import { IClusterModel, IDataset, IEmOptions } from '../types';
import * as MathUtils from '../utils/math';
import { MATRIX_SIZE_ERROR } from '../errors';
import * as Core from '../core/expmax_core';

/**
 * Expectation maximization using a gaussian mixture model
 * This class only holds the internal state of the
 * optimized gaussian mixture model (GMM) used for each clustering model.
 *
 * For instance:
 *
 * ```ts
 *  const dataset: IDataset = {
 *       'points': [
 *           [1.1,1],
 *           [1,2],
 *           [2,2],
 *           [2,1],
 *           [15,15],
 *           [15,16],
 *        ],
 *       'label':'test',
 *   };
 *   const opts: IEmOptions = {
 *       'clusterQt':2,
 *       'maxEpochs':1000,
 *       'threshold': 2e-16
 *   }
 *   const test = new ExpMax(dataset, opts);
 *   const chose = test.train()
 *   console.log(chose);
 * ```
 *
 * This program will output the likelihood of your data point belonging to each
 * distribution.
 *
 * Please mind that, in this class, the dataset is mutable for computation purpose.
 * Indeed, we don't want to fit a new, randomly generated GMM for
 * every slight change in our dataset.
 */
export default class ExpMax {

    /**
     * Clusters
     */
    private _clusters: IClusterModel[];

    /**
     * Data
     */
    private _dataset: IDataset;

    /**
     * Vector space dimension
     */
    private _vectorSpaceDim: number;

    /**
     * Options
     */
    private _opts: IEmOptions;

    constructor(dataset: IDataset, options: IEmOptions) {
        const {
            threshold = 2e-16,
            maxEpochs = 1000,
            clusterQt = 2,
        } = options;

        this._opts = { threshold, maxEpochs, clusterQt };
        this._dataset = dataset;
        this._vectorSpaceDim = dataset.points[0].length;
        this._clusters = Core.createRandomClusters(dataset, clusterQt, this._vectorSpaceDim);
    }

    /**
     * Update dataset
     * Mutates internal state
     * @param newDataset: IDataset
     * @returns IClusterModel[] | error
     */
    update(newDataset:IDataset) {
        this._dataset = newDataset;
        this.train();
        return this._clusters;
    }

    /**
     * Em Algorithm
     * @returns IClusterModel[]
     */
    train(): IClusterModel[] {
        this._clusters = Core.train(this._clusters, this._dataset, this._opts);
        return this._clusters;
    }

}