@mailwoman/normalize 4.16.1 → 5.0.0
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/out/abbreviations.d.ts +5 -5
- package/out/abbreviations.d.ts.map +1 -1
- package/out/abbreviations.js +5 -5
- package/out/abbreviations.js.map +1 -1
- package/out/cjk.d.ts.map +1 -1
- package/out/cjk.js.map +1 -1
- package/out/compute.d.ts.map +1 -1
- package/out/compute.js +3 -3
- package/out/compute.js.map +1 -1
- package/out/index.d.ts +1 -1
- package/out/index.js +1 -1
- package/out/nfc.d.ts +2 -2
- package/out/nfc.d.ts.map +1 -1
- package/out/nfc.js +6 -6
- package/out/nfc.js.map +1 -1
- package/out/offset-map.d.ts +6 -6
- package/out/offset-map.d.ts.map +1 -1
- package/out/offset-map.js +6 -6
- package/out/offset-map.js.map +1 -1
- package/out/punctuation.d.ts.map +1 -1
- package/out/punctuation.js.map +1 -1
- package/out/types.d.ts +4 -4
- package/out/whitespace.d.ts.map +1 -1
- package/out/whitespace.js.map +1 -1
- package/package.json +6 -6
package/out/abbreviations.d.ts
CHANGED
|
@@ -21,12 +21,12 @@ export interface AbbreviationResult {
|
|
|
21
21
|
}>;
|
|
22
22
|
}
|
|
23
23
|
/**
|
|
24
|
-
* Expand known abbreviations. Walks the input token-by-token (whitespace-delimited) and rewrites
|
|
25
|
-
*
|
|
26
|
-
*
|
|
24
|
+
* Expand known abbreviations. Walks the input token-by-token (whitespace-delimited) and rewrites matching tokens to
|
|
25
|
+
* their canonical long form. The output map points every char of the expanded form to its position in the original
|
|
26
|
+
* short form (first char of input token).
|
|
27
27
|
*
|
|
28
|
-
* Case rules: match case-insensitively. Output form preserves the dictionary's canonical casing
|
|
29
|
-
*
|
|
28
|
+
* Case rules: match case-insensitively. Output form preserves the dictionary's canonical casing (`St` → `Street`, `st`
|
|
29
|
+
* → `Street`, `ST` → `Street`).
|
|
30
30
|
*/
|
|
31
31
|
export declare function expandAbbreviations(input: string, locale?: string): AbbreviationResult;
|
|
32
32
|
//# sourceMappingURL=abbreviations.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"abbreviations.d.ts","sourceRoot":"","sources":["../abbreviations.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,YAAY,CAAA;
|
|
1
|
+
{"version":3,"file":"abbreviations.d.ts","sourceRoot":"","sources":["../abbreviations.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,YAAY,CAAA;AAkD3C,MAAM,WAAW,kBAAkB;IAClC,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,UAAU,EAAE,KAAK,CAAC;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,EAAE,EAAE,MAAM,CAAC;QAAC,EAAE,EAAE,SAAS,CAAA;KAAE,CAAC,CAAA;CAC9D;AAED;;;;;;;GAOG;AACH,wBAAgB,mBAAmB,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,CAAC,EAAE,MAAM,GAAG,kBAAkB,CAwDtF"}
|
package/out/abbreviations.js
CHANGED
|
@@ -50,12 +50,12 @@ function getDictionary(locale) {
|
|
|
50
50
|
return EN_US_DICT;
|
|
51
51
|
}
|
|
52
52
|
/**
|
|
53
|
-
* Expand known abbreviations. Walks the input token-by-token (whitespace-delimited) and rewrites
|
|
54
|
-
*
|
|
55
|
-
*
|
|
53
|
+
* Expand known abbreviations. Walks the input token-by-token (whitespace-delimited) and rewrites matching tokens to
|
|
54
|
+
* their canonical long form. The output map points every char of the expanded form to its position in the original
|
|
55
|
+
* short form (first char of input token).
|
|
56
56
|
*
|
|
57
|
-
* Case rules: match case-insensitively. Output form preserves the dictionary's canonical casing
|
|
58
|
-
*
|
|
57
|
+
* Case rules: match case-insensitively. Output form preserves the dictionary's canonical casing (`St` → `Street`, `st`
|
|
58
|
+
* → `Street`, `ST` → `Street`).
|
|
59
59
|
*/
|
|
60
60
|
export function expandAbbreviations(input, locale) {
|
|
61
61
|
const dict = getDictionary(locale);
|
package/out/abbreviations.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"abbreviations.js","sourceRoot":"","sources":["../abbreviations.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AASH,MAAM,UAAU,GAAqC;IACpD,kCAAkC;IAClC,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,OAAO,EAAE;IAC1B,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,OAAO,EAAE;IAC1B,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,MAAM,EAAE;IACzB,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,MAAM,EAAE;IACzB,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,kBAAkB;IAClB,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC5B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC7B,EAAE,IAAI,EAAE,MAAM,EAAE,EAAE,EAAE,WAAW,EAAE;IACjC,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE;IAC1B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE;IAC1B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,MAAM,EAAE,EAAE,EAAE,SAAS,EAAE;IAC/B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE;IAC9B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC5B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE;CAC9B,CAAA;AAED,MAAM,UAAU,GAAqC;IACpD,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,KAAK,EAAE;IACxB,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC5B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,WAAW,EAAE;IAChC,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE;IAC9B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;CAC5B,CAAA;AAED,SAAS,aAAa,CAAC,MAA0B;IAChD,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,OAAO,CAAC,CAAC,WAAW,EAAE,CAAA;
|
|
1
|
+
{"version":3,"file":"abbreviations.js","sourceRoot":"","sources":["../abbreviations.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AASH,MAAM,UAAU,GAAqC;IACpD,kCAAkC;IAClC,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,OAAO,EAAE;IAC1B,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,OAAO,EAAE;IAC1B,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,MAAM,EAAE;IACzB,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,MAAM,EAAE;IACzB,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,kBAAkB;IAClB,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC5B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC7B,EAAE,IAAI,EAAE,MAAM,EAAE,EAAE,EAAE,WAAW,EAAE;IACjC,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE;IAC1B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,MAAM,EAAE;IAC1B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,MAAM,EAAE,EAAE,EAAE,SAAS,EAAE;IAC/B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE;IAC9B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC5B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE;CAC9B,CAAA;AAED,MAAM,UAAU,GAAqC;IACpD,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,EAAE,KAAK,EAAE;IACxB,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,WAAW,EAAE;IAC/B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;IAC5B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,WAAW,EAAE;IAChC,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,OAAO,EAAE;IAC3B,EAAE,IAAI,EAAE,KAAK,EAAE,EAAE,EAAE,SAAS,EAAE;IAC9B,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE;CAC5B,CAAA;AAED,SAAS,aAAa,CAAC,MAA0B;IAChD,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,OAAO,CAAC,CAAC,WAAW,EAAE,CAAA;IAE5C,IAAI,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC;QAAE,OAAO,UAAU,CAAA;IAE1C,OAAO,UAAU,CAAA;AAClB,CAAC;AAQD;;;;;;;GAOG;AACH,MAAM,UAAU,mBAAmB,CAAC,KAAa,EAAE,MAAe;IACjE,MAAM,IAAI,GAAG,aAAa,CAAC,MAAM,CAAC,CAAA;IAClC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAA;IAExC,KAAK,MAAM,KAAK,IAAI,IAAI;QAAE,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,WAAW,EAAE,EAAE,KAAK,CAAC,EAAE,CAAC,CAAA;IAExE,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,UAAU,GAAuD,EAAE,CAAA;IAEzE,IAAI,CAAC,GAAG,CAAC,CAAA;IAET,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QACzB,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;QACpB,kFAAkF;QAClF,8DAA8D;QAC9D,MAAM,WAAW,GAAG,CAAC,CAAS,EAAE,EAAE,CAAC,kBAAkB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QAE7D,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC,EAAE,CAAC;YACtB,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YACZ,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACX,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACT,CAAC;QACD,MAAM,KAAK,GAAG,CAAC,CAAA;QAEf,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,IAAI,WAAW,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC;YAAE,CAAC,IAAI,CAAC,CAAA;QACzD,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC,CAAA;QACnC,MAAM,oBAAoB,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,KAAK,CAAA;QACvF,MAAM,SAAS,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,EAAE,CAAA;QACxD,MAAM,SAAS,GAAG,MAAM,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;QAEvC,IAAI,CAAC,SAAS,EAAE,CAAC;YAChB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACvC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,CAAA;gBACnB,GAAG,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,CAAA;YACpB,CAAC;YACD,SAAQ;QACT,CAAC;QAED,gEAAgE;QAChE,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3C,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAE,CAAC,CAAA;YACvB,GAAG,CAAC,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAA;QAChD,CAAC;QACD,UAAU,CAAC,IAAI,CAAC;YACf,IAAI,EAAE,oBAAoB;YAC1B,EAAE,EAAE,SAAS;YACb,EAAE,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE;SAClC,CAAC,CAAA;QAEF,4FAA4F;QAC5F,IAAI,CAAC,GAAG,KAAK,CAAC,MAAM,IAAI,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG;YAAE,CAAC,IAAI,CAAC,CAAA;IACjD,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,CAAA;AAC/C,CAAC"}
|
package/out/cjk.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"cjk.d.ts","sourceRoot":"","sources":["../cjk.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAIH,MAAM,WAAW,SAAS;IACzB,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,yFAAyF;IACzF,MAAM,EAAE,MAAM,CAAA;IACd,qDAAqD;IACrD,QAAQ,EAAE,MAAM,CAAA;CAChB;AAQD,wBAAgB,qBAAqB,CAAC,KAAK,EAAE,MAAM,GAAG,SAAS,
|
|
1
|
+
{"version":3,"file":"cjk.d.ts","sourceRoot":"","sources":["../cjk.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAIH,MAAM,WAAW,SAAS;IACzB,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,yFAAyF;IACzF,MAAM,EAAE,MAAM,CAAA;IACd,qDAAqD;IACrD,QAAQ,EAAE,MAAM,CAAA;CAChB;AAQD,wBAAgB,qBAAqB,CAAC,KAAK,EAAE,MAAM,GAAG,SAAS,CAsC9D"}
|
package/out/cjk.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"cjk.js","sourceRoot":"","sources":["../cjk.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAW7C,MAAM,eAAe,GAAG,MAAM,CAAA,CAAC,IAAI;AACnC,MAAM,aAAa,GAAG,MAAM,CAAA,CAAC,IAAI;AACjC,MAAM,kBAAkB,GAAG,MAAM,CAAA,CAAC,2BAA2B;AAC7D,MAAM,iBAAiB,GAAG,MAAM,CAAA;AAChC,MAAM,WAAW,GAAG,MAAM,CAAA,CAAC,IAAI;AAE/B,MAAM,UAAU,qBAAqB,CAAC,KAAa;IAClD,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,GAAG,GAAa,EAAE,CAAA;IAExB,gGAAgG;IAChG,oFAAoF;IACpF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAA;
|
|
1
|
+
{"version":3,"file":"cjk.js","sourceRoot":"","sources":["../cjk.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2BG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAW7C,MAAM,eAAe,GAAG,MAAM,CAAA,CAAC,IAAI;AACnC,MAAM,aAAa,GAAG,MAAM,CAAA,CAAC,IAAI;AACjC,MAAM,kBAAkB,GAAG,MAAM,CAAA,CAAC,2BAA2B;AAC7D,MAAM,iBAAiB,GAAG,MAAM,CAAA;AAChC,MAAM,WAAW,GAAG,MAAM,CAAA,CAAC,IAAI;AAE/B,MAAM,UAAU,qBAAqB,CAAC,KAAa;IAClD,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,IAAI,QAAQ,GAAG,CAAC,CAAA;IAChB,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,GAAG,GAAa,EAAE,CAAA;IAExB,gGAAgG;IAChG,oFAAoF;IACpF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,KAAK,CAAC,UAAU,CAAC,CAAC,CAAC,CAAA;QAEhC,IAAI,IAAI,KAAK,WAAW,EAAE,CAAC;YAC1B,QAAQ,IAAI,CAAC,CAAA;YACb,SAAQ,CAAC,yEAAyE;QACnF,CAAC;QAED,IAAI,IAAI,IAAI,eAAe,IAAI,IAAI,IAAI,aAAa,EAAE,CAAC;YACtD,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,YAAY,CAAC,IAAI,GAAG,kBAAkB,CAAC,CAAC,CAAA;YACxD,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACX,MAAM,IAAI,CAAC,CAAA;YACX,SAAQ;QACT,CAAC;QAED,IAAI,IAAI,KAAK,iBAAiB,EAAE,CAAC;YAChC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACb,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACX,MAAM,IAAI,CAAC,CAAA;YACX,SAAQ;QACT,CAAC;QACD,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC,CAAA;QACnB,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;IACZ,CAAC;IAED,IAAI,MAAM,KAAK,CAAC,IAAI,QAAQ,KAAK,CAAC,EAAE,CAAC;QACpC,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAA;IAC/E,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAA;AACrD,CAAC"}
|
package/out/compute.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"compute.d.ts","sourceRoot":"","sources":["../compute.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAOH,OAAO,KAAK,EAA0B,eAAe,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AAGxF,wBAAgB,SAAS,CAAC,GAAG,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,aAAa,GAAG,eAAe,
|
|
1
|
+
{"version":3,"file":"compute.d.ts","sourceRoot":"","sources":["../compute.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAOH,OAAO,KAAK,EAA0B,eAAe,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AAGxF,wBAAgB,SAAS,CAAC,GAAG,EAAE,MAAM,EAAE,IAAI,CAAC,EAAE,aAAa,GAAG,eAAe,CAiF5E"}
|
package/out/compute.js
CHANGED
|
@@ -8,7 +8,7 @@
|
|
|
8
8
|
*/
|
|
9
9
|
import { expandAbbreviations } from "./abbreviations.js";
|
|
10
10
|
import { applyCjkNormalization } from "./cjk.js";
|
|
11
|
-
import {
|
|
11
|
+
import { applyNFC } from "./nfc.js";
|
|
12
12
|
import { composeMaps, identityMap } from "./offset-map.js";
|
|
13
13
|
import { applyPunctuation } from "./punctuation.js";
|
|
14
14
|
import { collapseWhitespace } from "./whitespace.js";
|
|
@@ -17,8 +17,8 @@ export function normalize(raw, opts) {
|
|
|
17
17
|
let text = raw;
|
|
18
18
|
let map = identityMap(raw.length);
|
|
19
19
|
// 1. NFC
|
|
20
|
-
if (!opts?.
|
|
21
|
-
const r =
|
|
20
|
+
if (!opts?.skipNFC) {
|
|
21
|
+
const r = applyNFC(text);
|
|
22
22
|
text = r.text;
|
|
23
23
|
map = composeMaps(map, r.map);
|
|
24
24
|
transforms.push({ kind: "nfc", changed: r.changed });
|
package/out/compute.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"compute.js","sourceRoot":"","sources":["../compute.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,mBAAmB,EAAE,MAAM,oBAAoB,CAAA;AACxD,OAAO,EAAE,qBAAqB,EAAE,MAAM,UAAU,CAAA;AAChD,OAAO,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAA;AACnC,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAC1D,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AAEnD,OAAO,EAAE,kBAAkB,EAAE,MAAM,iBAAiB,CAAA;AAEpD,MAAM,UAAU,SAAS,CAAC,GAAW,EAAE,IAAoB;IAC1D,MAAM,UAAU,GAA6B,EAAE,CAAA;IAC/C,IAAI,IAAI,GAAG,GAAG,CAAA;IACd,IAAI,GAAG,GAAG,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IAEjC,SAAS;IACT,IAAI,CAAC,IAAI,EAAE,OAAO,EAAE,CAAC;QACpB,MAAM,CAAC,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAA;QACxB,IAAI,GAAG,CAAC,CAAC,IAAI,CAAA;QACb,GAAG,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAA;QAC7B,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAA;IACrD,CAAC;IAED,+FAA+F;IAC/F,+FAA+F;IAC/F,iGAAiG;IACjG,CAAC;QACA,MAAM,CAAC,GAAG,qBAAqB,CAAC,IAAI,CAAC,CAAA;
|
|
1
|
+
{"version":3,"file":"compute.js","sourceRoot":"","sources":["../compute.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,mBAAmB,EAAE,MAAM,oBAAoB,CAAA;AACxD,OAAO,EAAE,qBAAqB,EAAE,MAAM,UAAU,CAAA;AAChD,OAAO,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAA;AACnC,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAC1D,OAAO,EAAE,gBAAgB,EAAE,MAAM,kBAAkB,CAAA;AAEnD,OAAO,EAAE,kBAAkB,EAAE,MAAM,iBAAiB,CAAA;AAEpD,MAAM,UAAU,SAAS,CAAC,GAAW,EAAE,IAAoB;IAC1D,MAAM,UAAU,GAA6B,EAAE,CAAA;IAC/C,IAAI,IAAI,GAAG,GAAG,CAAA;IACd,IAAI,GAAG,GAAG,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,CAAA;IAEjC,SAAS;IACT,IAAI,CAAC,IAAI,EAAE,OAAO,EAAE,CAAC;QACpB,MAAM,CAAC,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAA;QACxB,IAAI,GAAG,CAAC,CAAC,IAAI,CAAA;QACb,GAAG,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAA;QAC7B,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,CAAA;IACrD,CAAC;IAED,+FAA+F;IAC/F,+FAA+F;IAC/F,iGAAiG;IACjG,CAAC;QACA,MAAM,CAAC,GAAG,qBAAqB,CAAC,IAAI,CAAC,CAAA;QAErC,IAAI,CAAC,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,CAAC,QAAQ,GAAG,CAAC,EAAE,CAAC;YACpC,IAAI,GAAG,CAAC,CAAC,IAAI,CAAA;YACb,GAAG,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAA;YAC7B,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,eAAe,EAAE,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,CAAC,CAAA;QACnF,CAAC;IACF,CAAC;IAED,iBAAiB;IACjB,CAAC;QACA,MAAM,CAAC,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAA;QAEhC,IAAI,CAAC,CAAC,YAAY,GAAG,CAAC,EAAE,CAAC;YACxB,IAAI,GAAG,CAAC,CAAC,IAAI,CAAA;YACb,GAAG,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAA;YAC7B,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,uBAAuB,EAAE,YAAY,EAAE,CAAC,CAAC,YAAY,EAAE,CAAC,CAAA;QACjF,CAAC;IACF,CAAC;IAED,gBAAgB;IAChB,CAAC;QACA,MAAM,CAAC,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAA;QAElC,IAAI,CAAC,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,KAAK,IAAI,CAAC,MAAM,EAAE,CAAC;YACjD,IAAI,GAAG,CAAC,CAAC,IAAI,CAAA;YACb,GAAG,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAA;YAC7B,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,qBAAqB,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;QAC/D,CAAC;IACF,CAAC;IAED,2FAA2F;IAC3F,gEAAgE;IAChE,IAAI,IAAI,EAAE,mBAAmB,EAAE,CAAC;QAC/B,MAAM,CAAC,GAAG,mBAAmB,CAAC,IAAI,EAAE,IAAI,CAAC,MAAM,CAAC,CAAA;QAEhD,IAAI,CAAC,CAAC,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YAC7B,IAAI,GAAG,CAAC,CAAC,IAAI,CAAA;YACb,GAAG,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAA;YAE7B,KAAK,MAAM,CAAC,IAAI,CAAC,CAAC,UAAU,EAAE,CAAC;gBAC9B,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,qBAAqB,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,CAAA;YACnF,CAAC;QACF,CAAC;IACF,CAAC;IAED,wBAAwB;IACxB,IAAI,IAAI,EAAE,QAAQ,EAAE,CAAC;QACpB,MAAM,EAAE,GAAG,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;QAE9C,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YACjB,IAAI,GAAG,EAAE,CAAA;YACT,sEAAsE;YACtE,UAAU,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,MAAM,EAAE,IAAI,CAAC,MAAM,IAAI,KAAK,EAAE,CAAC,CAAA;QACrE,CAAC;IACF,CAAC;IAED,OAAO,MAAM,CAAC,MAAM,CAAC;QACpB,GAAG;QACH,UAAU,EAAE,IAAI;QAChB,UAAU,EAAE,MAAM,CAAC,MAAM,CAAC,UAAU,CAA6B;QACjE,SAAS,EAAE,GAAG;QACd,aAAa,EAAE,IAAI,EAAE,MAAM;KAC3B,CAA2B,CAAA;AAC7B,CAAC"}
|
package/out/index.d.ts
CHANGED
|
@@ -15,7 +15,7 @@
|
|
|
15
15
|
export { expandAbbreviations } from "./abbreviations.js";
|
|
16
16
|
export { applyCjkNormalization, type CjkResult } from "./cjk.js";
|
|
17
17
|
export { normalize } from "./compute.js";
|
|
18
|
-
export {
|
|
18
|
+
export { applyNFC } from "./nfc.js";
|
|
19
19
|
export { composeMaps, identityMap } from "./offset-map.js";
|
|
20
20
|
export { applyPunctuation } from "./punctuation.js";
|
|
21
21
|
export type { NormalizationTransform, NormalizeOpts, NormalizedInput, SpanRange } from "./types.js";
|
package/out/index.js
CHANGED
|
@@ -15,7 +15,7 @@
|
|
|
15
15
|
export { expandAbbreviations } from "./abbreviations.js";
|
|
16
16
|
export { applyCjkNormalization } from "./cjk.js";
|
|
17
17
|
export { normalize } from "./compute.js";
|
|
18
|
-
export {
|
|
18
|
+
export { applyNFC } from "./nfc.js";
|
|
19
19
|
export { composeMaps, identityMap } from "./offset-map.js";
|
|
20
20
|
export { applyPunctuation } from "./punctuation.js";
|
|
21
21
|
export { collapseWhitespace } from "./whitespace.js";
|
package/out/nfc.d.ts
CHANGED
|
@@ -11,11 +11,11 @@
|
|
|
11
11
|
* map each output index to the start of its source sequence. Rare CJK edge cases involving
|
|
12
12
|
* variant selectors may produce off-by-one offsets — acceptable for v1.
|
|
13
13
|
*/
|
|
14
|
-
export interface
|
|
14
|
+
export interface NFCResult {
|
|
15
15
|
text: string;
|
|
16
16
|
/** `text[i]` came from `input[map[i]]`. */
|
|
17
17
|
map: number[];
|
|
18
18
|
changed: boolean;
|
|
19
19
|
}
|
|
20
|
-
export declare function
|
|
20
|
+
export declare function applyNFC(input: string): NFCResult;
|
|
21
21
|
//# sourceMappingURL=nfc.d.ts.map
|
package/out/nfc.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"nfc.d.ts","sourceRoot":"","sources":["../nfc.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAIH,MAAM,WAAW,SAAS;IACzB,IAAI,EAAE,MAAM,CAAA;IACZ,2CAA2C;IAC3C,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,OAAO,EAAE,OAAO,CAAA;CAChB;AAED,wBAAgB,QAAQ,CAAC,KAAK,EAAE,MAAM,GAAG,SAAS,
|
|
1
|
+
{"version":3,"file":"nfc.d.ts","sourceRoot":"","sources":["../nfc.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAIH,MAAM,WAAW,SAAS;IACzB,IAAI,EAAE,MAAM,CAAA;IACZ,2CAA2C;IAC3C,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,OAAO,EAAE,OAAO,CAAA;CAChB;AAED,wBAAgB,QAAQ,CAAC,KAAK,EAAE,MAAM,GAAG,SAAS,CAQjD"}
|
package/out/nfc.js
CHANGED
|
@@ -12,19 +12,19 @@
|
|
|
12
12
|
* variant selectors may produce off-by-one offsets — acceptable for v1.
|
|
13
13
|
*/
|
|
14
14
|
import { identityMap } from "./offset-map.js";
|
|
15
|
-
export function
|
|
15
|
+
export function applyNFC(input) {
|
|
16
16
|
const normalized = input.normalize("NFC");
|
|
17
17
|
if (normalized === input) {
|
|
18
18
|
return { text: input, map: identityMap(input.length), changed: false };
|
|
19
19
|
}
|
|
20
|
-
return { text: normalized, map:
|
|
20
|
+
return { text: normalized, map: estimateNFCMap(input, normalized), changed: true };
|
|
21
21
|
}
|
|
22
22
|
/**
|
|
23
|
-
* Estimate per-output-codepoint offsets. Walks both strings in parallel; emits the next source
|
|
24
|
-
*
|
|
25
|
-
*
|
|
23
|
+
* Estimate per-output-codepoint offsets. Walks both strings in parallel; emits the next source index for each output
|
|
24
|
+
* position. Imprecise for combining sequences but correct for length-equal NFC outputs (the common length-changing case
|
|
25
|
+
* is when a sequence shortens).
|
|
26
26
|
*/
|
|
27
|
-
function
|
|
27
|
+
function estimateNFCMap(input, output) {
|
|
28
28
|
const map = [];
|
|
29
29
|
let inIdx = 0;
|
|
30
30
|
for (let outIdx = 0; outIdx < output.length; outIdx++) {
|
package/out/nfc.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"nfc.js","sourceRoot":"","sources":["../nfc.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAS7C,MAAM,UAAU,QAAQ,CAAC,KAAa;IACrC,MAAM,UAAU,GAAG,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAA;
|
|
1
|
+
{"version":3,"file":"nfc.js","sourceRoot":"","sources":["../nfc.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;GAYG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAS7C,MAAM,UAAU,QAAQ,CAAC,KAAa;IACrC,MAAM,UAAU,GAAG,KAAK,CAAC,SAAS,CAAC,KAAK,CAAC,CAAA;IAEzC,IAAI,UAAU,KAAK,KAAK,EAAE,CAAC;QAC1B,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,CAAA;IACvE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,UAAU,EAAE,GAAG,EAAE,cAAc,CAAC,KAAK,EAAE,UAAU,CAAC,EAAE,OAAO,EAAE,IAAI,EAAE,CAAA;AACnF,CAAC;AAED;;;;GAIG;AACH,SAAS,cAAc,CAAC,KAAa,EAAE,MAAc;IACpD,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,IAAI,KAAK,GAAG,CAAC,CAAA;IAEb,KAAK,IAAI,MAAM,GAAG,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,EAAE,EAAE,CAAC;QACvD,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QACf,MAAM,KAAK,GAAG,MAAM,CAAC,WAAW,CAAC,MAAM,CAAE,CAAA;QACzC,MAAM,OAAO,GAAG,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;QAEtC,gGAAgG;QAChG,IAAI,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YAC1B,MAAM,IAAI,GAAG,KAAK,CAAC,WAAW,CAAC,KAAK,CAAE,CAAA;YACtC,KAAK,IAAI,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;YAE9B,OAAO,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;gBAC7B,MAAM,MAAM,GAAG,KAAK,CAAC,WAAW,CAAC,KAAK,CAAE,CAAA;gBAExC,IAAI,MAAM,IAAI,MAAM,IAAI,MAAM,IAAI,MAAM,EAAE,CAAC;oBAC1C,KAAK,IAAI,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;gBACjC,CAAC;qBAAM,CAAC;oBACP,MAAK;gBACN,CAAC;YACF,CAAC;QACF,CAAC;QAED,IAAI,OAAO,KAAK,CAAC;YAAE,MAAM,IAAI,CAAC,CAAA;IAC/B,CAAC;IAED,OAAO,GAAG,CAAA;AACX,CAAC"}
|
package/out/offset-map.d.ts
CHANGED
|
@@ -8,13 +8,13 @@
|
|
|
8
8
|
/** Identity map for an input of length `n`: `[0, 1, 2, ..., n-1]`. */
|
|
9
9
|
export declare function identityMap(n: number): number[];
|
|
10
10
|
/**
|
|
11
|
-
* Compose `inputMap` (input → raw) with `transformMap` (output → input) to produce `outputMap`
|
|
12
|
-
* (output → raw).
|
|
11
|
+
* Compose `inputMap` (input → raw) with `transformMap` (output → input) to produce `outputMap` (output → raw).
|
|
13
12
|
*
|
|
14
|
-
* @example
|
|
15
|
-
* raw
|
|
16
|
-
*
|
|
17
|
-
*
|
|
13
|
+
* @example
|
|
14
|
+
* // raw = "350 5th" (chars 0..7, double space at 3-4) // input = "350 5th" (identity from
|
|
15
|
+
* raw, length 8) // output = "350 5th" (whitespace collapsed, length 7) // inputMap =
|
|
16
|
+
* [0,1,2,3,4,5,6,7] // transformMap = [0,1,2,3,5,6,7] (output[3]=' ' came from input[3];
|
|
17
|
+
* output[4]='5' from input[5]) // composed = [0,1,2,3,5,6,7]
|
|
18
18
|
*/
|
|
19
19
|
export declare function composeMaps(inputMap: number[], transformMap: number[]): number[];
|
|
20
20
|
//# sourceMappingURL=offset-map.d.ts.map
|
package/out/offset-map.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"offset-map.d.ts","sourceRoot":"","sources":["../offset-map.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,sEAAsE;AACtE,wBAAgB,WAAW,CAAC,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,
|
|
1
|
+
{"version":3,"file":"offset-map.d.ts","sourceRoot":"","sources":["../offset-map.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,sEAAsE;AACtE,wBAAgB,WAAW,CAAC,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,CAM/C;AAED;;;;;;;;GAQG;AACH,wBAAgB,WAAW,CAAC,QAAQ,EAAE,MAAM,EAAE,EAAE,YAAY,EAAE,MAAM,EAAE,GAAG,MAAM,EAAE,CAShF"}
|
package/out/offset-map.js
CHANGED
|
@@ -13,13 +13,13 @@ export function identityMap(n) {
|
|
|
13
13
|
return m;
|
|
14
14
|
}
|
|
15
15
|
/**
|
|
16
|
-
* Compose `inputMap` (input → raw) with `transformMap` (output → input) to produce `outputMap`
|
|
17
|
-
* (output → raw).
|
|
16
|
+
* Compose `inputMap` (input → raw) with `transformMap` (output → input) to produce `outputMap` (output → raw).
|
|
18
17
|
*
|
|
19
|
-
* @example
|
|
20
|
-
* raw
|
|
21
|
-
*
|
|
22
|
-
*
|
|
18
|
+
* @example
|
|
19
|
+
* // raw = "350 5th" (chars 0..7, double space at 3-4) // input = "350 5th" (identity from
|
|
20
|
+
* raw, length 8) // output = "350 5th" (whitespace collapsed, length 7) // inputMap =
|
|
21
|
+
* [0,1,2,3,4,5,6,7] // transformMap = [0,1,2,3,5,6,7] (output[3]=' ' came from input[3];
|
|
22
|
+
* output[4]='5' from input[5]) // composed = [0,1,2,3,5,6,7]
|
|
23
23
|
*/
|
|
24
24
|
export function composeMaps(inputMap, transformMap) {
|
|
25
25
|
const out = new Array(transformMap.length);
|
package/out/offset-map.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"offset-map.js","sourceRoot":"","sources":["../offset-map.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,sEAAsE;AACtE,MAAM,UAAU,WAAW,CAAC,CAAS;IACpC,MAAM,CAAC,GAAG,IAAI,KAAK,CAAS,CAAC,CAAC,CAAA;
|
|
1
|
+
{"version":3,"file":"offset-map.js","sourceRoot":"","sources":["../offset-map.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,sEAAsE;AACtE,MAAM,UAAU,WAAW,CAAC,CAAS;IACpC,MAAM,CAAC,GAAG,IAAI,KAAK,CAAS,CAAC,CAAC,CAAA;IAE9B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE;QAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAA;IAEpC,OAAO,CAAC,CAAA;AACT,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,WAAW,CAAC,QAAkB,EAAE,YAAsB;IACrE,MAAM,GAAG,GAAG,IAAI,KAAK,CAAS,YAAY,CAAC,MAAM,CAAC,CAAA;IAElD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,CAAC,GAAG,YAAY,CAAC,CAAC,CAAE,CAAA;QAC1B,GAAG,CAAC,CAAC,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;IAC1B,CAAC;IAED,OAAO,GAAG,CAAA;AACX,CAAC"}
|
package/out/punctuation.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"punctuation.d.ts","sourceRoot":"","sources":["../punctuation.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAiBH,MAAM,WAAW,iBAAiB;IACjC,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,YAAY,EAAE,MAAM,CAAA;CACpB;AAED,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,GAAG,iBAAiB,
|
|
1
|
+
{"version":3,"file":"punctuation.d.ts","sourceRoot":"","sources":["../punctuation.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAiBH,MAAM,WAAW,iBAAiB;IACjC,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,YAAY,EAAE,MAAM,CAAA;CACpB;AAED,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,GAAG,iBAAiB,CA6BjE"}
|
package/out/punctuation.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"punctuation.js","sourceRoot":"","sources":["../punctuation.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAE7C,MAAM,YAAY,GAAG,IAAI,GAAG,CAAiB;IAC5C,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,YAAY;IACxB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,YAAY;IACxB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,+EAA+E;IAC3F,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,8DAA8D;IAC1E,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,+BAA+B;IAC7C,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,qBAAqB;CACjC,CAAC,CAAA;AAQF,MAAM,UAAU,gBAAgB,CAAC,KAAa;IAC7C,IAAI,OAAO,GAAG,KAAK,CAAA;IACnB,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,GAAG,GAAa,EAAE,CAAA;IAExB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;QACpB,MAAM,GAAG,GAAG,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAA;
|
|
1
|
+
{"version":3,"file":"punctuation.js","sourceRoot":"","sources":["../punctuation.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAE7C,MAAM,YAAY,GAAG,IAAI,GAAG,CAAiB;IAC5C,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,IAAI;IAChB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,YAAY;IACxB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,YAAY;IACxB,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,+EAA+E;IAC3F,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,8DAA8D;IAC1E,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,+BAA+B;IAC7C,CAAC,GAAG,EAAE,GAAG,CAAC,EAAE,qBAAqB;CACjC,CAAC,CAAA;AAQF,MAAM,UAAU,gBAAgB,CAAC,KAAa;IAC7C,IAAI,OAAO,GAAG,KAAK,CAAA;IACnB,IAAI,YAAY,GAAG,CAAC,CAAA;IACpB,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,GAAG,GAAa,EAAE,CAAA;IAExB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;QACpB,MAAM,GAAG,GAAG,YAAY,CAAC,GAAG,CAAC,EAAE,CAAC,CAAA;QAEhC,IAAI,GAAG,KAAK,SAAS,EAAE,CAAC;YACvB,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YACZ,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACZ,CAAC;aAAM,CAAC;YACP,OAAO,GAAG,IAAI,CAAA;YACd,YAAY,IAAI,CAAC,CAAA;YAEjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACrC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAE,CAAC,CAAA;gBACjB,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACZ,CAAC;QACF,CAAC;IACF,CAAC;IAED,IAAI,CAAC,OAAO,EAAE,CAAC;QACd,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,YAAY,EAAE,CAAC,EAAE,CAAA;IACxE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,YAAY,EAAE,CAAA;AACjD,CAAC"}
|
package/out/types.d.ts
CHANGED
|
@@ -34,9 +34,9 @@ export type NormalizationTransform = {
|
|
|
34
34
|
/**
|
|
35
35
|
* Result of running `normalize()` on a raw input string.
|
|
36
36
|
*
|
|
37
|
-
* `offsetMap[i]` is the index in `raw` from which `normalized[i]` came. For multi-character source
|
|
38
|
-
*
|
|
39
|
-
*
|
|
37
|
+
* `offsetMap[i]` is the index in `raw` from which `normalized[i]` came. For multi-character source sequences (NFC
|
|
38
|
+
* composition, whitespace collapse, abbreviation expansion), each output char points to the FIRST source char by
|
|
39
|
+
* convention.
|
|
40
40
|
*/
|
|
41
41
|
export interface NormalizedInput {
|
|
42
42
|
/** The input as the caller sent it. */
|
|
@@ -58,6 +58,6 @@ export interface NormalizeOpts {
|
|
|
58
58
|
/** Expand known abbreviations (`St` → `Street`, `NW` → `Northwest`, etc.). Default: false. */
|
|
59
59
|
expandAbbreviations?: boolean;
|
|
60
60
|
/** Skip Unicode NFC. Only use for debugging — production callers should leave on. */
|
|
61
|
-
|
|
61
|
+
skipNFC?: boolean;
|
|
62
62
|
}
|
|
63
63
|
//# sourceMappingURL=types.d.ts.map
|
package/out/whitespace.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"whitespace.d.ts","sourceRoot":"","sources":["../whitespace.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAOH,MAAM,WAAW,gBAAgB;IAChC,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,IAAI,EAAE,MAAM,CAAA;CACZ;AAED,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,MAAM,GAAG,gBAAgB,
|
|
1
|
+
{"version":3,"file":"whitespace.d.ts","sourceRoot":"","sources":["../whitespace.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAOH,MAAM,WAAW,gBAAgB;IAChC,IAAI,EAAE,MAAM,CAAA;IACZ,GAAG,EAAE,MAAM,EAAE,CAAA;IACb,IAAI,EAAE,MAAM,CAAA;CACZ;AAED,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,MAAM,GAAG,gBAAgB,CA6DlE"}
|
package/out/whitespace.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"whitespace.js","sourceRoot":"","sources":["../whitespace.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAE7C,MAAM,YAAY,GAAG,OAAO,CAAA;AAC5B,MAAM,SAAS,GAAG,WAAW,CAAA;AAQ7B,MAAM,UAAU,kBAAkB,CAAC,KAAa;IAC/C,IAAI,OAAO,GAAG,KAAK,CAAA;IACnB,IAAI,IAAI,GAAG,CAAC,CAAA;IACZ,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,IAAI,CAAC,GAAG,CAAC,CAAA;IAET,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QACzB,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;
|
|
1
|
+
{"version":3,"file":"whitespace.js","sourceRoot":"","sources":["../whitespace.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,OAAO,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAA;AAE7C,MAAM,YAAY,GAAG,OAAO,CAAA;AAC5B,MAAM,SAAS,GAAG,WAAW,CAAA;AAQ7B,MAAM,UAAU,kBAAkB,CAAC,KAAa;IAC/C,IAAI,OAAO,GAAG,KAAK,CAAA;IACnB,IAAI,IAAI,GAAG,CAAC,CAAA;IACZ,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,MAAM,GAAG,GAAa,EAAE,CAAA;IACxB,IAAI,CAAC,GAAG,CAAC,CAAA;IAET,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QACzB,MAAM,EAAE,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;QAEpB,IAAI,EAAE,KAAK,IAAI,IAAI,EAAE,KAAK,IAAI,EAAE,CAAC;YAChC,2CAA2C;YAC3C,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;YACZ,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACX,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACT,CAAC;QAED,IAAI,YAAY,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,CAAC;YAC3B,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;YACb,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACX,MAAM,KAAK,GAAG,CAAC,CAAA;YACf,CAAC,IAAI,CAAC,CAAA;YAEN,OAAO,CAAC,GAAG,KAAK,CAAC,MAAM,IAAI,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAE,CAAC;gBAAE,CAAC,IAAI,CAAC,CAAA;YAE/D,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,EAAE,CAAC;gBACnB,OAAO,GAAG,IAAI,CAAA;gBACd,IAAI,IAAI,CAAC,CAAA;YACV,CAAC;YACD,SAAQ;QACT,CAAC;QAED,yBAAyB;QACzB,IAAI,EAAE,KAAK,IAAI,IAAI,GAAG,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;YACjD,2EAA2E;QAC5E,CAAC;QACD,GAAG,CAAC,IAAI,CAAC,EAAE,CAAC,CAAA;QACZ,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;QACX,CAAC,IAAI,CAAC,CAAA;IACP,CAAC;IAED,wCAAwC;IACxC,IAAI,IAAI,GAAG,CAAC,CAAA;IAEZ,OAAO,IAAI,GAAG,GAAG,CAAC,MAAM,IAAI,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAE,CAAC;QAAE,IAAI,IAAI,CAAC,CAAA;IACjE,IAAI,KAAK,GAAG,GAAG,CAAC,MAAM,CAAA;IAEtB,OAAO,KAAK,GAAG,IAAI,IAAI,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAE,CAAC;QAAE,KAAK,IAAI,CAAC,CAAA;IAElE,IAAI,IAAI,GAAG,CAAC,IAAI,KAAK,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QACpC,OAAO,GAAG,IAAI,CAAA;IACf,CAAC;IACD,MAAM,UAAU,GAAG,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;IACzC,MAAM,UAAU,GAAG,GAAG,CAAC,KAAK,CAAC,IAAI,EAAE,KAAK,CAAC,CAAA;IAEzC,IAAI,CAAC,OAAO,IAAI,UAAU,CAAC,MAAM,KAAK,KAAK,CAAC,MAAM,EAAE,CAAC;QACpD,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,CAAA;IAChE,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,CAAA;AAC5D,CAAC"}
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@mailwoman/normalize",
|
|
3
|
-
"version": "
|
|
3
|
+
"version": "5.0.0",
|
|
4
4
|
"description": "Stage 1 of the runtime pipeline — deterministic input preprocessing (Unicode NFC, punctuation, whitespace, abbreviation). Pure functions, no ML.",
|
|
5
5
|
"license": "AGPL-3.0-only",
|
|
6
6
|
"repository": {
|
|
@@ -8,17 +8,17 @@
|
|
|
8
8
|
"url": "https://github.com/sister-software/mailwoman.git",
|
|
9
9
|
"directory": "normalize"
|
|
10
10
|
},
|
|
11
|
-
"type": "module",
|
|
12
|
-
"exports": {
|
|
13
|
-
"./package.json": "./package.json",
|
|
14
|
-
".": "./out/index.js"
|
|
15
|
-
},
|
|
16
11
|
"files": [
|
|
17
12
|
"out/**/*.js",
|
|
18
13
|
"out/**/*.js.map",
|
|
19
14
|
"out/**/*.d.ts",
|
|
20
15
|
"out/**/*.d.ts.map"
|
|
21
16
|
],
|
|
17
|
+
"type": "module",
|
|
18
|
+
"exports": {
|
|
19
|
+
"./package.json": "./package.json",
|
|
20
|
+
".": "./out/index.js"
|
|
21
|
+
},
|
|
22
22
|
"publishConfig": {
|
|
23
23
|
"access": "public"
|
|
24
24
|
}
|