Co-authored-by: aider (openrouter/openai/gpt-5.2-codex) <aider@aider.chat>

@@ -0,0 +53,4 @@
+        client.dataset_info(repo_id=repo_id)
+    except HfHubHTTPError as exc:
+        status_code = exc.response.status_code if exc.response else None
+        if status_code == 404:

@@ -0,0 +100,4 @@
+    return download_path
+
+
+def dataset_is_parquet_file(

@@ -0,0 +194,4 @@
+                rdf_format = "xml"
+
+            graph = Graph()
+            graph.parse(str(file_path), format=rdf_format)

@@ -0,0 +203,4 @@
+    return store
+
+
+def collect_parquet_strings(

@@ -1269,0 +1286,4 @@
+    base_dir: Path,
+) -> bool:
+    """Validate that source strings and parquet strings match (Level 2/3)."""
+    source_strings = collect_source_strings(dataset_id, base_dir)

@@ -1586,6 +1718,11 @@ def main() -> int:
                 console.print(f"[bold cyan]Removing {rdf_file}[/bold cyan]")
                 remove_dir(rdf_file.parent)
 
+        validation_ok = validate(datasets_to_process, args.base_dir)

@@ -1269,0 +1299,4 @@
+    else:
+        parquet_strings = _new_string_store()
+        for parquet_file in parquet_files:
+            parquet_strings = parquet_strings.union(

@@ -1269,0 +1303,4 @@
+                collect_parquet_strings(parquet_file)
+            )
+
+    missing_in_parquet = parquet_strings in source_strings

@@ -1269,0 +1322,4 @@
+            )
+
+    missing_in_parquet = is_sub_bloom_filter(parquet_strings, source_strings)
+    if missing_in_parquet:

@@ -1269,0 +1329,4 @@
+        return False
+
+    missing_in_source = is_sub_bloom_filter(source_strings, parquet_strings)
+    if missing_in_source:

@@ -1269,0 +1321,4 @@
+                collect_parquet_strings(parquet_file)
+            )
+
+    missing_in_parquet = is_sub_bloom_filter(parquet_strings, source_strings)

@@ -1269,0 +1328,4 @@
+        )
+        return False
+
+    missing_in_source = is_sub_bloom_filter(source_strings, parquet_strings)

@@ -0,0 +384,4 @@
+    count = 0
+    filter_fn = filter_fn or filter_string
+
+    for field in _generate_fields_from_source(dataset_id, base_dir):