OuterProduct SDK

outerproduct-sdk is the sole public Python package for OuterProduct. It owns the Workspace API, runtime serialization, Files, Unity Catalog adaptation, Flight SQL, and UC-governed object storage. OuterProductClient is the only public client and owns all UC and non-UC operations for one automatically selected Workspace.

Natural workspace API

op.init() reads OUTERPRODUCT_API_KEY and OUTERPRODUCT_BASE_URL, ensures workspaces/default, and returns an already-scoped client. Pass workspace= to select another Workspace; callers never create or unwrap a second client.

import outerproduct_sdk as op

client = op.init()
catalogs = client.list_catalogs()

base_image = (await client.list_images())[0]
image = await client.build_image(
    base_image.name,
    display_name="analytics",
    description="Pinned analytics dependencies",
    pypi_dependencies=("numpy==2.3.2", "polars>=1.33"),
)
created = await client.create_compute_environment(
    image.name,
    cpu=1,
)

compute_environment = client.get_runtime_handle(created.name)


async def add(
    runtime: op.OuterProductClient,
    left: int,
    right: int,
) -> int:
    return left + right


async def add_twice(
    runtime: op.OuterProductClient,
    left: int,
    right: int,
) -> int:
    pending = await runtime.submit(add, left, right)
    first = await pending.refresh()
    return await runtime.compute(add, first.result(), right)


run = await compute_environment.submit(add_twice, 20, 22)
result = await compute_environment.compute(add, 20, 22)

await client.write_file("results/answer.txt", str(result).encode())
answer = await client.read_file("results/answer.txt")

Images are immutable registry artifacts. build_image() layers canonical PyPI requirements onto an existing workspace image and returns only after the new digest is published. ComputeEnvironments are immutable snapshots. Their create_compute_environment(), get_compute_environment(), and list_compute_environments() lifecycle methods live directly on the client. submit() returns after the scheduler acknowledges a durable computation. compute() submits and polls to a terminal state, yielding to asyncio between polls. Natural computed functions are async and receive their runtime workspace as the first positional parameter. Inside a computed function, runtime.submit() and runtime.compute() start child computations in the same compute environment. Their result types remain Run[T] and T, respectively.

Unity Catalog CRUD remains part of the flat Workspace API, including on the client injected into managed computations. The flattened methods retain the generated Unity Catalog signatures for static type checkers. Temporary table/volume/path/model credentials, Files, Flight SQL, compute_environments, and runs are flat Workspace methods too. JSON literals, objects implementing the SDK serialization contract, and Pydantic models can cross run boundaries.

Storage boundaries

File operations are flat methods on the workspace-scoped client. ComputeEnvironment volumes map Unity Catalog volumes to container paths and are materialized for each managed invocation; they are not a provider-native POSIX mount or a write-back filesystem.

Use store_from_url, store_from_volume, store_from_table, or store_from_path for refresh-aware UC-governed object stores. download_s3_prefix(client, prefix, target) securely streams such a prefix into a local directory with bounded concurrency.

Packaging

The published distribution is one wheel with one native extension. Internal serialization and UC object-store Python sources are vendored under outerproduct_sdk._vendor; the wheel has no dependency on separately published OuterProduct client packages. Third-party runtime dependencies remain ordinary wheel dependencies.