Merge pull request #17969 from yuyichao/svgmeta

tacaswell · web-flow · commit 6958f8856e02 · 2020-07-21T20:04:27.000-04:00
FIX: Honor `'Date': None` in metadata
diff --git a/lib/matplotlib/backends/backend_svg.py b/lib/matplotlib/backends/backend_svg.py
@@ -360,7 +360,8 @@ def _write_metadata(self, metadata):
                                  'Expected str, date, datetime, or iterable '
                                  'of the same, not {!r}.'.format(type(date)))
             metadata['Date'] = '/'.join(dates)
-        else:
+        elif 'Date' not in metadata:
+            # Do not add `Date` if the user explicitly set `Date` to `None`
             # Get source date from SOURCE_DATE_EPOCH, if set.
             # See https://reproducible-builds.org/specs/source-date-epoch/
             date = os.getenv("SOURCE_DATE_EPOCH")
@@ -370,23 +371,30 @@ def _write_metadata(self, metadata):
             else:
                 metadata['Date'] = datetime.datetime.today().isoformat()
 
-        mid = writer.start('metadata')
-        writer.start('rdf:RDF', attrib={
+        mid = None
+        def ensure_metadata(mid):
+            if mid is not None:
+                return mid
+            mid = writer.start('metadata')
+            writer.start('rdf:RDF', attrib={
                 'xmlns:dc': "http://purl.org/dc/elements/1.1/",
                 'xmlns:cc': "http://creativecommons.org/ns#",
                 'xmlns:rdf': "http://www.w3.org/1999/02/22-rdf-syntax-ns#",
             })
-        writer.start('cc:Work')
+            writer.start('cc:Work')
+            return mid
 
         uri = metadata.pop('Type', None)
         if uri is not None:
+            mid = ensure_metadata(mid)
             writer.element('dc:type', attrib={'rdf:resource': uri})
 
         # Single value only.
         for key in ['title', 'coverage', 'date', 'description', 'format',
                     'identifier', 'language', 'relation', 'source']:
             info = metadata.pop(key.title(), None)
             if info is not None:
+                mid = ensure_metadata(mid)
                 writer.element(f'dc:{key}', text=info)
 
         # Multiple Agent values.
@@ -398,6 +406,7 @@ def _write_metadata(self, metadata):
             if isinstance(agents, str):
                 agents = [agents]
 
+            mid = ensure_metadata(mid)
             writer.start(f'dc:{key}')
             for agent in agents:
                 writer.start('cc:Agent')
@@ -411,14 +420,16 @@ def _write_metadata(self, metadata):
             if isinstance(keywords, str):
                 keywords = [keywords]
 
+            mid = ensure_metadata(mid)
             writer.start('dc:subject')
             writer.start('rdf:Bag')
             for keyword in keywords:
                 writer.element('rdf:li', text=keyword)
             writer.end('rdf:Bag')
             writer.end('dc:subject')
 
-        writer.close(mid)
+        if mid is not None:
+            writer.close(mid)
 
         if metadata:
             raise ValueError('Unknown metadata key(s) passed to SVG writer: ' +
diff --git a/lib/matplotlib/tests/test_backend_svg.py b/lib/matplotlib/tests/test_backend_svg.py
@@ -371,6 +371,70 @@ def test_svg_default_metadata(monkeypatch):
     # Type
     assert 'StillImage' in buf
 
+    # Now make sure all the default metadata can be cleared.
+    with BytesIO() as fd:
+        fig.savefig(fd, format='svg', metadata={'Date': None, 'Creator': None,
+                                                'Format': None, 'Type': None})
+        buf = fd.getvalue().decode()
+
+    # Creator
+    assert mpl.__version__ not in buf
+    # Date
+    assert '1970-08-16' not in buf
+    # Format
+    assert 'image/svg+xml' not in buf
+    # Type
+    assert 'StillImage' not in buf
+
+
+def test_svg_clear_default_metadata(monkeypatch):
+    # Makes sure that setting a default metadata to `None`
+    # removes the corresponding tag from the metadata.
+    monkeypatch.setenv('SOURCE_DATE_EPOCH', '19680801')
+
+    metadata_contains = {'creator': mpl.__version__, 'date': '1970-08-16',
+                         'format': 'image/svg+xml', 'type': 'StillImage'}
+
+    SVGNS = '{http://www.w3.org/2000/svg}'
+    RDFNS = '{http://www.w3.org/1999/02/22-rdf-syntax-ns#}'
+    CCNS = '{http://creativecommons.org/ns#}'
+    DCNS = '{http://purl.org/dc/elements/1.1/}'
+
+    fig, ax = plt.subplots()
+    for name in metadata_contains:
+        with BytesIO() as fd:
+            fig.savefig(fd, format='svg', metadata={name.title(): None})
+            buf = fd.getvalue().decode()
+
+        root = xml.etree.ElementTree.fromstring(buf)
+        work, = root.findall(f'./{SVGNS}metadata/{RDFNS}RDF/{CCNS}Work')
+        for key in metadata_contains:
+            data = work.findall(f'./{DCNS}{key}')
+            if key == name:
+                # The one we cleared is not there
+                assert not data
+                continue
+            # Everything else should be there
+            data, = data
+            xmlstr = xml.etree.ElementTree.tostring(data, encoding="unicode")
+            assert metadata_contains[key] in xmlstr
+
+
+def test_svg_clear_all_metadata():
+    # Makes sure that setting all default metadata to `None`
+    # removes the metadata tag from the output.
+
+    fig, ax = plt.subplots()
+    with BytesIO() as fd:
+        fig.savefig(fd, format='svg', metadata={'Date': None, 'Creator': None,
+                                                'Format': None, 'Type': None})
+        buf = fd.getvalue().decode()
+
+    SVGNS = '{http://www.w3.org/2000/svg}'
+
+    root = xml.etree.ElementTree.fromstring(buf)
+    assert not root.findall(f'./{SVGNS}metadata')
+
 
 def test_svg_metadata():
     single_value = ['Coverage', 'Identifier', 'Language', 'Relation', 'Source',
