@sepiariver/unique-set 2.0.2 → 2.0.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +1 -1
- package/dist/index.mjs +46 -50
- package/index.ts +61 -63
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
|
|
3
3
|
Extends the native `Set` class to deeply compare using [fast-deep-equal](https://www.npmjs.com/package/fast-deep-equal), with optional Bloom filter optimization.
|
|
4
4
|
|
|
5
|
-
Supports ESM and CommonJS. Thanks @sakgoyal for contributing to and instigating ESM support.
|
|
5
|
+
Supports ESM and CommonJS. Thanks [@sakgoyal](https://github.com/sakgoyal) for contributing to and instigating ESM support.
|
|
6
6
|
|
|
7
7
|
```js
|
|
8
8
|
import { BloomSet, UniqueSet } from '@sepiariver/unique-set';
|
package/dist/index.mjs
CHANGED
|
@@ -1,5 +1,48 @@
|
|
|
1
1
|
// index.ts
|
|
2
2
|
import equal from "fast-deep-equal/es6/index.js";
|
|
3
|
+
var serialize = (item) => {
|
|
4
|
+
if (typeof item === "number" && isNaN(item)) {
|
|
5
|
+
return "NaN";
|
|
6
|
+
}
|
|
7
|
+
if (item && typeof item === "object") {
|
|
8
|
+
if (Array.isArray(item)) {
|
|
9
|
+
return `[${item.map(serialize).join("")}]`;
|
|
10
|
+
} else {
|
|
11
|
+
return `{${Object.entries(item).sort(([a], [b]) => a.localeCompare(b)).map(([k, v]) => `${k}:${serialize(v)}`).join("")}}`;
|
|
12
|
+
}
|
|
13
|
+
}
|
|
14
|
+
return String(item);
|
|
15
|
+
};
|
|
16
|
+
var fnv1a = (str) => {
|
|
17
|
+
if (typeof str !== "string") {
|
|
18
|
+
str = String(str);
|
|
19
|
+
}
|
|
20
|
+
let hash = 2166136261;
|
|
21
|
+
for (let i = 0; i < str.length; i++) {
|
|
22
|
+
hash ^= str.charCodeAt(i);
|
|
23
|
+
hash = hash * 16777619 >>> 0;
|
|
24
|
+
}
|
|
25
|
+
return hash >>> 0;
|
|
26
|
+
};
|
|
27
|
+
var findNextPrime = (num) => {
|
|
28
|
+
if (num < 2) return 2;
|
|
29
|
+
if ((num & 1) === 0) num++;
|
|
30
|
+
while (!isPrime(num)) {
|
|
31
|
+
num += 2;
|
|
32
|
+
}
|
|
33
|
+
return num;
|
|
34
|
+
};
|
|
35
|
+
var isPrime = (num) => {
|
|
36
|
+
if (num < 2) return false;
|
|
37
|
+
if (num === 2 || num === 3) return true;
|
|
38
|
+
if ((num & 1) === 0) return false;
|
|
39
|
+
if (num % 3 === 0) return false;
|
|
40
|
+
const sqrt = Math.sqrt(num);
|
|
41
|
+
for (let i = 5; i <= sqrt; i += 6) {
|
|
42
|
+
if (num % i === 0 || num % (i + 2) === 0) return false;
|
|
43
|
+
}
|
|
44
|
+
return true;
|
|
45
|
+
};
|
|
3
46
|
var UniqueSet = class extends Set {
|
|
4
47
|
/*** @throws TypeError If the input is not iterable. */
|
|
5
48
|
constructor(iterable = []) {
|
|
@@ -62,7 +105,7 @@ var BloomSet = class extends Set {
|
|
|
62
105
|
if (typeof size !== "number" || size <= 0) {
|
|
63
106
|
size = 6553577;
|
|
64
107
|
}
|
|
65
|
-
this.#aSize =
|
|
108
|
+
this.#aSize = findNextPrime(size);
|
|
66
109
|
if (typeof hashCount !== "number" || hashCount <= 0) {
|
|
67
110
|
hashCount = 7;
|
|
68
111
|
}
|
|
@@ -73,45 +116,10 @@ var BloomSet = class extends Set {
|
|
|
73
116
|
}
|
|
74
117
|
}
|
|
75
118
|
/** @internal */
|
|
76
|
-
#findNextPrime(num) {
|
|
77
|
-
if (num < 2) return 2;
|
|
78
|
-
if (num % 2 === 0) num++;
|
|
79
|
-
while (!this.#isPrime(num)) {
|
|
80
|
-
num += 2;
|
|
81
|
-
}
|
|
82
|
-
return num;
|
|
83
|
-
}
|
|
84
|
-
/** @internal */
|
|
85
|
-
#isPrime(num) {
|
|
86
|
-
if (num < 2) return false;
|
|
87
|
-
if (num === 2 || num === 3) return true;
|
|
88
|
-
if (num % 2 === 0 || num % 3 === 0) return false;
|
|
89
|
-
const sqrt = Math.floor(Math.sqrt(num));
|
|
90
|
-
for (let i = 5; i <= sqrt; i += 6) {
|
|
91
|
-
if (num % i === 0 || num % (i + 2) === 0) return false;
|
|
92
|
-
}
|
|
93
|
-
return true;
|
|
94
|
-
}
|
|
95
|
-
/** @internal */
|
|
96
|
-
#serialize(item) {
|
|
97
|
-
if (typeof item === "number" && isNaN(item)) {
|
|
98
|
-
return "NaN";
|
|
99
|
-
}
|
|
100
|
-
if (item && typeof item === "object") {
|
|
101
|
-
const serialize = this.#serialize.bind(this);
|
|
102
|
-
if (Array.isArray(item)) {
|
|
103
|
-
return `[${item.map(serialize).join(",")}]`;
|
|
104
|
-
} else {
|
|
105
|
-
return `{${Object.entries(item).sort(([a], [b]) => a.localeCompare(b)).map(([k, v]) => `${k}:${serialize(v)}`).join(",")}}`;
|
|
106
|
-
}
|
|
107
|
-
}
|
|
108
|
-
return String(item);
|
|
109
|
-
}
|
|
110
|
-
/** @internal */
|
|
111
119
|
#hashes(item) {
|
|
112
120
|
const hashes = [];
|
|
113
|
-
const str =
|
|
114
|
-
let hash =
|
|
121
|
+
const str = serialize(item);
|
|
122
|
+
let hash = fnv1a(str);
|
|
115
123
|
for (let i = 0; i < this.#hashCount; i++) {
|
|
116
124
|
hash %= this.#aSize;
|
|
117
125
|
hashes.push(hash);
|
|
@@ -121,18 +129,6 @@ var BloomSet = class extends Set {
|
|
|
121
129
|
return hashes;
|
|
122
130
|
}
|
|
123
131
|
/** @internal */
|
|
124
|
-
#fnv1a(str) {
|
|
125
|
-
if (typeof str !== "string") {
|
|
126
|
-
str = String(str);
|
|
127
|
-
}
|
|
128
|
-
let hash = 2166136261;
|
|
129
|
-
for (let i = 0; i < str.length; i++) {
|
|
130
|
-
hash ^= str.charCodeAt(i);
|
|
131
|
-
hash = hash * 16777619 >>> 0;
|
|
132
|
-
}
|
|
133
|
-
return hash >>> 0;
|
|
134
|
-
}
|
|
135
|
-
/** @internal */
|
|
136
132
|
#setBits(hashes) {
|
|
137
133
|
for (const hash of hashes) {
|
|
138
134
|
const index = Math.floor(hash / 8);
|
package/index.ts
CHANGED
|
@@ -1,5 +1,63 @@
|
|
|
1
1
|
import equal from "fast-deep-equal/es6/index.js";
|
|
2
2
|
|
|
3
|
+
/** Utility functions */
|
|
4
|
+
|
|
5
|
+
const serialize = (item: any | number | object): string => {
|
|
6
|
+
if (typeof item === "number" && isNaN(item)) {
|
|
7
|
+
return "NaN";
|
|
8
|
+
}
|
|
9
|
+
|
|
10
|
+
if (item && typeof item === "object") {
|
|
11
|
+
if (Array.isArray(item)) {
|
|
12
|
+
return `[${item.map(serialize).join("")}]`;
|
|
13
|
+
} else {
|
|
14
|
+
return `{${Object.entries(item)
|
|
15
|
+
.sort(([a], [b]) => a.localeCompare(b))
|
|
16
|
+
.map(([k, v]) => `${k}:${serialize(v)}`)
|
|
17
|
+
.join("")}}`;
|
|
18
|
+
}
|
|
19
|
+
}
|
|
20
|
+
|
|
21
|
+
return String(item);
|
|
22
|
+
};
|
|
23
|
+
|
|
24
|
+
const fnv1a = (str: string) => {
|
|
25
|
+
if (typeof str !== "string") {
|
|
26
|
+
str = String(str);
|
|
27
|
+
}
|
|
28
|
+
let hash = 2166136261; // FNV offset basis for 32-bit
|
|
29
|
+
for (let i = 0; i < str.length; i++) {
|
|
30
|
+
hash ^= str.charCodeAt(i);
|
|
31
|
+
hash = (hash * 16777619) >>> 0; // Multiply by the FNV prime and ensure 32-bit unsigned
|
|
32
|
+
}
|
|
33
|
+
return hash >>> 0;
|
|
34
|
+
};
|
|
35
|
+
|
|
36
|
+
const findNextPrime = (num: number) => {
|
|
37
|
+
if (num < 2) return 2;
|
|
38
|
+
if ((num & 1) === 0) num++; // Odd numbers only
|
|
39
|
+
|
|
40
|
+
while (!isPrime(num)) {
|
|
41
|
+
num += 2; // Odd numbers only
|
|
42
|
+
}
|
|
43
|
+
|
|
44
|
+
return num;
|
|
45
|
+
};
|
|
46
|
+
|
|
47
|
+
const isPrime = (num: number): boolean => {
|
|
48
|
+
if (num < 2) return false;
|
|
49
|
+
if (num === 2 || num === 3) return true;
|
|
50
|
+
if ((num & 1) === 0) return false;
|
|
51
|
+
if (num % 3 === 0) return false;
|
|
52
|
+
|
|
53
|
+
const sqrt = Math.sqrt(num);
|
|
54
|
+
for (let i = 5; i <= sqrt; i += 6) {
|
|
55
|
+
if (num % i === 0 || num % (i + 2) === 0) return false;
|
|
56
|
+
}
|
|
57
|
+
|
|
58
|
+
return true;
|
|
59
|
+
};
|
|
60
|
+
|
|
3
61
|
/** A `Set` extension that ensures uniqueness of items using deep equality checks. */
|
|
4
62
|
export class UniqueSet<T> extends Set<T> {
|
|
5
63
|
/*** @throws TypeError If the input is not iterable. */
|
|
@@ -72,7 +130,7 @@ export class BloomSet<T> extends Set<T> {
|
|
|
72
130
|
if (typeof size !== "number" || size <= 0) {
|
|
73
131
|
size = 6553577; // Targeting < 1 collision per 100,000 elements, ~819 KB memory, needs 7 hashes
|
|
74
132
|
}
|
|
75
|
-
this.#aSize =
|
|
133
|
+
this.#aSize = findNextPrime(size);
|
|
76
134
|
|
|
77
135
|
if (typeof hashCount !== "number" || hashCount <= 0) {
|
|
78
136
|
hashCount = 7;
|
|
@@ -85,58 +143,11 @@ export class BloomSet<T> extends Set<T> {
|
|
|
85
143
|
}
|
|
86
144
|
}
|
|
87
145
|
|
|
88
|
-
/** @internal */
|
|
89
|
-
#findNextPrime(num: number) {
|
|
90
|
-
if (num < 2) return 2;
|
|
91
|
-
if (num % 2 === 0) num++; // Odd numbers only
|
|
92
|
-
|
|
93
|
-
while (!this.#isPrime(num)) {
|
|
94
|
-
num += 2; // Odd numbers only
|
|
95
|
-
}
|
|
96
|
-
|
|
97
|
-
return num;
|
|
98
|
-
}
|
|
99
|
-
|
|
100
|
-
/** @internal */
|
|
101
|
-
#isPrime(num: number) {
|
|
102
|
-
if (num < 2) return false;
|
|
103
|
-
if (num === 2 || num === 3) return true;
|
|
104
|
-
if (num % 2 === 0 || num % 3 === 0) return false;
|
|
105
|
-
|
|
106
|
-
const sqrt = Math.floor(Math.sqrt(num));
|
|
107
|
-
for (let i = 5; i <= sqrt; i += 6) {
|
|
108
|
-
if (num % i === 0 || num % (i + 2) === 0) return false;
|
|
109
|
-
}
|
|
110
|
-
|
|
111
|
-
return true;
|
|
112
|
-
}
|
|
113
|
-
|
|
114
|
-
/** @internal */
|
|
115
|
-
#serialize(item: T | number | object): string {
|
|
116
|
-
if (typeof item === "number" && isNaN(item)) {
|
|
117
|
-
return "NaN";
|
|
118
|
-
}
|
|
119
|
-
|
|
120
|
-
if (item && typeof item === "object") {
|
|
121
|
-
const serialize = this.#serialize.bind(this);
|
|
122
|
-
if (Array.isArray(item)) {
|
|
123
|
-
return `[${item.map(serialize).join(",")}]`;
|
|
124
|
-
} else {
|
|
125
|
-
return `{${Object.entries(item)
|
|
126
|
-
.sort(([a], [b]) => a.localeCompare(b))
|
|
127
|
-
.map(([k, v]) => `${k}:${serialize(v)}`)
|
|
128
|
-
.join(",")}}`;
|
|
129
|
-
}
|
|
130
|
-
}
|
|
131
|
-
|
|
132
|
-
return String(item);
|
|
133
|
-
}
|
|
134
|
-
|
|
135
146
|
/** @internal */
|
|
136
147
|
#hashes(item: T) {
|
|
137
148
|
const hashes: number[] = [];
|
|
138
|
-
const str =
|
|
139
|
-
let hash =
|
|
149
|
+
const str = serialize(item);
|
|
150
|
+
let hash = fnv1a(str); // Base hash
|
|
140
151
|
|
|
141
152
|
// Bloom into hashCount hash values
|
|
142
153
|
for (let i = 0; i < this.#hashCount; i++) {
|
|
@@ -151,19 +162,6 @@ export class BloomSet<T> extends Set<T> {
|
|
|
151
162
|
return hashes;
|
|
152
163
|
}
|
|
153
164
|
|
|
154
|
-
/** @internal */
|
|
155
|
-
#fnv1a(str: string) {
|
|
156
|
-
if (typeof str !== "string") {
|
|
157
|
-
str = String(str);
|
|
158
|
-
}
|
|
159
|
-
let hash = 2166136261; // FNV offset basis for 32-bit
|
|
160
|
-
for (let i = 0; i < str.length; i++) {
|
|
161
|
-
hash ^= str.charCodeAt(i);
|
|
162
|
-
hash = (hash * 16777619) >>> 0; // Multiply by the FNV prime and ensure 32-bit unsigned
|
|
163
|
-
}
|
|
164
|
-
return hash >>> 0;
|
|
165
|
-
}
|
|
166
|
-
|
|
167
165
|
/** @internal */
|
|
168
166
|
#setBits(hashes: number[]): void {
|
|
169
167
|
for (const hash of hashes) {
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@sepiariver/unique-set",
|
|
3
|
-
"version": "2.0.
|
|
3
|
+
"version": "2.0.3",
|
|
4
4
|
"description": "Extends the native Set class to deeply compare using fast-deep-equal, with optional Bloom filter optimization. This version exports 2 classes instead of a default, breaking b/c with version 1.",
|
|
5
5
|
"main": "dist/index.js",
|
|
6
6
|
"module": "dist/index.mjs",
|