Add support for HEIC files

Thanks to Maxime Morin ( https://www.maijin.fr/ )
for the patch.

4 files changed, 35 insertions, 0 deletions

diff --git a/libmat2/images.py b/libmat2/images.py
index 3da2673..9c24998 100644
--- a/libmat2/images.py
+++ b/libmat2/images.py
@@ -181,3 +181,23 @@ class PPMParser(abstract.AbstractParser):
                         line = re.sub(r"\s+", "", line, flags=re.UNICODE)
                         fout.write(line)
         return True
+
+class HEICParser(exiftool.ExiftoolParser):
+    mimetypes = {'image/heic'}
+    meta_allowlist = {'SourceFile', 'ExifToolVersion', 'FileName','Directory',
+            'FileSize', 'FileModifyDate', 'FileAccessDate',
+            'FileInodeChangeDate', 'FilePermissions', 'FileType',
+            'FileTypeExtension', 'MIMEType', 'MajorBrand', 'MinorVersion',
+            'CompatibleBrands','HandlerType', 'PrimaryItemReference',
+            'HEVCConfigurationVersion', 'GeneralProfileSpace',
+            'GeneralTierFlag', 'GeneralProfileIDC',
+            'GenProfileCompatibilityFlags', 'ConstraintIndicatorFlags',
+            'GeneralLevelIDC', 'MinSpatialSegmentationIDC',
+            'ParallelismType','ChromaFormat', 'BitDepthLuma', 'BitDepthChroma',
+            'NumTemporalLayers', 'TemporalIDNested', 'ImageWidth',
+            'ImageHeight', 'ImageSpatialExtent', 'ImagePixelDepth',
+            'AverageFrameRate', 'ConstantFrameRate', 'MediaDataSize',
+            'MediaDataOffset','ImageSize', 'Megapixels'}
+
+    def remove_all(self) -> bool:
+        return self._lightweight_cleanup()
diff --git a/libmat2/parser_factory.py b/libmat2/parser_factory.py
index 9965432..a539d12 100644
--- a/libmat2/parser_factory.py
+++ b/libmat2/parser_factory.py
@@ -11,6 +11,10 @@ T = TypeVar('T', bound='abstract.AbstractParser')
 mimetypes.add_type('application/epub+zip', '.epub')
 mimetypes.add_type('application/x-dtbncx+xml', '.ncx')  # EPUB Navigation Control XML File
 
+# This should be removed after we move to python3.10
+# https://github.com/python/cpython/commit/20a5b7e986377bdfd929d7e8c4e3db5847dfdb2d
+mimetypes.add_type('image/heic', '.heic')
+
 
 def __load_all_parsers():
     """ Loads every parser in a dynamic way """
diff --git a/tests/data/dirty.heic b/tests/data/dirty.heic
new file mode 100644
index 0000000..0dc531d
--- /dev/null
+++ b/tests/data/dirty.heic
diff --git a/tests/test_libmat2.py b/tests/test_libmat2.py
index 9e56969..b32005f 100644
--- a/tests/test_libmat2.py
+++ b/tests/test_libmat2.py
@@ -251,6 +251,12 @@ class TestGetMeta(unittest.TestCase):
         meta = p.get_meta()
         self.assertEqual(meta['Name'], 'I am so')
 
+    def test_heic(self):
+        p = images.HEICParser('./tests/data/dirty.heic')
+        meta = p.get_meta()
+        self.assertEqual(meta['ProfileCopyright'], 'Public Domain')
+        self.assertEqual(meta['ProfileDescription'], 'GIMP built-in sRGB')
+
 
 class TestRemovingThumbnails(unittest.TestCase):
     def test_odt(self):
@@ -504,6 +510,11 @@ class TestCleaning(unittest.TestCase):
                 'EncodingSettings': 'Lavf52.103.0',
             },
             'expected_meta': {},
+        },{
+            'name': 'heic',
+            'parser': images.HEICParser,
+            'meta': {},
+            'expected_meta': {},
         }
         ]