HashSet and Dictionary in TypeScript: export of Interface - dictionary

Question

I think I need some feedback on my collection classes - still learning typescript and javascript and these implementations can surely be improved. I am looking forward to any suggestion. I think I do use the generic types in a useful way, any advice here would be appreciated.
The answer I am looking for most is removing the duplicate IHashTable definition from the end of both snippets and moving it to its own file, I cannot get that done it seems. I am even unsure if this IS an interface in the first place. It compiles and works this way, as far as I can see.
The collection types are incomplete and only define the basic most function at the moment. Once I am sure I use the language and its features correct the other functions should not be too difficult.
Here is my HashSet:
import { IHashable } from "./IHashable"
export class HashSet<T extends IHashable> {
private _items: HashTable<T>;
public constructor() {
this._items = {};
}
public Add(key: T): void {
let str: string = key.GetHash();
if (this._items[str] == null) {
this._items[str] = key;
}
else {
throw new RangeError("Key '" + str + "' already exists.");
}
}
public Contains(key: T): boolean {
let str: string = key.GetHash();
return this._items[str] != null;
}
}
interface HashTable<T> {
[key: string]: T;
}
I wonder if I can avoid the checking-before-adding in a way. The javascript-dictionary this relies on does allow duplicates, so to avoid them there is no other way than to check myself?
This is my Dictionary:
import { IHashable } from "./IHashable"
export class Dictionary<T1 extends IHashable, T2> {
private _items: HashTable<KeyValuePair<T1, T2>>;
public constructor() {
this._items = {};
}
public Add(key: T1, value: T2) {
let str: string = key.GetHash();
if (this._items[str] == null) {
let kvp: KeyValuePair<T1, T2> = new KeyValuePair(key, value);
this._items[str] = kvp;
}
else {
throw new RangeError("Key '" + str + "' already exists.");
}
}
public ContainsKey(key: T1): boolean {
let str: string = key.GetHash();
return this._items[str] != null;
}
public Get(key: T1): T2 {
let str: string = key.GetHash();
let kvp: KeyValuePair<T1, T2> = this._items[str];
if (kvp == null) throw new RangeError("Key '" + str + "' not found")
return kvp.Value;
}
}
export class KeyValuePair<T1 extends IHashable, T2> {
private _key: T1;
private _value: T2;
public get Key(): T1 { return this._key; }
public get Value(): T2 { return this._value; }
public constructor(key: T1, value: T2) {
this._key = key;
this._value = value;
}
}
interface HashTable<T> {
[key: string]: T;
}
Both rely on a definition of IHashable (hashABLE and hashTABLE: I should find other names.)
export interface IHashable {
GetHash(): string;
}
The dictionary looks a bit strange, it "wraps" my dictionary into a new type KeyValuePair and then uses this in the javascript dictionary. What I hope to gain by doing this is get my own type for key, in and out, as long as it offers a string by which it can be indexed. - No idea if that makes sense or is completly wrong.
What I am missing is the count of items in the collection, a way to remove items, and a way to iterate over the keys and the values.
Regarding iterating over I will post another question with my implementation of a list and a ForEach over it, hoping iterating the keys or values might be possible in the same way.
Probably the most important question I forgot here: How could the GetHash-Method be build for an own class? I was going to have a static number on my classes, and count up by 1 in the constructor before assign this number to each instance. This would guarantee uniqueness... is there something better?
Thanks for any tip!
Ralf