Skip to content

Latest commit

 

History

History
51 lines (35 loc) · 2.56 KB

README.md

File metadata and controls

51 lines (35 loc) · 2.56 KB
Raw

Transformer

Build Status

Library that takes a set of start values, assumption and turns them into attributes useable for the ETM with help of the graph.

How Transfomer creates a dataset

The typical entrypoint to the library is through DatasetGenerator. This is initialized with a hash containing all the keys and values exported by ETLocal for a dataset. The generator creates a DatasetCast which describes the attributes for the dataset itself (stored in the Active Document file), such as the area key, parent dataset key, number of residences, etc.

Generator

The generator then runs through two steps to create or update a dataset in ETSource:

  1. DatasetGenerator::Scaler – Uses Atlas to create new scaled dataset. This is based on the "base_dataset" param provided when the generator was initialized. At this point, ETSource now contains the new dataset, with an ActiveDocument file, and numerious directories symlinked to the parent dataset (curves, demands, network, etc).

  2. DatasetGenerator::Analyzer – Responsible for taking the parameters with which the generator was initialized and updates the dataset; either by making changes to the Active Document file, or by preparing values to be written to the graph_values.yml file.

Analyzer

The analyzer itself runs through several steps to determine what changes must be made to the dataset.

  1. Caster.cast converts the params provided to the generator to a more structured form; the params are split into sub-hashes, depending on where the value will be persisted (for example, "area" contains values stored in the Active Document file).

    To create this structure, Caster creates a Template which allows setting area and graph values. Then it:

    1. Runs Caster::Exporter. Exporter takes the params provided by ETLocal and maps their keys and values to attributes stored in the graph_values.yml file. Any attribute which have a related "sparse graph query" have this query executed, with the resulting value added to the template's graph values.

    2. Runs AreaAttributes which takes the ETLocal params and extracts values which should be stored in the dataset Active Document. As in step 1, any attributes which have a query will execute that query and store the value in the template.

  2. Values for the dataset Active Document file are updated with those from the Caster.

  3. Values for the graph_values.yml file are set with those from the Caster.