Support Chinese next next week day - first commit (#3184)

Co-authored-by: Michael Wang (Centific Technologies Inc) <[email protected]>

.NET/Microsoft.Recognizers.Definitions.Common/Chinese/DateTimeDefinitions.cs

@@ -42,6 +42,8 @@ public static class DateTimeDefinitions
       public static readonly string DateThisRegex = $@"(这个|这一个|这|这一|本){WeekDayRegex}";
       public static readonly string DateLastRegex = $@"(上一个|上个|上一|上|最后一个|最后)(的)?{WeekDayRegex}";
       public static readonly string DateNextRegex = $@"(下一个|下个|下一|下)(的)?{WeekDayRegex}";
+      public static readonly string DateNextNextRegex = $@"(下下|下下[个個]){WeekDayRegex}";
+      public static readonly string DateLastLastRegex = $@"(上上|上上[个個]){WeekDayRegex}";
       public const string WeekWithWeekDayRangeRegex = @"^[.]";
       public const string WoMLastRegex = @"最后一";
       public const string WoMPreviousRegex = @"上个";
@@ -78,7 +80,8 @@ public static class DateTimeDefinitions
       public const string DatePeriodThisRegex = @"这个|这一个|这|这一|本";
       public const string DatePeriodLastRegex = @"上个|上一个|上|上一";
       public const string DatePeriodNextRegex = @"下个|下一个|下|下一";
-      public const string DatePeriodNextNextRegex = @"下下";
+      public const string DatePeriodNextNextRegex = @"下下|下下[个個]";
+      public const string DatePeriodLastLastRegex = @"上上|上上[个個]";
       public static readonly string RelativeMonthRegex = $@"(?<relmonth>({DatePeriodThisRegex}|{DatePeriodLastRegex}|{DatePeriodNextRegex})\s*月)";
       public const string HalfYearRegex = @"((?<firstHalf>(上|前)半年)|(?<secondHalf>(下|后)半年))";
       public static readonly string YearRegex = $@"(({YearNumRegex})(\s*年)?|({SimpleYearRegex})\s*年){HalfYearRegex}?";
@@ -90,7 +93,7 @@ public static class DateTimeDefinitions
       public static readonly string YearAndMonth = $@"({DatePeriodYearInCJKRegex}|{YearRegex}|(?<yearrel>明年|今年|去年))\s*({MonthRegex}|的?(?<cardinal>第一|第二|第三|第四|第五|第六|第七|第八|第九|第十|第十一|第十二|最后一)\s*个月\s*)";
       public static readonly string SimpleYearAndMonth = $@"({YearNumRegex}[/\\\-]{MonthNumRegex}\b$)";
       public static readonly string PureNumYearAndMonth = $@"({YearRegexInNumber}\s*[-\.\/]\s*{MonthNumRegex})|({MonthNumRegex}\s*\/\s*{YearRegexInNumber})";
-      public static readonly string OneWordPeriodRegex = $@"(((?<yearrel>(明|今|去)年)\s*)?{MonthRegex}|({DatePeriodThisRegex}|{DatePeriodLastRegex}|{DatePeriodNextNextRegex}|{DatePeriodNextRegex})(?<halfTag>半)?\s*(周末|周|月|年)|周末|(今|明|去|前|后)年(\s*{HalfYearRegex})?)";
+      public static readonly string OneWordPeriodRegex = $@"(((?<yearrel>(明|今|去)年)\s*)?{MonthRegex}|({DatePeriodThisRegex}|{DatePeriodLastLastRegex}|{DatePeriodLastRegex}|{DatePeriodNextNextRegex}|{DatePeriodNextRegex})(?<halfTag>半)?\s*([周週]末|[周週]|月|年)|[周週]末|(今|明|去|前|后)年(\s*{HalfYearRegex})?)";
       public const string LaterEarlyPeriodRegex = @"^[.]";
       public const string DatePointWithAgoAndLater = @"^[.]";
       public static readonly string WeekOfMonthRegex = $@"(?<wom>{MonthSuffixRegex}的(?<cardinal>第一|第二|第三|第四|第五|最后一)\s*周\s*)";
@@ -286,7 +289,8 @@ public static class DateTimeDefinitions
         };
       public static readonly IList<string> WeekendTerms = new List<string>
         {
-            @"周末"
+            @"周末",
+            @"週末"
         };
       public static readonly IList<string> WeekTerms = new List<string>
         {

.NET/Microsoft.Recognizers.Text.DateTime/Chinese/Extractors/ChineseDateExtractorConfiguration.cs

@@ -27,6 +27,10 @@ public class ChineseDateExtractorConfiguration : BaseDateTimeOptionsConfiguratio
 
         public static readonly Regex NextRegex = new Regex(DateTimeDefinitions.DateNextRegex, RegexFlags, RegexTimeOut);
 
+        public static readonly Regex NextNextRegex = new Regex(DateTimeDefinitions.DateNextNextRegex, RegexFlags, RegexTimeOut);
+
+        public static readonly Regex LastLastRegex = new Regex(DateTimeDefinitions.DateLastLastRegex, RegexFlags, RegexTimeOut);
+
         public static readonly Regex SpecialDayRegex = new Regex(DateTimeDefinitions.SpecialDayRegex, RegexFlags, RegexTimeOut);
 
         public static readonly Regex WeekDayOfMonthRegex = new Regex(DateTimeDefinitions.WeekDayOfMonthRegex, RegexFlags, RegexTimeOut);
@@ -78,7 +82,7 @@ public ChineseDateExtractorConfiguration(IDateTimeOptionsConfiguration config)
 
             ImplicitDateList = new List<Regex>
             {
-                LunarRegex, SpecialDayRegex, ThisRegex, LastRegex, NextRegex,
+                LunarRegex, SpecialDayRegex, ThisRegex, LastLastRegex, LastRegex, NextNextRegex, NextRegex,
                 WeekDayRegex, WeekDayOfMonthRegex, SpecialDate,
             };
 

.NET/Microsoft.Recognizers.Text.DateTime/Chinese/Extractors/ChineseDatePeriodExtractorConfiguration.cs

@@ -93,6 +93,7 @@ public class ChineseDatePeriodExtractorConfiguration : BaseDateTimeOptionsConfig
         public static readonly Regex DateUnitRegex = new Regex(DateTimeDefinitions.DateUnitRegex, RegexFlags, RegexTimeOut);
         public static readonly Regex LastRegex = new Regex(DateTimeDefinitions.DatePeriodLastRegex, RegexFlags, RegexTimeOut);
         public static readonly Regex NextNextRegex = new Regex(DateTimeDefinitions.DatePeriodNextNextRegex, RegexFlags, RegexTimeOut);
+        public static readonly Regex LastLastRegex = new Regex(DateTimeDefinitions.DatePeriodLastLastRegex, RegexFlags, RegexTimeOut);
         public static readonly Regex NextRegex = new Regex(DateTimeDefinitions.DatePeriodNextRegex, RegexFlags, RegexTimeOut);
         public static readonly Regex RelativeMonthRegex = new Regex(DateTimeDefinitions.RelativeMonthRegex, RegexFlags, RegexTimeOut);
         public static readonly Regex LaterEarlyPeriodRegex = new Regex(DateTimeDefinitions.LaterEarlyPeriodRegex, RegexFlags, RegexTimeOut);

.NET/Microsoft.Recognizers.Text.DateTime/Chinese/Parsers/ChineseDateParserConfiguration.cs

@@ -51,6 +51,8 @@ public ChineseDateParserConfiguration(ICJKCommonDateTimeParserConfiguration conf
             WeekDayAndDayRegex = ChineseDateExtractorConfiguration.WeekDayAndDayRegex;
             DurationRelativeDurationUnitRegex = ChineseDateExtractorConfiguration.DurationRelativeDurationUnitRegex;
             SpecialDayWithNumRegex = ChineseDateExtractorConfiguration.SpecialDayWithNumRegex;
+            NextNextRegex = ChineseDateExtractorConfiguration.NextNextRegex;
+            LastLastRegex = ChineseDateExtractorConfiguration.LastLastRegex;
 
             CardinalMap = config.CardinalMap;
             UnitMap = config.UnitMap;
@@ -94,6 +96,10 @@ public ChineseDateParserConfiguration(ICJKCommonDateTimeParserConfiguration conf
 
         public Regex NextRegex { get; }
 
+        public Regex NextNextRegex { get; }
+
+        public Regex LastLastRegex { get; }
+
         public Regex ThisRegex { get; }
 
         public Regex LastRegex { get; }

.NET/Microsoft.Recognizers.Text.DateTime/Chinese/Parsers/ChineseDatePeriodParserConfiguration.cs

@@ -38,6 +38,7 @@ public ChineseDatePeriodParserConfiguration(ICJKCommonDateTimeParserConfiguratio
             SimpleCasesRegex = ChineseDatePeriodExtractorConfiguration.SimpleCasesRegex;
             ThisRegex = ChineseDatePeriodExtractorConfiguration.ThisRegex;
             NextNextRegex = ChineseDatePeriodExtractorConfiguration.NextNextRegex;
+            LastLastRegex = ChineseDatePeriodExtractorConfiguration.LastLastRegex;
             NextRegex = ChineseDatePeriodExtractorConfiguration.NextRegex;
             LastRegex = ChineseDatePeriodExtractorConfiguration.LastRegex;
             YearToYear = ChineseDatePeriodExtractorConfiguration.YearToYear;
@@ -124,6 +125,8 @@ public ChineseDatePeriodParserConfiguration(ICJKCommonDateTimeParserConfiguratio
 
         public Regex NextNextRegex { get; }
 
+        public Regex LastLastRegex { get; }
+
         public Regex NextRegex { get; }
 
         public Regex LastRegex { get; }

.NET/Microsoft.Recognizers.Text.DateTime/Parsers/CJK/BaseCJKDateParser.cs

@@ -5,6 +5,7 @@
 using System.Collections.Generic;
 using System.Globalization;
 using System.Text.RegularExpressions;
+using Microsoft.Recognizers.Text.DateTime.Chinese;
 using Microsoft.Recognizers.Text.Utilities;
 using DateObject = System.DateTime;
 
@@ -355,11 +356,21 @@ protected DateTimeResolutionResult ParseImplicitDate(string text, DateObject ref
                 ret = MatchThisWeekday(text, referenceDate);
             }
 
+            if (!ret.Success)
+            {
+                ret = MatchNextNextWeekday(text, referenceDate);
+            }
+
             if (!ret.Success)
             {
                 ret = MatchNextWeekday(text, referenceDate);
             }
 
+            if (!ret.Success)
+            {
+                ret = MatchLastLastWeekday(text, referenceDate);
+            }
+
             if (!ret.Success)
             {
                 ret = MatchLastWeekday(text, referenceDate);
@@ -497,6 +508,28 @@ protected DateTimeResolutionResult MatchNextWeekday(string text, DateObject refe
             return result;
         }
 
+        protected DateTimeResolutionResult MatchNextNextWeekday(string text, DateObject reference)
+        {
+            var result = new DateTimeResolutionResult();
+            var cnConfig = this.config as ChineseDateParserConfiguration;
+            if (cnConfig != null)
+            {
+                var match = cnConfig.NextNextRegex.MatchExact(text, trim: true);
+                if (match.Success)
+                {
+                    var weekdayKey = match.Groups["weekday"].Value;
+                    var value = reference.Next((DayOfWeek)this.config.DayOfWeek[weekdayKey]);
+                    value = value.Next((DayOfWeek)this.config.DayOfWeek[weekdayKey]);
+
+                    result.Timex = DateTimeFormatUtil.LuisDate(value);
+                    result.FutureValue = result.PastValue = DateObject.MinValue.SafeCreateFromValue(value.Year, value.Month, value.Day);
+                    result.Success = true;
+                }
+            }
+
+            return result;
+        }
+
         protected DateTimeResolutionResult MatchThisWeekday(string text, DateObject reference)
         {
             var result = new DateTimeResolutionResult();
@@ -533,6 +566,29 @@ protected DateTimeResolutionResult MatchLastWeekday(string text, DateObject refe
             return result;
         }
 
+        protected DateTimeResolutionResult MatchLastLastWeekday(string text, DateObject reference)
+        {
+            var result = new DateTimeResolutionResult();
+            var cnConfig = this.config as ChineseDateParserConfiguration;
+            if (cnConfig != null)
+            {
+                var match = cnConfig.LastLastRegex.MatchExact(text, trim: true);
+
+                if (match.Success)
+                {
+                    var weekdayKey = match.Groups["weekday"].Value;
+                    var value = reference.Last((DayOfWeek)this.config.DayOfWeek[weekdayKey]);
+                    value = value.Last((DayOfWeek)this.config.DayOfWeek[weekdayKey]);
+
+                    result.Timex = DateTimeFormatUtil.LuisDate(value);
+                    result.FutureValue = result.PastValue = DateObject.MinValue.SafeCreateFromValue(value.Year, value.Month, value.Day);
+                    result.Success = true;
+                }
+            }
+
+            return result;
+        }
+
         protected DateTimeResolutionResult MatchWeekdayAlone(string text, DateObject reference)
         {
             var result = new DateTimeResolutionResult();

.NET/Microsoft.Recognizers.Text.DateTime/Parsers/CJK/BaseCJKDatePeriodParser.cs

@@ -1023,8 +1023,10 @@ private DateTimeResolutionResult ParseOneWordPeriod(string text, DateObject refe
                 }
 
                 // In Chinese, "下" means next, "下下周" means next next week, "下下周末" means next next weekend, need to check whether the text match "下下"
+                // "上" means last, "上上周" means last last week, "上上周末" means last last weekend, need to check whether the text match "上上"
                 ChineseDatePeriodParserConfiguration config = this.config as ChineseDatePeriodParserConfiguration;
                 bool nextNextMatch = config == null ? false : config.NextNextRegex.Match(trimmedText).Success;
+                bool lastlastMatch = config == null ? false : config.LastLastRegex.Match(trimmedText).Success;
 
                 var nextMatch = this.config.NextRegex.Match(trimmedText);
                 var lastMatch = this.config.LastRegex.Match(trimmedText);
@@ -1081,6 +1083,11 @@ private DateTimeResolutionResult ParseOneWordPeriod(string text, DateObject refe
                         // If it is Chinese "下下周" (next next week), "下下周末" (next next weekend), then swift is 2
                         swift = 2;
                     }
+                    else if (lastlastMatch)
+                    {
+                        // If it is Chinese "上上周" (last last week), "上上周末" (last last weekend), then swift is -2
+                        swift = -2;
+                    }
                     else if (nextMatch.Success)
                     {
                         if (nextMatch.Groups[Constants.AfterGroupName].Success)

Patterns/Chinese/Chinese-DateTime.yaml

@@ -48,6 +48,12 @@ DateLastRegex: !nestedRegex
 DateNextRegex: !nestedRegex
   def: (下一个|下个|下一|下)(的)?{WeekDayRegex}
   references: [WeekDayRegex]
+DateNextNextRegex: !nestedRegex
+  def: (下下|下下[个個]){WeekDayRegex}
+  references: [WeekDayRegex]
+DateLastLastRegex: !nestedRegex
+  def: (上上|上上[个個]){WeekDayRegex}
+  references: [WeekDayRegex]
 WeekWithWeekDayRangeRegex: !simpleRegex
   # TODO: modify below regex according to the counterpart in Japanese
   def: ^[.]
@@ -150,7 +156,9 @@ DatePeriodLastRegex: !simpleRegex
 DatePeriodNextRegex: !simpleRegex
   def: 下个|下一个|下|下一
 DatePeriodNextNextRegex: !simpleRegex
-  def: 下下
+  def: 下下|下下[个個]
+DatePeriodLastLastRegex: !simpleRegex
+  def: 上上|上上[个個]
 RelativeMonthRegex: !nestedRegex
   def: (?<relmonth>({DatePeriodThisRegex}|{DatePeriodLastRegex}|{DatePeriodNextRegex})\s*月)
   references: [DatePeriodThisRegex, DatePeriodLastRegex, DatePeriodNextRegex]
@@ -183,8 +191,8 @@ PureNumYearAndMonth: !nestedRegex
   def: ({YearRegexInNumber}\s*[-\.\/]\s*{MonthNumRegex})|({MonthNumRegex}\s*\/\s*{YearRegexInNumber})
   references: [YearRegexInNumber, MonthNumRegex]
 OneWordPeriodRegex: !nestedRegex
-  def: (((?<yearrel>(明|今|去)年)\s*)?{MonthRegex}|({DatePeriodThisRegex}|{DatePeriodLastRegex}|{DatePeriodNextNextRegex}|{DatePeriodNextRegex})(?<halfTag>半)?\s*(周末|周|月|年)|周末|(今|明|去|前|后)年(\s*{HalfYearRegex})?)
-  references: [MonthRegex, DatePeriodThisRegex, DatePeriodLastRegex, DatePeriodNextNextRegex, DatePeriodNextRegex, HalfYearRegex]
+  def: (((?<yearrel>(明|今|去)年)\s*)?{MonthRegex}|({DatePeriodThisRegex}|{DatePeriodLastLastRegex}|{DatePeriodLastRegex}|{DatePeriodNextNextRegex}|{DatePeriodNextRegex})(?<halfTag>半)?\s*([周週]末|[周週]|月|年)|[周週]末|(今|明|去|前|后)年(\s*{HalfYearRegex})?)
+  references: [MonthRegex, DatePeriodThisRegex, DatePeriodLastLastRegex, DatePeriodLastRegex, DatePeriodNextNextRegex, DatePeriodNextRegex, HalfYearRegex]
 LaterEarlyPeriodRegex: !simpleRegex
   # TODO: modify below regex according to the counterpart in Japanese
   def: ^[.]
@@ -567,6 +575,7 @@ WeekendTerms: !list
   types: [ string ]
   entries:
     - 周末
+    - 週末
 WeekTerms: !list
   types: [ string ]
   entries:

Specs/DateTime/Chinese/DateExtractor.json

@@ -245,6 +245,30 @@
       }
     ]
   },
+  {
+    "Input": "马拉松在下下周日举行",
+    "NotSupported": "java, javascript, python",
+    "Results": [
+      {
+        "Text": "下下周日",
+        "Type": "date",
+        "Start": 4,
+        "Length": 4
+      }
+    ]
+  },
+  {
+    "Input": "任務是在上上個週三完成的",
+    "NotSupported": "java, javascript, python",
+    "Results": [
+      {
+        "Text": "上上個週三",
+        "Type": "date",
+        "Start": 4,
+        "Length": 5
+      }
+    ]
+  },
   {
     "Input": "下次的12号",
     "Results": [

Specs/DateTime/Chinese/DateParser.json

@@ -581,6 +581,54 @@
       }
     ]
   },
+  {
+    "Input": "马拉松在下下周日举行",
+    "Context": {
+      "ReferenceDateTime": "2024-11-15T00:00:00"
+    },
+    "NotSupported": "java, javascript, python",
+    "Results": [
+      {
+        "Text": "下下周日",
+        "Type": "date",
+        "Value": {
+          "Timex": "2024-12-01",
+          "FutureResolution": {
+            "date": "2024-12-01"
+          },
+          "PastResolution": {
+            "date": "2024-12-01"
+          }
+        },
+        "Start": 4,
+        "Length": 4
+      }
+    ]
+  },
+  {
+    "Input": "任務是在上上個週三完成的",
+    "Context": {
+      "ReferenceDateTime": "2024-11-15T00:00:00"
+    },
+    "NotSupported": "java, javascript, python",
+    "Results": [
+      {
+        "Text": "上上個週三",
+        "Type": "date",
+        "Value": {
+          "Timex": "2024-10-30",
+          "FutureResolution": {
+            "date": "2024-10-30"
+          },
+          "PastResolution": {
+            "date": "2024-10-30"
+          }
+        },
+        "Start": 4,
+        "Length": 5
+      }
+    ]
+  },
   {
     "Input": "12号",
     "Context": {

Specs/DateTime/Chinese/DatePeriodExtractor.json

@@ -100,6 +100,30 @@
       }
     ]
   },
+  {
+    "Input": "我是上上周回来的",
+    "NotSupported": "java, javascript, python",
+    "Results": [
+      {
+        "Text": "上上周",
+        "Type": "daterange",
+        "Start": 2,
+        "Length": 3
+      }
+    ]
+  },
+  {
+    "Input": "你上上個週末干嘛了",
+    "NotSupported": "java, javascript, python",
+    "Results": [
+      {
+        "Text": "上上個週末",
+        "Type": "daterange",
+        "Start": 1,
+        "Length": 5
+      }
+    ]
+  },
   {
     "Input": "下个月完工",
     "Results": [