Python

Create a module

fluentci init --sdk python
# Or add to an existing workspace:
fluentci modules add ci --sdk python

The SDK starter creates its own source and build configuration. The following sample belongs in plugin.py. It prints a greeting, then inspects the current Git workspace; run it from a Git repository.

Pipeline example

import extism
from fluentci import pipeline


@extism.plugin_fn
def hello() -> None:
    output = (
        pipeline("hello")
        .exec("echo", "Hello from FluentCI")
        .exec("git", "status", "--short")
        .stdout()
    )
    extism.output_str(output)

Chain .exec() calls; .stdout() runs the pipeline and returns its output.

Toolchain and build

Install uv and the Extism Python/Binaryen tools required by the SDK (extism-py, wasm-merge, and wasm-opt). The starter's pyproject.toml points at a local .sdk/python package and requires Python 3.11 or newer.

The default CLI build synchronizes dependencies and invokes the fluentci-plugin compiler for the selected entry, producing dist/plugin.wasm. A registered existing module can specify an entry; the default is plugin.py.

Entry and execution

The small generated starter uses @extism.plugin_fn to export hello and writes a greeting with extism.output_str. The pipeline example above expands that job with host commands. Builder chaining describes the pipeline; stdout() performs execution and returns output.

The guest environment is not a general host CPython installation. Use SDK host operations for supported filesystem/process capabilities rather than assuming arbitrary native extensions and host Python packages are embedded in Wasm.

Run and inspect

fluentci run main hello
fluentci modules main --json

Use the configured module name instead of main if it was added as ci. CLI invocation builds the source module before running it. The resulting artifact is normally dist/plugin.wasm.

Continue with workspace modules for custom builds and artifact paths, or publishing to share the compiled plugin.