@ocelotjungle/collections

This package contains 7 collections that are subtypes of a standard Set and Map, and one complex collection DataPond, which can be used to manage some micro-lakes of data, relating entities to each other.

Collections

ComplexSet<K - key, I - item>

This collection acts like a common Set except ComplexSet allows to use object values. It compares equality not of items, but of something returned by selector (keys). Actually, this is just a map with selector. Creating a ComplexSet with item type { foo: number, bar: string } and key type number. First argument is a selector, that selects foo from item type. In this case, foo is a key and it will refleсt equality of items, so if a.foo == b.foo, then a == b. If updateExisting is true, then adding an item with already existing key will replace the existing item.

new ComplexSet<number, { foo: number, bar: string }>(({ foo }) => foo, [updateExisting])
  .add({ foo: 42, bar: "question" })
  .add({ foo: 42, bar: "not question" }) // replaces previous if updateExisting
  .add({ foo: 100, bar: "IOO" })

EnsuredMap<K - key, V - value>

Does nothing except get returns X instead of X | undefined. No real checking, no safety, only types. Use it only when you are absolutely sure here is a value for key.

new EnsuredMap<number, string>([entries])
  .set(42, "question") // { 42 => "question" }
  .get(42) // "question"
  .get(43) // undefined as V

ArrayEnsuredMap<K - key, V - value>

The same as EnsuredMap, but stores V[]. Also allows to push/clear array by a key. Other array functions are not implemented.

new ArrayEnsuredMap<number, string>([entries])
  .set(42, []) // { 42 => [] }
  .push(42, "question") // { 42 => ["question"] }
  .clear(42) // { 42 => [] }
  .clear() // { }

StrongMap<K - key, V - value>

This collection is a map that guarantees an existence of required key by using a callback that creates some default V.

new StrongMap<number, string>(() => "default value", [entries])
  .set(42, "question") // { 42 => "question" }
  .get(42) // "question"
  .get(43) // "default value"

ArrayStrongMap<K - key, V - value>

The same as StrongMap, but stores V[] and default value is always an empty array. Also allows to push/clear array by a key. Other array functions are not implemented.

new ArrayStrongMap<number, string>([entries])
  .push(42, "foo") // { 42 => ["foo"] }
  .get(42) // ["foo"]
  .get(43) // []

ComplexKeyMap<K - key, V - value>

This collection is a Map that allows to correctly use complex objects. When you use get, it won't compare object's reference equality, but its structure. Also, likely StrongMap, guarantees that get returns a defined value. Requires passing default value callback.

new ComplexKeyMap<{ foo: number, bar: string }, boolean>(() => false, [entries])
  .set({ foo: 42, bar: "question" }, true) // { { foo: 42, bar: "question" } => true }
  .get({ foo: 42, bar: "question" }) // true
  .get(43) // false

ArrayComplexKeyMap<K - key, V - value>

The same as ComplexKeyMap, but stores V[] and default value is always an empty array. Also allows to push/clear array by a key. Other array functions are not implemented.

new ArrayComplexKeyMap<{ foo: number, bar: string }, boolean>([entries])
  .push({ foo: 42, bar: "question" }, true) // { { foo: 42, bar: "question" } => [true] }
  .get({ foo: 42, bar: "question" }) // [true]
  .get({ foo: 43, bar: "no value" }) // []

DataPond<P - pond scheme>

This collection is like a micro-database with preconfigured JOINs. It realizes a crude state machine on types for user's convenience (that makes impossible to correctly extend this class, preserving the state machine).

Creating

DataPond.create<P>(idExtractors)

• P is a scheme of data, it looks like { [name]: object }
• idExtractors is an object like { [name]: extractor }
• extractor is a selector function that accepts corresponding type object and returns some identifier

Example:

DataPond.create<{ foo: { bar: number, baz: Date } }>({ foo: ({ bar }) => bar })

• foo is a pond entity, it's type { bar: number, baz: Date }
• ({ bar }) => bar is an extractor, it selects bar from foo

Using

Add entities (add(name, entity))

Initially the pond has state 0 (next DataPond<0>), that means no cache. Adding one entity returns DataPond<1>, that means there's owner in the cache. Adding one more entity returns DataPond<2>, that means there're owner and item in the cache. Adding more entities won't affect state of the pond. The pond "remembers" only last 2 added entities.

Relate entities (relate(item, owner) or relate({ owner, item }))

owner and item can be either pointers to the entity ([entity_name, entity_id]) or arrays of pointers or even undefined, if applicable (depending on cache size).

DataPond<0>.relate:

• requires both owner and item;
• returns DataPond<0>;

DataPond<1>.relate:

• requires either:
  • passing item in relate(item, owner) and in this case owner is optional (if pass no owner, will be used owner from cache)
  • or passing owner or item in relate({ owner, item }) and in this case the second argument will be optional (will be used from cache);
• returns DataPond<1>, that allows add one entity and then call multiple relate in chain with different arguments;

DataPond<2>.relate:

• requires no arguments and if called without them, will use owner and item from cache;
• passing one or both arguments will lead to using them instead of cache;
• returns DataPond<0>;

Fetching entities (fetch(name, id) or fetch(name, id, array_of_names))

fetch(name, id) returns entity with corresponding name and id. fetch(name, id, array_of_names) acts like fetch(name, id), but augments the result with related entities, which names are in array_of_names: { _[name]: related_entity[] }

Examples

type Foo = { id: number, value: string };
type Bar = { name: string };
type Schema = { foo: Foo, bar: Bar };

// Creating
const dp = new DataPond<Schema>({
    foo: ({ id }) => id,
    bar: ({ name }) => name,
});

// Adding

dp.add("foo", { id: 42, value: "question" });
dp.add("bar", { name: "baz" });
dp.relate(["bar", "baz"], ["foo", 42]); // relating item to owner, foo-42 owns bar-baz

dp
    .add("foo", { id: 43, value: "not question" }) // state 1, owner in cache
    .relate(["bar", "baz"]) // relates item to cached owner, foo-43 also owns bar-baz

dp
    .add("foo", { id: 44, value: "idk" }) // state 1, owner in cache
    .add("bar", { name: "buzz" }) // state 2, owner and item in cache
    .relate() // relates cached item to cached owner, foo-44 owns bar-buzz

// Fetching

dp.fetch("foo", 42) // { id: 42, value: "question" }
dp.fetch("foo", 44, ["bar"]) // { id: 44, value: "idk", _bar: [{ name: "buzz" }] }