thereadme/pyhumandate
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

fix: support day of month in the format MONTH DAY

Browse Source
This commit is contained in:
Mahdi Dibaiee 2018-01-19 17:08:25 +03:30
parent cde8b3bf89
commit 9bdd1eb47a
2 changed files with 17 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
humandate/index.py
View File

@ -44,6 +44,8 @@ symbols = ['!', ',', '@', '#', '$', '%', '^', '&', '*', '?']
transformations = [dict(zip((x + 's' for x in multipliers.keys()), multipliers.keys()))] transformations = [dict(zip((x + 's' for x in multipliers.keys()), multipliers.keys()))]
transformations.append(dict(zip(direction.keys(), (x + ' ' + '1' for x in direction.keys())))) transformations.append(dict(zip(direction.keys(), (x + ' ' + '1' for x in direction.keys()))))
# TODO: find a way to distinguish this `second` from the other `second`
#transformations.append(dict(zip(['first', 'second', 'third', 'fourth', 'fifth', 'sixth', 'seventh', 'eighth', 'ninth', 'tenth'], map(str, range(1, 10)))))
transformations.append({ 'tomorrow': 'next day', 'yesterday': 'past 1 day' }) transformations.append({ 'tomorrow': 'next day', 'yesterday': 'past 1 day' })
transformations.append({ 'next': '1', 'upcoming': '1', 'following': '1', 'a': '1', 'an': '1' }) transformations.append({ 'next': '1', 'upcoming': '1', 'following': '1', 'a': '1', 'an': '1' })
@ -59,14 +61,23 @@ def tokenize(string):
pattern = re.compile(r'\b(' + '|'.join(ts.keys()) + r')\b') pattern = re.compile(r'\b(' + '|'.join(ts.keys()) + r')\b')
string = pattern.sub(lambda x: ts[x.group()], string) string = pattern.sub(lambda x: ts[x.group()], string)
space_pattern = re.compile(r'\s{2,}')
string = space_pattern.sub(' ', string)
words = string.lower().split(' ') words = string.lower().split(' ')
n = len(words) n = len(words)
filtered = [] filtered = []
for (i, w) in enumerate(words): for (i, w) in enumerate(words):
has_next = i + 1 < n
has_prev = i > 0
has_prev_f = len(filtered) > 0
if w in keywords: if w in keywords:
filtered.append(w) filtered.append(w)
elif w.isdigit() and ((i + 1 < n and words[i + 1] in keywords) or (i > 0 and words[i - 1] == time_prefix)): elif w.isdigit() and ((has_next and words[i + 1] in keywords) or (has_prev and words[i - 1] == time_prefix)):
filtered.append(int(w)) filtered.append(int(w))
elif w.isdigit() and has_prev_f and filtered[len(filtered) - 1] in months:
filtered.insert(len(filtered) - 1, int(w))
elif w[0:-1].isdigit(): elif w[0:-1].isdigit():
postfix = next((x for x in postfixes if w.endswith(x)), None) postfix = next((x for x in postfixes if w.endswith(x)), None)
if postfix is not None: if postfix is not None:

7
humandate/tests.py
View File

@ -1,7 +1,7 @@
import unittest import unittest
import datetime import datetime
from datetime import timedelta from datetime import timedelta
from index import parse_date, days, months, years from index import parse_date, days, months, years, tokenize
class HumanDateTests(unittest.TestCase): class HumanDateTests(unittest.TestCase):
def setUp(self): def setUp(self):
@ -52,7 +52,10 @@ class HumanDateTests(unittest.TestCase):
tests = { tests = {
'12 jan': (12, 1), '12 jan': (12, 1),
'28 february': (28, 2), '28 february': (28, 2),
'1th december': (1, 12) '1th december': (1, 12),
'jan 12': (12, 1),
'february 28': (28, 2),
'december 1th': (1, 12),
} }
for (k, (day, m)) in tests.items(): for (k, (day, m)) in tests.items():