For Consideration

[ntjohnson1]

This mostly just shuffles code around but the result is a big headache to review. I wanted to propose a concrete example in case you think it makes any of the documentation stuff nicer. If it looks like more work than it is worth then we can just close.

If the initial POC looks nice then we can figure out how to resolve GCP.

Pros:

• Eliminates most of the ambiguity around file name shadowing class names
• code is a little more logically organized
• everywhere but gcp the flattening already works and keeps standard user facing interface the same

Cons:

• For users directly pointing to the source files this is definitely a breaking change
• need to dig into gcp to see if I can preform some flattening there for either a no change or a less ugly one
• decomposition is quite long, we could probably just remove that level and just straight to pyttb.cp vs pyttb.tucker

---

📚 Documentation preview 📚: https://pyttb--453.org.readthedocs.build/en/453/