@microsoft/recognizers-text-sequence - npm Package Compare versions

Comparing version 1.1.4 to 1.2.7

dist/types/recognizers-text-sequence.d.ts

@@ -5,3 +5,4 @@ export { default as SequenceRecognizer, SequenceOptions, recognizePhoneNumber, recognizeIpAddress, recognizeMention, recognizeHashtag, recognizeEmail, recognizeURL, recognizeGUID } from "./sequence/sequenceRecognizer";
 export { BaseSequenceExtractor, BasePhoneNumberExtractor, BaseIpExtractor, BaseMentionExtractor, BaseHashtagExtractor, BaseEmailExtractor, BaseURLExtractor, BaseGUIDExtractor } from "./sequence/extractors";
-export { PhoneNumberExtractor, IpExtractor, MentionExtractor, HashtagExtractor, EmailExtractor, URLExtractor, GUIDExtractor } from "./sequence/english/extractors";
+export { EnglishPhoneNumberExtractorConfiguration, EnglishIpExtractorConfiguration, MentionExtractor, HashtagExtractor, EmailExtractor, EnglishURLExtractorConfiguration, GUIDExtractor } from "./sequence/english/extractors";
+export { ChineseURLExtractorConfiguration, ChinesePhoneNumberExtractorConfiguration, ChineseIpExtractorConfiguration } from "./sequence/chinese/extractors";
 export { BaseSequenceParser, BaseIpParser } from "./sequence/parsers";
@@ -8,0 +9,0 @@ export { PhoneNumberParser, IpParser, MentionParser, HashtagParser, EmailParser, URLParser, GUIDParser } from "./sequence/english/parsers";

dist/types/resources/baseGUID.d.ts

 export declare namespace BaseGUID {
-    const GUIDRegexElement = "([A-Fa-f0-9]{8}(-[A-Fa-f0-9]{4}){3}-[A-Fa-f0-9]{12})";
+    const GUIDRegexElement = "(([a-f0-9]{8}(-[a-f0-9]{4}){3}-[a-f0-9]{12})|([a-f0-9]{32}))";
     const GUIDRegex: string;
 }

dist/types/resources/basePhoneNumbers.d.ts

 export declare namespace BasePhoneNumbers {
     const NumberReplaceToken = "@builtin.phonenumber";
-    const GeneralPhoneNumberRegex = "(\\b(((\\d[\\s]?){7,15}\\d))(?!-)\\b)|(\\(\\d{5}\\)\\s?\\d{5,6})|\\+\\d{2}\\(\\d\\)\\d{10}";
-    const BRPhoneNumberRegex = "(((\\B\\(\\s?))\\d{2,3}(\\s?\\))|(\\b\\d{2,3}))\\s?\\d{4,5}-?\\d{3,5}(?!-)\\b";
-    const UKPhoneNumberRegex = "(((\\b(00)|\\B\\+)\\s?)?(\\b\\d{2}\\s?)?((\\s?\\(0\\)[-\\s]?|\\b|(?<=(\\b^#)\\d{2}))\\d{2,5}|\\(0\\d{3,4}\\))[/-]?\\s?(\\d{5,8}|\\d{3,4}[-\\s]?\\d{3,4})(?!-)\\b)";
-    const DEPhoneNumberRegex = "((\\+\\d{2}\\s?((\\(0\\))?\\d\\s?)?|\\b)(\\d{2,4}\\s?[-/]?[\\s\\d]{7,10}\\d)(?!-)\\b)";
-    const USPhoneNumberRegex = "((((\\B\\+)|\\b)1(\\s|-)?)|\\b)?(\\(\\d{3}\\)[-\\s]?|\\b\\d{3}\\s?[-\\.]?\\s?)\\d{3}\\s?[-\\.]?\\s?\\d{4}(\\s?(x|X|ext)\\s?\\d{3,5})?(?!-)\\b";
-    const CNPhoneNumberRegex = "((\\b00\\s?)?\\+?86\\s?-?\\s?)?(((\\b|(?<=86))\\d{2,5}\\s?-?\\s?|\\(\\d{2,5}\\)\\s?)\\d{4}\\s?-?\\s?\\d{4}(\\s?-?\\s?\\d{4})?|(\\b|(?<=86))\\d{3}\\s?-?\\s?\\d{4}\\s?-?\\s?\\d{4})(?!-)\\b";
-    const DKPhoneNumberRegex = "((\\(\\s?(\\+\\s?|00)45\\s?\\)\\s?)|(((?<!\\d)\\+\\s?|\\b00)45\\s?)|\\b)(\\s?\\(0\\)\\s?)?((\\d{8})|(\\d{4}\\s?-?\\s?\\d{4,6})|((\\d{2}[\\s-]){3}\\d{2})|(\\d{2}\\s?-?\\s?\\d{3}\\s?-?\\s?\\d{3}))(?!-)\\b";
-    const ITPhoneNumberRegex = "((\\(\\s?(\\+\\s?|00)39\\s?\\)\\s?)|(((?<!\\d)\\+\\s?|\\b00)39\\s?)|\\b)((0[\\d\\s-]{4,12}\\d)|(3[\\d\\s-]{7,12}\\d))(?!-)\\b";
-    const NLPhoneNumberRegex = "((((\\(\\s?(\\+\\s?|00)31\\s?\\)\\s?)|(((?<!\\d)\\+\\s?|\\b00)31\\s?))?(((\\b|(?<=31))0?\\d{1,3}|\\(\\s?0?\\d{1,3}\\s?\\)|\\(0\\)[-\\s]?\\d{1,3})\\s?-?[\\d\\s]{5,11}\\d))|\\b\\d{10,12})(?!-)\\b";
-    const SpecialPhoneNumberRegex = "\\b(\\d{3,4}[/-]\\d{1,4}[/-]\\d{3,4})\\b";
+    const WordBoundariesRegex = "\\b";
+    const NonWordBoundariesRegex = "\\B";
+    const EndWordBoundariesRegex = "\\b";
+    const PreCheckPhoneNumberRegex = "(\\d{1,4}.){2,4}\\s?\\d{2,3}";
+    const GeneralPhoneNumberRegex: (WordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const BRPhoneNumberRegex: (WordBoundariesRegex: string, NonWordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const UKPhoneNumberRegex: (WordBoundariesRegex: string, NonWordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const DEPhoneNumberRegex: (WordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const USPhoneNumberRegex: (WordBoundariesRegex: string, NonWordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const CNPhoneNumberRegex: (WordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const DKPhoneNumberRegex: (WordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const ITPhoneNumberRegex: (WordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const NLPhoneNumberRegex: (WordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const SpecialPhoneNumberRegex: (WordBoundariesRegex: string, EndWordBoundariesRegex: string) => string;
+    const NoAreaCodeUSPhoneNumberRegex = "(?<!(-|-\\s|\\d|\\)|\\)\\s|\\.))[2-9]\\d{2}\\s?[-\\.]\\s?\\d{4}(?!(-\\s?\\d))\\b";
+    const InternationDialingPrefixRegex = "0(0|11)$";
     const TypicalDeductionRegexList: string[];
-    const PhoneNumberMaskRegex = "([0-9A-E]{2}(\\s[0-9A-E]{2}){7})";
+    const PhoneNumberMaskRegex = "([0-9a-e]{2}(\\s[0-9a-e]{2}){7})";
     const CountryCodeRegex = "^(\\(\\s?(\\+\\s?|00)\\d{1,3}\\s?\\)|(\\+\\s?|00)\\d{1,3})";
     const AreaCodeIndicatorRegex = "\\(";
     const FormatIndicatorRegex = "(\\s|-|/|\\.)+";
-    const SeparatorCharList: string[];
+    const ColonMarkers: string[];
+    const ColonPrefixCheckRegex = "(([a-z])\\s*$)";
+    const SpecialBoundaryMarkers: string[];
+    const BoundaryMarkers: string[];
+    const ForbiddenPrefixMarkers: string[];
+    const ForbiddenSuffixMarkers: string[];
 }

dist/types/resources/baseURL.d.ts

@@ -8,5 +8,6 @@ export declare namespace BaseURL {
     const UrlRegex: string;
-    const UrlRegex2 = "((ht|f)tp(s?)\\:\\/\\/|www\\.)[0-9a-zA-Z]([-.\\w]*[0-9a-zA-Z])*(\\.(?<Tld>[0-9a-zA-Z]+))+(:(0-9)*)*(\\/?)([a-zA-Z0-9\\-\\.\\?\\,\\'\\/\\\\\\+&amp;%\\$#_=@]*)?";
+    const UrlRegex2 = "((ht|f)tp(s?)\\:\\/\\/|www\\.)[-a-zA-Z0-9@:%._\\+~#=]{1,256}\\.(?<Tld>[a-zA-Z0-9()]{1,6})\\b([-a-zA-Z0-9()@:%_\\+.~#?&//=]*)";
     const IpUrlRegex: string;
+    const AmbiguousTimeTerm = "^(1?[0-9]|2[0-3]).[ap]m$";
     const TldList: string[];
 }

dist/types/sequence/constants.d.ts

@@ -15,7 +15,7 @@ export declare class Constants {
     static readonly PHONE_NUMBER_REGEX_SPECIAL: string;
-    static readonly MENTION_REGEX: string;
-    static readonly HASHTAG_REGEX: string;
-    static readonly EMAIL_REGEX: string;
-    static readonly URL_REGEX: string;
-    static readonly GUID_REGEX: string;
+    static readonly MENTION_REGEX = "Mention";
+    static readonly HASHTAG_REGEX = "Hashtag";
+    static readonly EMAIL_REGEX = "Email";
+    static readonly URL_REGEX = "Url";
+    static readonly GUID_REGEX = "Guid";
 }

dist/types/sequence/english/extractors.d.ts

@@ -1,7 +0,14 @@
-import { BasePhoneNumberExtractor, BaseIpExtractor, BaseMentionExtractor, BaseHashtagExtractor, BaseEmailExtractor, BaseURLExtractor, BaseGUIDExtractor } from "../extractors";
-import { ExtractResult } from "@microsoft/recognizers-text";
-export declare class PhoneNumberExtractor extends BasePhoneNumberExtractor {
-    extract(source: string): Array<ExtractResult>;
+import { IIpExtractorConfiguration, BaseMentionExtractor, BaseHashtagExtractor, BaseEmailExtractor, IURLExtractorConfiguration, BaseGUIDExtractor, IPhoneNumberExtractorConfiguration } from "../extractors";
+export declare class EnglishPhoneNumberExtractorConfiguration implements IPhoneNumberExtractorConfiguration {
+    readonly WordBoundariesRegex: string;
+    readonly NonWordBoundariesRegex: string;
+    readonly EndWordBoundariesRegex: string;
+    readonly ColonPrefixCheckRegex: string;
+    readonly ForbiddenPrefixMarkers: string[];
+    constructor();
 }
-export declare class IpExtractor extends BaseIpExtractor {
+export declare class EnglishIpExtractorConfiguration implements IIpExtractorConfiguration {
+    readonly Ipv4Regex: RegExp;
+    readonly Ipv6Regex: RegExp;
+    constructor();
 }
@@ -14,5 +21,8 @@ export declare class MentionExtractor extends BaseMentionExtractor {
 }
-export declare class URLExtractor extends BaseURLExtractor {
+export declare class EnglishURLExtractorConfiguration implements IURLExtractorConfiguration {
+    readonly UrlRegex: RegExp;
+    readonly IpUrlRegex: RegExp;
+    constructor();
 }
 export declare class GUIDExtractor extends BaseGUIDExtractor {
 }

dist/types/sequence/english/parsers.d.ts

@@ -15,2 +15,3 @@ import { BaseSequenceParser, BaseIpParser } from "../parsers";
     continueFormatIndicatorDeductionScore: number;
+    wrongFormatIndicatorDeductionScore: number;
     maxFormatIndicatorNum: number;
@@ -21,2 +22,4 @@ maxLengthAwardNum: number;
     pureDigitLengthLimit: number;
+    completeBracketRegex: RegExp;
+    singleBracketRegex: RegExp;
     tailSameDigitRegex: RegExp;
@@ -40,2 +43,12 @@ pureDigitRegex: RegExp;
 export declare class GUIDParser extends BaseSequenceParser {
+    scoreUpperLimit: number;
+    scoreLowerLimit: number;
+    baseScore: number;
+    noBoundaryPenalty: number;
+    noFormatPenalty: number;
+    pureDigitPenalty: number;
+    pureDigitRegex: RegExp;
+    formatRegex: RegExp;
+    ScoreGUID(GUIDText: string): number;
+    parse(extResult: ExtractResult): ParseResult;
 }

dist/types/sequence/extractors.d.ts

@@ -1,16 +0,29 @@
-import { IExtractor, ExtractResult } from "@microsoft/recognizers-text";
+import { IExtractor, ExtractResult, Match } from "@microsoft/recognizers-text";
 export declare abstract class BaseSequenceExtractor implements IExtractor {
     abstract regexes: Map<RegExp, string>;
     protected extractType: string;
-    extract(source: string): Array<ExtractResult>;
+    extract(source: string): ExtractResult[];
+    isValidMatch(match: Match): boolean;
 }
+export interface IPhoneNumberExtractorConfiguration {
+    WordBoundariesRegex: string;
+    NonWordBoundariesRegex: string;
+    EndWordBoundariesRegex: string;
+    ColonPrefixCheckRegex: string;
+    ForbiddenPrefixMarkers: string[];
+}
 export declare class BasePhoneNumberExtractor extends BaseSequenceExtractor {
     regexes: Map<RegExp, string>;
-    constructor();
-    extract(source: string): Array<ExtractResult>;
+    config: IPhoneNumberExtractorConfiguration;
+    constructor(config: IPhoneNumberExtractorConfiguration);
+    extract(source: string): ExtractResult[];
 }
+export interface IIpExtractorConfiguration {
+    Ipv4Regex: RegExp;
+    Ipv6Regex: RegExp;
+}
 export declare class BaseIpExtractor extends BaseSequenceExtractor {
     regexes: Map<RegExp, string>;
-    constructor();
-    extract(source: string): Array<ExtractResult>;
+    constructor(config: IIpExtractorConfiguration);
+    extract(source: string): ExtractResult[];
     isLetterOrDigit(c: string): boolean;
@@ -30,5 +43,11 @@ }
 }
+export interface IURLExtractorConfiguration {
+    UrlRegex: RegExp;
+    IpUrlRegex: RegExp;
+}
 export declare class BaseURLExtractor extends BaseSequenceExtractor {
     regexes: Map<RegExp, string>;
-    constructor();
+    ambiguousTimeTerm: RegExp;
+    constructor(config: IURLExtractorConfiguration);
+    isValidMatch(match: Match): boolean;
 }
@@ -35,0 +54,0 @@ export declare class BaseGUIDExtractor extends BaseSequenceExtractor {

dist/types/sequence/models.d.ts

 import { IModel, ModelResult, IExtractor, IParser } from "@microsoft/recognizers-text";
 export declare abstract class AbstractSequenceModel implements IModel {
-    readonly abstract modelTypeName: string;
+    abstract readonly modelTypeName: string;
     protected readonly extractor: IExtractor;
@@ -31,2 +31,3 @@ protected readonly parser: IParser;
     modelTypeName: string;
+    parse(query: string): ModelResult[];
 }

dist/types/sequence/parsers.d.ts

@@ -7,3 +7,3 @@ import { IParser, ExtractResult, ParseResult } from "@microsoft/recognizers-text";
     parse(extResult: ExtractResult): ParseResult;
-    private dropLeadingZeros(text);
+    private dropLeadingZeros;
 }

dist/types/sequence/sequenceRecognizer.d.ts

 import { Recognizer, IModel, ModelResult } from "@microsoft/recognizers-text";
 export declare enum SequenceOptions {
-    None = 0,
+    None = 0
 }
-export declare function recognizePhoneNumber(query: string, culture: string, options?: SequenceOptions): Array<ModelResult>;
-export declare function recognizeIpAddress(query: string, culture: string, options?: SequenceOptions): Array<ModelResult>;
-export declare function recognizeMention(query: string, culture: string, options?: SequenceOptions): Array<ModelResult>;
-export declare function recognizeHashtag(query: string, culture: string, options?: SequenceOptions): Array<ModelResult>;
-export declare function recognizeEmail(query: string, culture: string, options?: SequenceOptions): Array<ModelResult>;
-export declare function recognizeURL(query: string, culture: string, options?: SequenceOptions): Array<ModelResult>;
-export declare function recognizeGUID(query: string, culture: string, options?: SequenceOptions): Array<ModelResult>;
+export declare function recognizePhoneNumber(query: string, culture: string, options?: SequenceOptions): ModelResult[];
+export declare function recognizeIpAddress(query: string, culture: string, options?: SequenceOptions): ModelResult[];
+export declare function recognizeMention(query: string, culture: string, options?: SequenceOptions): ModelResult[];
+export declare function recognizeHashtag(query: string, culture: string, options?: SequenceOptions): ModelResult[];
+export declare function recognizeEmail(query: string, culture: string, options?: SequenceOptions): ModelResult[];
+export declare function recognizeURL(query: string, culture: string, options?: SequenceOptions): ModelResult[];
+export declare function recognizeGUID(query: string, culture: string, options?: SequenceOptions): ModelResult[];
 export default class SequenceRecognizer extends Recognizer<SequenceOptions> {
@@ -13,0 +13,0 @@ constructor(culture: string, options?: SequenceOptions, lazyInitialization?: boolean);

package.json

 {
   "name": "@microsoft/recognizers-text-sequence",
-  "version": "1.1.4",
+  "version": "1.2.7",
   "description": "recognizers-text-sequence provides robust recognition and resolution of series entities like phone numbers and IP addresses.",
@@ -27,3 +27,3 @@ "author": "Microsoft Corp.",
   "engines": {
-    "node": ">=6.0.0"
+    "node": ">=10.3.0"
   },
@@ -36,5 +36,5 @@ "scripts": {
   "dependencies": {
-    "@microsoft/recognizers-text": "~1.1.4",
+    "@microsoft/recognizers-text": "~1.2.7",
     "grapheme-splitter": "^1.0.2"
   }
 }

No alert changes

Improved metrics

Total package byte prevSize

1025078

increased by6.14%

Number of package files

24

increased by20%

Lines of code

22428

increased by5.13%

Dependency changes

• + Added@microsoft/recognizers-text@1.2.7(transitive)

• - Removed@microsoft/recognizers-text@1.1.4(transitive)

• Updated@microsoft/recognizers-text@~1.2.7