extract wikilinks.

test/Wikiq_Unit_Test.py

@@ -831,3 +831,44 @@ def test_no_wikitext_columns():
     assert "citations" not in test.columns, "citations column should NOT exist without --citations flag"
 
     print(f"No wikitext columns test passed! {len(test)} rows processed")
+
+
+def test_wikilinks():
+    """Test that --wikilinks extracts internal wikilinks correctly."""
+    import mwparserfromhell
+
+    tester = WikiqTester(SAILORMOON, "wikilinks", in_compression="7z", out_format="parquet")
+
+    try:
+        tester.call_wikiq("--wikilinks", "--text", "--fandom-2020")
+    except subprocess.CalledProcessError as exc:
+        pytest.fail(exc.stderr.decode("utf8"))
+
+    test = pd.read_parquet(tester.output + f"/{SAILORMOON}.parquet")
+
+    # Verify wikilinks column exists
+    assert "wikilinks" in test.columns, "wikilinks column should exist"
+
+    # Verify column has list/array type
+    assert test["wikilinks"].apply(lambda x: x is None or hasattr(x, '__len__')).all()
+
+    # Verify extraction matches mwparserfromhell for sample rows
+    rows_with_links = test[test["wikilinks"].apply(lambda x: x is not None and len(x) > 0)]
+    if len(rows_with_links) > 0:
+        sample = rows_with_links.head(5)
+        for idx, row in sample.iterrows():
+            text = row["text"]
+            if text:
+                wikicode = mwparserfromhell.parse(text)
+                expected = []
+                for link in wikicode.filter_wikilinks():
+                    title = str(link.title).strip()
+                    display_text = str(link.text).strip() if link.text else None
+                    expected.append({"title": title, "text": display_text})
+
+                actual = list(row["wikilinks"])
+                # Convert to comparable format (pandas may read as dicts or named tuples)
+                actual_dicts = [{"title": item["title"], "text": item["text"]} for item in actual]
+                assert actual_dicts == expected, f"Row {idx}: wikilinks mismatch"
+
+    print(f"Wikilinks test passed! {len(test)} rows processed")