sophia_iri/
_regex.rs

1use lazy_static::lazy_static;
2use regex::Regex;
3
4/// Check whether `ns` and `suffix` concatenate into a valid (absolute or relative) IRI reference.
5///
6/// # Performances
7/// In the current implementation, a new String is allocated if `suffix`
8/// is not `None`.
9/// Future implementations may be smarter about this.
10#[inline]
11#[must_use]
12pub fn is_valid_suffixed_iri_ref(ns: &str, suffix: Option<&str>) -> bool {
13    match suffix {
14        None => is_valid_iri_ref(ns),
15        Some(suffix) => {
16            let mut buffer = String::with_capacity(ns.len() + suffix.len());
17            buffer.push_str(ns);
18            buffer.push_str(suffix);
19            is_valid_iri_ref(&buffer)
20        }
21    }
22}
23
24/// Check whether `txt` is a valid (absolute or relative) IRI reference.
25#[inline]
26#[must_use]
27pub fn is_valid_iri_ref(txt: &str) -> bool {
28    IRI_REGEX.is_match(txt) || IRELATIVE_REF_REGEX.is_match(txt)
29}
30
31/// Check whether `txt` is an absolute IRI reference.
32#[inline]
33#[must_use]
34pub fn is_absolute_iri_ref(txt: &str) -> bool {
35    IRI_REGEX.is_match(txt)
36}
37
38/// Check whether `txt` is a relative IRI reference.
39#[inline]
40#[must_use]
41pub fn is_relative_iri_ref(txt: &str) -> bool {
42    IRELATIVE_REF_REGEX.is_match(txt)
43}
44
45lazy_static! {
46    /// Match an absolute IRI reference.
47    pub(crate) static ref IRI_REGEX: Regex = Regex::new(r"(?x)^
48        #scheme
49       ( # CAPTURE scheme
50        [A-Za-z] [-A-Za-z0-9+.]*
51       )
52        :
53        #ihier_part
54        (?: #iauthority + ipath_abempty
55          //
56         ( # CAPTURE iauthority
57          (?: # iuserinfo
58            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:]
59          |
60            %[0-9a-fA-F]{2}
61          )*
62          @
63          )?
64          # ihost
65          (?: # ip_literal
66             \[
67            (?: # ipv6address
68              (?:
69                (?:[0-9a-fA-F]{1,4}:){6}
70                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
71              |
72                ::
73                (?:[0-9a-fA-F]{1,4}:){5}
74                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
75              |
76                (?:[0-9a-fA-F]{1,4})?
77                ::
78                (?:[0-9a-fA-F]{1,4}:){4}
79                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
80              |
81                (?:(?:[0-9a-fA-F]{1,4}:){0,1}:[0-9a-fA-F]{1,4})?
82                ::
83                (?:[0-9a-fA-F]{1,4}:){3}
84                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
85              |
86                (?:(?:[0-9a-fA-F]{1,4}:){0,2}:[0-9a-fA-F]{1,4})?
87                ::
88                (?:[0-9a-fA-F]{1,4}:){2}
89                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
90              |
91                (?:(?:[0-9a-fA-F]{1,4}:){0,3}:[0-9a-fA-F]{1,4})?
92                ::
93                [0-9a-fA-F]{1,4}:
94                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
95              |
96                (?:(?:[0-9a-fA-F]{1,4}:){0,4}:[0-9a-fA-F]{1,4})?
97                ::
98                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
99              |
100                (?:(?:[0-9a-fA-F]{1,4}:){0,5}:[0-9a-fA-F]{1,4})?
101                ::
102                [0-9a-fA-F]{1,4}
103              |
104                (?:(?:[0-9a-fA-F]{1,4}:){0,6}:[0-9a-fA-F]{1,4})?
105                ::
106              )
107            | # ipvfuture
108              v[0-9a-fA-F]+ \. [-A-Za-z0-9._~!$&'()*+,;=:]+
109            )
110             \]
111          | # ipv4address
112            (?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5])) (?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3}
113          | # ireg_name
114              (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=]
115              | %[0-9a-fA-F]{2}
116              )*
117          )
118          (?:
119            :
120            [0-9]* # port
121          )?
122         )
123          #ipath_abempty
124         ( # CAPTURE ipath_abempty
125          (?:
126            /
127            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
128            | %[0-9a-fA-F]{2}
129            )*
130          )*
131         )
132        | #ipath_absolute
133         ( # CAPTURE ipath_absolute
134          /
135          (?:
136            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
137            | %[0-9a-fA-F]{2}
138            )*
139            (?:
140              /
141              (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
142              | %[0-9a-fA-F]{2}
143              )*
144            )*
145          )?
146         )
147        | #ipath_rootless
148         ( # CAPTURE ipath_rootless
149          (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
150          | %[0-9a-fA-F]{2}
151          )+
152          (?:
153            /
154            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
155            | %[0-9a-fA-F]{2}
156            )*
157          )*
158         )
159        )? # optional because of ipath_empty
160        (?: # ?iquery
161          \?
162         ( # CAPTURE iquery
163          (?:
164            [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@'\u{E000}-\u{F8FF}\u{F0000}-\u{FFFFD}\u{100000}-\u{10FFFD}/?]
165            | %[0-9a-fA-F]{2}
166          )*
167         )
168        )?
169        (?: # #ifragment
170          \#
171         ( # CAPTURE ifragment
172          (?:
173            [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@/?]
174            | %[0-9a-fA-F]{2}
175          )*
176         )
177        )?
178    $").unwrap();
179
180    /// Match a relative IRI reference.
181    pub(crate) static ref IRELATIVE_REF_REGEX: Regex = Regex::new(r"(?x)^
182        #irelative_part
183        (?: #iauthority + ipath_abempty
184          //
185         ( # CAPTURE iauthority
186          (?: # iuserinfo
187            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:]
188          |
189            %[0-9a-fA-F]{2}
190          )*
191          @
192          )?
193          # ihost
194          (?: # ip_literal
195             \[
196            (?: # ipv6address
197              (?:
198                (?:[0-9a-fA-F]{1,4}:){6}
199                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
200              |
201                ::
202                (?:[0-9a-fA-F]{1,4}:){5}
203                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
204              |
205                (?:[0-9a-fA-F]{1,4})?
206                ::
207                (?:[0-9a-fA-F]{1,4}:){4}
208                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
209              |
210                (?:(?:[0-9a-fA-F]{1,4}:){0,1}:[0-9a-fA-F]{1,4})?
211                ::
212                (?:[0-9a-fA-F]{1,4}:){3}
213                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
214              |
215                (?:(?:[0-9a-fA-F]{1,4}:){0,2}:[0-9a-fA-F]{1,4})?
216                ::
217                (?:[0-9a-fA-F]{1,4}:){2}
218                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
219              |
220                (?:(?:[0-9a-fA-F]{1,4}:){0,3}:[0-9a-fA-F]{1,4})?
221                ::
222                [0-9a-fA-F]{1,4}:
223                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
224              |
225                (?:(?:[0-9a-fA-F]{1,4}:){0,4}:[0-9a-fA-F]{1,4})?
226                ::
227                (?:[0-9a-fA-F]{1,4}:[0-9a-fA-F]{1,4}|(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))(?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3})
228              |
229                (?:(?:[0-9a-fA-F]{1,4}:){0,5}:[0-9a-fA-F]{1,4})?
230                ::
231                [0-9a-fA-F]{1,4}
232              |
233                (?:(?:[0-9a-fA-F]{1,4}:){0,6}:[0-9a-fA-F]{1,4})?
234                ::
235              )
236            | # ipvfuture
237              v[0-9a-fA-F]+ \. [-A-Za-z0-9._~!$&'()*+,;=:]+
238            )
239             \]
240          | # ipv4address
241            (?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5])) (?:\.(?:[0-9]|(?:[1-9][0-9])|(?:1[0-9]{2})|(?:2[0-4][0-9])|(?:25[0-5]))){3}
242          | # ireg_name
243              (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=]
244              | %[0-9a-fA-F]{2}
245              )*
246          )
247          (?:
248            :
249            [0-9]* # port
250          )?
251         )
252          #ipath_abempty
253         ( # CAPTURE ipath_abempty
254          (?:
255            /
256            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
257            | %[0-9a-fA-F]{2}
258            )*
259          )*
260         )
261        | #ipath_absolute
262         ( # CAPTURE ipath_absolute
263          /
264          (?:
265            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
266            | %[0-9a-fA-F]{2}
267            )*
268            (?:
269              /
270              (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
271              | %[0-9a-fA-F]{2}
272              )*
273            )*
274          )?
275         )
276        | #ipath_noscheme
277         ( # CAPTURE ipath_noscheme
278          (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=@]
279          | %[0-9a-fA-F]{2}
280          )+
281          (?:
282            /
283            (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@]
284            | %[0-9a-fA-F]{2}
285            )*
286          )*
287         )
288        )? # optional because of ipath_empty
289        (?: # ?iquery
290          \?
291         ( # CAPTURE iquery
292          (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@'\u{E000}-\u{F8FF}\u{F0000}-\u{FFFFD}\u{100000}-\u{10FFFD}/?]
293          | %[0-9a-fA-F]{2}
294          )*
295         )
296        )?
297        (?: # #ifragment
298            \#
299         ( # CAPTURE ifragment
300          (?: [-A-Za-z0-9._~\u{A0}-\u{D7FF}\u{F900}-\u{FDCF}\u{FDF0}-\u{FFEF}\u{10000}-\u{1FFFD}\u{20000}-\u{2FFFD}\u{30000}-\u{3FFFD}\u{40000}-\u{4FFFD}\u{50000}-\u{5FFFD}\u{60000}-\u{6FFFD}\u{70000}-\u{7FFFD}\u{80000}-\u{8FFFD}\u{90000}-\u{9FFFD}\u{A0000}-\u{AFFFD}\u{B0000}-\u{BFFFD}\u{C0000}-\u{CFFFD}\u{D0000}-\u{DFFFD}\u{E1000}-\u{EFFFD}!$&'()*+,;=:@/?]
301          | %[0-9a-fA-F]{2}
302          )*
303         )
304        )?
305    $").unwrap();
306}
307
308#[cfg(test)]
309mod test {
310    use super::*;
311    use crate::test::{NEGATIVE_IRIS, POSITIVE_IRIS};
312
313    #[test]
314    fn regex_abs() {
315        for (txt, parsed) in POSITIVE_IRIS {
316            assert_eq!(IRI_REGEX.is_match(txt), parsed.0);
317        }
318        for txt in NEGATIVE_IRIS {
319            assert!(!IRI_REGEX.is_match(txt));
320        }
321    }
322
323    #[test]
324    fn regex_rel() {
325        for (txt, parsed) in POSITIVE_IRIS {
326            assert_eq!(IRELATIVE_REF_REGEX.is_match(txt), !parsed.0);
327        }
328        for txt in NEGATIVE_IRIS {
329            assert!(!IRELATIVE_REF_REGEX.is_match(txt));
330        }
331    }
332}
sophia_iri/_regex.rs

sophia_iri/
_regex.rs
