error during superanimal inference step

[ralphpeterson]

I'm experimenting with AmadeusGPT on a group of untracked rats. I got up to the first query in the YourData.ipynb notebook and am receiving this error. It looks like the model name is missing during the inference step. I'm running DLC v3.0.0rc10 and AmadeusGPT v0.1.3 on a MacBook w/ Apple M4 Max chip (maybe not compatible?). Thanks for your help!

-Ralph

query = "Plot the trajectory of the rats using the animal center and color it by time"
qa_message = amadeus.step(query)
parse_result(amadeus, qa_message)

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[18], line 2
      1 query = "Plot the trajectory of the rats using the animal center and color it by time"
----> 2 qa_message = amadeus.step(query)
      3 parse_result(amadeus, qa_message)

File /opt/anaconda3/envs/amadeusgpt/lib/python3.10/site-packages/amadeusgpt/main.py:157, in AMADEUS.step(self, user_query)
    154 integration_module_names = self.match_integration_module(user_query)
    156 self.sandbox.update_matched_integration_modules(integration_module_names)
--> 157 qa_message = self.sandbox.llm_step(user_query)
    159 return qa_message

File /opt/anaconda3/envs/amadeusgpt/lib/python3.10/site-packages/amadeusgpt/programs/sandbox.py:498, in Sandbox.llm_step(self, user_query)
    495 if len(self.meta_info) > 0:
    496     qa_message.meta_info = self.meta_info
--> 498 qa_message = self.llms["code_generator"].speak(self, qa_message)
    499 # cache the resulted qa message for future use
    501 self.message_cache[user_query] = qa_message

File /opt/anaconda3/envs/amadeusgpt/lib/python3.10/site-packages/amadeusgpt/analysis_objects/llm.py:303, in CodeGenerationLLM.speak(self, sandbox, qa_message, share_video_file)
    301 behavior_analysis = sandbox.analysis_dict[identifier]
    302 scene_image = behavior_analysis.visual_manager.get_scene_image()
--> 303 keypoint_names = behavior_analysis.animal_manager.get_keypoint_names()
    304 object_names = behavior_analysis.object_manager.get_object_names()
    305 animal_names = behavior_analysis.animal_manager.get_animal_names()        

File /opt/anaconda3/envs/amadeusgpt/lib/python3.10/site-packages/amadeusgpt/managers/animal_manager.py:379, in AnimalManager.get_keypoint_names(self)
    375 """
    376 Get the names of the bodyparts. This is used to index the keypoints for a specific bodypart.
    377 """
    378 # this is to initialize
--> 379 self.get_keypoints()
    381 return self.full_keypoint_names

File /opt/anaconda3/envs/amadeusgpt/lib/python3.10/site-packages/amadeusgpt/managers/animal_manager.py:318, in AnimalManager.get_keypoints(self)
    316 if not os.path.exists(self.keypoint_file_path):
    317     print(f"going to inference video with {self.superanimal_name}")
--> 318     video_inference_superanimal(
    319         videos=[self.video_file_path],
    320         superanimal_name=self.superanimal_name,
    321         max_individuals=self.max_individuals,
    322         video_adapt=False,
    323     )
    325 if os.path.exists(self.keypoint_file_path):
    326     self.init_pose()

TypeError: video_inference_superanimal() missing 1 required positional argument: 'model_name'