Refactor check_auto_docstring using AST (#41432)

* refactor check_auto_docstring with AST

* use dataclass for ASTIndexes

* simplify and improve readability

* fix missing imports

* fix modular

* fix modular issues

Author: yonigozlan

src/transformers/models/glm4v/modeling_glm4v.py

@@ -1418,14 +1418,11 @@ def forward(
         pixel_values_videos: Optional[torch.FloatTensor] = None,
         image_grid_thw: Optional[torch.LongTensor] = None,
         video_grid_thw: Optional[torch.LongTensor] = None,
-        rope_deltas: Optional[torch.LongTensor] = None,
         cache_position: Optional[torch.LongTensor] = None,
         logits_to_keep: Union[int, torch.Tensor] = 0,
         **kwargs: Unpack[TransformersKwargs],
     ) -> Union[tuple, Glm4vCausalLMOutputWithPast]:
         r"""
-        rope_deltas (`torch.LongTensor` of shape `(batch_size, )`, *optional*):
-            The rope index difference between sequence length and multimodal rope.
         labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
             Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
             config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored

src/transformers/models/glm4v/modular_glm4v.py

@@ -1341,14 +1341,11 @@ def forward(
         pixel_values_videos: Optional[torch.FloatTensor] = None,
         image_grid_thw: Optional[torch.LongTensor] = None,
         video_grid_thw: Optional[torch.LongTensor] = None,
-        rope_deltas: Optional[torch.LongTensor] = None,
         cache_position: Optional[torch.LongTensor] = None,
         logits_to_keep: Union[int, torch.Tensor] = 0,
         **kwargs: Unpack[TransformersKwargs],
     ) -> Union[tuple, Glm4vCausalLMOutputWithPast]:
         r"""
-        rope_deltas (`torch.LongTensor` of shape `(batch_size, )`, *optional*):
-            The rope index difference between sequence length and multimodal rope.
         labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
             Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
             config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored

src/transformers/models/glm4v_moe/modeling_glm4v_moe.py

@@ -1638,8 +1638,6 @@ def forward(
         **kwargs: Unpack[TransformersKwargs],
     ) -> Union[tuple, Glm4vMoeCausalLMOutputWithPast]:
         r"""
-        rope_deltas (`torch.LongTensor` of shape `(batch_size, )`, *optional*):
-            The rope index difference between sequence length and multimodal rope.
         labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
             Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
             config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored