Make EDTF parser available as undate formatter; handle 5+ digit years

Author: rlskoeser

src/undate/dateformat/edtf/__init__.py

@@ -0,0 +1,26 @@
+from typing import Dict, List, Union
+
+from lark.exceptions import UnexpectedCharacters
+
+from undate.undate import Undate, UndateInterval
+from undate.dateformat.base import BaseDateFormat
+from undate.dateformat.edtf.parser import edtf_parser
+from undate.dateformat.edtf.transformer import EDTFTransformer
+
+
+class EDTFDateFormat(BaseDateFormat):
+    name: str = "EDTF"
+
+    def __init__(self):
+        self.transformer = EDTFTransformer()
+
+    def parse(self, value: str) -> Union[Undate, UndateInterval]:
+        # parse the input string, then transform to undate object
+        try:
+            parsetree = edtf_parser.parse(value)
+            return self.transformer.transform(parsetree)
+        except UnexpectedCharacters as err:
+            raise ValueError("Parsing failed due to UnexpectedCharacters: %s" % err)
+
+    # def to_string(self, undate: Undate) -> str:
+    # TODO: how do we leverage the parser for this?

src/undate/dateformat/edtf/transformer.py

@@ -63,8 +63,8 @@ def day_unspecified(self, items):
     def date_level1(self, items):
         return self.date(items)
 
-    def year_fivedigitsplus(self, token):
+    def year_fivedigitsplus(self, items):
         # strip off the leading Y and convert to integer
-        # TODO: undate is currently limited to 4-digit years
-        # (datetime max year of 9999)
-        return tok.update(int(token[:1]))
+        token = items[0]
+        year = int(token.value.lstrip("Y"))
+        return Tree(data="year", children=[year])

src/undate/dateformat/iso8601.py

@@ -1,6 +1,7 @@
+from typing import Dict, List, Union
+
 from undate.dateformat.base import BaseDateFormat
 from undate.undate import Undate, UndateInterval
-from typing import Dict, List, Union
 
 
 class ISO8601DateFormat(BaseDateFormat):

tests/test_dateformat/edtf/test_edtf_transformer.py

@@ -13,12 +13,10 @@
     ("1001-03-30", Undate(1001, 3, 30)),
     ("1000/2000", UndateInterval(Undate(1000), Undate(2000))),
     ("1000-01/2000-05-01", UndateInterval(Undate(1000, 1), Undate(2000, 5, 1))),
-    # # level 1
-    # NOTE: undate currently doesn't most of the level 1 functionality
-    # NOTE: undate currently doesn't support years beyond 9999 (datetime.MAXYEAR)
-    # ("Y17000002", Undate(17000002)),
+    # level 1
+    ("Y17000002", Undate(17000002)),
     # "2001-21",  # spring 2001
-    # # qualifiers
+    # qualifiers TODO - not yet supported by undate
     # "1984?",
     # "2004-06~",
     # "2004-06-11%",

tests/test_dateformat/test_edtf.py

@@ -0,0 +1,39 @@
+import pytest
+
+from undate.dateformat.edtf import EDTFDateFormat
+from undate.undate import Undate, UndateInterval
+
+
+class TestEDTFDateFormat:
+    def test_parse_singledate(self):
+        assert EDTFDateFormat().parse("2002") == Undate(2002)
+        assert EDTFDateFormat().parse("1991-05") == Undate(1991, 5)
+        assert EDTFDateFormat().parse("1991-05-03") == Undate(1991, 5, 3)
+        # unknown dates are not strictly equal, but string comparison should match
+        assert str(EDTFDateFormat().parse("201X")) == str(Undate("201X"))
+        assert str(EDTFDateFormat().parse("2004-XX")) == str(Undate(2004, "XX"))
+        # missing year but month/day known
+        # assert EDTFDateFormat().parse("--05-03") == Undate(month=5, day=3)
+
+    def test_parse_singledate_unequal(self):
+        assert EDTFDateFormat().parse("2002") != Undate(2003)
+        assert EDTFDateFormat().parse("1991-05") != Undate(1991, 6)
+        assert EDTFDateFormat().parse("1991-05-03") != Undate(1991, 5, 4)
+        # missing year but month/day known
+        # - does EDTF not support this or is parsing logic incorrect?
+        # assert EDTFDateFormat().parse("XXXX-05-03") != Undate(month=5, day=4)
+
+    def test_parse_invalid(self):
+        with pytest.raises(ValueError):
+            assert EDTFDateFormat().parse("1991-5") == Undate(1991, 5)
+
+    def test_parse_range(self):
+        assert EDTFDateFormat().parse("1800/1900") == UndateInterval(
+            Undate(1800), Undate(1900)
+        )
+
+    # def test_to_string(self):
+    #     # NOTE: iso8601 to_string currently tested more thoroughly
+    #     # in undate str tests; may want to move those tests here
+    #     assert EDTFDateFormat().to_string(Undate(900)) == "0900"
+    #     assert EDTFDateFormat().to_string(Undate(33)) == "0033"