Package

A .cloacina package is the distributable unit: a package.toml manifest plus source. Once built, the same package runs under the embedded library, the daemon, or the server unchanged — it is the seam that lets you move a Workflow or Computation Graph between the two doors.

Mental model

• A package wraps your code with a package.toml and a conventional source layout.
• The loop is new → validate → pack → upload, via cloacinactl package.
• The server loads it (the reconciler registers its workflows, triggers, graphs).

Interfaces

The two languages package differently:

Python

A Python package is a module tree under workflow/, declaring tasks with bare @cloaca.task decorators (no WorkflowBuilder — the loader supplies the workflow context):

my-pipeline/
├── package.toml
└── workflow/
    └── my_pipeline/
        ├── __init__.py
        └── tasks.py        # bare @cloaca.task decorators

Rust

A Rust package ships source with a Cargo.toml + src/lib.rs; the server compiles it on load. Build/check locally with cloacinactl package build.

cloacinactl package new my-pipeline --lang python   # scaffold
cloacinactl package validate my-pipeline            # check format
cloacinactl package pack my-pipeline                # build the .cloacina archive
cloacinactl --tenant acme package upload my-pipeline-0.1.0.cloacina

Key facts

• Portable: one artifact runs embedded / daemon / server.
• Python: bare decorators only; a WorkflowBuilder inside a package fails to load.
• Rust: compiled on the server at load time.

See also

• Creating Your First Package · Packaging Python Workflows
• Run the Service — where packages are deployed.