backup/pywb
1
0
Fork 0
mirror of https://github.com/webrecorder/pywb.git synced 2025-03-24 06:59:52 +01:00
Code Issues Projects Releases Activity

fuzzymatcher: fix fuzzymatcher to remove '~' from prefix match, per changes from #250

Browse Source
This commit is contained in:
Ilya Kreymer 2017-10-13 11:37:03 -07:00
parent 056aed085c
commit f851d4b473
2 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
pywb/warcserver/index/fuzzymatcher.py
View File

@ -17,7 +17,7 @@ FuzzyRule = namedtuple('FuzzyRule',
# ============================================================================ # ============================================================================
class FuzzyMatcher(object): class FuzzyMatcher(object):
DEFAULT_FILTER = ['~urlkey:{0}'] DEFAULT_FILTER = ['urlkey:{0}']
DEFAULT_MATCH_TYPE = 'prefix' DEFAULT_MATCH_TYPE = 'prefix'
DEFAULT_REPLACE_AFTER = '?' DEFAULT_REPLACE_AFTER = '?'

4
pywb/warcserver/index/test/test_fuzzymatcher.py
View File

@ -36,7 +36,7 @@ class TestFuzzy(object):
return params return params
def get_expected(self, url, mime='text/html', filters=None): def get_expected(self, url, mime='text/html', filters=None):
filters = filters or ['~urlkey:'] filters = filters or ['urlkey:']
exp = [{'filter': filters, exp = [{'filter': filters,
'is_fuzzy': True, 'is_fuzzy': True,
'urlkey': canonicalize(url), 'urlkey': canonicalize(url),
@ -107,7 +107,7 @@ class TestFuzzy(object):
actual_url = 'http://youtube.com/get_video_info?a=d&html5=true&___abc=125&video_id=ABCD&id=1234' actual_url = 'http://youtube.com/get_video_info?a=d&html5=true&___abc=125&video_id=ABCD&id=1234'
params = self.get_params(url, actual_url) params = self.get_params(url, actual_url)
cdx_iter, errs = self.fuzzy(self.source, params) cdx_iter, errs = self.fuzzy(self.source, params)
filters = ['~urlkey:html5=true', '~urlkey:video_id=abcd'] filters = ['urlkey:html5=true', 'urlkey:video_id=abcd']
assert list(cdx_iter) == self.get_expected(url=actual_url, filters=filters) assert list(cdx_iter) == self.get_expected(url=actual_url, filters=filters)
def test_no_fuzzy_custom_rule_video_id_diff(self): def test_no_fuzzy_custom_rule_video_id_diff(self):