Co-authored-by: aider (openrouter/anthropic/claude-sonnet-4) <aider@aider.chat>

@@ -0,0 +143,4 @@
+            context.auth_result = context.tracker.authenticate()
+        except Exception as e:
+            context.auth_result = False
+            context.error = str(e)

@@ -0,0 +44,4 @@
+
+  Scenario: Column letter conversion for column AB
+    When I convert column number 28 to letter
+    Then the column letter result should be "AB"

@@ -0,0 +41,4 @@
+
+  Scenario: Convenience methods return success status
+    When I mark "wordnet" as in progress
+    Then the mark_in_progress method should return True

@@ -0,0 +26,4 @@
+      | Dataset Name | Status      | Error |
+      | wordnet      | NOT STARTED |       |
+    When I attempt to mark "nonexistent" as in progress
+    Then the update should fail

@@ -0,0 +49,4 @@
+    Given an authenticated sheets tracker
+    And a sheet with empty rows
+    When I get the status of "wordnet"
+    Then the status should be None

@@ -0,0 +68,4 @@
+    And a sheet with only one row (headers only)
+    When I get the status of "wordnet"
+    Then the status should be None
+    And the row should be None

@@ -0,0 +76,4 @@
+      | Dataset Name | Status      | Error |
+      | wordnet      | NOT STARTED |       |
+    When I attempt to update status for "nonexistent-dataset"
+    Then the update should return False

@@ -0,0 +25,4 @@
+    And a sheet with headers "ID, Status, Error"
+    When the tracker loads headers
+    Then dataset_name_col should be identified as 0
+    And status_col should be 1

@@ -0,0 +32,4 @@
+    And a sheet with headers "Dataset Name, State, Error"
+    When the tracker loads headers
+    Then dataset_name_col should be 0
+    And status_col should be identified as 1

@@ -0,0 +39,4 @@
+    And a sheet with headers "Dataset Name, Status, Error Message"
+    When the tracker loads headers
+    Then dataset_name_col should be 0
+    And error_col should be identified as 2

@@ -0,0 +45,4 @@
+  Scenario: Process when tracker not authenticated
+    Given an unauthenticated sheets tracker
+    When I check if "wordnet" should be processed
+    Then the decision should be to process

@@ -0,0 +50,4 @@
+
+  Scenario: Error message truncation
+    When I reset "wordnet" to not started with error message longer than 500 characters
+    Then the error for "wordnet" should be at most 500 characters

@@ -0,0 +55,4 @@
+      | wordnet      | NOT STARTED |       |
+    When I attempt to update status for "nonexistent"
+    Then the update should fail gracefully
+    And no exception should be raised

@@ -0,0 +62,4 @@
+      | Dataset Name | Status      | Error |
+      | wordnet      | NOT STARTED |       |
+    When I attempt to update status with very long value
+    Then the update should handle the long value correctly

@@ -0,0 +7,4 @@
+    Given an authenticated sheets tracker
+
+  Scenario: Update status with only status column present (no error column)
+    Given a sheet with headers "Dataset Name, Status"

@@ -0,0 +1,2646 @@
+"""Step definitions for Google Sheets tracker tests."""

@@ -0,0 +95,4 @@
+            if header.strip().lower() in ["status", "processing status"]:
+                context.tracker.status_col = idx
+            elif header.strip().lower() in ["error", "error message"]:
+                context.tracker.error_col = idx

@@ -0,0 +2630,4 @@
+
+
+
+

@@ -523,2 +582,2 @@
-    graph = Graph()
-    graph.parse(str(file_path), format=rdf_format)
+    # Check if this is GeoNames format (special handling required)
+    if rdf_format in ("xml", "application/rdf+xml") and _detect_geonames_format(file_path):

@@ -828,0 +965,4 @@
+                                            subject=subject,
+                                            predicate=predicate,
+                                            object=obj,
+                                            object_type="literal",  # TSV typically has literals

@@ -828,0 +970,4 @@
+                                            object_language=None
+                                        ))
+                            except Exception as e:
+                                logger.debug(f"Skipping malformed TSV line {line_num}: {e}")

@@ -840,0 +990,4 @@
+            # Create train/test split if requested
+            if config.create_train_test_split:
+                assert isinstance(dataset, Dataset), "Expected Dataset instance"
+                train_test = dataset.train_test_split(test_size=config.test_size, seed=42)

@@ -1160,2 +1324,2 @@
-                            f"{num_workers} workers[/yellow]"
-                        )
+                # Handle compressed files
+                is_gzip = config.input_path.suffix == ".gz" or str(config.input_path).endswith(".gz")

@@ -1162,0 +1327,4 @@
+
+                try:
+                    if is_gzip:
+                        file_obj = gzip.open(config.input_path, "rt", encoding="utf-8", errors="ignore") # noqa: SIM115

@@ -1165,3 +1334,1 @@
-                            f"[yellow]Using standard RDF parser for "
-                            f"{config.rdf_format} (single-threaded)[/yellow]"
-                        )
+                        file_obj = open(config.input_path, encoding="utf-8", errors="ignore") # noqa: SIM115

@@ -1168,0 +1350,4 @@
+                    finally:
+                        file_obj.close()
+                except Exception as e:
+                    logger.warning(f"Error detecting GeoNames format: {e}, using generic parser")

@@ -1249,0 +1435,4 @@
+            # Create train/test split if requested
+            if config.create_train_test_split:
+                assert isinstance(dataset, Dataset), "Expected Dataset instance"
+                train_test = dataset.train_test_split(test_size=config.test_size, seed=42)

@@ -1278,0 +1473,4 @@
+                # Use single process for small datasets to avoid index errors
+                # Only use multiprocessing for datasets with > 1000 examples
+                if total_triples > 1000 and num_workers > 1:
+                    dataset_dict.save_to_disk(str(config.output_path), num_proc=num_workers)