include wc validator in the pytest suite for the exported html

.github/workflows/test.yml

@@ -32,6 +32,7 @@ jobs:
     - name: init playwright nbconvert-a11y
       run: |
         playwright install --with-deps chromium
+        npm install vnu-jar
         pip install -e.
         doit copy
     - name: test with pytest

test-environment.yml

@@ -7,14 +7,14 @@ dependencies:
   - openjdk
   - pip
   - pip:
-    - html5validator
     - pytest-playwright
     - nbval
     - accessible-pygments
     - requests-cache
     - markdown-it-py[plugins,linkify]
     - python-slugify
     - html5lib
+  - nodejs
   - playwright
   - nbconvert
   - pytest

tests/test_w3c.py

@@ -0,0 +1,92 @@
+# requires node
+# requires jvm
+
+import itertools, operator, functools, collections, exceptiongroup, re
+import pathlib, json, subprocess, shlex
+
+EXCLUDE = re.compile(
+    """or with a “role” attribute whose value is “table”, “grid”, or “treegrid”.$"""
+    # https://github.com/validator/validator/issues/1125
+)
+
+
+@functools.lru_cache(1)
+def vnu_jar():
+    VNU_JAR = (
+        pathlib.Path(subprocess.check_output(shlex.split("npm root vnu-jar")).strip().decode())
+        / "vnu-jar/build/dist/vnu.jar"
+    )
+    assert VNU_JAR.exists()
+    return VNU_JAR
+
+
+def validate_html(*files: pathlib.Path) -> dict:
+    return json.loads(
+        subprocess.check_output(
+            shlex.split(f"java -jar {vnu_jar()} --stdout --format json --exit-zero-always")
+            + list(files)
+        ).decode()
+    )
+
+
+def organize_validator_results(results):
+    collect = collections.defaultdict(functools.partial(collections.defaultdict, list))
+    for (error, msg), group in itertools.groupby(
+        results["messages"], key=operator.itemgetter("type", "message")
+    ):
+        for item in group:
+            collect[error][msg].append(item)
+    return collect
+
+def raise_if_errors(results, exclude=EXCLUDE):
+    collect = organize_validator_results(results)
+    exceptions = []
+    for msg in collect["error"]:
+        if not exclude or not exclude.search(msg):
+            exceptions.append(exceptiongroup.ExceptionGroup(msg, [Exception(x["extract"]) for x in collect["error"][msg]]))
+    if exceptions:
+            raise exceptiongroup.ExceptionGroup("nu validator errors", exceptions)
+
+
+import dataclasses
+from json import dumps, loads
+from logging import getLogger
+from pathlib import Path
+
+import exceptiongroup
+from test_nbconvert_html5 import exporter
+
+
+from pytest import fixture, mark
+
+HERE = Path(__file__).parent
+NOTEBOOKS = HERE / "notebooks"
+EXPORTS = HERE / "exports"
+HTML = EXPORTS / "html"
+LOGGER = getLogger(__name__)
+VALIDATOR = EXPORTS / "validator"
+
+# it would be possible to test loaded baseline documents with playwright.
+# export the resting state document and pass them to the validator. 
+# this would be better validate widgets.
+
+@mark.parametrize(
+    "notebook",
+    list(
+        x
+        for x in NOTEBOOKS.glob("*.ipynb")
+        if x.name not in {"Imaging_Sky_Background_Estimation.ipynb"}
+    ),
+)
+def test_baseline_w3c(page, exporter, notebook):
+    target = HTML / notebook.with_suffix(".html").name
+    target.parent.mkdir(exist_ok=True, parents=True)
+    target.write_text(exporter.from_filename(notebook)[0])
+
+    result = validate_html(target)
+    VALIDATOR.mkdir(parents=True, exist_ok=True)
+    audit = VALIDATOR / notebook.with_suffix(".json").name
+    LOGGER.info(f"""writing {audit} with {len(result.get("violations", ""))} violations""")
+    audit.write_text(dumps(result))
+
+    raise_if_errors(result)