Recursive Weighted Average

A package for computing the weighted average of all leaf nodes of a tree with the following shape:

                  Root = [...]
                       ^
     +-----------------+----------------------+
     |                 |                      |
     |                 |                      |
     +                 +                      +
Ingredient1,1        Ingredient1,2       Ingredient1,3
{                    {                   {
  weight: 2,           weight: 5,          weight: 10
  score: 10            score: 3            ingredients: [...]
}                    }                   }                ^
                                   +----------------------+----+
                                   |                           |
                                   +                           +
                              Ingredient1,3,1             Ingredient1,3,2
                              {                           {
                                weight: 5,                  weight: 3,
                                ingredients: [...]          score: 8
                              }                ^          }
                           +-------------------+----+
                           |                        |
                           |                        |
                           +                        +
                      Ingredient1,3,1,1        Ingredient1,3,1,2
                      {                        {
                        weight: 0,               weight: 3,
                        score: 5                 score: 1
                      }                        }

There are also helper functions that may also be useful to flatten the tree and normalize the tree across heights.

Fairly Important Disclaimer

The package has working helper functions (that are actually what CarLabs needs currently)... It does not have a working function to compute... the weighted average. I bet I tricked you with this package name.

Installation

npm i --save @carlabs/recursive-weighted-average

Usage

  const rwa = require('@carlabs/recursive-weighted-average');

  const tree = [
  {
    weight: 1,
    trim_scores: [
      {
        score: 1
      },
      {
        score: 2
      }
    ]
  },
  {
    weight: 9,
    ingredients: [
      {
        weight: 10,
        trim_scores: [
          {
            score: 1
          },
          {
            score: 2
          }
        ]
      },
      {
        weight: 20,
        trim_scores: [
          {
            score: 1
          },
          {
            score: 2
          }
        ]
      }
    ]
  }
];

const treeWithNormalizedWeights = rwa.normalizeWeights(tree);
const flattenedAndNormalizedTree = rwa.flattenTree(treeWithNormalizedWeights);

Configuration

Currently, you can configure the names of the fields: weight, scores, children, as well as the getter function for the score:

  const myConfig = {
    node: {
      weight: 'myWeightFieldName',
      score: 'myScoreFieldName',
      children: 'myChildNodeName',
      getScore: x => x,
    },
  };

  // functions accept a second parameter for the config object
  const treeWithNormalizedWeights = rwa.normalizeWeights(tree, myConfig);

The defaults are set to:

  const defaultConfig = {
    node: {
      weight: 'weight',
      score: 'trim_scores',
      children: 'ingredients',
      getScore: x => x,
    }
  };

Development

Fairly standard Node setup: eslint, prettier... simple tests

To update the NPM package:

1. make changes
2. stage and commit changes
3. run npm version patch (or minor or major) - this will run tests, bump the version, and push the changes to GitHub