Unhashable type numpy error

[Potajito]

Hi!
I'm trying to run the script, but I'm getting the numpy error "TypeError: unhashable type: 'numpy.ndarray'". This is on latest version, tried in windows and ubuntu, same issue. Python 3.10 and 3.11. The script seems to see the screen, but fails to continue.
Any idea?
Thanks!

INFO main.py 56 Running...
INFO deploy_agent.py 94 INITIALIZATION [self.run]: Loading Model
INFO deploy_agent.py 97 INITIALIZATION [self.run]: Create SuperAutoPetsEnv Object
INFO deploy_agent.py 104 CV SYSTEM [self.run]: Detect the Pets and Food in the Shop Slots
INFO deploy_agent.py 106 CV SYSTEM [self.run]: Calls [image_detection.find_the_animals]
INFO deploy_agent.py 111 CV SYSTEM [self.run]: The detected Pets and Food in the Shop is : ['fish', 'horse', 'pig', 'honey']
INFO deploy_agent.py 112 GAME ENGINE [self.run]: Set Environment Shop = detected Pets and Food
INFO deploy_agent.py 121 GAME ENGINE [self.run]: Get the best action to make for the given state from the loaded model
Traceback (most recent call last):
  File "/home/pablo/super-ml-pets/main.py", line 63, in <module>
    main()
  File "/home/pablo/super-ml-pets/main.py", line 57, in main
    run(ret)
  File "/home/pablo/super-ml-pets/src/deploy_agent.py", line 128, in run
    " Current Team and Shop \n{}".format(s[action][0]))
TypeError: unhashable type: 'numpy.ndarray'

[andreped]

I'm currently on Easter holiday, but I could take a look after the weekend.

I believe the deployment used to work just fine. Might be that we should change the numpy version, or similar.

Could also be that the deployment was not done according to the documentations. Could you take another look to ensure you have followed every single step (see here).

Could also be that the game's UI has changed drastically from last time resulting in unexpected behaviour.

[andreped]

BTW: Did you observe any mouse movements?

I launched the game to check, and I noticed that the icons of all animals seem to have changed. The UI also has been modified slightly. Hence, we might have to update the machine vision system to work with the new version of the game.

[andreped]

I was unable to debug this right now, as I only have a macbook and ran into the issue #54.

[Potajito]

Yeah, I'm leaning towards UI changes. I reverted animal icons to classic style, and that helped the script to see them, but I think it has troubles with the interface. No mouse movements over here. Gonna tinker a bit with the vision system on my side see what I can come up. CHeers!

[andreped]

Oh, it is possible to do that! Great!

I could see if it works on my old ass windows laptop. Hang on a bit.

If you fork the repo, you could adjust it as you'd like. If you manage to get it working, you could make a PR and I could try it on my end :]

[andreped]

I tried changing it but it still looks like this? I'm using the web app without being logged in. Did it work for you?

[Potajito]

Its a bit counter intuitive to change the animal graphics. You do it in the main menu, customize, pets, change all, classic.
Gonna try and fork it this weekend if I got time and play with it a bit. Also gonna try my hand at fixing the 1080p res requirement, should be that hard.

[andreped]

Its a bit counter intuitive to change the animal graphics. You do it in the main menu, customize, pets, change all, classic.

Yeah, I tried. Did not seem like it changed the animals, but it did change the UI in general, as far as I could see.

Also gonna try my hand at fixing the 1080p res requirement, should be that hard.

I made an issue regarding that which I could assign to you if you want, see issue #17. I made a new comment there regarding what you likely need to change to get this working (see here).

[andreped]

I managed to reproduce this issue. Note that you can increase verbosity by adding --verbose 2 to the run command.

From my verbose below, you can see this line INFO deploy_agent.py 111 CV SYSTEM [self.run]: The detected Pets and Food in the Shop is : ['apple'], which means that it was unable to find any pets. Hence, this would indicate that the templates of all animals needs updating.

Could you send me a screenshot how it looks before and after changing style of animals/UI?

This is my verbose:

(venv) PS C:\Users\andrp\workspace\super-ml-pets> python .\main.py --task deploy --infer_model .\models\model_sap_gym_sb3_180822_checkpoint_2175_steps --verbose 2
DEBUG main.py 41 Namespace(batch_size=512, finetune=None, gamma=0.99, infer_model='.\\models\\model_sap_gym_sb3_180822_checkpoint_2175_steps', infer_pversion='3.7', learning_rate=0.0003, model_name='rl_model', nb_games=10000, nb_retries=1, nb_steps=10000, save_freq=10000, task='deploy', verbose=2)
DEBUG __init__.py 277 matplotlib data path: C:\Users\andrp\workspace\super-ml-pets\venv\lib\site-packages\matplotlib\mpl-data
DEBUG __init__.py 277 CONFIGDIR=C:\Users\andrp\.matplotlib
DEBUG __init__.py 1447 interactive is False
DEBUG __init__.py 1448 platform is win32
[...]
DEBUG __init__.py 277 CACHEDIR=C:\Users\andrp\.matplotlib
DEBUG font_manager.py 1439 Using fontManager instance from C:\Users\andrp\.matplotlib\fontlist-v330.json
INFO main.py 53 Pausing...
INFO main.py 56 Running...
INFO deploy_agent.py 94 INITIALIZATION [self.run]: Loading Model
INFO deploy_agent.py 97 INITIALIZATION [self.run]: Create SuperAutoPetsEnv Object
INFO deploy_agent.py 104 CV SYSTEM [self.run]: Detect the Pets and Food in the Shop Slots
INFO deploy_agent.py 106 CV SYSTEM [self.run]: Calls [image_detection.find_the_animals]
INFO deploy_agent.py 111 CV SYSTEM [self.run]: The detected Pets and Food in the Shop is : ['apple']
INFO deploy_agent.py 112 GAME ENGINE [self.run]: Set Environment Shop = detected Pets and Food
INFO deploy_agent.py 121 GAME ENGINE [self.run]: Get the best action to make for the given state from the loaded model
Traceback (most recent call last):
  File ".\main.py", line 63, in <module>
    main()
  File ".\main.py", line 57, in main
    run(ret)
  File "C:\Users\andrp\workspace\super-ml-pets\src\deploy_agent.py", line 128, in run
    " Current Team and Shop \n{}".format(s[action][0]))
TypeError: unhashable type: 'numpy.ndarray'

[andreped]

A fix for this was added by @laughinggaschambers in dd96345.

See #55 (comment) for more information.