Difference between array.py and m_array.py

[ttzytt]

I cloned the GitHub repository and tried to run a few pieces of examples from the documentation:

class MyScene(Scene):
    def construct(self):
        arr = MArray(self, [1, 2, 3])
        self.play(Create(arr))
        self.wait(1)

An error just occurred:

│ E:\pythons\3.11\Lib\site-packages\manim_data_structures\array.py:470 in __init__                 │
│                                                                                                  │
│   467 │   │                                                                                      │
│   468 │   │   print("arr ", arr)                                                                 │
│   469 │   │                                                                                      │
│ ❱ 470 │   │   for v in arr:                                                                      │
│   471 │   │   │   self.__append_elem(v, mob_square_args, mob_value_args, mob_index_args)         │
│   472 │                                                                                          │
│   473 │   def update_elem_value(self, index: int, value, mob_value_args: dict = {}) -> Text:     │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: 'MyScene' object is not iterable

It seems like the package I downloaded from pip is different from what I cloned from GitHub. The parameter of __init__ function for m_array.py (from the Github repo), and array.py is different. One of them uses scene as a parameter but the other does not.

After deleting the parameter of self, everything worked correctly.

So I'm wondering what caused this inconsistency. Is it because the package from pip is not updated? Or just because I'm not using this correctly?

[drageelr]

Hello @ttzytt
From version 0.1.6 onward, the Scene needs to be passed as the first argument to MArray and other such classes. The package is published to PyPi using the PYPI workflow defined on the main branch, so imo the repo's main branch is consistent with the latest package published.

Moreover, the file array.py was renamed to m_array.py in the release 0.1.5 (as mentioned in the changelog). Since you mentioned that these two files differ I think that you haven't updated the main branch for your clone.

Try pulling the latest commits from the main branch and let me know if the error still persists.

[drageelr]

@ttzytt Just re-read your comment. I think your repo is updated but the pip package is not, kindly update it to at least 0.1.6 and the error would be resolved.

[ttzytt]

@drageelr Thanks a lot. Now I can run the code correctly by using sys.path.append() to add the path of the cloned repo. Since the repo is updated, so I directly used it without the pip package.

[drageelr]

You can also run poetry shell while in the repo's directory and execute pip install . to install the package from the repo. However, make sure to invoke manim via poetry shell.