Score store should allow multiple assessment types

[richford]

• Change scores getter to scoresByTask and make it nested by task
• Create an allScores getter that is the union of scoresByTask
• Create a helper function getTaskSpecificScores that takes in a single row of scores (i.e., a run) and generates a ROAR score, normed scores, comparison scores, etc. that are specific to each task. Something like

const getTaskSpecificScores = (run) => {
    if (run.taskId === 'swr') {
        return processSWRRun(run);
    } else if (run.taskId === 'pa') {
        return processPARun(run);
    } // etc.
}

and then, also in the helper file, we have to create the aforesaid functions processSWRRun and processPARun. This also requires that processXXXRun conforms to a specific API: input is a run and output is an object like

{
    roarScore: Number,
    subScores: {
        // These subtask IDs will differ in between tasks
        fsm: Number,
        lsm: Number,
        del: Number,
    },
    comparisonPercentile: Number,
    comparisonType: String (e.g. "Woodcock Johnson"),
    normedPercentile: Number,
    classifications: {
        // Here each field can differ depending on the task
        support: String,
        automaticity: String,
        fsm: String,
        lsm: String,
    }
}

[richford]

This is a meta-issue. We'll have to open up individual issues for each of the processABCRun functions.