def _path_trigram_filter(path, is_case_sensitive):
"""Return an ES filter clause that returns docs whose paths match the
given path all the way to their ends.
If a given path starts with a /, the user is explicitly requesting a match
starting at the root level.
"""
if path.startswith('/'):
path = path[1:] # Leading slashes aren't stored in the index.
regex = '^{0}$' # Insist it start at the beginning.
else:
regex = '(/|^){0}$' # Start at any path segment.
return es_regex_filter(
regex_grammar.parse(
regex.format(
re.escape(
path.encode('ascii', 'backslashreplace')
)
)
),
'path',
is_case_sensitive
)
def test_python_visitor():
"""Make sure we can render out Python regexes from parse trees.
There's just one difference from JS so far.
"""
eq_(PythonRegexVisitor().visit(regex_grammar.parse(r'\a')), r'\a')
def test_python_visitor():
"""Make sure we can render out Python regexes from parse trees.
There's just one difference from JS so far.
"""
eq_(PythonRegexVisitor().visit(regex_grammar.parse(r'\a')), r'\a')
def filter(self):
glob = self._term["arg"]
try:
return es_regex_filter(
regex_grammar.parse(glob_to_regex(glob)), "path", is_case_sensitive=self._term["case_sensitive"]
)
except NoTrigrams:
raise BadTerm("Path globs need at least 3 literal characters in a row " "for speed.")
def js_eq(regex, expected):
"""Assert that taking a regex apart into an AST and then building a JS
regex from it matches the expected result.
:arg regex: A string representing a regex pattern
:arg expected: What to compare the reconstructed regex against
"""
eq_(JsRegexVisitor().visit(regex_grammar.parse(regex)), expected)
def js_eq(regex, expected):
"""Assert that taking a regex apart into an AST and then building a JS
regex from it matches the expected result.
:arg regex: A string representing a regex pattern
:arg expected: What to compare the reconstructed regex against
"""
eq_(JsRegexVisitor().visit(regex_grammar.parse(regex)), expected)
def filter(self):
glob = self._term['arg']
try:
return es_regex_filter(
regex_grammar.parse(glob_to_regex(glob)),
'path',
is_case_sensitive=self._term['case_sensitive'])
except NoTrigrams:
raise BadTerm('Path globs need at least 3 literal characters in a row '
'for speed.')
def filter(self):
glob = self._term['arg']
try:
return es_regex_filter(
regex_grammar.parse(glob_to_regex(glob)),
'path',
is_case_sensitive=self._term['case_sensitive'])
except NoTrigrams:
raise BadTerm('Path globs need at least 3 literal characters in a row '
'for speed.')
def _regex_filter(self, path_seg_property_name, no_trigrams_error_text):
"""Return an ES regex filter that matches this filter's glob against the
path segment at path_seg_property_name.
"""
glob = self._term['arg']
try:
return es_regex_filter(
regex_grammar.parse(glob_to_regex(glob)),
path_seg_property_name,
is_case_sensitive=self._term['case_sensitive'])
except NoTrigrams:
raise BadTerm(no_trigrams_error_text)
def _regex_filter(self, path_seg_property_name, no_trigrams_error_text):
"""Return an ES regex filter that matches this filter's glob against the
path segment at path_seg_property_name.
"""
glob = self._term['arg']
try:
return es_regex_filter(
regex_grammar.parse(glob_to_regex(glob)),
path_seg_property_name,
is_case_sensitive=self._term['case_sensitive'])
except NoTrigrams:
raise BadTerm(no_trigrams_error_text)
def _path_trigram_filter(path, is_case_sensitive):
"""Return an ES filter clause that returns docs whose paths match the
given path all the way to their ends.
If a given path starts with a /, the user is explicitly requesting a match
starting at the root level.
"""
if path.startswith('/'):
path = path[1:] # Leading slashes aren't stored in the index.
regex = '^{0}$' # Insist it start at the beginning.
else:
regex = '(/|^){0}$' # Start at any path segment.
return es_regex_filter(regex_grammar.parse(regex.format(re.escape(path))),
'path', is_case_sensitive)
def __init__(self, term, enabled_plugins):
"""Compile the Python equivalent of the regex so we don't have to lean
on the regex cache during highlighting.
Python's regex cache is naive: after it hits 100, it just clears: no
LRU.
"""
super(RegexpFilter, self).__init__(term, enabled_plugins)
try:
self._parsed_regex = regex_grammar.parse(term['arg'])
except ParseError:
raise BadTerm('Invalid regex.')
self._compiled_regex = (
re.compile(PythonRegexVisitor().visit(self._parsed_regex),
flags=0 if self._term['case_sensitive'] else re.I))
def __init__(self, term, enabled_plugins):
"""Compile the Python equivalent of the regex so we don't have to lean
on the regex cache during highlighting.
Python's regex cache is naive: after it hits 100, it just clears: no
LRU.
"""
super(RegexpFilter, self).__init__(term, enabled_plugins)
try:
self._parsed_regex = regex_grammar.parse(term['arg'])
except ParseError:
raise BadTerm('Invalid regex.')
self._compiled_regex = (
re.compile(PythonRegexVisitor().visit(self._parsed_regex),
flags=0 if self._term['case_sensitive'] else re.I))
def test_parse_regexp():
regex_grammar.parse('hello+ dolly')
regex_grammar.parse('hello+|hi')
regex_grammar.parse(r'(hello|hi) dolly')
regex_grammar.parse(r'(hello|hi|) dolly')
regex_grammar.parse(r'(hello||hi) dolly')
regex_grammar.parse(r'|hello|hi')
regex_grammar.parse(ur'aböut \d{2}')
assert_raises(ParseError, regex_grammar.parse, '[smoo')
# This isn't supported yet, so it's better to throw an error than to
# quietly misinterpret the user's intent:
assert_raises(ParseError, regex_grammar.parse, '(?:hi)')
def visit_regex(regex):
return SubstringTreeVisitor().visit(regex_grammar.parse(regex))
def test_parse_regexp():
regex_grammar.parse('hello+ dolly')
regex_grammar.parse('hello+|hi')
regex_grammar.parse(r'(hello|hi) dolly')
regex_grammar.parse(r'(hello|hi|) dolly')
regex_grammar.parse(r'(hello||hi) dolly')
regex_grammar.parse(r'|hello|hi')
regex_grammar.parse(ur'aböut \d{2}')
assert_raises(ParseError, regex_grammar.parse, '[smoo')
# This isn't supported yet, so it's better to throw an error than to
# quietly misinterpret the user's intent:
assert_raises(ParseError, regex_grammar.parse, '(?:hi)')
def visit_regex(regex):
return SubstringTreeVisitor().visit(regex_grammar.parse(regex))
