Merge pull request #236 from khaeru/ci/2025-w29

Adjust ECB data expectations in 2 tests

Author: khaeru

doc/api.rst

@@ -71,6 +71,11 @@ SDMX-ML
 
 .. automodule:: sdmx.format.xml
    :members:
+   :exclude-members: validate_xml
+
+.. automodule:: sdmx.format.xml.common
+   :members:
+   :exclude-members: install_schemas, validate_xml
 
 ``message``: SDMX messages
 ==========================

doc/whatsnew.rst

@@ -6,6 +6,10 @@ What's new?
 Next release
 ============
 
+- :func:`.install_schemas` and :func:`.construct_schema` fetch, store, and use a local copy of :file:`xhtml1-strict.dsd` (:pull:`236`, :issue:`235`).
+  This enables use of :func:`.validate_xml`
+  with lxml version 6.0.0 (`released 2025-06-26 <https://lxml.de/6.0/changes-6.0.0.html>`__)
+  for SDMX-ML messages containing XHTML values.
 - Correct a broken link to :ref:`im` in the README (:pull:`233`; thanks :gh-user:`econometricsfanboy` for :issue:`232`).
 
 .. _2.22.0:

sdmx/format/xml/common.py

@@ -189,11 +189,11 @@ def construct_schema(
     if schema_etree is None:
         raise FileNotFoundError(f"Could not find XSD files in {schema_dir}")
 
-    # Modify the schema by inserting an <xs:import > reference to the XHTML schema URL
+    # Modify the schema by inserting an <xs:import > reference to the XHTML schema file
     elem = etree.Element(
         "{http://www.w3.org/2001/XMLSchema}import",
         namespace="http://www.w3.org/1999/xhtml",
-        schemaLocation="http://www.w3.org/2002/08/xhtml/xhtml1-strict.xsd",
+        schemaLocation="xhtml1-strict.xsd",
     )
     schema_etree.getroot().insert(0, elem)
 
@@ -269,6 +269,10 @@ def _extracted_zipball(version: Version, force: bool = False) -> Path:
             # Unpack the entire archive
             zf.extractall(target.parent)
 
+    # Fetch a copy of the XHTML1 XSD, which is missing from the SDMX bundle
+    resp = requests.get(url="http://www.w3.org/2002/08/xhtml/xhtml1-strict.xsd")
+    result.joinpath("schemas", "xhtml1-strict.xsd").write_bytes(resp.content)
+
     return result
 
 

sdmx/testing/__init__.py

@@ -263,6 +263,7 @@ def mock_gh_api():
     mock = responses.RequestsMock(assert_all_requests_are_fired=False)
     mock.add_passthru(re.compile(rf"{base}/zipball/\w+"))
     mock.add_passthru(re.compile(r"https://codeload.github.com/\w+"))
+    mock.add_passthru(re.compile(r"http://www.w3.org/\w+"))
 
     for v in "2.1", "3.0", "3.0.0":
         mock.get(

sdmx/tests/test_client.py

@@ -204,7 +204,7 @@ def test_preview_data(self) -> None:
 
         # A filter can be provided, resulting in fewer keys
         keys = ECB.preview_data("EXR", {"CURRENCY": "CAD+CHF+CNY"})
-        N = 30
+        N = 33
         assert N >= len(keys)
 
         # Result can be converted to pandas object

sdmx/tests/test_docs.py

@@ -233,7 +233,7 @@ def test_doc_usage_data():
 
     series_keys = list(data.series)
 
-    assert len(series_keys) == 16
+    assert len(series_keys) == 22
 
     series_keys[5]
 

sdmx/tests/writer/test_writer_xml.py

@@ -314,7 +314,7 @@ def test_ErrorMessage(errormessage):
     sdmx.to_xml(errormessage)
 
 
-@pytest.mark.usefixtures("tmp_path", "specimen")
+@pytest.mark.usefixtures("tmp_path", "installed_schemas", "specimen")
 class RoundTripTests(ABC):
     """Abstract base class for tests that read-write-read SDMX-ML."""
 
@@ -332,6 +332,7 @@ def rw_test(
         # Unpack fixture using the request_fixture
         tmp_path = request.getfixturevalue("tmp_path")
         specimen = request.getfixturevalue("specimen")
+        schemas_dir = request.getfixturevalue("installed_schemas")
 
         if structure_id:
             # Read {D,Metad}ataStructure from file
@@ -350,7 +351,7 @@ def rw_test(
         data = io.BytesIO(sdmx.to_xml(msg0, pretty_print=True))
 
         # Validate using XSD
-        assert not validate or validate_xml(data), "Invalid SDMX-ML"
+        assert not validate or validate_xml(data, schemas_dir), "Invalid SDMX-ML"
 
         # Contents can be read again
         try: