remove parquet output support

Nate has given up on getting wikiq to output parquet directly because it
makes too many unpredictable large memory allocations. The workflow now
outputs JSONL in a single pass and then uses spark (wikiq_spark) to index
it as parquet in a second pass.

Remove the parquet output path along with the machinery that existed only
to support it: checkpoint files, resume temp-file merging, namespace
partitioning, and file rotation (--partition-namespaces,
--max-revisions-per-file). Resume support remains for JSONL output, where
the resume point is derived from the last complete line of the output
file. Also remove the parquet tests and baseline files.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
(cherry picked from commit 72a8851373)

test/Wikiq_Unit_Test.py

@@ -440,7 +440,7 @@ def test_external_links_only():
     # Verify citations column does NOT exist
     assert "citations" not in test.columns, "citations column should NOT exist when only --external-links is used"
 
-    # Verify column has list/array type (pandas reads parquet lists as numpy arrays)
+    # Verify column has list/array type (pandas reads pyarrow lists as numpy arrays)
     assert test["external_links"].apply(lambda x: x is None or hasattr(x, '__len__')).all(), \
         "external_links should be a list/array type or None"
 
@@ -495,7 +495,7 @@ def test_citations_only():
     # Verify external_links column does NOT exist
     assert "external_links" not in test.columns, "external_links column should NOT exist when only --citations is used"
 
-    # Verify column has list/array type (pandas reads parquet lists as numpy arrays)
+    # Verify column has list/array type (pandas reads pyarrow lists as numpy arrays)
     assert test["citations"].apply(lambda x: x is None or hasattr(x, '__len__')).all(), \
         "citations should be a list/array type or None"
 
@@ -545,7 +545,7 @@ def test_external_links_and_citations():
     assert "external_links" in test.columns, "external_links column should exist"
     assert "citations" in test.columns, "citations column should exist"
 
-    # Verify both columns have list/array types (pandas reads parquet lists as numpy arrays)
+    # Verify both columns have list/array types (pandas reads pyarrow lists as numpy arrays)
     assert test["external_links"].apply(lambda x: x is None or hasattr(x, '__len__')).all(), \
         "external_links should be a list/array type or None"
     assert test["citations"].apply(lambda x: x is None or hasattr(x, '__len__')).all(), \
@@ -753,37 +753,4 @@ def test_headings():
     print(f"Headings test passed! {len(test)} rows processed")
 
 
-def test_parquet_output():
-    """Test that Parquet output format works correctly."""
-    tester = WikiqTester(SAILORMOON, "parquet_output", in_compression="7z", out_format="parquet")
-
-    try:
-        tester.call_wikiq("--fandom-2020")
-    except subprocess.CalledProcessError as exc:
-        pytest.fail(exc.stderr.decode("utf8"))
-
-    # Verify output file exists
-    assert os.path.exists(tester.output), f"Parquet output file should exist at {tester.output}"
-
-    # Read and verify content
-    test = pd.read_parquet(tester.output)
-
-    # Verify expected columns exist
-    assert "revid" in test.columns
-    assert "articleid" in test.columns
-    assert "title" in test.columns
-    assert "namespace" in test.columns
-
-    # Verify row count matches JSONL output
-    tester_jsonl = WikiqTester(SAILORMOON, "parquet_compare", in_compression="7z", out_format="jsonl")
-    try:
-        tester_jsonl.call_wikiq("--fandom-2020")
-    except subprocess.CalledProcessError as exc:
-        pytest.fail(exc.stderr.decode("utf8"))
-
-    test_jsonl = pd.read_json(tester_jsonl.output, lines=True)
-    assert len(test) == len(test_jsonl), f"Parquet and JSONL should have same row count: {len(test)} vs {len(test_jsonl)}"
-
-    print(f"Parquet output test passed! {len(test)} rows")
-