I have a specific use-case where I'm saving my objects to disk (using @dataclasses_json.to_json()). This works well except for FlyteFile - in the case of FlyteFiles it saves the

path

(/tmp/easfdasdfas) and not the

_remote_source

(gs://flyte-asdfdasfd/asdfa). Is there a Flyte way to serialize objects to json and storing the

_remote_source

for FlyteFile?

I think my might not be the best way but you can use

thanks!! my object is a bit nested though (i.e. my FlyteFiles are a couple of levels down). Is there a way to do this automatically? (Perhaps I can use Flyte's serialization method, somehow?)

I'm working on a PR to customize FlyteFile's serialization behaviour in Flyte

I had this same problem, solved it like:

class FlyteFileSerialize(SerializationStrategy):
    """Class to serialize FlyteFile and maintain the remote path or source."""

    def serialize(self, v: FlyteFile) -> dict[str, str]:
        return {"path": str(v.remote_path or v.remote_source or v.path)}

    def deserialize(self, v: dict[str, str]) -> FlyteFile:
        return FlyteFile(v["path"])

@dataclass
class ExampleDataclass:
    class Config:
        """Config defining serialization for mashumaro."""

        serialization_strategy: ClassVar = {
            FlyteFile: FlyteFileSerialize(),
        }

You can checkout my PR here

nice, I like the

SerializableType

abstraction so we can extend to other non-mashumaro supported types!

Will merge it tonight

sorry, for the late reply, but appreciate it! In the end, I went for a slightly different solution by basically storing all longer strings in a file, instead of serializing whole objects.

Oh this is a nice solution too!