specific")
+def test_ocr_ommisions() -> None:
+    # Create the command that will list all the installed languages in the container
+    # image.
+    runtime = Container.get_runtime()
+    command = [runtime, "run", Container.CONTAINER_NAME, "tesseract", "--list-langs"]
+
+    # Run the command, strip any extra whitespace, and remove the following first line
+    # from the result:
+    #
+    #     List of available languages in "/usr/share/tessdata/" ...
+    installed_langs = set(
+        subprocess.run(command, text=True, check=True, stdout=subprocess.PIPE)
+        .stdout.strip()
+        .split("\n")[1:]
+    )
+
+    # Remove the "osd" and "equ" languages from the list of installed languages, since
+    # they are not an actual language. Read more in:
+    # https://pyimagesearch.com/2021/11/15/tesseract-page-segmentation-modes-psms-explained-how-to-improve-your-ocr-accuracy/
+    installed_langs -= {"osd", "equ"}
+
+    # Grab the languages that Dangerzone offers to the user through the GUI/CLI.
+    offered_langs = set(DangerzoneCore(Container(True)).ocr_languages.values())
+
+    # Ensure that both the installed languages and the ones we offer to the user are the
+    # same.
+    assert installed_langs == offered_langs