Kanit Wongsuphasawat

- [x] Make build a method under cql.schema.build (rather than a static method cql.schema.Schema.build() - [ ] Currently build calculates all fields at the same time. If we can calculate...

For example, we know min and max are not correct. Maybe other stats are not correct too.

Allow supplying schema and statistics as a part of the json

1

Sometimes default type inference (from datalib) might be incorrect. - [x] We should add `schema` property to `Query` as an object that takes key-value mapping (from `field: string` to a...

- [ ] correlation (spearman / pearson) - [ ] mutual information

- [ ] Extract and test `hasRequiredPropertyAsEnumSpec` in `satisfy` of `EncodingConstraintModel` and `SpecConstraintModel` ## Replicating Compass ### Gen - [ ] aggregate.test.ts - [ ] encodings.test.ts Run `npm run cover`...

- [x] ~~EnumSpecIndexStructure is a mess `IndexTuple` class looks confusing.~~ Done in #161 - [ ] Currently we generate properties by property type rather than generate properties for one encoding...

Reduce redundancy for checking type of properties

1

Spec constraint's `satisfy` can become more effective when `requireAllProperties` is `true`. Currently we repeatedly checking if we have all the requires properties.

(Inspired from Wills & Wilkinson's AutoVis) > We iteratively apply Tukey’s ladder of powers1,17 in order to make 1D distributions as symmetric as possible.