Added unit test for ixt3, ixt4 and ixt-sec registries

Author: manusimidt

tests/test_transformation.py

@@ -17,44 +17,153 @@
 testTransforms = {
     "http://www.xbrl.org/inlineXBRL/transformation/2015-02-26": [
         # [format,value,expected]
-        ['booleanfalse', 'no', 'false'],
+        ['booleanfalse', 'nope', 'false'],
+
         ['booleantrue', 'yeah', 'true'],
 
         ['datedaymonth', '2.12', '--12-02'],
+        ['datedaymonth', '1.1', '--01-01'],
+        ['datedaymonth', '12.11', '--11-12'],
+
         ['datedaymonthen', '2. December', '--12-02'],
+        ['datedaymonthen', '2 Sept.', '--09-02'],
+        ['datedaymonthen', '14.    april', '--04-14'],
+
         ['datedaymonthyear', '2.12.2021', '2021-12-02'],
+        ['datedaymonthyear', '1.1.99', '1999-01-01'],
+        ['datedaymonthyear', '18. 02 2022', '2022-02-18'],
+
         ['datedaymonthyearen', '02. December 2021', '2021-12-02'],
+        ['datedaymonthyearen', '13. Dec. 21', '2021-12-13'],
+        ['datedaymonthyearen', '1 Feb 99', '1999-02-01'],
+
         ['datemonthday', '1.2', '--01-02'],
+        ['datemonthday', '12-1', '--12-01'],
+        ['datemonthday', '1.30', '--01-30'],
+
         ['datemonthdayen', 'Jan 02', '--01-02'],
+        ['datemonthdayen', 'February 13', '--02-13'],
+        ['datemonthdayen', 'sept. 1', '--09-01'],
+
         ['datemonthdayyear', '12-30-2021', '2021-12-30'],
+        ['datemonthdayyear', '2-16-22', '2022-02-16'],
+        ['datemonthdayyear', '2-1-2019', '2019-02-01'],
+
         ['datemonthdayyearen', 'March 31, 2021', '2021-03-31'],
-        ['dateyearmonthday', '2021.12.31', '2021-12-31'],
+        ['datemonthdayyearen', 'Dec. 31, 22', '2022-12-31'],
+        ['datemonthdayyearen', 'april 12 2021', '2021-04-12'],
 
         ['datemonthyear', '12 2021', '2021-12'],
+        ['datemonthyear', '1 22', '2022-01'],
+        ['datemonthyear', '02-1999', '1999-02'],
+
         ['datemonthyearen', 'December 2021', '2021-12'],
+        ['datemonthyearen', 'apr. 22', '2022-04'],
+        ['datemonthyearen', 'Sept. 2000', '2000-09'],
+
+        ['dateyearmonthday', '2021.12.31', '2021-12-31'],
+        ['dateyearmonthday', '2021 1  31', '2021-01-31'],
+        ['dateyearmonthday', '22-1-1', '2022-01-01'],
+
         ['dateyearmonthen', '2021 December', '2021-12'],
+        ['dateyearmonthen', '22 sept.', '2022-09'],
+        ['dateyearmonthen', '21.apr.', '2021-04'],
 
         ['nocontent', 'Bla bla', ''],
+
         ['numcommadecimal', '1.499,99', '1499.99'],
+        ['numcommadecimal', '100*499,999', '100499.999'],
+        ['numcommadecimal', '0,5', '0.5'],
+
         ['numdotdecimal', '1,499.99', '1499.99'],
+        ['numdotdecimal', '1*499', '1499'],
+        ['numdotdecimal', '1,000,000.5', '1000000.5'],
 
-        ['zerodash', '-', '0'],
+        ['zerodash', '--', '0'],
     ],
     "http://www.xbrl.org/inlineXBRL/transformation/2020-02-12": [
         # [format,value,expected]
+        ['date-day-month', '1.1', '--01-01'],
+        ['date-day-month', '31-12', '--12-31'],
+        ['date-day-month', '27*2', '--02-27'],
+
+        ['date-day-month-year', '1-2-20', '2020-02-01'],
+        ['date-day-month-year', '1-02-20', '2020-02-01'],
+        ['date-day-month-year', '01 02 2020', '2020-02-01'],
+
+        ['date-day-monthname-en', '1. sept.', '--09-01'],
+        ['date-day-monthname-en', '01. sep.', '--09-01'],
+        ['date-day-monthname-en', '30 August', '--08-30'],
+
+        ['date-day-monthname-year-en', '30 August 22', '2022-08-30'],
+        ['date-day-monthname-year-en', '01 Aug 22', '2022-08-01'],
+        ['date-day-monthname-year-en', '1 Aug 2022', '2022-08-01'],
+
+        ['date-month-day', '1 31', '--01-31'],
+        ['date-month-day', '01-31', '--01-31'],
+        ['date-month-day', '12.1', '--12-01'],
+
+        ['date-month-day-year', '12. 1 22', '2022-12-01'],
+        ['date-month-day-year', '01/12/2022', '2022-01-12'],
+        ['date-month-day-year', '01.12.2022', '2022-01-12'],
+
+        ['date-month-year', '1*22', '2022-01'],
+        ['date-month-year', '01  22', '2022-01'],
+        ['date-month-year', '12.2022', '2022-12'],
+
+        ['date-monthname-day-en', 'April/1', '--04-01'],
+        ['date-monthname-day-en', 'Sept./20', '--09-20'],
+        ['date-monthname-day-en', 'december 31', '--12-31'],
+
+        ['date-monthname-day-year-en', 'december 31, 22', '2022-12-31'],
+        ['date-monthname-day-year-en', 'dec. 31, 2022', '2022-12-31'],
+        ['date-monthname-day-year-en', 'dec. 1, 2022', '2022-12-01'],
+
+        ['date-year-month', '99/1', '1999-01'],
+        ['date-year-month', '2022 - 12', '2022-12'],
+        ['date-year-month', '2022 -/ 1', '2022-01'],
+
+        ['date-year-month-day', '   22-1-2 ', '2022-01-02'],
+        ['date-year-month-day', ' 2022/1/2 ', '2022-01-02'],
+        ['date-year-month-day', '  22/01/02 ', '2022-01-02'],
+
+        ['date-year-monthname-en', '22/december', '2022-12'],
+        ['date-year-monthname-en', '22/dec.', '2022-12'],
+        ['date-year-monthname-en', '2022-dec', '2022-12'],
+
+        ['fixed-empty', 'some text', ''],
+        ['fixed-false', 'some text', 'false'],
+        ['fixed-true', 'some text', 'true'],
+        ['fixed-zero', 'some text', '0'],
+
+        ['num-comma-decimal', '1.499,99', '1499.99'],
+        ['num-comma-decimal', '100*499,999', '100499.999'],
+        ['num-comma-decimal', '0,5', '0.5'],
+
+        ['num-dot-decimal', '1,499.99', '1499.99'],
+        ['num-dot-decimal', '1*499', '1499'],
+        ['num-dot-decimal', '1,000,000.5', '1000000.5'],
+
     ],
     "http://www.sec.gov/inlineXBRL/transformation/2015-08-31": [
         # [format,value,expected]
+        ['durwordsen', 'Five years, two months', 'P5Y2M0D'],
+        ['durwordsen', '9 years, 2 months', 'P9Y2M0D'],
+        ['durwordsen', '12 days', 'P0Y0M12D'],
+        ['durwordsen', 'ONE MONTH AND THREE DAYS', 'P0Y1M3D'],
+
         ['numwordsen', 'no', '0'],
         ['numwordsen', 'None', '0'],
         ['numwordsen', 'nineteen hundred forty-four', '1944'],
         ['numwordsen', 'Seventy Thousand and one', '70001'],
+
         ['boolballotbox', '☐', 'false'],
+        ['boolballotbox', '☐', 'false'],
         ['boolballotbox', '☑', 'true'],
+        ['boolballotbox', '☑', 'true'],
         ['boolballotbox', '☒', 'true'],
-        ['durwordsen', 'Five years, two months', 'P5Y2M0D'],
-        ['durwordsen', '9 years, 2 months', 'P9Y2M0D'],
-        ['durwordsen', '12 days', 'P0Y0M12D']
+        ['boolballotbox', '☒', 'true'],
+
     ]
 }
 
@@ -69,7 +178,7 @@ def test_normalize(self):
             for i, testCase in enumerate(testTransforms[namespace]):
                 formatCode, testInput, expected = testCase
                 testOutput = normalize(namespace, formatCode, testInput)
-                self.assertEqual(expected, testOutput, msg=f'Failed at test elem {i}')
+                self.assertEqual(expected, testOutput, msg=f'Failed at test elem {i} of registry {namespace}')
 
 
 if __name__ == '__main__':

xbrl/transformations/__init__.py

@@ -91,7 +91,7 @@ def notImplemented(arg: str) -> str:
 
 def yearNorm(year: str) -> str:
     if len(year) == 4: return year
-    if len(year) == 2: return '19' if int(year) > 60 else '20' + year
+    if len(year) == 2: return '19' + year if int(year) > 55 else '20' + year
     raise TransformationException(f'Could not normalize "{year}" to a year')
 
 
@@ -115,7 +115,7 @@ def dateDayMonthEN(arg: str) -> str:
 def dateDayMonthYear(arg: str) -> str:
     # (D)D*(M)M*(Y)Y(YY) -> YYYY-MM-DD
     seg = re.split(r'[^\d]+', arg)  # split at any char sequence that is not a digit
-    return f"{seg[2].zfill(2)}-{seg[1].zfill(2)}-{seg[0].zfill(2)}"
+    return f"{yearNorm(seg[2])}-{seg[1].zfill(2)}-{seg[0].zfill(2)}"
 
 
 def dateDayMonthYearEN(arg: str) -> str:
@@ -151,7 +151,7 @@ def dateMonthDayYearEN(arg: str) -> str:
 def dateMonthYear(arg: str) -> str:
     # (M)M*(Y)Y(YY) -> YYYY-MM
     seg = re.split(r'[^\d]+', arg)  # split at any char sequence that is not a digit
-    return f"{seg[1].zfill(2)}-{seg[0].zfill(2)}"
+    return f"{yearNorm(seg[1])}-{seg[0].zfill(2)}"
 
 
 def dateMonthYearEN(arg: str) -> str:
@@ -180,7 +180,7 @@ def dateYearMonthEN(arg: str) -> str:
 
 def numCommaDecimal(arg: str) -> str:
     # nnn*nnn*nnn,n -> nnnnnnnnn.n
-    arg = re.sub(r'(\s|-|\.)', '', arg)
+    arg = re.sub(r'[^\d,]+', '', arg)  # remove all chars that are not a digit and not a comma
     return arg.replace(',', '.')