Nested conversion

What happens at construction

When you build an AttributeDict, every contained mapping is recursively converted into an AttributeDict — including mappings inside lists and tuples :

from attributedict import AttributeDict

x = AttributeDict({
 "database": {
 "host": "localhost",
 "ports": [5432, {"tls": True}],
 }
})

x.database.host # "localhost"
x.database.ports[0] # 5432
x.database.ports[1].tls # True

What is NOT converted

• set / frozenset are left untouched.
• Any non-container value is left as-is (identity preserved).
• An already-converted AttributeDict passed as a value is shared (idempotent conversion — shallow-copy semantics).

Cycle safety

Conversion is cycle-safe : a mapping that is being converted is reused rather than re-converted, so self-referential structures terminate and the cycle is preserved:

inner = {}
outer = {"self": inner}
inner["back"] = outer

d = AttributeDict(outer)
d.self.back is d # True — the cycle is preserved

Cost

Construction is O(n) in the number of contained items. See performance for measured data.