Writing Custom Flyte Types

Flyte is a strongly-typed framework for authoring tasks and workflows. But there are situations when the existing set of types do not directly work. In fact, this is true with any programming language!

Just like a programming language enabling higher-level concepts to describe user-specific objects - like classes in Python/Java/C++, struct in C/Golang, etc., Flytekit allows modeling user classes. The idea is to make an interface that is more productive for the use case, but write a transformer that transforms the user-defined type to one of the generic constructs in Flyte’s type system.

This example will try to model an example user-defined dataset and show how it can be integrated seamlessly with Flytekit’s typing engine.

The video below will walk you through the example.

First, we import the dependencies.

import os
import tempfile
import typing
from typing import Type

from flytekit import (
    Blob,
    BlobMetadata,
    BlobType,
    FlyteContext,
    Literal,
    LiteralType,
    Scalar,
    task,
    workflow,
)
from flytekit.extend import TypeEngine, TypeTransformer

Note

FlyteContext is used to access a random local directory.

Defined type here represents a list of files on the disk. We will refer to it as MyDataset.

class MyDataset(object):
    """
    MyDataset is a collection of files. In Flyte, this maps to a multi-part blob or directory
    """

    def __init__(self, base_dir: str = None):
        if base_dir is None:
            self._tmp_dir = tempfile.TemporaryDirectory()
            self._base_dir = self._tmp_dir.name
            self._files = []
        else:
            self._base_dir = base_dir
            files = os.listdir(base_dir)
            self._files = [os.path.join(base_dir, f) for f in files]

    @property
    def base_dir(self) -> str:
        return self._base_dir

    @property
    def files(self) -> typing.List[str]:
        return self._files

    def new_file(self, name: str) -> str:
        new_file = os.path.join(self._base_dir, name)
        self._files.append(new_file)
        return new_file

MyDataset represents a set of files locally, but when a workflow consists of multiple steps, we want the data to flow between the different steps. To achieve this, it is necessary to explain how the data will be transformed to Flyte’s remote references. To do this, we create a new instance of TypeTransformer, for the type MyDataset as follows:

Note

The TypeTransformer is a Generic abstract base class. The Generic type argument refers to the actual object that we want to work with. In this case, it is the MyDataset object.

class MyDatasetTransformer(TypeTransformer[MyDataset]):
    _TYPE_INFO = BlobType(
        format="binary", dimensionality=BlobType.BlobDimensionality.MULTIPART
    )

    def __init__(self):
        super(MyDatasetTransformer, self).__init__(
            name="mydataset-transform", t=MyDataset
        )

    def get_literal_type(self, t: Type[MyDataset]) -> LiteralType:
        """
        This is useful to tell the Flytekit type system that ``MyDataset`` actually refers to what corresponding type.
        In this example, we say its of format binary (do not try to introspect) and there is more than one file in it.
        """
        return LiteralType(blob=self._TYPE_INFO)

    def to_literal(
        self,
        ctx: FlyteContext,
        python_val: MyDataset,
        python_type: Type[MyDataset],
        expected: LiteralType,
    ) -> Literal:
        """
        This method is used to convert from the given python type object ``MyDataset`` to the Literal representation.
        """
        # Step 1: let's upload all the data into a remote place recommended by Flyte
        remote_dir = ctx.file_access.get_random_remote_directory()
        ctx.file_access.upload_directory(python_val.base_dir, remote_dir)
        # Step 2: let's return a pointer to this remote_dir in the form of a Literal
        return Literal(
            scalar=Scalar(
                blob=Blob(uri=remote_dir, metadata=BlobMetadata(type=self._TYPE_INFO))
            )
        )

    def to_python_value(
        self, ctx: FlyteContext, lv: Literal, expected_python_type: Type[MyDataset]
    ) -> MyDataset:
        """
        In this method, we want to be able to re-hydrate the custom object from Flyte Literal value.
        """
        # Step 1: let's download remote data locally
        local_dir = ctx.file_access.get_random_local_directory()
        ctx.file_access.download_directory(lv.scalar.blob.uri, local_dir)
        # Step 2: create the ``MyDataset`` object
        return MyDataset(base_dir=local_dir)

Before we can use MyDataset in our tasks, we need to let Flytekit know that MyDataset should be considered as a valid type. This is done using TypeEngine’s register method.

TypeEngine.register(MyDatasetTransformer())

The new type should be ready to use! Let us write an example generator and consumer for this new datatype.

@task
def generate() -> MyDataset:
    d = MyDataset()
    for i in range(3):
        fp = d.new_file(f"x{i}")
        with open(fp, "w") as f:
            f.write(f"Contents of file{i}")

    return d


@task
def consume(d: MyDataset) -> str:
    s = ""
    for f in d.files:
        with open(f) as fp:
            s += fp.read()
            s += "\n"
    return s


@workflow
def wf() -> str:
    return consume(d=generate())

We can run this workflow locally and test it. Even when you run it locally, Flytekit will exercise the entire path.

if __name__ == "__main__":
    print(wf())

Out:

Contents of file1
Contents of file0
Contents of file2