1#![cfg_attr(test, feature(test))]
2#![cfg_attr(doc, feature(doc_cfg))]
3
4#[cfg(feature = "censor")]
5pub(crate) mod banned;
6#[cfg(feature = "censor")]
7pub(crate) mod buffer_proxy_iterator;
8#[cfg(feature = "censor")]
9pub(crate) mod censor;
10#[cfg(feature = "censor")]
11pub(crate) mod feature_cell;
12#[cfg(feature = "censor")]
13pub(crate) mod mtch;
14#[cfg(feature = "censor")]
15pub(crate) mod replacements;
16#[cfg(feature = "censor")]
17pub(crate) mod trie;
18#[cfg(feature = "censor")]
19pub(crate) mod typ;
20
21#[cfg(feature = "context")]
22pub(crate) mod context;
23
24#[cfg(feature = "pii")]
25mod pii;
26#[cfg(feature = "width")]
27pub(crate) mod width;
28
29#[cfg(feature = "censor")]
30pub use banned::Banned;
31#[cfg(feature = "censor")]
32pub use replacements::Replacements;
33#[cfg(feature = "censor")]
34pub use trie::Trie;
35
36#[cfg(feature = "width")]
37pub use width::{trim_to_width, width, width_str, width_str_max_unbroken, WordBreak};
38
39#[cfg(feature = "censor")]
40pub use typ::Type;
41
42#[cfg(feature = "censor")]
43pub use censor::{Censor, CensorIter, CensorStr};
44
45#[cfg(feature = "censor")]
47pub(crate) type Map<K, V> = rustc_hash::FxHashMap<K, V>;
48
49#[cfg(feature = "censor")]
50pub(crate) type Set<V> = rustc_hash::FxHashSet<V>;
51
52#[cfg(feature = "customize")]
53#[allow(deprecated)]
54pub use censor::add_word;
55
56#[cfg(all(feature = "context", feature = "width"))]
57pub use context::ContextWordBreakOptions;
58#[cfg(feature = "context")]
59pub use context::{
60 BlockReason, Context, ContextProcessingOptions, ContextRateLimitOptions,
61 ContextRepetitionLimitOptions,
62};
63
64#[cfg(feature = "pii")]
65pub use pii::censor_and_analyze_pii;
66
67pub fn trim_whitespace(s: &str) -> &str {
70 s.trim_start_matches(|c| is_whitespace(c) || matches!(c, '\u{0488}' | '\u{0489}'))
74 .trim_end_matches(is_whitespace)
75}
76
77pub fn is_whitespace(c: char) -> bool {
80 use finl_unicode::categories::CharacterCategories;
81
82 c.is_whitespace()
90 || c.is_other()
91 || c.is_format()
92 || matches!(
93 c,
94 '\u{115F}' | '\u{1160}' | '\u{20DD}' | '\u{2800}' | '\u{3164}' | '\u{FFA0}' | '\u{FFFC}'
95 )
96}
97
98#[cfg(test)]
99mod tests {
100 #![allow(unused_imports)]
101 extern crate test;
102
103 #[test]
104 fn trim_whitespace() {
105 assert_eq!(crate::trim_whitespace("\u{0020}\u{00A0}\u{2000}\u{2001}\u{2002}\u{2004}\u{2005}\u{2006}\u{2007}\u{2008}\u{2009}\u{200A}\u{200B}\u{200C}\u{200D}\u{2028}\u{205F}\u{3000}"), "");
107
108 assert_eq!(
110 crate::trim_whitespace(" \u{1160} \u{2800} abc \u{3164} \u{FFA0} \t \u{115F} \n "),
111 "abc"
112 );
113
114 assert_eq!(
116 crate::trim_whitespace(
117 "\u{FFF9}\u{FFFA}\u{FFFB}\u{FFFC}\u{0488}\u{1160}\u{0489}\u{1160}\u{0488}\u{1160}\u{0489}abc\u{0488}\u{0489}"
118 ),
119 "abc\u{0488}\u{0489}"
120 )
121 }
122
123 #[test]
124 fn is_whitespace() {
125 assert!(crate::is_whitespace(' '));
126 assert!(crate::is_whitespace('\u{2800}'));
127 assert!(!crate::is_whitespace('a'));
128 }
129}
130
131doc_comment::doctest!("../README.md");