Fix the test fetcher

src/transformers/utils/import_utils.py

@@ -2378,6 +2378,7 @@ def flatten_dict(_dict, previous_key=None):
     return flattened_import_structure
 
 
+@lru_cache()
 def define_import_structure(module_path: str, prefix: str = None) -> IMPORT_STRUCTURE_T:
     """
     This method takes a module_path as input and creates an import structure digestible by a _LazyModule.

utils/tests_fetcher.py

@@ -736,7 +736,7 @@ def get_module_dependencies(module_fname: str, cache: Dict[str, List[str]] = Non
             # the object is fully defined in the __init__)
             if module.endswith("__init__.py"):
                 # So we get the imports from that init then try to find where our objects come from.
-                new_imported_modules = extract_imports(module, cache=cache)
+                new_imported_modules = dict(extract_imports(module, cache=cache))
 
                 # Add imports via `define_import_structure` after the #35167 as we remove explicit import in `__init__.py`
                 from transformers.utils.import_utils import define_import_structure
@@ -749,9 +749,15 @@ def get_module_dependencies(module_fname: str, cache: Dict[str, List[str]] = Non
                         # We replace with os.path.sep so that it's Windows-compatible
                         _module = _module.replace(".", os.path.sep)
                         _module = module.replace("__init__.py", f"{_module}.py")
-                        new_imported_modules.append((_module, list(_imports)))
-
-                for new_module, new_imports in new_imported_modules:
+                        if _module not in new_imported_modules:
+                            new_imported_modules[_module] = list(_imports)
+                        else:
+                            original_imports = new_imported_modules[_module]
+                            for potential_new_item in list(_imports):
+                                if potential_new_item not in original_imports:
+                                    new_imported_modules[_module].append(potential_new_item)
+
+                for new_module, new_imports in new_imported_modules.items():
                     if any(i in new_imports for i in imports):
                         if new_module not in dependencies:
                             new_modules.append((new_module, [i for i in new_imports if i in imports]))
@@ -1041,18 +1047,17 @@ def infer_tests_to_run(
     """
     if not test_all:
         modified_files = get_modified_python_files(diff_with_last_commit=diff_with_last_commit)
-        reverse_map = create_reverse_dependency_map()
-        impacted_files = modified_files.copy()
-        for f in modified_files:
-            if f in reverse_map:
-                impacted_files.extend(reverse_map[f])
     else:
-        impacted_files = modified_files = [
-            str(k) for k in PATH_TO_TESTS.glob("*/*") if str(k).endswith(".py") and "test_" in str(k)
-        ]
+        modified_files = [str(k) for k in PATH_TO_TESTS.glob("*/*") if str(k).endswith(".py") and "test_" in str(k)]
         print("\n### test_all is TRUE, FETCHING ALL FILES###\n")
     print(f"\n### MODIFIED FILES ###\n{_print_list(modified_files)}")
 
+    reverse_map = create_reverse_dependency_map()
+    impacted_files = modified_files.copy()
+    for f in modified_files:
+        if f in reverse_map:
+            impacted_files.extend(reverse_map[f])
+
     # Remove duplicates
     impacted_files = sorted(set(impacted_files))
     print(f"\n### IMPACTED FILES ###\n{_print_list(impacted_files)}")