Skip to content

Pipelines¤

End-to-end workflows that turn raw Azure timeseries into actionable production insights. Each pipeline starts with just three inputs:

  1. Azure connection config (connection string, SAS URL, or AAD credentials)
  2. UUID list (the signal identifiers for your use case)
  3. Time range (start and end timestamps)

Common Pattern¤

Every pipeline is a single, reusable Pipeline object: data is loaded once, then .transform steps clean the signal and .detect steps branch off KPI tables.

graph LR
    A[Azure Blob Storage] --> B[Load by UUIDs + Time Range]
    B --> C["Pipeline — .transform steps clean the signal"]
    C --> D["Pipeline — .detect steps branch off analytics"]
    D --> E["PipelineResult — .data + .events"]

See the Pipeline guide for the step types, sentinels, and debugging tools every pipeline below uses.

Available Pipelines¤

  • OEE Dashboard

    Machine state, part counters, and reject signals into daily OEE breakdown by shift with availability, performance, and quality components.

    Signals: 4 UUIDs

  • Cycle Time Analysis

    Cycle triggers and part numbers into cycle time statistics, slow cycle detection, trend analysis, and golden cycle comparison.

    Signals: 3 UUIDs

  • Downtime Pareto

    Machine state and reason codes into Pareto analysis, shift-level downtime comparison, and availability trends.

    Signals: 2 UUIDs

  • Quality & SPC

    Measurement signals with tolerances into outlier detection, SPC rule checks, control charts, and Cp/Cpk capability trending.

    Signals: 1+ measurement UUIDs

  • Process Engineering

    Setpoint, actual value, and process state signals into setpoint adherence, startup detection, control loop health, and stability scores.

    Signals: 3 UUIDs

  • Pipeline

    Chain transforms, segmentation, feature computation, and detectors into a single reusable Pipeline. From raw timeseries to ML-ready feature tables and event logs.

    Signals: N process parameters + 1 order signal

Prerequisites¤

All pipelines require:

pip install ts-shape
pip install azure-storage-blob   # for Azure loaders

For detailed module documentation, see the API Reference or the Guides.