echogarden 0.3.0 → 0.3.1
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.
|
@@ -9,23 +9,23 @@ export function getNormalizationMapForSpeech(words, language) {
|
|
|
9
9
|
const fourDigitYearPattern = /^[0-9][0-9][0-9][0-9]$/;
|
|
10
10
|
const fourDigitDecadePattern = /^[0-9][0-9][0-9]0s$/;
|
|
11
11
|
const fourDigitYearRangePattern = /^[0-9][0-9][0-9][0-9][\-\–][0-9][0-9][0-9][0-9]$/;
|
|
12
|
-
const
|
|
12
|
+
const wordsPrecedingAYear = [
|
|
13
13
|
"in", "since", "©",
|
|
14
14
|
"january", "february", "march", "april", "may", "june", "july", "august", "september", "october", "november", "december",
|
|
15
15
|
"jan", "feb", "mar", "apr", "may", "jun", "jul", "aug", "sep", "oct", "nov", "dec"
|
|
16
16
|
];
|
|
17
17
|
for (let wordIndex = 0; wordIndex < words.length; wordIndex++) {
|
|
18
18
|
const word = words[wordIndex];
|
|
19
|
-
const
|
|
19
|
+
const lowerCaseWord = word.toLocaleLowerCase();
|
|
20
20
|
const nextWords = words.slice(wordIndex + 1);
|
|
21
21
|
const nextWord = nextWords[0];
|
|
22
|
-
if (
|
|
22
|
+
if (wordsPrecedingAYear.includes(lowerCaseWord) &&
|
|
23
23
|
fourDigitYearPattern.test(nextWord)) {
|
|
24
24
|
const normalizedString = normalizeFourDigitYearString(nextWord);
|
|
25
25
|
normalizationMap.set(wordIndex + 1, normalizedString);
|
|
26
26
|
wordIndex += 1;
|
|
27
27
|
}
|
|
28
|
-
else if (
|
|
28
|
+
else if (lowerCaseWord == "the" &&
|
|
29
29
|
fourDigitDecadePattern.test(nextWord)) {
|
|
30
30
|
const normalizedString = normalizeFourDigitDecadeString(nextWord);
|
|
31
31
|
normalizationMap.set(wordIndex + 1, normalizedString);
|
|
@@ -56,10 +56,10 @@ export function normalizeFourDigitYearString(yearString) {
|
|
|
56
56
|
return normalizedString;
|
|
57
57
|
}
|
|
58
58
|
export function normalizeFourDigitDecadeString(decadeString) {
|
|
59
|
-
const firstTwoDigitsValue =
|
|
60
|
-
const secondTwoDigitsValue =
|
|
59
|
+
const firstTwoDigitsValue = parseInt(decadeString.substring(0, 2));
|
|
60
|
+
const secondTwoDigitsValue = parseInt(decadeString.substring(2, 4));
|
|
61
61
|
let normalizedString;
|
|
62
|
-
if (firstTwoDigitsValue >= 10 && firstTwoDigitsValue
|
|
62
|
+
if (firstTwoDigitsValue >= 10 && firstTwoDigitsValue >= 10) {
|
|
63
63
|
normalizedString = `${firstTwoDigitsValue} ${secondTwoDigitsValue}s`;
|
|
64
64
|
}
|
|
65
65
|
else {
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"TextNormalizer.js","sourceRoot":"","sources":["../../src/nlp/TextNormalizer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,MAAM,wBAAwB,CAAA;AAE7D,MAAM,UAAU,4BAA4B,CAAC,KAAe,EAAE,QAAgB;IAC7E,QAAQ,GAAG,oBAAoB,CAAC,QAAQ,CAAC,CAAA;IAEzC,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAkB,CAAA;IAElD,IAAI,QAAQ,IAAI,IAAI,EAAE;QACrB,OAAO,gBAAgB,CAAA;KACvB;IAED,MAAM,aAAa,GAAG,mBAAmB,CAAA;IAEzC,MAAM,oBAAoB,GAAG,wBAAwB,CAAA;IACrD,MAAM,sBAAsB,GAAG,qBAAqB,CAAA;IAEpD,MAAM,yBAAyB,GAAG,kDAAkD,CAAA;IAEpF,MAAM,
|
|
1
|
+
{"version":3,"file":"TextNormalizer.js","sourceRoot":"","sources":["../../src/nlp/TextNormalizer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,oBAAoB,EAAE,MAAM,wBAAwB,CAAA;AAE7D,MAAM,UAAU,4BAA4B,CAAC,KAAe,EAAE,QAAgB;IAC7E,QAAQ,GAAG,oBAAoB,CAAC,QAAQ,CAAC,CAAA;IAEzC,MAAM,gBAAgB,GAAG,IAAI,GAAG,EAAkB,CAAA;IAElD,IAAI,QAAQ,IAAI,IAAI,EAAE;QACrB,OAAO,gBAAgB,CAAA;KACvB;IAED,MAAM,aAAa,GAAG,mBAAmB,CAAA;IAEzC,MAAM,oBAAoB,GAAG,wBAAwB,CAAA;IACrD,MAAM,sBAAsB,GAAG,qBAAqB,CAAA;IAEpD,MAAM,yBAAyB,GAAG,kDAAkD,CAAA;IAEpF,MAAM,mBAAmB,GAAG;QAC3B,IAAI,EAAE,OAAO,EAAE,GAAG;QAClB,SAAS,EAAE,UAAU,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,WAAW,EAAE,SAAS,EAAE,UAAU,EAAE,UAAU;QACxH,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK;KAClF,CAAA;IAED,KAAK,IAAI,SAAS,GAAG,CAAC,EAAE,SAAS,GAAG,KAAK,CAAC,MAAM,EAAE,SAAS,EAAE,EAAE;QAC9D,MAAM,IAAI,GAAG,KAAK,CAAC,SAAS,CAAC,CAAA;QAC7B,MAAM,aAAa,GAAG,IAAI,CAAC,iBAAiB,EAAE,CAAA;QAE9C,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC,CAAA;QAC5C,MAAM,QAAQ,GAAG,SAAS,CAAC,CAAC,CAAC,CAAA;QAE7B,IACC,mBAAmB,CAAC,QAAQ,CAAC,aAAa,CAAC;YAC3C,oBAAoB,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE;YACrC,MAAM,gBAAgB,GAAG,4BAA4B,CAAC,QAAQ,CAAC,CAAA;YAE/D,gBAAgB,CAAC,GAAG,CAAC,SAAS,GAAG,CAAC,EAAE,gBAAgB,CAAC,CAAA;YAErD,SAAS,IAAI,CAAC,CAAA;SACd;aAAM,IACN,aAAa,IAAI,KAAK;YACtB,sBAAsB,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE;YAEvC,MAAM,gBAAgB,GAAG,8BAA8B,CAAC,QAAQ,CAAC,CAAA;YAEjE,gBAAgB,CAAC,GAAG,CAAC,SAAS,GAAG,CAAC,EAAE,gBAAgB,CAAC,CAAA;YAErD,SAAS,IAAI,CAAC,CAAA;SACd;aAAM,IAAI,yBAAyB,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE;YAChD,MAAM,eAAe,GAAG,4BAA4B,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;YAC1E,MAAM,aAAa,GAAG,4BAA4B,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;YAExE,MAAM,gBAAgB,GAAG,GAAG,eAAe,OAAO,aAAa,EAAE,CAAA;YAEjE,gBAAgB,CAAC,GAAG,CAAC,SAAS,EAAE,gBAAgB,CAAC,CAAA;SACjD;KACD;IAED,OAAO,gBAAgB,CAAA;AACxB,CAAC;AAED,MAAM,UAAU,4BAA4B,CAAC,UAAkB;IAC9D,MAAM,mBAAmB,GAAG,UAAU,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAClE,MAAM,oBAAoB,GAAG,UAAU,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAEnE,IAAI,gBAAwB,CAAA;IAE5B,IAAI,mBAAmB,IAAI,EAAE,IAAI,oBAAoB,IAAI,EAAE,EAAE;QAC5D,gBAAgB,GAAG,GAAG,mBAAmB,IAAI,oBAAoB,EAAE,CAAA;KACnE;SAAM,IAAI,mBAAmB,IAAI,EAAE,IAAI,mBAAmB,GAAG,EAAE,IAAI,CAAC,IAAI,oBAAoB,GAAG,EAAE,EAAE;QACnG,gBAAgB,GAAG,GAAG,mBAAmB,OAAO,oBAAoB,EAAE,CAAA;KACtE;SAAM;QACN,gBAAgB,GAAG,UAAU,CAAA;KAC7B;IAED,OAAO,gBAAgB,CAAA;AACxB,CAAC;AAED,MAAM,UAAU,8BAA8B,CAAC,YAAoB;IAClE,MAAM,mBAAmB,GAAG,QAAQ,CAAC,YAAY,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAClE,MAAM,oBAAoB,GAAG,QAAQ,CAAC,YAAY,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAEnE,IAAI,gBAAwB,CAAA;IAE5B,IAAI,mBAAmB,IAAI,EAAE,IAAI,mBAAmB,IAAI,EAAE,EAAE;QAC3D,gBAAgB,GAAG,GAAG,mBAAmB,IAAI,oBAAoB,GAAG,CAAA;KACpE;SAAM;QACN,gBAAgB,GAAG,YAAY,CAAA;KAC/B;IAED,OAAO,gBAAgB,CAAA;AACxB,CAAC"}
|