Limitations of Typedpy

Nothing is perfect, and Typedpy has several trade-offs and limitations:

Not Incorporated into IDE

Typedpy is not supported by the IDE code analysis, so there are no warnings/errors that are specific to Typedpy in the IDE.

Errors are found at run-time. There are workaround, like creating a hybrid Structure/dataclass, as explain in What About a Hybrid? ,

but it is still a limitation.

However, since version 2.14, Typedpy allows to easily generate Python Interface files. These files are read automatically

by Pycharm, and will result it correct intellisense.

Not Optimized for Speed

Typedpy is not optimized for speed, especially when dealing with immutables. If speed is absolutely crucial, you need to be aware of that.

Implicit Arbitrary Class Wrappers Are not Pickalable

Starting with version 2.1, Typedpy supports using arbitrary classes as field types, as described in Implicit Wrapping of Arbitrary Classes As Field.

However, if you use the implicit conversion feature in your structure, your data instance cannot be pickled.

Typedpy will raise an appropriate exception if you try to do it, so you will be protected from silent failures.

Type Hints Usage Can be Confusing

Typedpy Field types should not be used in type hints instead of the Python types they refer to.

The reason is that the Field class is not the same as the type of the content. For example,

Array[String] is not the same as list[str], and String is not the same as str.

To demonstrate, if we have the following structure:

class Example1(Structure):
    names_by_id: Map[String, Array[String]]
    ...

The following type hint for the return value is wrong:

def  get_names_for_id(id: str, example: Example) -> Array[String]:

The correct return type should be list[str]. However, if we want to keep the type hints consistent

with the Structure definition, we can define it as follows:

class Example1(Structure):
    names_by_id: dict[str, list[str]]
    ...

Typedpy automatically converts it to its own internal class:Field types, but you can use type hints as usual.

def  get_names_for_id(id: str, example: Example) -> list[str]:
     ...

Serialization Mappers

Combining functions transformations cannot be followed by other functions and key mapping.