Merge pull request #45 from artshumrc/pytest-format-tests

Reformat tests to use Pytest style; fix infinite comparison for OneOfASet

.github/workflows/ci.yml

@@ -18,7 +18,7 @@ jobs:
             - uses: actions/checkout@v4
 
             - name: Set up Python ${{ matrix.python-version }}
-              uses: actions/setup-python@v4
+              uses: actions/setup-python@v5
               with:
                 python-version: ${{ matrix.python-version }}
                 cache: 'pip'

edtf/natlang/tests.py

@@ -1,26 +1,30 @@
-import unittest
+import pytest
 from edtf.natlang.en import text_to_edtf
 
+# TODO update the tests and code to test and output the new spec
+
 # where examples are tuples, the second item is the normalised output
-EXAMPLES = (
-    ('active late 17th-19th centuries', '16xx/18xx'), # ignoring 'late' for now
-    ('active 17-19th Centuries', '16xx/18xx'), # ignoring 'late' for now
+@pytest.mark.parametrize("input_text,expected_output", [
+    # Ignoring 'late' for simplicity in these examples
+    ('active late 17th-19th centuries', '16xx/18xx'),
+    ('active 17-19th Centuries', '16xx/18xx'),
 
     # Unrecognised values
     ('', None),
     ('this isn\'t a date', None),
 
-    # Explicity rejected values that would otherwise be badly converted
+    # Explicitly rejected values that would otherwise be badly converted
     ('23rd Dynasty', None),
 
-    ('90', '1990'),  # implied century
+    # Implied century and specific years
+    ('90', '1990'), # Implied century
     ('1860', '1860'),
     ('the year 1800', '1800'),
     ('the year 1897', '1897'),
     ('January 2008', '2008-01'),
     ('January 12, 1940', '1940-01-12'),
 
-    # uncertain/approximate
+    # Uncertain or approximate dates
     ('1860?', '1860?'),
     ('1862 (uncertain)', '1862?'),
     ('maybe 1862', '1862?'),
@@ -31,11 +35,11 @@
     ('~ Feb 1812', '1812-02~'),
     ('circa Feb 1812', '1812-02~'),
     ('Feb 1812 approx', '1812-02~'),
-    ('c1860', '1860~'),  # different abbreviations
-    ('c.1860', '1860~'),  # with or without .
+    ('c1860', '1860~'), # Different abbreviations
+    ('c.1860', '1860~'), # With or without .
     ('ca1860', '1860~'),
     ('ca.1860', '1860~'),
-    ('c 1860', '1860~'),  # with or without space
+    ('c 1860', '1860~'), # With or without space
     ('c. 1860', '1860~'),
     ('ca. 1860', '1860~'),
     ('approx 1860', '1860~'),
@@ -44,75 +48,72 @@
     ('approximately 1860', '1860~'),
     ('about 1860', '1860~'),
     ('about Spring 1849', '1849-21~'),
-    ('notcirca 1860', '1860'),  # avoid words containing circa
-    ('attica 1802', '1802'),
-    # avoid false positive circa at the end of preceding word
-    ('attic. 1802', '1802'),  # avoid false positive circa
+    ('notcirca 1860', '1860'),  # Avoid words containing 'circa'
+    ('attica 1802', '1802'), # Avoid false positive 'circa' at the end of preceding word
+    ('attic. 1802', '1802'), # Avoid false positive 'circa'
 
-    # masked precision
-    ('1860s', '186x'),  # 186x has decade precision, 186u has year precision.
+    # Masked precision
+    ('1860s', '186x'), # 186x has decade precision, 186u has year precision.
 
-    # masked precision + uncertainty
+    # Masked precision + uncertainty
     ('ca. 1860s', '186x~'),
     ('c. 1860s', '186x~'),
     ('Circa 1840s', '184x~'),
     ('circa 1840s', '184x~'),
     ('ca. 1860s?', '186x?~'),
     ('uncertain: approx 1862', '1862?~'),
 
-    # masked precision with first decade (ambiguous)
-    ('1800s', '18xx'),  # without additional uncertainty, use the century
-    ('2000s', '20xx'),  # without additional uncertainty, use the century
-    ('c1900s', '190x~'),  # if there's additional uncertainty, use the decade
-    ('c1800s?', '180x?~'),  # if there's additional uncertainty, use the decade
+    # Ambiguous masked precision for centuries and decades
+    ('1800s', '18xx'), # Without additional uncertainty, use the century
+    ('2000s', '20xx'), # Without additional uncertainty, use the century
+    ('c1900s', '190x~'), # If there's additional uncertainty, use the decade
+    ('c1800s?', '180x?~'), # If there's additional uncertainty, use the decade
 
-    # unspecified
+    # Unspecified dates
     ('January 12', 'uuuu-01-12'),
     ('January', 'uuuu-01'),
     ('10/7/2008', '2008-10-07'),
     ('7/2008', '2008-07'),
 
-    # seasons
+    # Seasons mapped to specific codes
     ('Spring 1872', '1872-21'),
     ('Summer 1872', '1872-22'),
     ('Autumn 1872', '1872-23'),
     ('Fall 1872', '1872-23'),
     ('Winter 1872', '1872-24'),
 
-    # before/after
+    # Dates relative to known events (before/after)
     ('earlier than 1928', 'unknown/1928'),
     ('before 1928', 'unknown/1928'),
     ('after 1928', '1928/unknown'),
     ('later than 1928', '1928/unknown'),
     ('before January 1928', 'unknown/1928-01'),
     ('before 18 January 1928', 'unknown/1928-01-18'),
 
-    # before/after approx
+    # Approximations combined with before/after
     ('before approx January 18 1928', 'unknown/1928-01-18~'),
     ('before approx January 1928', 'unknown/1928-01~'),
     ('after approx January 1928', '1928-01~/unknown'),
     ('after approx Summer 1928', '1928-22~/unknown'),
 
-    # before/after and uncertain/unspecificed
+    # Before and after with uncertain / unspecified components
     ('after about the 1920s', '192x~/unknown'),
     ('before about the 1900s', 'unknown/190x~'),
     ('before the 1900s', 'unknown/19xx'),
 
-    # unspecified
+    # Specifying unspecified components within a date
     # ('decade in 1800s', '18ux'), #too esoteric
     # ('decade somewhere during the 1800s', '18ux'), #lengthier. Keywords are 'in' or 'during'
-    ('year in the 1860s', '186u'),
-    # 186x has decade precision, 186u has year precision.
-    ('year in the 1800s', '18xu'),
+    ('year in the 1860s', '186u'), # 186x has decade precision 
+    ('year in the 1800s', '18xu'), # 186u has year precision
     ('year in about the 1800s', '180u~'),
     ('month in 1872', '1872-uu'),
     ('day in Spring 1849', '1849-21-uu'),
     ('day in January 1872', '1872-01-uu'),
     ('day in 1872', '1872-uu-uu'),
     ('birthday in 1872', '1872'),
-    # avoid false positive at end of preceding word
 
-    # centuries
+    # Handling centuries with approximation and uncertainty
     ('1st century', '00xx'),
     ('10c', '09xx'),
     ('19th century', '18xx'),
@@ -126,7 +127,7 @@
     ('19c?', '18xx?'),
     ('c.19c?', '18xx?~'),
 
-    # BC/AD
+    # BC/AD dating
     ('1 AD', '0001'),
     ('17 CE', '0017'),
     ('127 CE', '0127'),
@@ -136,18 +137,17 @@
     ('c127 CE', '0127~'),
     ('c1270 CE', '1270~'),
     ('c64 BCE', '-0064~'),
-    ('2nd century bc', '-01xx'),  # -200 to -101
+    ('2nd century bc', '-01xx'), # -200 to -101
     ('2nd century bce', '-01xx'),
     ('2nd century ad', '01xx'),
     ('2nd century ce', '01xx'),
 
-    # c-c-c-combo
-    # just showing off now...
+    # Combining uncertainties and approximations in creative ways
     ('a day in about Spring 1849?', '1849-21-uu?~'),
 
-    # simple ranges. Not all of these results are correct EDTF, but
-    # this is as good as the EDTF implementation and simple natural
-    # language parser we have.
+    # Simple date ranges, showcasing both the limitations and capabilities of the parser
+    # Not all of these results are correct EDTF, but this is as good as the EDTF implementation
+    # and simple natural language parser we have.
     ('1851-1852', '1851/1852'),
     ('1851-1852; printed 1853-1854', '1851/1852'),
     ('1851-52', '1851/1852'),
@@ -156,7 +156,6 @@
     ('1857-mid 1860s', '1857/186x'),
     ('1858/1860', '[1858, 1860]'),
     ('1860s-1870s', '186x/187x'),
-    ('1861, printed 1869', '1861'),
     ('1910-30', '1910/1930'),
     ('active 1910-30', '1910/1930'),
     ('1861-67', '1861/1867'),
@@ -174,16 +173,13 @@
     ('1900; 1973', '1900'),
     ('1900; printed 1912', '1900'),
     ('1915 late - autumn 1916', '1915/1916-23'),
-
-    ('1915, from Camerawork, October 1916', '1915'), # should be {1915, 1916-10}
+    ('1915, from Camerawork, October 1916', '1915'),  # should be {1915, 1916-10}
     ('1920s -early 1930s', '192x/193x'),
     ('1930s, printed early 1960s', '193x'), # should be something like {193x, 196x},
-    # though those forms aren't explicitly supported in the spec.
     ('1932, printed 1976 by Gunther Sander', '1932'), # should be {1932, 1976}
-    ('1938, printed 1940s-1950s', '1938'), # should be something like {1938, 194x-195x}
-
-
+    ('1938, printed 1940s-1950s', '1938') # should be something like {1938, 194x-195x}
 
+    # Uncertain and approximate on different parts of the date
     # for these to work we need to recast is_uncertain and is_approximate
     # such that they work on different parts. Probably worth rolling our own
     # dateparser at this point.
@@ -194,22 +190,13 @@
     # ('a day in about Spring in about 1849', '1849~-21~-uu'),
     # ('maybe January in some year in about the 1830s', '183u~-01?'),
     # ('about July? in about 1849', '1849~-07?~'),
-)
-
-
-class TestLevel0(unittest.TestCase):
-    def test_natlang(self):
-        """
-        For each of the examples, establish that:
-            - the unicode of the parsed object is acceptably equal to the EDTF string
-            - the parsed object is a subclass of EDTFObject
-        :return:
-        """
-        for i, o in EXAMPLES:
-            e = text_to_edtf(i)
-            print("%s => %s" % (i, e))
-            self.assertEqual(e, o)
+])
 
+def test_natlang(input_text, expected_output):
+    """
+    Test natural language conversion to EDTF format:
+    Verify that the conversion from text to EDTF format matches the expected output.
+    """
+    result = text_to_edtf(input_text)
+    assert result == expected_output, f"Failed for input: {input_text}"
 
-if __name__ == '__main__':
-    unittest.main()

edtf/parser/parser_classes.py

@@ -713,10 +713,18 @@ def __str__(self):
         return "[%s]" % (", ".join([str(o) for o in self.objects]))
 
     def _strict_date(self, lean):
+        strict_dates = [x._strict_date(lean) for x in self.objects]
+        # Accounting for possible 'inf' and '-inf' values
         if lean == LATEST:
-            return max([x._strict_date(lean) for x in self.objects])
+            if any(isinstance(d, float) and d == float('inf') for d in strict_dates):
+                return float('inf')
+            else:
+                return max((d for d in strict_dates if not isinstance(d, float)), default=float('inf'))
         else:
-            return min([x._strict_date(lean) for x in self.objects])
+            if any(isinstance(d, float) and d == float('-inf') for d in strict_dates):
+                return float('-inf')
+            else:
+                return min((d for d in strict_dates if not isinstance(d, float)), default=float('-inf'))
 
 
 class MultipleDates(EDTFObject):