ETL upon Time Series Property Values

Ask the Community
,
1

I am having difficulty working with Time Series Properties in either a Python or TypeScript FOOs Repository environment, where I seek too:

  1. Retrieve all points from the property for further use
  2. Apply transformation to the datatype returned in (1) and restructure to interface with Workshop on the front-end.

I currently have a ontology object that possesses a tsp attribute, in which I seek to retrieve all of the key:value pairs visible in the associated Time Series Sync.

From here, I want to apply additional computation to generate ACF & PACF values, then transformation this derived data into the TwoDimesionalAggregation datatype that the Chary XY widget in Workshop can display visually.

I am in communication with my enrollments client support team, whom have recommended I make use of the OSDK & Developer Console. But unfortunately the later is currently disabled on the Foundry enrollment I am working within.

I have achieved minimally what I desire in a TypeScript FOO Repo, by way of synthetic value generation. However the specific computation component here pushed me towards Python.

@Function()
public async generateSyntethicACF(): Promise<TwoDimensionalAggregation<string>> {
    const buckets = Array.from({ length: 52 }, (_, i) => ({
        key: (i + 1).toString(),
        value: Math.ransom() * 2 - 1
    }));
    return { buckets: buckets };

Looking for guidance from anyone who has achieved something similar on platform.

2

While I haven’t done this exact kind of time series manipulation, I might suggest that you look at working with the FoundryTS python library (full API docs here) to derive additional ACF and PACF time series in your pipeline for the relevant data, rather than calculating it “on the fly” in a Typescript or Python function.

Someone else might have a better approach - mostly just wanted to make sure you got the pointer to the FoundryTS library info.