Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.11/site-packages/rfc3986/misc.py: 84%

1# Copyright (c) 2014 Rackspace 

2# Licensed under the Apache License, Version 2.0 (the "License"); 

3# you may not use this file except in compliance with the License. 

4# You may obtain a copy of the License at 

5# 

6# http://www.apache.org/licenses/LICENSE-2.0 

7# 

8# Unless required by applicable law or agreed to in writing, software 

9# distributed under the License is distributed on an "AS IS" BASIS, 

10# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 

11# implied. 

12# See the License for the specific language governing permissions and 

13# limitations under the License. 

14""" 

15Module containing compiled regular expressions and constants. 

16 

17This module contains important constants, patterns, and compiled regular 

18expressions for parsing and validating URIs and their components. 

19""" 

20 

21import re 

22import typing as t 

23 

24from . import abnf_regexp 

25 

26 

27class URIReferenceBase(t.NamedTuple): 

28 """The namedtuple used as a superclass of URIReference and IRIReference.""" 

29 

30 scheme: t.Optional[str] 

31 authority: t.Optional[str] 

32 path: t.Optional[str] 

33 query: t.Optional[str] 

34 fragment: t.Optional[str] 

35 

36 

37important_characters = { 

38 "generic_delimiters": abnf_regexp.GENERIC_DELIMITERS, 

39 "sub_delimiters": abnf_regexp.SUB_DELIMITERS, 

40 # We need to escape the '*' in this case 

41 "re_sub_delimiters": abnf_regexp.SUB_DELIMITERS_RE, 

42 "unreserved_chars": abnf_regexp.UNRESERVED_CHARS, 

43 # We need to escape the '-' in this case: 

44 "re_unreserved": abnf_regexp.UNRESERVED_RE, 

45} 

46 

47# For details about delimiters and reserved characters, see: 

48# http://tools.ietf.org/html/rfc3986#section-2.2 

49GENERIC_DELIMITERS = abnf_regexp.GENERIC_DELIMITERS_SET 

50SUB_DELIMITERS = abnf_regexp.SUB_DELIMITERS_SET 

51RESERVED_CHARS = abnf_regexp.RESERVED_CHARS_SET 

52# For details about unreserved characters, see: 

53# http://tools.ietf.org/html/rfc3986#section-2.3 

54UNRESERVED_CHARS = abnf_regexp.UNRESERVED_CHARS_SET 

55NON_PCT_ENCODED = abnf_regexp.NON_PCT_ENCODED_SET 

56 

57URI_MATCHER = re.compile(abnf_regexp.URL_PARSING_RE) 

58 

59SUBAUTHORITY_MATCHER = re.compile( 

60 ( 

61 "^(?:(?P<userinfo>{})@)?" # userinfo 

62 "(?P<host>{})" # host 

63 "(?::(?P<port>{}))?$" # port 

64 ).format( 

65 abnf_regexp.USERINFO_RE, abnf_regexp.HOST_PATTERN, abnf_regexp.PORT_RE 

66 ) 

67) 

68 

69 

70HOST_MATCHER = re.compile("^" + abnf_regexp.HOST_RE + "$") 

71IPv4_MATCHER = re.compile("^" + abnf_regexp.IPv4_RE + "$") 

72IPv6_MATCHER = re.compile(r"^\[" + abnf_regexp.IPv6_ADDRZ_RFC4007_RE + r"\]$") 

73 

74# Used by host validator 

75IPv6_NO_RFC4007_MATCHER = re.compile(r"^\[%s\]$" % (abnf_regexp.IPv6_ADDRZ_RE)) 

76 

77# Matcher used to validate path components 

78PATH_MATCHER = re.compile(abnf_regexp.PATH_RE) 

79 

80 

81# ################################## 

82# Query and Fragment Matcher Section 

83# ################################## 

84 

85QUERY_MATCHER = re.compile(abnf_regexp.QUERY_RE) 

86 

87FRAGMENT_MATCHER = QUERY_MATCHER 

88 

89# Scheme validation, see: http://tools.ietf.org/html/rfc3986#section-3.1 

90SCHEME_MATCHER = re.compile(f"^{abnf_regexp.SCHEME_RE}$") 

91 

92RELATIVE_REF_MATCHER = re.compile( 

93 r"^%s(\?%s)?(#%s)?$" 

94 % ( 

95 abnf_regexp.RELATIVE_PART_RE, 

96 abnf_regexp.QUERY_RE, 

97 abnf_regexp.FRAGMENT_RE, 

98 ) 

99) 

100 

101# See http://tools.ietf.org/html/rfc3986#section-4.3 

102ABSOLUTE_URI_MATCHER = re.compile( 

103 r"^%s:%s(\?%s)?$" 

104 % ( 

105 abnf_regexp.COMPONENT_PATTERN_DICT["scheme"], 

106 abnf_regexp.HIER_PART_RE, 

107 abnf_regexp.QUERY_RE[1:-1], 

108 ) 

109) 

110 

111# ############### 

112# IRIs / RFC 3987 

113# ############### 

114 

115IRI_MATCHER = re.compile(abnf_regexp.URL_PARSING_RE, re.UNICODE) 

116 

117ISUBAUTHORITY_MATCHER = re.compile( 

118 ( 

119 "^(?:(?P<userinfo>{})@)?" # iuserinfo 

120 "(?P<host>{})" # ihost 

121 ":?(?P<port>{})?$" # port 

122 ).format( 

123 abnf_regexp.IUSERINFO_RE, abnf_regexp.IHOST_RE, abnf_regexp.PORT_RE 

124 ), 

125 re.UNICODE, 

126) 

127 

128 

129# Path merger as defined in http://tools.ietf.org/html/rfc3986#section-5.2.3 

130def merge_paths(base_uri: URIReferenceBase, relative_path: str) -> str: 

131 """Merge a base URI's path with a relative URI's path.""" 

132 if base_uri.path is None and base_uri.authority is not None: 

133 return "/" + relative_path 

134 else: 

135 path = base_uri.path or "" 

136 index = path.rfind("/") 

137 return path[:index] + "/" + relative_path 

138 

139 

140UseExisting: t.Final[t.Any] = object()