Handle unsupported SVG files gracefully

CHANGELOG.md

@@ -4,6 +4,7 @@
 - Added support for scene based granular video code processing
 - Added ffprobe video metadata extraction (duration, fps, width, height, language)
 - Added support str and Path objects for all file inpunts
+- Handle unsupported SVG files gracefully
 
 ## 0.6.2 - 2024-06-13
 - Update and relax dependencies

docs/changelog.md

@@ -4,6 +4,7 @@
 - Added support for scene based granular video code processing
 - Added ffprobe video metadata extraction (duration, fps, width, height, language)
 - Added support str and Path objects for all file inpunts
+- Handle unsupported SVG files gracefully
 
 ## 0.6.2 - 2024-06-13
 - Update and relax dependencies

iscc_sdk/cli.py

@@ -95,8 +95,12 @@ def create(file: Path):
     """Create ISCC-CODE for single FILE."""
     log.remove()
     if file.is_file() and file.exists():
-        result = idk.code_iscc(file.as_posix())
-        typer.echo(result.json(indent=2))
+        try:
+            result = idk.code_iscc(file.as_posix())
+            typer.echo(result.json(indent=2))
+        except idk.IsccUnsupportedMediatype as e:
+            typer.echo(e)
+            raise typer.Exit(code=1)
     else:
         typer.echo(f"Invalid file path {file}")
         raise typer.Exit(code=1)

iscc_sdk/image.py

@@ -189,6 +189,8 @@ def image_meta_embed(fp, meta):
         cmdf += f"set Xmp.dc.creator {meta.creator}\n"
     if meta.rights:
         cmdf += f"set Xmp.dc.rights {meta.rights}\n"
+    if meta.identifier:
+        cmdf += f"set Xmp.dc.identifier {meta.identifier}\n"
 
     # Create temp filepaths
     tempdir = Path(tempfile.mkdtemp())

iscc_sdk/mediatype.py

@@ -169,9 +169,9 @@ def mediatype_to_mode(mime_type):
     # type: (str) -> str
     """Get perceptual processing mode from mimetype.
 
-    :param str mime_type: RFC-6838 mediatype string
-    :return str: Processing mode ("text", "image", "audio", "video")
-    :raise ValueError: if no matching processing mode was found.
+    :param mime_type: RFC-6838 mediatype string
+    :return Processing mode ("text", "image", "audio", "video")
+    :raise IsccUnsupportedMediatype: if no matching processing mode was found.
     """
 
     mime_type = mediatype_clean(mime_type)
@@ -181,7 +181,7 @@ def mediatype_to_mode(mime_type):
 
     # Fallback to guess mode by top-level type
     mode = mime_type.split("/")[0]
-    if mode in ["text", "image", "audio", "video"]:
+    if mode in ["text", "image", "audio", "video"] and mime_type not in UNSUPPORTED_MEDIATYPES:
         log.warning(f"Guessing perceptual mode from {mime_type}")
         return mode
 
@@ -202,6 +202,7 @@ def mediatype_to_mode(mime_type):
 mimetypes.add_type("image/heic", ".heic")
 mimetypes.add_type("image/avif", ".avif")
 
+
 SUPPORTED_MEDIATYPES = {
     # Text Formats
     "application/rtf": {"mode": "text", "ext": "rtf"},
@@ -278,6 +279,12 @@ def mediatype_to_mode(mime_type):
     "video/x-ms-wmv": {"mode": "video", "ext": "wmv"},
 }
 
+# Signals eplixitly unsupported mediatypes that fail processing
+UNSUPPORTED_MEDIATYPES = {
+    "image/svg+xml": {"mode": "image", "ext": "svg"},
+}
+
+
 MEDIATYPE_NORM = {
     "audio/x-aiff": "audio/aiff",
     "audio/x-wav": "audio/wav",

iscc_sdk/metadata.py

@@ -1,5 +1,6 @@
 """*Metadata handling functions*"""
 
+from loguru import logger as log
 from pathlib import Path
 from typing import Optional
 

tests/conftest.py

@@ -115,6 +115,18 @@ def epub_file(tmp_path_factory):
     return dst.as_posix()
 
 
+@pytest.fixture(scope="module")
+def svg_file(tmp_path_factory) -> str:
+    dst = tmp_path_factory.mktemp("data") / "image.svg"
+    svg_content = """
+    <svg xmlns="http://www.w3.org/2000/svg" width="100" height="100">
+        <circle cx="50" cy="50" r="40" stroke="black" stroke-width="3" fill="red" />
+    </svg>
+    """
+    dst.write_text(svg_content)
+    return dst.as_posix()
+
+
 @pytest.fixture(scope="module")
 def asset_tree(tmp_path_factory):
     dst = Path(tmp_path_factory.mktemp("data"))

tests/test_cli.py

@@ -68,6 +68,12 @@ def test_cli_create():
     }
 
 
+def test_cli_create_unsupported(svg_file):
+    result = runner.invoke(app, ["create", svg_file])
+    assert result.exit_code == 1
+    assert "Unsupported mediatype image/svg+xml" in result.stdout
+
+
 def test_cli_batch_no_arg():
     result = runner.invoke(app, ["batch"])
     assert result.exit_code == 2

tests/test_metadata.py

@@ -1,3 +1,5 @@
+import pytest
+
 import iscc_sdk as idk
 
 
@@ -10,6 +12,11 @@ def test_extract_metadata(jpg_file):
     }
 
 
+def test_extract_metadata_unsupported(svg_file):
+    with pytest.raises(idk.IsccUnsupportedMediatype):
+        idk.extract_metadata(svg_file)
+
+
 def test_embed_metadata(jpg_file):
     meta = idk.IsccMeta(name="Some Title", description="Some Description")
     new_file = idk.embed_metadata(jpg_file, meta)
@@ -22,6 +29,19 @@ def test_embed_metadata(jpg_file):
     }
 
 
+def test_metadata_identifier_field_image(jpg_file):
+    meta = idk.IsccMeta(name="Some Title", description="Some Description", identifier="abcdefghijk")
+    new_file = idk.embed_metadata(jpg_file, meta)
+    assert idk.extract_metadata(new_file).dict() == {
+        "name": "Some Title",
+        "description": "Some Description",
+        "creator": "Some Cat Lover",
+        "height": 133,
+        "width": 200,
+        "identifier": "abcdefghijk",
+    }
+
+
 def test_embed_metadata_unsupported(doc_file):
     meta = idk.IsccMeta(name="Some Title", description="Some Description")
     new_file = idk.embed_metadata(doc_file, meta)