michael_sayapin/holiday-cn
8
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Implement parsing (WIP)

Browse Source
This commit is contained in:
NateScarlet 2019-03-08 00:23:52 +08:00
parent 7a2d9af66e
commit 5f68f773a7
No known key found for this signature in database
GPG Key ID: 5C242793B070309C
1 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
fetch_holidays.py
View File

@ -86,7 +86,7 @@ class SentenceParser:
def _extract_dates_2(self, value): def _extract_dates_2(self, value):
match = re.match( match = re.match(
r'(?:(\d+)年)?(?:(\d+)月)(\d+)日(?:至|-)(?:(\d+)年)?(?:(\d+)月)?(\d+)日', value) r'(?:(\d+)年)?(?:(\d+)月)(\d+)日(?:至|-|—)(?:(\d+)年)?(?:(\d+)月)?(\d+)日', value)
if match: if match:
groups = [_cast_int(i) for i in match.groups()] groups = [_cast_int(i) for i in match.groups()]
assert len(groups) == 6, groups assert len(groups) == 6, groups
@ -171,8 +171,12 @@ class SentenceParser:
def parse_holiday_description(description: str, year: int): def parse_holiday_description(description: str, year: int):
date_memory = set()
for i in re.split('，|。', description): for i in re.split('，|。', description):
for j in SentenceParser(i, year).parse(): for j in SentenceParser(i, year).parse():
if j['date'] in date_memory:
continue
date_memory.add(j['date'])
yield j yield j