distillate 0.14.0 → 0.14.1

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.
Files changed (9) hide show
  1. package/dist/cuckoo/index.cjs +19 -8
  2. package/dist/cuckoo/index.js +19 -8
  3. package/dist/distinct-BhxDB8mW.js +61 -0
  4. package/dist/distinct-Ck5PKuCS.cjs +66 -0
  5. package/dist/fuse/index.cjs +5 -10
  6. package/dist/fuse/index.js +5 -10
  7. package/dist/index.cjs +1 -1
  8. package/dist/index.js +1 -1
  9. package/package.json +7 -7
package/dist/cuckoo/index.cjs CHANGED
@@ -2,6 +2,7 @@ Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
2
2
  const require_hasher = require("../hasher-DPXmP7SH.cjs");
3
3
  const require_serialize = require("../serialize-CHikZ4GH.cjs");
4
4
  const require_params = require("../params-UTZbJ22c.cjs");
5
+ const require_distinct = require("../distinct-Ck5PKuCS.cjs");
5
6
  function cuckooSizing(n, epsilon) {
6
7
  require_params.assertProbability(epsilon, "epsilon");
7
8
  const f = Math.max(4, Math.ceil(Math.log2(8 / epsilon)));
@@ -54,17 +55,21 @@ var CuckooFilter = class CuckooFilter {
54
55
  });
55
56
  }
56
57
  static from(keys, epsilon, options = {}) {
57
- const seen = new Set();
58
- const distinct = [];
58
+ const distinct = new require_distinct.DistinctHashes(4);
59
59
  for (const key of keys) {
60
60
  require_hasher.hash128KeyInto(key, options.seed ?? 0, HASH);
61
- const id = `${String(HASH.w0)},${String(HASH.w1)},${String(HASH.w2)},${String(HASH.w3)}`;
62
- if (seen.has(id)) continue;
63
- seen.add(id);
64
- distinct.push(key);
61
+ distinct.add(HASH);
62
+ }
63
+ const f = CuckooFilter.create(Math.max(1, distinct.size), epsilon, options);
64
+ const words = distinct.words;
65
+ for (let at = 0; at < words.length; at += 4) {
66
+ HASH.w0 = words[at] ?? 0;
67
+ HASH.w1 = words[at + 1] ?? 0;
68
+ HASH.w2 = words[at + 2] ?? 0;
69
+ HASH.w3 = words[at + 3] ?? 0;
70
+ f.#derive();
71
+ f.#insert();
65
72
  }
66
- const f = CuckooFilter.create(Math.max(1, distinct.length), epsilon, options);
67
- for (const k of distinct) f.add(k);
68
73
  return f;
69
74
  }
70
75
  static fromBytes(bytes) {
@@ -134,6 +139,9 @@ var CuckooFilter = class CuckooFilter {
134
139
  }
135
140
  add(key) {
136
141
  this.#hash(key);
142
+ this.#insert();
143
+ }
144
+ #insert() {
137
145
  if (this.#replace(this.#i1, 0, this.#fp) || this.#replace(this.#i2, 0, this.#fp)) {
138
146
  this.#count++;
139
147
  return;
@@ -202,6 +210,9 @@ var CuckooFilter = class CuckooFilter {
202
210
  }
203
211
  #hash(key) {
204
212
  require_hasher.hash128KeyInto(key, this.#seed, HASH);
213
+ this.#derive();
214
+ }
215
+ #derive() {
205
216
  this.#fp = HASH.w1 >>> 32 - this.#f || 1;
206
217
  this.#i1 = require_hasher.reduce(HASH.w0, this.#buckets);
207
218
  this.#i2 = this.#alt(this.#i1, this.#fp);
package/dist/cuckoo/index.js CHANGED
@@ -1,6 +1,7 @@
1
1
  import { i as hash128KeyInto, u as reduce } from "../hasher-DOdmZ-LI.js";
2
2
  import { a as TruncatedError, c as assertBodyLength, d as bytesEqual, f as fromJSONEnvelope, g as writeFrame, h as toJSONEnvelope, i as SerializationError, l as assertMinBodyLength, m as readHeader, n as ChecksumError, o as UnknownHashVariantError, r as ReservedBitsError, s as UnknownVersionError, t as BadMagicError, u as assertParamsPadding } from "../serialize-CMSYFym-.js";
3
3
  import { i as assertProbability, o as assertUint32, r as assertPositiveInt, t as ParamError } from "../params-CeajSrwv.js";
4
+ import { t as DistinctHashes } from "../distinct-BhxDB8mW.js";
4
5
  function cuckooSizing(n, epsilon) {
5
6
  assertProbability(epsilon, "epsilon");
6
7
  const f = Math.max(4, Math.ceil(Math.log2(8 / epsilon)));
@@ -53,17 +54,21 @@ var CuckooFilter = class CuckooFilter {
53
54
  });
54
55
  }
55
56
  static from(keys, epsilon, options = {}) {
56
- const seen = new Set();
57
- const distinct = [];
57
+ const distinct = new DistinctHashes(4);
58
58
  for (const key of keys) {
59
59
  hash128KeyInto(key, options.seed ?? 0, HASH);
60
- const id = `${String(HASH.w0)},${String(HASH.w1)},${String(HASH.w2)},${String(HASH.w3)}`;
61
- if (seen.has(id)) continue;
62
- seen.add(id);
63
- distinct.push(key);
60
+ distinct.add(HASH);
61
+ }
62
+ const f = CuckooFilter.create(Math.max(1, distinct.size), epsilon, options);
63
+ const words = distinct.words;
64
+ for (let at = 0; at < words.length; at += 4) {
65
+ HASH.w0 = words[at] ?? 0;
66
+ HASH.w1 = words[at + 1] ?? 0;
67
+ HASH.w2 = words[at + 2] ?? 0;
68
+ HASH.w3 = words[at + 3] ?? 0;
69
+ f.#derive();
70
+ f.#insert();
64
71
  }
65
- const f = CuckooFilter.create(Math.max(1, distinct.length), epsilon, options);
66
- for (const k of distinct) f.add(k);
67
72
  return f;
68
73
  }
69
74
  static fromBytes(bytes) {
@@ -133,6 +138,9 @@ var CuckooFilter = class CuckooFilter {
133
138
  }
134
139
  add(key) {
135
140
  this.#hash(key);
141
+ this.#insert();
142
+ }
143
+ #insert() {
136
144
  if (this.#replace(this.#i1, 0, this.#fp) || this.#replace(this.#i2, 0, this.#fp)) {
137
145
  this.#count++;
138
146
  return;
@@ -201,6 +209,9 @@ var CuckooFilter = class CuckooFilter {
201
209
  }
202
210
  #hash(key) {
203
211
  hash128KeyInto(key, this.#seed, HASH);
212
+ this.#derive();
213
+ }
214
+ #derive() {
204
215
  this.#fp = HASH.w1 >>> 32 - this.#f || 1;
205
216
  this.#i1 = reduce(HASH.w0, this.#buckets);
206
217
  this.#i2 = this.#alt(this.#i1, this.#fp);
package/dist/distinct-BhxDB8mW.js ADDED
@@ -0,0 +1,61 @@
1
+ //#region src/core/distinct.ts
2
+ const START = 1024;
3
+ var DistinctHashes = class {
4
+ #width;
5
+ #words;
6
+ #table = new Int32Array(2 * START);
7
+ #size = 0;
8
+ constructor(width) {
9
+ this.#width = width;
10
+ this.#words = new Uint32Array(START * width);
11
+ }
12
+ get size() {
13
+ return this.#size;
14
+ }
15
+ get words() {
16
+ return this.#words.subarray(0, this.#size * this.#width);
17
+ }
18
+ add(hash) {
19
+ const width = this.#width;
20
+ const w0 = hash.w0 >>> 0;
21
+ const w1 = hash.w1 >>> 0;
22
+ const w2 = hash.w2 >>> 0;
23
+ const w3 = hash.w3 >>> 0;
24
+ if ((this.#size + 1) * width > this.#words.length) this.#grow();
25
+ const words = this.#words;
26
+ const mask = this.#table.length - 1;
27
+ let slot = w0 & mask;
28
+ for (;;) {
29
+ const held = this.#table[slot] ?? 0;
30
+ if (held === 0) break;
31
+ const at = (held - 1) * width;
32
+ if (words[at] === w0 && words[at + 1] === w1 && (width === 2 || words[at + 2] === w2 && words[at + 3] === w3)) return false;
33
+ slot = slot + 1 & mask;
34
+ }
35
+ const at = this.#size * width;
36
+ words[at] = w0;
37
+ words[at + 1] = w1;
38
+ if (width === 4) {
39
+ words[at + 2] = w2;
40
+ words[at + 3] = w3;
41
+ }
42
+ this.#size++;
43
+ this.#table[slot] = this.#size;
44
+ return true;
45
+ }
46
+ #grow() {
47
+ const words = new Uint32Array(this.#words.length * 2);
48
+ words.set(this.#words);
49
+ this.#words = words;
50
+ this.#table = new Int32Array(this.#table.length * 2);
51
+ for (let i = 0; i < this.#size; i++) this.#place(i);
52
+ }
53
+ #place(i) {
54
+ const mask = this.#table.length - 1;
55
+ let slot = (this.#words[i * this.#width] ?? 0) & mask;
56
+ while ((this.#table[slot] ?? 0) !== 0) slot = slot + 1 & mask;
57
+ this.#table[slot] = i + 1;
58
+ }
59
+ };
60
+ //#endregion
61
+ export { DistinctHashes as t };
package/dist/distinct-Ck5PKuCS.cjs ADDED
@@ -0,0 +1,66 @@
1
+ //#region src/core/distinct.ts
2
+ const START = 1024;
3
+ var DistinctHashes = class {
4
+ #width;
5
+ #words;
6
+ #table = new Int32Array(2 * START);
7
+ #size = 0;
8
+ constructor(width) {
9
+ this.#width = width;
10
+ this.#words = new Uint32Array(START * width);
11
+ }
12
+ get size() {
13
+ return this.#size;
14
+ }
15
+ get words() {
16
+ return this.#words.subarray(0, this.#size * this.#width);
17
+ }
18
+ add(hash) {
19
+ const width = this.#width;
20
+ const w0 = hash.w0 >>> 0;
21
+ const w1 = hash.w1 >>> 0;
22
+ const w2 = hash.w2 >>> 0;
23
+ const w3 = hash.w3 >>> 0;
24
+ if ((this.#size + 1) * width > this.#words.length) this.#grow();
25
+ const words = this.#words;
26
+ const mask = this.#table.length - 1;
27
+ let slot = w0 & mask;
28
+ for (;;) {
29
+ const held = this.#table[slot] ?? 0;
30
+ if (held === 0) break;
31
+ const at = (held - 1) * width;
32
+ if (words[at] === w0 && words[at + 1] === w1 && (width === 2 || words[at + 2] === w2 && words[at + 3] === w3)) return false;
33
+ slot = slot + 1 & mask;
34
+ }
35
+ const at = this.#size * width;
36
+ words[at] = w0;
37
+ words[at + 1] = w1;
38
+ if (width === 4) {
39
+ words[at + 2] = w2;
40
+ words[at + 3] = w3;
41
+ }
42
+ this.#size++;
43
+ this.#table[slot] = this.#size;
44
+ return true;
45
+ }
46
+ #grow() {
47
+ const words = new Uint32Array(this.#words.length * 2);
48
+ words.set(this.#words);
49
+ this.#words = words;
50
+ this.#table = new Int32Array(this.#table.length * 2);
51
+ for (let i = 0; i < this.#size; i++) this.#place(i);
52
+ }
53
+ #place(i) {
54
+ const mask = this.#table.length - 1;
55
+ let slot = (this.#words[i * this.#width] ?? 0) & mask;
56
+ while ((this.#table[slot] ?? 0) !== 0) slot = slot + 1 & mask;
57
+ this.#table[slot] = i + 1;
58
+ }
59
+ };
60
+ //#endregion
61
+ Object.defineProperty(exports, "DistinctHashes", {
62
+ enumerable: true,
63
+ get: function() {
64
+ return DistinctHashes;
65
+ }
66
+ });
package/dist/fuse/index.cjs CHANGED
@@ -1,6 +1,7 @@
1
1
  Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
2
2
  const require_hasher = require("../hasher-DPXmP7SH.cjs");
3
3
  const require_serialize = require("../serialize-CHikZ4GH.cjs");
4
+ const require_distinct = require("../distinct-Ck5PKuCS.cjs");
4
5
  //#region src/fuse/fuse.ts
5
6
  const ARITY = 3;
6
7
  const TYPE_FUSE8 = 3;
@@ -131,18 +132,12 @@ function buildFingerprints(fp, hashes, params, maxAttempts = 100) {
131
132
  throw new BinaryFuseBuildError("binary fuse construction failed");
132
133
  }
133
134
  function buildState(keys, alloc) {
134
- const hashList = [];
135
- const seen = new Set();
135
+ const distinct = new require_distinct.DistinctHashes(2);
136
136
  for (const key of keys) {
137
137
  require_hasher.hash128KeyInto(key, 0, scratchHash);
138
- const lo = scratchHash.w0 >>> 0;
139
- const hi = scratchHash.w1 >>> 0;
140
- const id = `${String(lo)},${String(hi)}`;
141
- if (seen.has(id)) continue;
142
- seen.add(id);
143
- hashList.push(lo, hi);
138
+ distinct.add(scratchHash);
144
139
  }
145
- const size = hashList.length / 2;
140
+ const size = distinct.size;
146
141
  const params = computeParams(size);
147
142
  if (size === 0) return {
148
143
  fp: alloc(0),
@@ -150,7 +145,7 @@ function buildState(keys, alloc) {
150
145
  params,
151
146
  size: 0
152
147
  };
153
- const hashes = Uint32Array.from(hashList);
148
+ const hashes = distinct.words;
154
149
  const fp = alloc(params.arrayLength);
155
150
  return {
156
151
  fp,
package/dist/fuse/index.js CHANGED
@@ -1,5 +1,6 @@
1
1
  import { i as hash128KeyInto, n as RLO, o as mul64, r as fmix64, t as RHI } from "../hasher-DOdmZ-LI.js";
2
2
  import { a as TruncatedError, c as assertBodyLength, d as bytesEqual, f as fromJSONEnvelope, g as writeFrame, h as toJSONEnvelope, i as SerializationError, l as assertMinBodyLength, m as readHeader, n as ChecksumError, o as UnknownHashVariantError, r as ReservedBitsError, s as UnknownVersionError, t as BadMagicError } from "../serialize-CMSYFym-.js";
3
+ import { t as DistinctHashes } from "../distinct-BhxDB8mW.js";
3
4
  //#region src/fuse/fuse.ts
4
5
  const ARITY = 3;
5
6
  const TYPE_FUSE8 = 3;
@@ -130,18 +131,12 @@ function buildFingerprints(fp, hashes, params, maxAttempts = 100) {
130
131
  throw new BinaryFuseBuildError("binary fuse construction failed");
131
132
  }
132
133
  function buildState(keys, alloc) {
133
- const hashList = [];
134
- const seen = new Set();
134
+ const distinct = new DistinctHashes(2);
135
135
  for (const key of keys) {
136
136
  hash128KeyInto(key, 0, scratchHash);
137
- const lo = scratchHash.w0 >>> 0;
138
- const hi = scratchHash.w1 >>> 0;
139
- const id = `${String(lo)},${String(hi)}`;
140
- if (seen.has(id)) continue;
141
- seen.add(id);
142
- hashList.push(lo, hi);
137
+ distinct.add(scratchHash);
143
138
  }
144
- const size = hashList.length / 2;
139
+ const size = distinct.size;
145
140
  const params = computeParams(size);
146
141
  if (size === 0) return {
147
142
  fp: alloc(0),
@@ -149,7 +144,7 @@ function buildState(keys, alloc) {
149
144
  params,
150
145
  size: 0
151
146
  };
152
- const hashes = Uint32Array.from(hashList);
147
+ const hashes = distinct.words;
153
148
  const fp = alloc(params.arrayLength);
154
149
  return {
155
150
  fp,
package/dist/index.cjs CHANGED
@@ -1,6 +1,6 @@
1
1
  Object.defineProperty(exports, Symbol.toStringTag, { value: "Module" });
2
2
  //#endregion
3
3
  //#region src/index.ts
4
- const VERSION = "0.14.0";
4
+ const VERSION = "0.14.1";
5
5
  //#endregion
6
6
  exports.VERSION = VERSION;
package/dist/index.js CHANGED
@@ -1,5 +1,5 @@
1
1
  //#endregion
2
2
  //#region src/index.ts
3
- const VERSION = "0.14.0";
3
+ const VERSION = "0.14.1";
4
4
  //#endregion
5
5
  export { VERSION };
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "distillate",
3
- "version": "0.14.0",
3
+ "version": "0.14.1",
4
4
  "description": "Probabilistic data structures for JavaScript. Space-efficient, approximate-membership filters (Bloom, Blocked Bloom, Binary Fuse) with tunable error and a portable binary format; zero dependencies, universal.",
5
5
  "keywords": [
6
6
  "bloom-filter",
@@ -113,17 +113,17 @@
113
113
  ],
114
114
  "devDependencies": {
115
115
  "@arethetypeswrong/cli": "^0.18.5",
116
- "@microsoft/api-extractor": "^7.59.1",
117
- "@types/node": "^22.20.3",
118
- "@vitest/coverage-v8": "^5.0.1",
119
- "fast-check": "^4.10.1",
116
+ "@microsoft/api-extractor": "^7.59.2",
117
+ "@types/node": "^22.20.4",
118
+ "@vitest/coverage-v8": "^5.0.2",
119
+ "fast-check": "^4.10.2",
120
120
  "mitata": "^1.0.34",
121
121
  "publint": "^0.3.24",
122
122
  "tsdown": "^0.23.0",
123
- "tsx": "^4.23.13",
123
+ "tsx": "^4.23.15",
124
124
  "typedoc": "^0.28.20",
125
125
  "typescript": "^5.9.3",
126
- "vitest": "^5.0.1"
126
+ "vitest": "^5.0.2"
127
127
  },
128
128
  "engines": {
129
129
  "node": ">=22"