Merge pull request #40 from kanedata/issue-39

more permissive date format acceptance. Fix for issue #39
kanedata · Dec 19, 2023 · c8b8e8c · c8b8e8c
2 parents 9f149a2 + be7981c
commit c8b8e8c
Show file tree

Hide file tree

Showing 3 changed files with 43 additions and 52 deletions.
diff --git a/src/ixbrlparse/__about__.py b/src/ixbrlparse/__about__.py
@@ -1 +1 @@
-__version__ = "0.8.1"
+__version__ = "0.8.2"
diff --git a/src/ixbrlparse/components/formats.py b/src/ixbrlparse/components/formats.py
@@ -115,6 +115,7 @@ class ixtNumDotDecimal(ixbrlFormat):  # noqa: N801
 
 
 DATE_ORDINAL_SUFFIX_REGEX = re.compile(r"([0-9]{1,2})(st|nd|rd|th)\b")
+DATE_NON_ALPHANUMERIC_REGEX = re.compile(r"[\/\.\-\\–— ]")  # noqa: RUF001
 
 
 class ixtDateFormat(ixbrlFormat):  # noqa: N801
@@ -131,6 +132,9 @@ def parse_value(self, value: Union[str, int, float]) -> Optional[datetime.date]:
             value = value.lower()
             # remove ordinal suffixes with regex
             value = DATE_ORDINAL_SUFFIX_REGEX.sub(r"\1", value)
+            # replace non-alphanumeric characters with dashes
+            value = DATE_NON_ALPHANUMERIC_REGEX.sub("-", value)
+
             date_formats = self._get_date_formats()
             error: Optional[Exception] = None
             for date_format in date_formats:
@@ -154,71 +158,65 @@ class ixtDateLongUK(ixtDateFormat):  # noqa: N801
         "ixt:datelonguk",
         "ixt:datedaymonthyearen",
     )
-    date_format = ("%d %B %Y", "%d %B %y")
+    date_format = ("%d-%B-%Y", "%d-%B-%y")
 
 
 class ixtDateLongUS(ixtDateFormat):  # noqa: N801
     format_names = (
         "datelongus",
         "ixt:datelongus",
     )
-    date_format = ("%B %d, %Y", "%B %d, %y")
+    date_format = ("%B-%d,-%Y", "%B-%d,-%y")
 
 
 class ixtDateShortUK(ixtDateFormat):  # noqa: N801
     format_names = (
         "dateshortuk",
         "ixt:dateshortuk",
     )
-    date_format = ("%d %b %Y", "%d %b %y")
+    date_format = ("%d-%b-%Y", "%d-%b-%y")
 
 
 class ixtDateShortUS(ixtDateFormat):  # noqa: N801
     format_names = (
         "dateshortus",
         "ixt:dateshortus",
     )
-    date_format = ("%b %d, %Y", "%b %d, %y")
+    date_format = ("%b-%d,-%Y", "%b-%d,-%y")
 
 
 class ixtDateDayMonthYear(ixtDateFormat):  # noqa: N801
     format_names = (
         "datedaymonthyear",
         "ixt:datedaymonthyear",
-    )
-    date_format = ("%d.%m.%Y", "%d.%m.%y")
-
-
-class ixtDateSlashEU(ixtDateFormat):  # noqa: N801
-    format_names = (
         "dateslasheu",
         "ixt:dateslasheu",
+        "datedoteu",
+        "ixt:datedoteu",
     )
-    date_format = ("%d/%m/%Y", "%d/%m/%y")
+    date_format = ("%d-%m-%Y", "%d-%m-%y")
 
 
 class ixtDateSlashUS(ixtDateFormat):  # noqa: N801
     format_names = (
         "dateslashus",
         "ixt:dateslashus",
+        "datedotus",
+        "ixt:datedotus",
     )
-    date_format = ("%m/%d/%Y", "%m/%d/%y")
+    date_format = ("%m-%d-%Y", "%m-%d-%y")
 
 
-class ixtDateDotEU(ixtDateFormat):  # noqa: N801
-    format_names = (
-        "datedoteu",
-        "ixt:datedoteu",
-    )
-    date_format = ("%d.%m.%y", "%d.%m.%Y")
+class ixtDateDotEU(ixtDateDayMonthYear):  # noqa: N801
+    pass
 
 
-class ixtDateDotUS(ixtDateFormat):  # noqa: N801
-    format_names = (
-        "datedotus",
-        "ixt:datedotus",
-    )
-    date_format = ("%m.%d.%y", "%m.%d.%Y")
+class ixtDateSlashEU(ixtDateDayMonthYear):  # noqa: N801
+    pass
+
+
+class ixtDateDotUS(ixtDateSlashUS):  # noqa: N801
+    pass
 
 
 @hookimpl
@@ -236,8 +234,5 @@ def ixbrl_add_formats() -> List[Type[ixbrlFormat]]:
         ixtDateShortUK,
         ixtDateShortUS,
         ixtDateDayMonthYear,
-        ixtDateSlashEU,
         ixtDateSlashUS,
-        ixtDateDotEU,
-        ixtDateDotUS,
     ]
diff --git a/tests/test_formats.py b/tests/test_formats.py
@@ -5,14 +5,11 @@
 from ixbrlparse.components.formats import (
     ixbrlFormat,
     ixtDateDayMonthYear,
-    ixtDateDotEU,
-    ixtDateDotUS,
     ixtDateFormat,
     ixtDateLongUK,
     ixtDateLongUS,
     ixtDateShortUK,
     ixtDateShortUS,
-    ixtDateSlashEU,
     ixtDateSlashUS,
     ixtFixedFalse,
     ixtFixedTrue,
@@ -27,27 +24,26 @@
 @pytest.mark.parametrize(
     "dateclass, datestring, expecteddate, errordate",
     (
-        (ixtDateFormat, "2019-01-05", date(2019, 1, 5), "04/05/2019"),
-        (ixtDateSlashEU, "05/01/2019", date(2019, 1, 5), "2019-05-04"),
-        (ixtDateSlashEU, "05/01/19", date(2019, 1, 5), "2019-05-04"),
-        (ixtDateLongUK, "05 January 2019", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateLongUK, "05 January 19", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateLongUS, "January 05, 2019", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateLongUS, "January 05, 19", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateShortUK, "05 Jan 2019", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateShortUK, "05 Jan 19", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateShortUS, "Jan 05, 2019", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateShortUS, "Jan 05, 19", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateDayMonthYear, "05.01.2019", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateDayMonthYear, "05.01.19", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateSlashEU, "05/01/2019", date(2019, 1, 5), "2019-05-04"),
-        (ixtDateSlashEU, "05/01/19", date(2019, 1, 5), "2019-05-04"),
-        (ixtDateSlashUS, "01/05/2019", date(2019, 1, 5), "2019-05-04"),
-        (ixtDateSlashUS, "01/05/19", date(2019, 1, 5), "2019-05-04"),
-        (ixtDateDotEU, "05.01.2019", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateDotEU, "05.01.19", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateDotUS, "01.05.2019", date(2019, 1, 5), "05/04/2019"),
-        (ixtDateDotUS, "01.05.19", date(2019, 1, 5), "05/04/2019"),
+        (ixtDateFormat, "2019-01-05", date(2019, 1, 5), "0400502019"),
+        (ixtDateLongUK, "05 January 2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateLongUK, "05 January 19", date(2019, 1, 5), "0400502019"),
+        (ixtDateLongUS, "January 05, 2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateLongUS, "January 05, 19", date(2019, 1, 5), "0400502019"),
+        (ixtDateShortUK, "05 Jan 2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateShortUK, "05 Jan 19", date(2019, 1, 5), "0400502019"),
+        (ixtDateShortUS, "Jan 05, 2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateShortUS, "Jan 05, 19", date(2019, 1, 5), "0400502019"),
+        (ixtDateDayMonthYear, "05/01/2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateDayMonthYear, "05.01.2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateDayMonthYear, "05.01.19", date(2019, 1, 5), "0400502019"),
+        (ixtDateDayMonthYear, "05/01/2019", date(2019, 1, 5), "2019005004"),
+        (ixtDateDayMonthYear, "05/01/19", date(2019, 1, 5), "2019005004"),
+        (ixtDateSlashUS, "01/05/2019", date(2019, 1, 5), "2019005004"),
+        (ixtDateSlashUS, "01/05/19", date(2019, 1, 5), "2019005004"),
+        (ixtDateDayMonthYear, "05.01.2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateDayMonthYear, "05.01.19", date(2019, 1, 5), "0400502019"),
+        (ixtDateSlashUS, "01.05.2019", date(2019, 1, 5), "0400502019"),
+        (ixtDateSlashUS, "01.05.19", date(2019, 1, 5), "0400502019"),
     ),
 )
 def test_date_formats(dateclass, datestring, expecteddate, errordate):