Merge pull request #109 from ecatkins/strict_fix

Improve strict=True

datefinder/__init__.py

@@ -177,6 +177,7 @@ def extract_date_strings_inner(self, text, text_start=0, strict=False):
             # digits_modifiers = captures.get('digits_modifiers')
             # days = captures.get('days')
             months = captures.get("months")
+            years = captures.get("years")
             # timezones = captures.get('timezones')
             # delimiters = captures.get('delimiters')
             # time_periods = captures.get('time_periods')
@@ -190,6 +191,11 @@ def extract_date_strings_inner(self, text, text_start=0, strict=False):
                     len(digits) == 2
                 ):  # 19 February 2013 year 09:10
                     complete = True
+                elif (len(years)==1) and (len(digits)==2): #09/06/2018
+                    complete = True
+
+                elif (len(years)==1) and (len(months)==1) and (len(digits)==1): # '19th day of May, 2015'
+                    complete = True
 
                 if not complete:
                     continue

tests/test_extract_date_strings.py

@@ -32,7 +32,7 @@ def test_extract_date_strings(date_string, expected_match_date_string):
     ['the Friday after next Tuesday the 20th', ''], # no matches
     ['This Tuesday March 2015 in the evening', ''], # no matches
     ['They said it was on 01-03-2015', 'on 01-03-2015'], # 3 digits strict match
-    ['May 20th 2015 is nowhere near the other date', 'May 20 2015'], # one month two digit match
+    ['May 20 2015 is nowhere near the other date', 'May 20 2015'], # one month two digit match
 ])
 def test_extract_date_strings_with_strict_option(date_string, expected_match_date_string):
     """

tests/test_find_dates_strict.py

@@ -0,0 +1,32 @@
+import pytest
+import datefinder
+from datetime import datetime
+import pytz
+import sys
+import logging
+logging.basicConfig(level=logging.DEBUG, stream=sys.stdout)
+logger = logging.getLogger(__name__)
+
+today = datetime.today()
+
+
+@pytest.mark.parametrize('input_text, expected_date', [
+
+    ('June 2018', []),
+    ('09/06/18',  datetime(2018, 9, 6)),
+    ('09/06/2018', datetime(2018, 9, 6)),
+    ('recorded: 03/14/2008', datetime(2008, 3, 14)),
+    ('19th day of May, 2015', datetime(2015, 5, 19)),
+    ('19th day of May', [])
+    
+])
+def test_find_date_strings_strict(input_text, expected_date):
+    if isinstance(expected_date,list):
+        matches = list(datefinder.find_dates(input_text, strict=True))
+        assert matches == expected_date
+
+    else:
+        return_date = None
+        for return_date in datefinder.find_dates(input_text, strict=True):
+            assert return_date == expected_date
+        assert return_date is not None, 'Did not find date for test line: "{}"'.format(input_text) # handles dates