1 | // This file is part of ICU4X. For terms of use, please see the file |
2 | // called LICENSE at the top level of the ICU4X source tree |
3 | // (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ). |
4 | |
5 | use crate::provider::{AndListV1Marker, ErasedListV1Marker, OrListV1Marker, UnitListV1Marker}; |
6 | use crate::ListError; |
7 | use crate::ListLength; |
8 | use core::fmt::{self, Write}; |
9 | use icu_provider::prelude::*; |
10 | use writeable::*; |
11 | |
12 | #[cfg (doc)] |
13 | extern crate writeable; |
14 | |
15 | /// A formatter that renders sequences of items in an i18n-friendly way. See the |
16 | /// [crate-level documentation](crate) for more details. |
17 | #[derive (Debug)] |
18 | pub struct ListFormatter { |
19 | data: DataPayload<ErasedListV1Marker>, |
20 | length: ListLength, |
21 | } |
22 | |
23 | macro_rules! constructor { |
24 | ($name: ident, $name_any: ident, $name_buffer: ident, $name_unstable: ident, $marker: ty, $doc: literal) => { |
25 | icu_provider::gen_any_buffer_data_constructors!( |
26 | locale: include, |
27 | style: ListLength, |
28 | error: ListError, |
29 | #[doc = concat!("Creates a new [`ListFormatter`] that produces a " , $doc, "-type list using compiled data." )] |
30 | /// |
31 | /// See the [CLDR spec](https://unicode.org/reports/tr35/tr35-general.html#ListPatterns) for |
32 | /// an explanation of the different types. |
33 | /// |
34 | /// ✨ *Enabled with the `compiled_data` Cargo feature.* |
35 | /// |
36 | /// [📚 Help choosing a constructor](icu_provider::constructors) |
37 | functions: [ |
38 | $name, |
39 | $name_any, |
40 | $name_buffer, |
41 | $name_unstable, |
42 | Self |
43 | ] |
44 | ); |
45 | |
46 | #[doc = icu_provider::gen_any_buffer_unstable_docs!(UNSTABLE, Self::$name)] |
47 | pub fn $name_unstable( |
48 | provider: &(impl DataProvider<$marker> + ?Sized), |
49 | locale: &DataLocale, |
50 | length: ListLength, |
51 | ) -> Result<Self, ListError> { |
52 | let data = provider |
53 | .load(DataRequest { |
54 | locale, |
55 | metadata: Default::default(), |
56 | })? |
57 | .take_payload()?.cast(); |
58 | Ok(Self { data, length }) |
59 | } |
60 | }; |
61 | } |
62 | |
63 | impl ListFormatter { |
64 | constructor!( |
65 | try_new_and_with_length, |
66 | try_new_and_with_length_with_any_provider, |
67 | try_new_and_with_length_with_buffer_provider, |
68 | try_new_and_with_length_unstable, |
69 | AndListV1Marker, |
70 | "and" |
71 | ); |
72 | constructor!( |
73 | try_new_or_with_length, |
74 | try_new_or_with_length_with_any_provider, |
75 | try_new_or_with_length_with_buffer_provider, |
76 | try_new_or_with_length_unstable, |
77 | OrListV1Marker, |
78 | "or" |
79 | ); |
80 | constructor!( |
81 | try_new_unit_with_length, |
82 | try_new_unit_with_length_with_any_provider, |
83 | try_new_unit_with_length_with_buffer_provider, |
84 | try_new_unit_with_length_unstable, |
85 | UnitListV1Marker, |
86 | "unit" |
87 | ); |
88 | |
89 | /// Returns a [`Writeable`] composed of the input [`Writeable`]s and the language-dependent |
90 | /// formatting. |
91 | /// |
92 | /// The [`Writeable`] is annotated with [`parts::ELEMENT`] for input elements, |
93 | /// and [`parts::LITERAL`] for list literals. |
94 | /// |
95 | /// # Example |
96 | /// |
97 | /// ``` |
98 | /// use icu::list::*; |
99 | /// # use icu::locid::locale; |
100 | /// # use writeable::*; |
101 | /// let formatteur = ListFormatter::try_new_and_with_length( |
102 | /// &locale!("fr" ).into(), |
103 | /// ListLength::Wide, |
104 | /// ) |
105 | /// .unwrap(); |
106 | /// let pays = ["Italie" , "France" , "Espagne" , "Allemagne" ]; |
107 | /// |
108 | /// assert_writeable_parts_eq!( |
109 | /// formatteur.format(pays.iter()), |
110 | /// "Italie, France, Espagne et Allemagne" , |
111 | /// [ |
112 | /// (0, 6, parts::ELEMENT), |
113 | /// (6, 8, parts::LITERAL), |
114 | /// (8, 14, parts::ELEMENT), |
115 | /// (14, 16, parts::LITERAL), |
116 | /// (16, 23, parts::ELEMENT), |
117 | /// (23, 27, parts::LITERAL), |
118 | /// (27, 36, parts::ELEMENT), |
119 | /// ] |
120 | /// ); |
121 | /// ``` |
122 | pub fn format<'a, W: Writeable + 'a, I: Iterator<Item = W> + Clone + 'a>( |
123 | &'a self, |
124 | values: I, |
125 | ) -> FormattedList<'a, W, I> { |
126 | FormattedList { |
127 | formatter: self, |
128 | values, |
129 | } |
130 | } |
131 | |
132 | /// Returns a [`String`] composed of the input [`Writeable`]s and the language-dependent |
133 | /// formatting. |
134 | pub fn format_to_string<W: Writeable, I: Iterator<Item = W> + Clone>( |
135 | &self, |
136 | values: I, |
137 | ) -> alloc::string::String { |
138 | self.format(values).write_to_string().into_owned() |
139 | } |
140 | } |
141 | |
142 | /// The [`Part`]s used by [`ListFormatter`]. |
143 | pub mod parts { |
144 | use writeable::Part; |
145 | |
146 | /// The [`Part`] used by [`FormattedList`](super::FormattedList) to mark the part of the string that is an element. |
147 | /// |
148 | /// * `category`: `"list"` |
149 | /// * `value`: `"element"` |
150 | pub const ELEMENT: Part = Part { |
151 | category: "list" , |
152 | value: "element" , |
153 | }; |
154 | |
155 | /// The [`Part`] used by [`FormattedList`](super::FormattedList) to mark the part of the string that is a list literal, |
156 | /// such as ", " or " and ". |
157 | /// |
158 | /// * `category`: `"list"` |
159 | /// * `value`: `"literal"` |
160 | pub const LITERAL: Part = Part { |
161 | category: "list" , |
162 | value: "literal" , |
163 | }; |
164 | } |
165 | |
166 | /// The [`Writeable`] implementation that is returned by [`ListFormatter::format`]. See |
167 | /// the [`writeable`] crate for how to consume this. |
168 | #[derive (Debug)] |
169 | pub struct FormattedList<'a, W: Writeable + 'a, I: Iterator<Item = W> + Clone + 'a> { |
170 | formatter: &'a ListFormatter, |
171 | values: I, |
172 | } |
173 | |
174 | impl<'a, W: Writeable + 'a, I: Iterator<Item = W> + Clone + 'a> Writeable |
175 | for FormattedList<'a, W, I> |
176 | { |
177 | fn write_to_parts<V: PartsWrite + ?Sized>(&self, sink: &mut V) -> fmt::Result { |
178 | macro_rules! literal { |
179 | ($lit:ident) => { |
180 | sink.with_part(parts::LITERAL, |l| l.write_str($lit)) |
181 | }; |
182 | } |
183 | macro_rules! value { |
184 | ($val:expr) => { |
185 | sink.with_part(parts::ELEMENT, |e| $val.write_to_parts(e)) |
186 | }; |
187 | } |
188 | |
189 | let mut values = self.values.clone(); |
190 | |
191 | if let Some(first) = values.next() { |
192 | if let Some(second) = values.next() { |
193 | if let Some(third) = values.next() { |
194 | // Start(values[0], middle(..., middle(values[n-3], End(values[n-2], values[n-1]))...)) = |
195 | // start_before + values[0] + start_between + (values[1..n-3] + middle_between)* + |
196 | // values[n-2] + end_between + values[n-1] + end_after |
197 | |
198 | let (start_before, start_between, _) = self |
199 | .formatter |
200 | .data |
201 | .get() |
202 | .start(self.formatter.length) |
203 | .parts(&second); |
204 | |
205 | literal!(start_before)?; |
206 | value!(first)?; |
207 | literal!(start_between)?; |
208 | value!(second)?; |
209 | |
210 | let mut next = third; |
211 | |
212 | for next_next in values { |
213 | let (_, between, _) = self |
214 | .formatter |
215 | .data |
216 | .get() |
217 | .middle(self.formatter.length) |
218 | .parts(&next); |
219 | literal!(between)?; |
220 | value!(next)?; |
221 | next = next_next; |
222 | } |
223 | |
224 | let (_, end_between, end_after) = self |
225 | .formatter |
226 | .data |
227 | .get() |
228 | .end(self.formatter.length) |
229 | .parts(&next); |
230 | literal!(end_between)?; |
231 | value!(next)?; |
232 | literal!(end_after) |
233 | } else { |
234 | // Pair(values[0], values[1]) = pair_before + values[0] + pair_between + values[1] + pair_after |
235 | let (before, between, after) = self |
236 | .formatter |
237 | .data |
238 | .get() |
239 | .pair(self.formatter.length) |
240 | .parts(&second); |
241 | literal!(before)?; |
242 | value!(first)?; |
243 | literal!(between)?; |
244 | value!(second)?; |
245 | literal!(after) |
246 | } |
247 | } else { |
248 | value!(first) |
249 | } |
250 | } else { |
251 | Ok(()) |
252 | } |
253 | } |
254 | |
255 | fn writeable_length_hint(&self) -> LengthHint { |
256 | let mut count = 0; |
257 | let item_length = self |
258 | .values |
259 | .clone() |
260 | .map(|w| { |
261 | count += 1; |
262 | w.writeable_length_hint() |
263 | }) |
264 | .sum::<LengthHint>(); |
265 | item_length |
266 | + self |
267 | .formatter |
268 | .data |
269 | .get() |
270 | .size_hint(self.formatter.length, count) |
271 | } |
272 | } |
273 | |
274 | impl<'a, W: Writeable + 'a, I: Iterator<Item = W> + Clone + 'a> core::fmt::Display |
275 | for FormattedList<'a, W, I> |
276 | { |
277 | fn fmt(&self, f: &mut core::fmt::Formatter<'_>) -> core::fmt::Result { |
278 | self.write_to(sink:f) |
279 | } |
280 | } |
281 | |
282 | #[cfg (all(test, feature = "datagen" ))] |
283 | mod tests { |
284 | use super::*; |
285 | use writeable::{assert_writeable_eq, assert_writeable_parts_eq}; |
286 | |
287 | fn formatter(length: ListLength) -> ListFormatter { |
288 | ListFormatter { |
289 | data: DataPayload::from_owned(crate::patterns::test::test_patterns()), |
290 | length, |
291 | } |
292 | } |
293 | |
294 | #[test ] |
295 | fn test_slices() { |
296 | let formatter = formatter(ListLength::Wide); |
297 | let values = ["one" , "two" , "three" , "four" , "five" ]; |
298 | |
299 | assert_writeable_eq!(formatter.format(values[0..0].iter()), "" ); |
300 | assert_writeable_eq!(formatter.format(values[0..1].iter()), "one" ); |
301 | assert_writeable_eq!(formatter.format(values[0..2].iter()), "$one;two+" ); |
302 | assert_writeable_eq!(formatter.format(values[0..3].iter()), "@one:two.three!" ); |
303 | assert_writeable_eq!( |
304 | formatter.format(values[0..4].iter()), |
305 | "@one:two,three.four!" |
306 | ); |
307 | |
308 | assert_writeable_parts_eq!( |
309 | formatter.format(values.iter()), |
310 | "@one:two,three,four.five!" , |
311 | [ |
312 | (0, 1, parts::LITERAL), |
313 | (1, 4, parts::ELEMENT), |
314 | (4, 5, parts::LITERAL), |
315 | (5, 8, parts::ELEMENT), |
316 | (8, 9, parts::LITERAL), |
317 | (9, 14, parts::ELEMENT), |
318 | (14, 15, parts::LITERAL), |
319 | (15, 19, parts::ELEMENT), |
320 | (19, 20, parts::LITERAL), |
321 | (20, 24, parts::ELEMENT), |
322 | (24, 25, parts::LITERAL) |
323 | ] |
324 | ); |
325 | } |
326 | |
327 | #[test ] |
328 | fn test_into_iterator() { |
329 | let formatter = formatter(ListLength::Wide); |
330 | |
331 | let mut vecdeque = std::collections::vec_deque::VecDeque::<u8>::new(); |
332 | vecdeque.push_back(10); |
333 | vecdeque.push_front(48); |
334 | |
335 | assert_writeable_parts_eq!( |
336 | formatter.format(vecdeque.iter()), |
337 | "$48;10+" , |
338 | [ |
339 | (0, 1, parts::LITERAL), |
340 | (1, 3, parts::ELEMENT), |
341 | (3, 4, parts::LITERAL), |
342 | (4, 6, parts::ELEMENT), |
343 | (6, 7, parts::LITERAL), |
344 | ] |
345 | ); |
346 | } |
347 | |
348 | #[test ] |
349 | fn test_iterator() { |
350 | let formatter = formatter(ListLength::Wide); |
351 | |
352 | assert_writeable_parts_eq!( |
353 | formatter.format(core::iter::repeat(5).take(2)), |
354 | "$5;5+" , |
355 | [ |
356 | (0, 1, parts::LITERAL), |
357 | (1, 2, parts::ELEMENT), |
358 | (2, 3, parts::LITERAL), |
359 | (3, 4, parts::ELEMENT), |
360 | (4, 5, parts::LITERAL), |
361 | ] |
362 | ); |
363 | } |
364 | |
365 | #[test ] |
366 | fn test_conditional() { |
367 | let formatter = formatter(ListLength::Narrow); |
368 | |
369 | assert_writeable_eq!(formatter.format(["Beta" , "Alpha" ].iter()), "Beta :o Alpha" ); |
370 | } |
371 | |
372 | macro_rules! test { |
373 | ($locale:literal, $type:ident, $(($input:expr, $output:literal),)+) => { |
374 | let f = ListFormatter::$type( |
375 | &icu::locid::locale!($locale).into(), |
376 | ListLength::Wide |
377 | ).unwrap(); |
378 | $( |
379 | assert_writeable_eq!(f.format($input.iter()), $output); |
380 | )+ |
381 | }; |
382 | } |
383 | |
384 | #[test ] |
385 | fn test_basic() { |
386 | test !("fr" , try_new_or_with_length, (["A" , "B" ], "A ou B" ),); |
387 | } |
388 | |
389 | #[test ] |
390 | fn test_spanish() { |
391 | test !( |
392 | "es" , |
393 | try_new_and_with_length, |
394 | (["x" , "Mallorca" ], "x y Mallorca" ), |
395 | (["x" , "Ibiza" ], "x e Ibiza" ), |
396 | (["x" , "Hidalgo" ], "x e Hidalgo" ), |
397 | (["x" , "Hierva" ], "x y Hierva" ), |
398 | ); |
399 | |
400 | test !( |
401 | "es" , |
402 | try_new_or_with_length, |
403 | (["x" , "Ibiza" ], "x o Ibiza" ), |
404 | (["x" , "Okinawa" ], "x u Okinawa" ), |
405 | (["x" , "8 más" ], "x u 8 más" ), |
406 | (["x" , "8" ], "x u 8" ), |
407 | (["x" , "87 más" ], "x u 87 más" ), |
408 | (["x" , "87" ], "x u 87" ), |
409 | (["x" , "11 más" ], "x u 11 más" ), |
410 | (["x" , "11" ], "x u 11" ), |
411 | (["x" , "110 más" ], "x o 110 más" ), |
412 | (["x" , "110" ], "x o 110" ), |
413 | (["x" , "11.000 más" ], "x u 11.000 más" ), |
414 | (["x" , "11.000" ], "x u 11.000" ), |
415 | (["x" , "11.000,92 más" ], "x u 11.000,92 más" ), |
416 | (["x" , "11.000,92" ], "x u 11.000,92" ), |
417 | ); |
418 | |
419 | test !( |
420 | "es-AR" , |
421 | try_new_and_with_length, |
422 | (["x" , "Ibiza" ], "x e Ibiza" ), |
423 | ); |
424 | } |
425 | |
426 | #[test ] |
427 | fn test_hebrew() { |
428 | test !( |
429 | "he" , |
430 | try_new_and_with_length, |
431 | (["x" , "יפו" ], "x ויפו" ), |
432 | (["x" , "Ibiza" ], "x ו‑Ibiza" ), |
433 | ); |
434 | } |
435 | } |
436 | |