meet error when run the command "luxai-s3 main.py main.py --output=replay.html"

[plantonic]

The error message below was shown when I try to run two simple agent to test the installation of the code. Anyone met similar problems?

(venv) PS D:\study\Kaggle-Lux-AI-s3\kits\python> luxai-s3 main.py main.py --output=replay.html player_0, main.py: Traceback (most recent call last): player_0, main.py:   File "D:\study\Kaggle-Lux-AI-s3\kits\python\main.py", line 6, in player_0, main.py:     import numpy as np player_0, main.py: ModuleNotFoundError: No module named 'numpy' player_1, main.py: Traceback (most recent call last): player_1, main.py:   File "D:\study\Kaggle-Lux-AI-s3\kits\python\main.py", line 6, in player_1, main.py:     import numpy as np player_1, main.py: ModuleNotFoundError: No module named 'numpy' player_0, main.py: cannot parse action '' player_1, main.py: cannot parse action '' Traceback (most recent call last):   File "D:\study\Kaggle-Lux-AI-s3\venv\Lib\site-packages\jax_src\pjit.py", line 616, in _infer_params_impl     avals.append(shaped_abstractify(a))                  ^^^^^^^^^^^^^^^^^^^^^   File "D:\study\Kaggle-Lux-AI-s3\venv\Lib\site-packages\jax_src\api_util.py", line 606, in shaped_abstractify     return handler(x) if handler is not None else _shaped_abstractify_slow(x)            ^^^^^^^^^^   File "D:\study\Kaggle-Lux-AI-s3\venv\Lib\site-packages\jax_src\api_util.py", line 617, in _numpy_array_abstractify     dtypes.check_valid_dtype(dtype)   File "D:\study\Kaggle-Lux-AI-s3\venv\Lib\site-packages\jax_src\dtypes.py", line 739, in check_valid_dtype     raise TypeError(f"Dtype {dtype} is not a valid JAX array " TypeError: Dtype object is not a valid JAX array type. Only arrays of numeric types are supported by JAX.   The above exception was the direct cause of the following exception:   Traceback (most recent call last):   File "D:\study\Kaggle-Lux-AI-s3\venv\Scripts\luxai-s3-script.py", line 33, in     sys.exit(load_entry_point('luxai-s3', 'console_scripts', 'luxai-s3')())              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^   File "d:\study\kaggle-lux-ai-s3\src\luxai_runner\cli.py", line 129, in main     results = asyncio.run(eps.run())     return self._loop.run_until_complete(task)            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^   File "D:\apps\python3.11\Lib\asyncio\base_events.py", line 653, in run_until_complete     return future.result()            ^^^^^^^^^^^^^^^   File "d:\study\kaggle-lux-ai-s3\src\luxai_runner\episode.py", line 174, in run     new_state_obs, rewards, terminations, truncations, infos = self.env.step(                                                                ^^^^^^^^^^^^^^   File "d:\study\kaggle-lux-ai-s3\src\luxai_s3\wrappers.py", line 137, in step     obs, reward, terminated, truncated, info = self.env.step(action)                                                ^^^^^^^^^^^^^^^^^^^^^   File "d:\study\kaggle-lux-ai-s3\src\luxai_s3\wrappers.py", line 87, in step     obs, self.state, reward, terminated, truncated, info = self.jax_env.step(                                                            ^^^^^^^^^^^^^^^^^^ TypeError: Error interpreting argument to <function LuxAIS3Env.step at 0x000001AAF7F7DA80> as an abstract array. The problematic value is of type <class 'numpy.ndarray'> and was passed to the function at path action['player_0']. This typically means that a jit-wrapped function was called with a non-array argument, and this argument was not marked as static using the static_argnums or static_argnames parameters of jax.jit.     -------------------- For simplicity, JAX has removed its internal frames from the traceback of the following exception. Set JAX_TRACEBACK_FILTERING=off to include these.

[RonnyGN]

Yes, the same exact issue has occured with me too. I have struck a huge wall and I don't know what to do next.

[StoneT2000]

It looks like you don't have numpy installed? how did you install the environment.

[AbdiHaryadi]

I think luxai-s3 doesn't support virtual environment dependencies. I have to install numpy outside of it to solve this problem.

[KirillTushin]

What helps me:
Pass the path to the virtual environment before starting the program:

export PYTHONPATH=$(python -c "import sys; print(':'.join(sys.path))")

For windows:

$env:PYTHONPATH = (python -c "import sys; print(';'.join(sys.path))")