Add validation for synset definitions and examples (#230)

- W305 blank definition
- W306 blank example
- W307 repeated definition

Resolves #151

CHANGELOG.md

@@ -11,6 +11,9 @@
 * `Relation` class ([#216])
 * `Sense.relation_map()` method ([#216])
 * `Synset.relation_map()` method ([#167], [#216])
+* `W305` blank definition on synset validation ([#151])
+* `W306` blank example on synset validation ([#151])
+* `W307` repeated definition on synset validation ([#151])
 
 ## Fixed
 
@@ -687,6 +690,7 @@ abandoned, but this is an entirely new codebase.
 [#146]: https://github.com/goodmami/wn/issues/146
 [#147]: https://github.com/goodmami/wn/issues/147
 [#148]: https://github.com/goodmami/wn/issues/148
+[#151]: https://github.com/goodmami/wn/issues/151
 [#152]: https://github.com/goodmami/wn/issues/152
 [#154]: https://github.com/goodmami/wn/issues/154
 [#155]: https://github.com/goodmami/wn/issues/155

tests/data/W305-0.xml

@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE LexicalResource SYSTEM "http://globalwordnet.github.io/schemas/WN-LMF-1.0.dtd">
+<LexicalResource xmlns:dc="http://purl.org/dc/elements/1.1/">
+
+<!-- blank definition in synset -->
+
+  <Lexicon id="test-w305"
+           label="Testing W305"
+           language="en"
+           email="[email protected]"
+           license="https://creativecommons.org/licenses/by/4.0/"
+           version="1">
+
+    <LexicalEntry id="test-w305-foo-n">
+      <Lemma partOfSpeech="n" writtenForm="foo" />
+      <Sense id="test-w305-foo-n" synset="test-w305-01-n" />
+    </LexicalEntry>
+
+    <Synset id="test-w305-01-n" ili="i12345" partOfSpeech="n">
+      <Definition>
+
+      </Definition>
+    </Synset>
+
+  </Lexicon>
+
+</LexicalResource>

tests/data/W306-0.xml

@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE LexicalResource SYSTEM "http://globalwordnet.github.io/schemas/WN-LMF-1.0.dtd">
+<LexicalResource xmlns:dc="http://purl.org/dc/elements/1.1/">
+
+<!-- blank example in synset -->
+
+  <Lexicon id="test-w306"
+           label="Testing W306"
+           language="en"
+           email="[email protected]"
+           license="https://creativecommons.org/licenses/by/4.0/"
+           version="1">
+
+    <LexicalEntry id="test-w306-foo-n">
+      <Lemma partOfSpeech="n" writtenForm="foo" />
+      <Sense id="test-w306-foo-n" synset="test-w306-01-n" />
+    </LexicalEntry>
+
+    <Synset id="test-w306-01-n" ili="i12345" partOfSpeech="n">
+      <Example>
+
+      </Example>
+    </Synset>
+
+  </Lexicon>
+
+</LexicalResource>

tests/data/W307-0.xml

@@ -0,0 +1,30 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE LexicalResource SYSTEM "http://globalwordnet.github.io/schemas/WN-LMF-1.0.dtd">
+<LexicalResource xmlns:dc="http://purl.org/dc/elements/1.1/">
+
+<!-- repeated definition in synset -->
+
+  <Lexicon id="test-w307"
+           label="Testing W307"
+           language="en"
+           email="[email protected]"
+           license="https://creativecommons.org/licenses/by/4.0/"
+           version="1">
+
+    <LexicalEntry id="test-w307-foo-n">
+      <Lemma partOfSpeech="n" writtenForm="foo" />
+      <Sense id="test-w307-foo-1-n" synset="test-w307-01-n" />
+      <Sense id="test-w307-foo-2-n" synset="test-w307-02-n" />
+    </LexicalEntry>
+
+    <Synset id="test-w307-01-n" ili="i12345" partOfSpeech="n">
+      <Definition>foo</Definition>
+    </Synset>
+
+    <Synset id="test-w307-02-n" ili="i12346" partOfSpeech="n">
+      <Definition>foo</Definition>
+    </Synset>
+
+  </Lexicon>
+
+</LexicalResource>

tests/validate_test.py

@@ -1,18 +1,24 @@
-from pathlib import Path
+import pytest
 
 from wn import lmf
 from wn.validate import validate
 
+tests = [
+    ("E101", 0),
+    ("E101", 1),
+    ("E101", 2),
+    ("E101", 3),
+    ("W305", 0),
+    ("W306", 0),
+    ("W307", 0),
+]
+test_ids = [f"{code}-{i}" for code, i in tests]
 
-def _assert_invalid(select: str, path: Path) -> None:
+
+@pytest.mark.parametrize("code,i", tests, ids=test_ids)
+def test_validate(datadir, code: str, i: int) -> None:
+    path = datadir / f"{code}-{i}.xml"
     lex = lmf.load(path, progress_handler=None)["lexicons"][0]
-    report = validate(lex, select=[select], progress_handler=None)
+    report = validate(lex, select=[code], progress_handler=None)
     print(report)
-    assert len(report[select]["items"]) > 0
-
-
-def test_E101(datadir):
-    _assert_invalid("E101", datadir / "E101-0.xml")
-    _assert_invalid("E101", datadir / "E101-1.xml")
-    _assert_invalid("E101", datadir / "E101-2.xml")
-    _assert_invalid("E101", datadir / "E101-3.xml")
+    assert len(report[code]["items"]) > 0

wn/validate.py

@@ -16,6 +16,9 @@
 W302  ILI is repeated across synsets.
 W303  Proposed ILI is missing a definition.
 W304  Existing ILI has a spurious definition.
+W305  Synset has a blank definition.
+W306  Synset has a blank example.
+W307  Synset repeats an existing definition.
 E401  Relation target is missing or invalid.
 W402  Relation type is invalid for the source and target.
 W403  Redundant relation between source and target.
@@ -125,6 +128,34 @@ def _spurious_ili_definition(lex: lmf.Lexicon, ids: _Ids) -> _Result:
             if ss['ili'] and ss['ili'] != 'in' and ss.get('ili_definition')}
 
 
+def _blank_synset_definition(lex: lmf.Lexicon, ids: _Ids) -> _Result:
+    """synset has a blank definition"""
+    return {
+        ss['id']: {} for ss in _synsets(lex)
+        if any(dfn["text"].strip() == "" for dfn in ss.get("definitions", []))
+    }
+
+def _blank_synset_example(lex: lmf.Lexicon, ids: _Ids) -> _Result:
+    """synset has a blank example"""
+    return {
+        ss['id']: {} for ss in _synsets(lex)
+        if any(ex["text"].strip() == "" for ex in ss.get("examples", []))
+    }
+
+
+def _repeated_synset_definition(lex: lmf.Lexicon, ids: _Ids) -> _Result:
+    """synset repeats an existing definition"""
+    repeated = _multiples(
+        dfn["text"]
+        for ss in _synsets(lex)
+        for dfn in ss.get("definitions", [])
+    )
+    return {
+        ss["id"]: {} for ss in _synsets(lex)
+        if any(dfn["text"] in repeated for dfn in ss.get("definitions", []))
+    }
+
+
 def _missing_relation_target(lex: lmf.Lexicon, ids: _Ids) -> _Result:
     """relation target is missing or invalid"""
     result = {s['id']: {'type': r['relType'], 'target': r['target']}
@@ -253,6 +284,9 @@ def _get_dc_type(r: lmf.Relation) -> Optional[str]:
     'W302': _repeated_ili,
     'W303': _missing_ili_definition,
     'W304': _spurious_ili_definition,
+    'W305': _blank_synset_definition,
+    'W306': _blank_synset_example,
+    'W307': _repeated_synset_definition,
     # 400 - relations
     'E401': _missing_relation_target,
     'W402': _invalid_relation_type,