cf-basic-signal 0.3.3

Cogniflow signal processing step package (FIFO windowing and averages) implemented in C++.

cf.basic.signal StepPackage

Signal-processing steps packaged for Cogniflow (FIFO buffer and statistical averages).

Note: Dev/demo/benchmark steps (InlineSource, OpcuaPhSource, OmpReductionStep, BudgetAwareStep) have moved to cf-basic-dev. This package now contains only core signal-processing steps.

Steps

Step	Description
cfbs:FifoWindowBufferStep	Sliding FIFO window buffer
cfbs:AverageStep	Arithmetic / geometric / harmonic mean with standard error

Files

• steps.nq - single manifest (package metadata + ontology + implementation bindings)
• src/cf_basic_signal/cpp/steps.cpp - C++ implementations (plugin backend)
• examples/golden_fifo_avg.json - minimal signal pipeline (InlineSource from cf-basic-dev -> FIFO -> Average) (legacy)

Pipeline backend plugin

The package registers a C++ plugin artifact described in steps.nq:

• C++ ops (impl="cpp"): fifo_window_buffer, average (registered via capsule register_cpp_op)
• No Python fallbacks; the C++ extension is required for pipeline_backend usage

Plugin metadata (artifact path, list/resolve symbols, certification) is stored under the cf:PluginArtifact node in steps.nq.

Quick check

1. Install package (editable or wheel) so the cogniflow.steps entry point exposes steps.nq.

2. Ingest step packages via cf_ontology and verify:

import os
from cf_ontology import ingest_installed_step_packages, OntologyManager

os.environ["CF_ENABLE_STEP_PACKAGES"] = "1"
ingest_installed_step_packages()
om = OntologyManager()
assert om.get_step_info("cfbs:FifoWindowBufferStep")
assert om.get_step_info("cfbs:AverageStep")

3. (Legacy) Run the example pipeline via your runner. Expected output:
   • golden_fifo_avg: window [4,5,6], mean 5.0, small SE.

Published distribution name:

pip install cf-basic-signal

Native build prerequisites

cf_basic_signal is built with scikit-build-core and CMake and requires the published SDK package at build time:

• CPython 3.14
• CMake on PATH
• a Windows C++ toolchain compatible with that CMake installation
• access to cf-pipeline-sdk from the package index

The package-local CMake currently generates the signature header through python -m cogniflow_pipeline_sdk.siggen and copies the built plugin artifact into the installed package's bin/ directory.

Publishing

cf_basic_signal is published with the dedicated Windows workflow:

• Workflow: .github/workflows/cf_basic_signal_windows_publish.yml
• Package directory: sandcastle/cf_basic_steps/cf_basic_signal
• PyPI tag: cf-basic-signal-v<version>
• TestPyPI tag: cf-basic-signal-v<version>-test

Local preflight:

powershell -ExecutionPolicy Bypass -File scripts/mimic_windows_python_publish_workflow.ps1 `
  -WorkflowFile .github/workflows/cf_basic_signal_windows_publish.yml `
  -PackageDir sandcastle/cf_basic_steps/cf_basic_signal `
  -PythonExe py `
  -PythonVersion 3.14

Queue a dry-run dispatch:

powershell -ExecutionPolicy Bypass -File scripts/queue_windows_python_publish_workflow.ps1 `
  -WorkflowFile .github/workflows/cf_basic_signal_windows_publish.yml `
  -PackageDir sandcastle/cf_basic_steps/cf_basic_signal `
  -PublishTarget testpypi `
  -Ref main `
  -RequireLocalPass `
  -DryRun

Watch queued runs:

gh run list --workflow cf_basic_signal_windows_publish.yml --limit 10
gh run watch --exit-status