Change how extensions and fallback priorities interact

This increases the priority of matching file extensions and mimetypes above non-matching values, even if the fallback is high priority.
girder · Jun 6, 2023 · eb7e920 · eb7e920
1 parent 736a82d
commit eb7e920
Show file tree

Hide file tree

Showing 4 changed files with 40 additions and 17 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,10 @@
 # Change Log
 
+## 1.22.3
+
+### Changes
+- Change how extensions and fallback priorities interact ([#1192](../../pull/1192))
+
 ## 1.22.2
 
 ### Changes

diff --git a/large_image/tilesource/__init__.py b/large_image/tilesource/__init__.py
@@ -63,16 +63,17 @@ def loadTileSources(entryPointName='large_image.source', sourceDict=AvailableTil
                 'Failed to loaded tile source %s' % entryPoint.name)
 
 
-def getSortedSourceList(availableSources, pathOrUri, *args, **kwargs):
+def getSortedSourceList(availableSources, pathOrUri, mimeType=None, *args, **kwargs):
     """
     Get an ordered list of sources where earlier sources are more likely to
     work for a specified path or uri.
 
     :param availableSources: an ordered dictionary of sources to try.
     :param pathOrUri: either a file path or a fixed source via
         large_image://<source>.
-    :returns: a list of (clash, priority, sourcename) for sources where
-        sourcename is a key in availableSources.
+    :param mimeType: the mimetype of the file, if known.
+    :returns: a list of (clash, fallback, priority, sourcename) for sources
+        where sourcename is a key in availableSources.
     """
     uriWithoutProtocol = str(pathOrUri).split('://', 1)[-1]
     isLargeImageUri = str(pathOrUri).startswith('large_image://')
@@ -84,8 +85,14 @@ def getSortedSourceList(availableSources, pathOrUri, *args, **kwargs):
     for sourceName in availableSources:
         sourceExtensions = availableSources[sourceName].extensions
         priority = sourceExtensions.get(None, SourcePriority.MANUAL)
+        fallback = True
+        if (mimeType and getattr(availableSources[sourceName], 'mimeTypes', None) and
+                mimeType in availableSources[sourceName].mimeTypes):
+            fallback = False
+            priority = min(priority, availableSources[sourceName].mimeTypes[mimeType])
         for ext in extensions:
             if ext in sourceExtensions:
+                fallback = False
                 priority = min(priority, sourceExtensions[ext])
         if isLargeImageUri and sourceName == uriWithoutProtocol:
             priority = SourcePriority.NAMED
@@ -94,11 +101,11 @@ def getSortedSourceList(availableSources, pathOrUri, *args, **kwargs):
         propertiesClash = any(
             getattr(availableSources[sourceName], k, False) != v
             for k, v in properties.items())
-        sourceList.append((propertiesClash, priority, sourceName))
+        sourceList.append((propertiesClash, fallback, priority, sourceName))
     return sourceList
 
 
-def getSourceNameFromDict(availableSources, pathOrUri, *args, **kwargs):
+def getSourceNameFromDict(availableSources, pathOrUri, mimeType=None, *args, **kwargs):
     """
     Get a tile source based on a ordered dictionary of known sources and a path
     name or URI.  Additional parameters are passed to the tile source and can
@@ -107,11 +114,12 @@ def getSourceNameFromDict(availableSources, pathOrUri, *args, **kwargs):
     :param availableSources: an ordered dictionary of sources to try.
     :param pathOrUri: either a file path or a fixed source via
         large_image://<source>.
+    :param mimeType: the mimetype of the file, if known.
     :returns: the name of a tile source that can read the input, or None if
         there is no such source.
     """
-    sourceList = getSortedSourceList(availableSources, pathOrUri, *args, **kwargs)
-    for _clash, _priority, sourceName in sorted(sourceList):
+    sourceList = getSortedSourceList(availableSources, pathOrUri, mimeType, *args, **kwargs)
+    for _clash, _fallback, _priority, sourceName in sorted(sourceList):
         if availableSources[sourceName].canRead(pathOrUri, *args, **kwargs):
             return sourceName
 
@@ -173,18 +181,23 @@ def canRead(*args, **kwargs):
     return False
 
 
-def canReadList(*args, **kwargs):
+def canReadList(pathOrUri, mimeType=None, *args, **kwargs):
     """
     Check if large_image can read a path or uri via each source.
 
+    :param pathOrUri: either a file path or a fixed source via
+        large_image://<source>.
+    :param mimeType: the mimetype of the file, if known.
     :returns: A list of tuples of (source name, canRead).
     """
     if not len(AvailableTileSources):
         loadTileSources()
-    sourceList = getSortedSourceList(AvailableTileSources, *args, **kwargs)
+    sourceList = getSortedSourceList(
+        AvailableTileSources, pathOrUri, mimeType, *args, **kwargs)
     result = []
-    for _clash, _priority, sourceName in sorted(sourceList):
-        result.append((sourceName, AvailableTileSources[sourceName].canRead(*args, **kwargs)))
+    for _clash, _fallback, _priority, sourceName in sorted(sourceList):
+        result.append((sourceName, AvailableTileSources[sourceName].canRead(
+            pathOrUri, *args, **kwargs)))
     return result
 
 

diff --git a/test/datastore.py b/test/datastore.py
@@ -30,6 +30,9 @@
     # Tiff with floating point pixels
     # Source: d042-353.crop.small.float32.tif
     'd042-353.crop.small.float32.tif': 'sha512:ae05dbe6f3330c912893b203b55db27b0fdf3222a0e7f626d372c09668334494d07dc1d35533670cfac51b588d2292eeee7431317741fdb4cbb281c28a289115',  # noqa
+    # JPEG with progressive compression and restart markers
+    # Source: d042-353.crop.small.jpg
+    'd042-353.crop.small.jpg': 'sha512:1353646637c1fae266b87312698aa39eca0311222c3a1399b60efdc13bfe55e2f3db59da005da945dd7e9e816f31ccd18846dd72744faac75215074c3d87414f',  # noqa
     # Tiff with JP2k compression
     # Source: huron.image2_jpeg2k.tif
     'huron.image2_jpeg2k.tif': 'sha512:eaba877079c86f0603b2a636a44d57832cdafe6d43a449121f575f0d43f69b8a17fa619301b066ece1c11050b41d687400b27407c404d827fd2c132d99e669ae',  # noqa

diff --git a/test/test_source_base.py b/test/test_source_base.py
@@ -38,12 +38,12 @@
     },
     'dummy': {'any': True, 'skipTiles': r''},
     'gdal': {
-        'read': r'\.(jpeg|jp2|ptif|scn|svs|tif.*|qptiff)$',
+        'read': r'\.(jpg|jpeg|jp2|ptif|scn|svs|tif.*|qptiff)$',
         'noread': r'(huron\.image2_jpeg2k|sample_jp2k_33003|TCGA-DU-6399|\.(ome.tiff|nc)$)',
         'skipTiles': r'\.*nc$',
     },
     'mapnik': {
-        'read': r'\.(jpeg|jp2|ptif|nc|scn|svs|tif.*|qptiff)$',
+        'read': r'\.(jpg|jpeg|jp2|ptif|nc|scn|svs|tif.*|qptiff)$',
         'noread': r'(huron\.image2_jpeg2k|sample_jp2k_33003|TCGA-DU-6399|\.(ome.tiff)$)',
         # we should only test this with a projection
         'skipTiles': r'',
@@ -64,12 +64,12 @@
         'skipTiles': r'one_layer_missing',
     },
     'pil': {
-        'read': r'\.(jpeg|png|tif.*)$',
-        'noread': r'(G10-3|JK-kidney|d042-353|huron|one_layer_missing|US_Geo|extraoverview' + (
+        'read': r'\.(jpg|jpeg|png|tif.*)$',
+        'noread': r'(G10-3|JK-kidney|d042-353.*tif|huron|one_layer_missing|US_Geo|extraoverview' + (
             r'|sample.*ome' if sys.version_info < (3, 7) else r'') + r')',
     },
     'rasterio': {
-        'read': r'\.(jpeg|jp2|ptif|scn|svs|tif.*|qptiff)$',
+        'read': r'\.(jpg|jpeg|jp2|ptif|scn|svs|tif.*|qptiff)$',
         'noread': r'(huron\.image2_jpeg2k|sample_jp2k_33003|TCGA-DU-6399|\.(ome.tiff|nc)$)',
         'python': sys.version_info >= (3, 8),
     },
@@ -82,7 +82,7 @@
         'skipTiles': r'(sample_image\.ptif|one_layer_missing_tiles)'},
     'tifffile': {
         'read': r'',
-        'noread': r'\.(nc|nd2|yml|yaml|json|czi|png|jpeg|jp2|dcm)$',
+        'noread': r'\.(nc|nd2|yml|yaml|json|czi|png|jpg|jpeg|jp2|dcm)$',
         'python': sys.version_info >= (3, 7),
     },
     'vips': {
@@ -113,9 +113,11 @@ def testCanRead():
     testDir = os.path.dirname(os.path.realpath(__file__))
     imagePath = os.path.join(testDir, 'test_files', 'yb10kx5k.png')
     assert large_image.canRead(imagePath) is False
+    assert large_image.canRead(imagePath, mimeType='image/png') is False
 
     imagePath = datastore.fetch('sample_image.ptif')
     assert large_image.canRead(imagePath) is True
+    assert large_image.canRead(imagePath, mimeType='image/png') is True
 
 
 @pytest.mark.parametrize('source', [k for k, v in SourceAndFiles.items() if not v.get('any')])