refactor(tavern/_core/pytest/file.py): Improve handling of initial YAML document in test file

- The loop variable `i` has been renamed to `document_idx` to better reflect its purpose, which is to track the index of the YAML document being processed.
- A comment has been added to explain the implicit behavior when the first YAML document has no "stages" but may have other properties like "name".
- The variable `merge_down` is now correctly initialized and used based on the index of the document being processed.

Author: michaelboulton

tavern/_core/pytest/file.py

@@ -366,12 +366,15 @@ def collect(self) -> Iterator[YamlItem]:
             raise exceptions.BadSchemaError from e
 
         merge_down = None
-        for i, test_spec in enumerate(all_tests):
+        # Iterate over yaml documents and tests
+        for document_idx, test_spec in enumerate(all_tests):
             if not test_spec:
                 logger.warning("Empty document in input file '%s'", self.path)
                 continue
 
-            if i == 0 and not test_spec.get("stages"):
+            if document_idx == 0 and not test_spec.get("stages"):
+                # If it's the first document and there were no tests, its implicitly the
+                # 'defaults' for the file and will be merged with all lower documents
                 if test_spec.get("name"):
                     logger.warning("initial block had no stages, but had a name")
                 merge_down = test_spec