Solution 3: Custom Clone Rules

Completed idea

deepcopy(self) clones everything uniformly. But sometimes clone rules are not uniform.

The solution is to build the clone manually, applying different copy strategies per field.

---

Solution

def clone(self, **changes) -> "Dashboard":
    cloned = Dashboard(
        title=self.title,
        owner=self.owner,
        layout=self.layout,
        data_source=self.data_source,
        widgets=deepcopy(self.widgets),
        filters=deepcopy(self.filters),
        refresh_minutes=self.refresh_minutes,
        cache={},
    )

    for field_name, value in changes.items():
        if not hasattr(cloned, field_name):
            raise AttributeError(f"Dashboard has no field: {field_name}")
        setattr(cloned, field_name, value)

    return cloned

---

The three clone strategies

Independent copy   -> widgets, filters
    deepcopy(self.widgets), deepcopy(self.filters)
    Each clone owns its own widgets and filter state.

Shared reference   -> data_source
    data_source=self.data_source
    All clones point to the same DataSource.
    This is intentional: the source is shared configuration.

Reset              -> cache
    cache={}
    Each clone starts with an empty cache.
    Inheriting old cached results would be wrong.

---

Why not deepcopy(self) here

deepcopy(self) would also deep-copy data_source.

That would give each clone its own separate DataSource, breaking the sharing invariant.

It would also copy the original’s cache contents into the clone, which should start fresh.

When different fields have different copy semantics, a manual clone is the right tool.

---

Back to exercise · Solution script