cry-search 2.0.2 → 2.0.3

package/dist/index.cjs

@@ -89,6 +89,7 @@ __export(exports_src, {
   createTokensMap: () => createTokensMap,
   createSearchLinkedMetadata: () => createSearchLinkedMetadataNumeric,
   createSearchArrayMetadata: () => createSearchArrayMetadataNumeric,
+  compareTokenized: () => compareTokenized,
   clearTokensMap: () => clearTokensMap,
   binarySearchUint32: () => binarySearchUint32,
   arrayToSearchableData: () => arrayToSearchableData,
@@ -319,6 +320,22 @@ function prepareStringForSearch(input) {
   const tokens = tokenize(sanitised, DEFAULT_TOKENIZE_CONFIG);
   return tokens.sort();
 }
+// src/utils/compareTokenized.ts
+function compareTokenized(a, b) {
+  const aTokens = new Set(prepareStringForSearch(a));
+  const bTokens = new Set(prepareStringForSearch(b));
+  if (aTokens.size === 0 && bTokens.size === 0)
+    return 1;
+  let intersection = 0;
+  for (const token of aTokens) {
+    if (bTokens.has(token))
+      intersection++;
+  }
+  const union = aTokens.size + bTokens.size - intersection;
+  if (union === 0)
+    return 0;
+  return intersection / union;
+}
 // src/utils/phoneNormalization.ts
 var RE_ASCII_DIGITS = /^[0-9]+$/;
 var RE_PHONE_LIKE_QUERY = /^[\d+()/\-. ]+$/;

package/dist/index.d.cts

@@ -11,6 +11,7 @@ export { preprocessString } from './utils/preprocessString';
 export { sanitiseString } from './utils/sanitiseString';
 export { tokenize } from './utils/tokenize';
 export { normalizeDates } from './utils/normalizeDates';
+export { compareTokenized } from './utils/compareTokenized';
 export { defaultPhoneNormalizer, defaultQueryPhoneNormalizer, resolvePhoneQueryNormalizer, } from './utils/phoneNormalization';
 export { tokens, getTokenCount, getTokenById, findTokenId, registerToken, getOrCreateTokenId, createTokensMap, clearTokensMap, resetTokenRegistry, } from './utils/tokenRegistry';
 export { createSearchArrayMetadataNumeric as createSearchArrayMetadata, prepareObjectForSearchNumeric as prepareObjectForSearch, encodeTokenWithMode, decodeTokenId, decodeMatchMode, TOKEN_ID_MASK, MATCH_MODE_BITS, MATCH_MODE_PREFIXES, PREFIX_TO_MATCH_MODE, isDeletedOrBlocked, } from './common/numeric/createSearchArrayMetadata';

package/dist/index.d.ts

@@ -11,6 +11,7 @@ export { preprocessString } from './utils/preprocessString';
 export { sanitiseString } from './utils/sanitiseString';
 export { tokenize } from './utils/tokenize';
 export { normalizeDates } from './utils/normalizeDates';
+export { compareTokenized } from './utils/compareTokenized';
 export { defaultPhoneNormalizer, defaultQueryPhoneNormalizer, resolvePhoneQueryNormalizer, } from './utils/phoneNormalization';
 export { tokens, getTokenCount, getTokenById, findTokenId, registerToken, getOrCreateTokenId, createTokensMap, clearTokensMap, resetTokenRegistry, } from './utils/tokenRegistry';
 export { createSearchArrayMetadataNumeric as createSearchArrayMetadata, prepareObjectForSearchNumeric as prepareObjectForSearch, encodeTokenWithMode, decodeTokenId, decodeMatchMode, TOKEN_ID_MASK, MATCH_MODE_BITS, MATCH_MODE_PREFIXES, PREFIX_TO_MATCH_MODE, isDeletedOrBlocked, } from './common/numeric/createSearchArrayMetadata';

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AACA;;;GAGG;AAKH,YAAY,EAEV,EAAE,EACF,WAAW,EACX,qBAAqB,EACrB,cAAc,EAEd,YAAY,EACZ,sBAAsB,EACtB,qBAAqB,EAErB,gBAAgB,EAChB,SAAS,EACT,eAAe,EACf,oBAAoB,EACpB,oBAAoB,EAEpB,UAAU,EAEV,2BAA2B,EAC3B,kBAAkB,EAClB,0BAA0B,EAE1B,oBAAoB,EAEpB,UAAU,EACV,WAAW,EACX,gBAAgB,EAChB,YAAY,EACZ,6BAA6B,EAE7B,mBAAmB,EACnB,gBAAgB,EAChB,YAAY,EACZ,iBAAiB,GAClB,MAAM,SAAS,CAAC;AAGjB,YAAY,EAAE,qBAAqB,IAAI,cAAc,EAAE,MAAM,SAAS,CAAC;AAGvE,OAAO,EAAE,mBAAmB,EAAE,oBAAoB,EAAE,MAAM,SAAS,CAAC;AAGpE,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,SAAS,CAAC;AAKvE,OAAO,EAAE,sBAAsB,EAAE,MAAM,gCAAgC,CAAC;AACxE,OAAO,EAAE,gBAAgB,EAAE,MAAM,0BAA0B,CAAC;AAC5D,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EACL,sBAAsB,EACtB,2BAA2B,EAC3B,2BAA2B,GAC5B,MAAM,4BAA4B,CAAC;AAKpC,OAAO,EACL,MAAM,EACN,aAAa,EACb,YAAY,EACZ,WAAW,EACX,aAAa,EACb,kBAAkB,EAClB,eAAe,EACf,cAAc,EACd,kBAAkB,GACnB,MAAM,uBAAuB,CAAC;AAO/B,OAAO,EACL,gCAAgC,IAAI,yBAAyB,EAC7D,6BAA6B,IAAI,sBAAsB,EACvD,mBAAmB,EACnB,aAAa,EACb,eAAe,EACf,aAAa,EACb,eAAe,EACf,mBAAmB,EACnB,oBAAoB,EACpB,kBAAkB,GACnB,MAAM,4CAA4C,CAAC;AAEpD,OAAO,EACL,iCAAiC,IAAI,0BAA0B,EAC/D,KAAK,mCAAmC,GACzC,MAAM,6CAA6C,CAAC;AAGrD,OAAO,EACL,2BAA2B,IAAI,oBAAoB,EACnD,gCAAgC,IAAI,yBAAyB,EAC7D,+BAA+B,IAAI,wBAAwB,EAC3D,oCAAoC,IAAI,6BAA6B,GACtE,MAAM,uCAAuC,CAAC;AAG/C,OAAO,EACL,4BAA4B,IAAI,YAAY,EAC5C,4BAA4B,IAAI,qBAAqB,EACrD,gCAAgC,IAAI,yBAAyB,EAC7D,qCAAqC,IAAI,qBAAqB,GAC/D,MAAM,+BAA+B,CAAC;AAGvC,OAAO,EACL,yBAAyB,EACzB,+BAA+B,GAChC,MAAM,qCAAqC,CAAC;AAG7C,OAAO,EAAE,iBAAiB,EAAE,MAAM,oCAAoC,CAAC;AAGvE,OAAO,EACL,8BAA8B,IAAI,uBAAuB,EACzD,0CAA0C,IAAI,mCAAmC,GAClF,MAAM,0CAA0C,CAAC;AAGlD,OAAO,EACL,wBAAwB,IAAI,iBAAiB,EAC7C,wBAAwB,IAAI,iBAAiB,EAC7C,uBAAuB,IAAI,gBAAgB,EAC3C,uBAAuB,IAAI,gBAAgB,EAC3C,gCAAgC,IAAI,yBAAyB,GAC9D,MAAM,6CAA6C,CAAC;AAGrD,OAAO,EACL,6BAA6B,EAC7B,qBAAqB,EACrB,kBAAkB,EAElB,kBAAkB,EAClB,8BAA8B,EAC9B,sBAAsB,EACtB,KAAK,kBAAkB,GACxB,MAAM,2BAA2B,CAAC;AAGnC,OAAO,EACL,oBAAoB,IAAI,aAAa,EACrC,oBAAoB,EACpB,QAAQ,EACR,qBAAqB,EACrB,QAAQ,EACR,eAAe,EACf,KAAK,YAAY,EACjB,KAAK,oBAAoB,EACzB,KAAK,uBAAuB,GAC7B,MAAM,gCAAgC,CAAC;AAKxC,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AACA;;;GAGG;AAKH,YAAY,EAEV,EAAE,EACF,WAAW,EACX,qBAAqB,EACrB,cAAc,EAEd,YAAY,EACZ,sBAAsB,EACtB,qBAAqB,EAErB,gBAAgB,EAChB,SAAS,EACT,eAAe,EACf,oBAAoB,EACpB,oBAAoB,EAEpB,UAAU,EAEV,2BAA2B,EAC3B,kBAAkB,EAClB,0BAA0B,EAE1B,oBAAoB,EAEpB,UAAU,EACV,WAAW,EACX,gBAAgB,EAChB,YAAY,EACZ,6BAA6B,EAE7B,mBAAmB,EACnB,gBAAgB,EAChB,YAAY,EACZ,iBAAiB,GAClB,MAAM,SAAS,CAAC;AAGjB,YAAY,EAAE,qBAAqB,IAAI,cAAc,EAAE,MAAM,SAAS,CAAC;AAGvE,OAAO,EAAE,mBAAmB,EAAE,oBAAoB,EAAE,MAAM,SAAS,CAAC;AAGpE,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,SAAS,CAAC;AAKvE,OAAO,EAAE,sBAAsB,EAAE,MAAM,gCAAgC,CAAC;AACxE,OAAO,EAAE,gBAAgB,EAAE,MAAM,0BAA0B,CAAC;AAC5D,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAC5C,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,gBAAgB,EAAE,MAAM,0BAA0B,CAAC;AAC5D,OAAO,EACL,sBAAsB,EACtB,2BAA2B,EAC3B,2BAA2B,GAC5B,MAAM,4BAA4B,CAAC;AAKpC,OAAO,EACL,MAAM,EACN,aAAa,EACb,YAAY,EACZ,WAAW,EACX,aAAa,EACb,kBAAkB,EAClB,eAAe,EACf,cAAc,EACd,kBAAkB,GACnB,MAAM,uBAAuB,CAAC;AAO/B,OAAO,EACL,gCAAgC,IAAI,yBAAyB,EAC7D,6BAA6B,IAAI,sBAAsB,EACvD,mBAAmB,EACnB,aAAa,EACb,eAAe,EACf,aAAa,EACb,eAAe,EACf,mBAAmB,EACnB,oBAAoB,EACpB,kBAAkB,GACnB,MAAM,4CAA4C,CAAC;AAEpD,OAAO,EACL,iCAAiC,IAAI,0BAA0B,EAC/D,KAAK,mCAAmC,GACzC,MAAM,6CAA6C,CAAC;AAGrD,OAAO,EACL,2BAA2B,IAAI,oBAAoB,EACnD,gCAAgC,IAAI,yBAAyB,EAC7D,+BAA+B,IAAI,wBAAwB,EAC3D,oCAAoC,IAAI,6BAA6B,GACtE,MAAM,uCAAuC,CAAC;AAG/C,OAAO,EACL,4BAA4B,IAAI,YAAY,EAC5C,4BAA4B,IAAI,qBAAqB,EACrD,gCAAgC,IAAI,yBAAyB,EAC7D,qCAAqC,IAAI,qBAAqB,GAC/D,MAAM,+BAA+B,CAAC;AAGvC,OAAO,EACL,yBAAyB,EACzB,+BAA+B,GAChC,MAAM,qCAAqC,CAAC;AAG7C,OAAO,EAAE,iBAAiB,EAAE,MAAM,oCAAoC,CAAC;AAGvE,OAAO,EACL,8BAA8B,IAAI,uBAAuB,EACzD,0CAA0C,IAAI,mCAAmC,GAClF,MAAM,0CAA0C,CAAC;AAGlD,OAAO,EACL,wBAAwB,IAAI,iBAAiB,EAC7C,wBAAwB,IAAI,iBAAiB,EAC7C,uBAAuB,IAAI,gBAAgB,EAC3C,uBAAuB,IAAI,gBAAgB,EAC3C,gCAAgC,IAAI,yBAAyB,GAC9D,MAAM,6CAA6C,CAAC;AAGrD,OAAO,EACL,6BAA6B,EAC7B,qBAAqB,EACrB,kBAAkB,EAElB,kBAAkB,EAClB,8BAA8B,EAC9B,sBAAsB,EACtB,KAAK,kBAAkB,GACxB,MAAM,2BAA2B,CAAC;AAGnC,OAAO,EACL,oBAAoB,IAAI,aAAa,EACrC,oBAAoB,EACpB,QAAQ,EACR,qBAAqB,EACrB,QAAQ,EACR,eAAe,EACf,KAAK,YAAY,EACjB,KAAK,oBAAoB,EACzB,KAAK,uBAAuB,GAC7B,MAAM,gCAAgC,CAAC;AAKxC,OAAO,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC"}

package/dist/index.js

@@ -212,6 +212,22 @@ function prepareStringForSearch(input) {
   const tokens = tokenize(sanitised, DEFAULT_TOKENIZE_CONFIG);
   return tokens.sort();
 }
+// src/utils/compareTokenized.ts
+function compareTokenized(a, b) {
+  const aTokens = new Set(prepareStringForSearch(a));
+  const bTokens = new Set(prepareStringForSearch(b));
+  if (aTokens.size === 0 && bTokens.size === 0)
+    return 1;
+  let intersection = 0;
+  for (const token of aTokens) {
+    if (bTokens.has(token))
+      intersection++;
+  }
+  const union = aTokens.size + bTokens.size - intersection;
+  if (union === 0)
+    return 0;
+  return intersection / union;
+}
 // src/utils/phoneNormalization.ts
 var RE_ASCII_DIGITS = /^[0-9]+$/;
 var RE_PHONE_LIKE_QUERY = /^[\d+()/\-. ]+$/;
@@ -2171,6 +2187,7 @@ export {
   createTokensMap,
   createSearchLinkedMetadataNumeric as createSearchLinkedMetadata,
   createSearchArrayMetadataNumeric as createSearchArrayMetadata,
+  compareTokenized,
   clearTokensMap,
   binarySearchUint32,
   arrayToSearchableData,

package/dist/utils/compareTokenized.d.ts

@@ -0,0 +1,24 @@
+/**
+ * Compares two strings by their tokenized forms and returns the Jaccard
+ * similarity of their token sets: a number between `0` (no tokens in common)
+ * and `1` (identical token sets). Order-independent and case/diacritics
+ * insensitive, because both sides go through the same search pipeline as the
+ * rest of the library (normalizeDates → preprocess → sanitise → tokenize with
+ * the default replacements/removals).
+ *
+ * Both empty strings are considered identical (`1`); an empty vs non-empty
+ * string yields `0`.
+ *
+ * @param a - First string to compare.
+ * @param b - Second string to compare.
+ * @returns Similarity ratio `0..1` (matching tokens / union of tokens).
+ *
+ * @example
+ * ```typescript
+ * compareTokenized('Janez Novak', 'novak janez'); // 1
+ * compareTokenized('Janes Novak', 'novak');        // 0.5
+ * compareTokenized('Janez Novak', 'ana');          // 0
+ * ```
+ */
+export declare function compareTokenized(a: string, b: string): number;
+//# sourceMappingURL=compareTokenized.d.ts.map

package/dist/utils/compareTokenized.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"compareTokenized.d.ts","sourceRoot":"","sources":["../../src/utils/compareTokenized.ts"],"names":[],"mappings":"AAEA;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,wBAAgB,gBAAgB,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,CAc7D"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "cry-search",
-  "version": "2.0.2",
+  "version": "2.0.3",
   "description": "A fast, memory-efficient search library for large datasets with support for tokenized matching, linked collections, and per-field match modes.",
   "type": "module",
   "main": "./dist/index.js",