1 | /*! |
2 | Represent an [XML 1.0](https://www.w3.org/TR/xml/) document as a read-only tree. |
3 | |
4 | The root point of the documentations is [`Document::parse`]. |
5 | |
6 | You can find more details in the [README] and the [parsing doc]. |
7 | |
8 | The tree structure itself is a heavily modified <https://github.com/causal-agent/ego-tree> |
9 | License: ISC. |
10 | |
11 | [`Document::parse`]: struct.Document.html#method.parse |
12 | [README]: https://github.com/RazrFalcon/roxmltree/blob/master/README.md |
13 | [parsing doc]: https://github.com/RazrFalcon/roxmltree/blob/master/docs/parsing.md |
14 | */ |
15 | |
16 | #![no_std ] |
17 | #![forbid (unsafe_code)] |
18 | #![warn (missing_docs)] |
19 | #![warn (missing_copy_implementations)] |
20 | #![warn (missing_debug_implementations)] |
21 | |
22 | extern crate alloc; |
23 | |
24 | #[cfg (feature = "std" )] |
25 | extern crate std; |
26 | |
27 | use core::cmp::Ordering; |
28 | use core::fmt; |
29 | use core::hash::{Hash, Hasher}; |
30 | use core::num::NonZeroU32; |
31 | use core::ops::Range; |
32 | |
33 | use alloc::vec::Vec; |
34 | |
35 | mod parse; |
36 | mod tokenizer; |
37 | |
38 | #[cfg (test)] |
39 | mod tokenizer_tests; |
40 | |
41 | pub use crate::parse::*; |
42 | |
43 | /// The <http://www.w3.org/XML/1998/namespace> URI. |
44 | pub const NS_XML_URI: &str = "http://www.w3.org/XML/1998/namespace" ; |
45 | /// The prefix 'xml', which is by definition bound to NS_XML_URI |
46 | const NS_XML_PREFIX: &str = "xml" ; |
47 | |
48 | /// The <http://www.w3.org/2000/xmlns/> URI. |
49 | pub const NS_XMLNS_URI: &str = "http://www.w3.org/2000/xmlns/" ; |
50 | /// The string 'xmlns', which is used to declare new namespaces |
51 | const XMLNS: &str = "xmlns" ; |
52 | |
53 | /// Position in text. |
54 | /// |
55 | /// Position indicates a row/line and a column in the original text. Starting from 1:1. |
56 | #[allow (missing_docs)] |
57 | #[derive (Clone, Copy, PartialEq, Eq, Hash, Debug)] |
58 | pub struct TextPos { |
59 | pub row: u32, |
60 | pub col: u32, |
61 | } |
62 | |
63 | impl TextPos { |
64 | /// Constructs a new `TextPos`. |
65 | pub fn new(row: u32, col: u32) -> TextPos { |
66 | TextPos { row, col } |
67 | } |
68 | } |
69 | |
70 | impl fmt::Display for TextPos { |
71 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
72 | write!(f, " {}: {}" , self.row, self.col) |
73 | } |
74 | } |
75 | |
76 | /// An XML tree container. |
77 | /// |
78 | /// A tree consists of [`Nodes`]. |
79 | /// There are no separate structs for each node type. |
80 | /// So you should check the current node type yourself via [`Node::node_type()`]. |
81 | /// There are only [5 types](enum.NodeType.html): |
82 | /// Root, Element, PI, Comment and Text. |
83 | /// |
84 | /// As you can see there are no XML declaration and CDATA types. |
85 | /// The XML declaration is basically skipped, since it doesn't contain any |
86 | /// valuable information (we support only UTF-8 anyway). |
87 | /// And CDATA will be converted into a Text node as is, without |
88 | /// any preprocessing (you can read more about it |
89 | /// [here](https://github.com/RazrFalcon/roxmltree/blob/master/docs/parsing.md)). |
90 | /// |
91 | /// Also, the Text node data can be accessed from the text node itself or from |
92 | /// the parent element via [`Node::text()`] or [`Node::tail()`]. |
93 | /// |
94 | /// [`Nodes`]: struct.Node.html |
95 | /// [`Node::node_type()`]: struct.Node.html#method.node_type |
96 | /// [`Node::text()`]: struct.Node.html#method.text |
97 | /// [`Node::tail()`]: struct.Node.html#method.tail |
98 | pub struct Document<'input> { |
99 | /// An original data. |
100 | /// |
101 | /// Required for `text_pos_at` methods. |
102 | text: &'input str, |
103 | nodes: Vec<NodeData<'input>>, |
104 | attributes: Vec<AttributeData<'input>>, |
105 | namespaces: Namespaces<'input>, |
106 | } |
107 | |
108 | impl<'input> Document<'input> { |
109 | /// Returns the root node. |
110 | /// |
111 | /// # Examples |
112 | /// |
113 | /// ``` |
114 | /// let doc = roxmltree::Document::parse("<e/>" ).unwrap(); |
115 | /// assert!(doc.root().is_root()); |
116 | /// assert!(doc.root().first_child().unwrap().has_tag_name("e" )); |
117 | /// ``` |
118 | #[inline ] |
119 | pub fn root<'a>(&'a self) -> Node<'a, 'input> { |
120 | Node { |
121 | id: NodeId::new(0), |
122 | d: &self.nodes[0], |
123 | doc: self, |
124 | } |
125 | } |
126 | |
127 | /// Returns the node of the tree with the given NodeId. |
128 | /// |
129 | /// Note: NodeId::new(0) represents the root node |
130 | /// |
131 | /// # Examples |
132 | /// |
133 | /// ``` |
134 | /// let doc = roxmltree::Document::parse("\ |
135 | /// <p> |
136 | /// text |
137 | /// </p> |
138 | /// " ).unwrap(); |
139 | /// |
140 | /// use roxmltree::NodeId; |
141 | /// assert_eq!(doc.get_node(NodeId::new(0)).unwrap(), doc.root()); |
142 | /// assert_eq!(doc.get_node(NodeId::new(1)), doc.descendants().find(|n| n.has_tag_name("p" ))); |
143 | /// assert_eq!(doc.get_node(NodeId::new(2)), doc.descendants().find(|n| n.is_text())); |
144 | /// assert_eq!(doc.get_node(NodeId::new(3)), None); |
145 | /// ``` |
146 | #[inline ] |
147 | pub fn get_node<'a>(&'a self, id: NodeId) -> Option<Node<'a, 'input>> { |
148 | self.nodes.get(id.get_usize()).map(|data| Node { |
149 | id, |
150 | d: data, |
151 | doc: self, |
152 | }) |
153 | } |
154 | |
155 | /// Returns the root element of the document. |
156 | /// |
157 | /// Unlike `root`, will return a first element node. |
158 | /// |
159 | /// The root element always exists. |
160 | /// |
161 | /// # Examples |
162 | /// |
163 | /// ``` |
164 | /// let doc = roxmltree::Document::parse("<!-- comment --><e/>" ).unwrap(); |
165 | /// assert!(doc.root_element().has_tag_name("e" )); |
166 | /// ``` |
167 | #[inline ] |
168 | pub fn root_element<'a>(&'a self) -> Node<'a, 'input> { |
169 | // `expect` is safe, because the `Document` is guarantee to have at least one element. |
170 | self.root() |
171 | .first_element_child() |
172 | .expect("XML documents must contain a root element" ) |
173 | } |
174 | |
175 | /// Returns an iterator over document's descendant nodes. |
176 | /// |
177 | /// Shorthand for `doc.root().descendants()`. |
178 | #[inline ] |
179 | pub fn descendants(&self) -> Descendants<'_, 'input> { |
180 | self.root().descendants() |
181 | } |
182 | |
183 | /// Calculates `TextPos` in the original document from position in bytes. |
184 | /// |
185 | /// **Note:** this operation is expensive. |
186 | /// |
187 | /// # Examples |
188 | /// |
189 | /// ``` |
190 | /// use roxmltree::*; |
191 | /// |
192 | /// let doc = Document::parse("\ |
193 | /// <!-- comment --> |
194 | /// <e/>" |
195 | /// ).unwrap(); |
196 | /// |
197 | /// assert_eq!(doc.text_pos_at(10), TextPos::new(1, 11)); |
198 | /// assert_eq!(doc.text_pos_at(9999), TextPos::new(2, 5)); |
199 | /// ``` |
200 | #[inline ] |
201 | pub fn text_pos_at(&self, pos: usize) -> TextPos { |
202 | tokenizer::Stream::new(self.text).gen_text_pos_from(pos) |
203 | } |
204 | |
205 | /// Returns the input text of the original document. |
206 | /// |
207 | /// # Examples |
208 | /// |
209 | /// ``` |
210 | /// use roxmltree::*; |
211 | /// |
212 | /// let doc = Document::parse("<e/>" ).unwrap(); |
213 | /// |
214 | /// assert_eq!(doc.input_text(), "<e/>" ); |
215 | /// ``` |
216 | #[inline ] |
217 | pub fn input_text(&self) -> &'input str { |
218 | self.text |
219 | } |
220 | } |
221 | |
222 | impl<'input> fmt::Debug for Document<'input> { |
223 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
224 | if !self.root().has_children() { |
225 | return write!(f, "Document []" ); |
226 | } |
227 | |
228 | macro_rules! writeln_indented { |
229 | ($depth:expr, $f:expr, $fmt:expr) => { |
230 | for _ in 0..$depth { write!($f, " " )?; } |
231 | writeln!($f, $fmt)?; |
232 | }; |
233 | ($depth:expr, $f:expr, $fmt:expr, $($arg:tt)*) => { |
234 | for _ in 0..$depth { write!($f, " " )?; } |
235 | writeln!($f, $fmt, $($arg)*)?; |
236 | }; |
237 | } |
238 | |
239 | fn print_into_iter< |
240 | T: fmt::Debug, |
241 | E: ExactSizeIterator<Item = T>, |
242 | I: IntoIterator<Item = T, IntoIter = E>, |
243 | >( |
244 | prefix: &str, |
245 | data: I, |
246 | depth: usize, |
247 | f: &mut fmt::Formatter, |
248 | ) -> Result<(), fmt::Error> { |
249 | let data = data.into_iter(); |
250 | if data.len() == 0 { |
251 | return Ok(()); |
252 | } |
253 | |
254 | writeln_indented!(depth, f, " {}: [" , prefix); |
255 | for v in data { |
256 | writeln_indented!(depth + 1, f, " {:?}" , v); |
257 | } |
258 | writeln_indented!(depth, f, "]" ); |
259 | |
260 | Ok(()) |
261 | } |
262 | |
263 | fn print_children( |
264 | parent: Node, |
265 | depth: usize, |
266 | f: &mut fmt::Formatter, |
267 | ) -> Result<(), fmt::Error> { |
268 | for child in parent.children() { |
269 | if child.is_element() { |
270 | writeln_indented!(depth, f, "Element {{" ); |
271 | writeln_indented!(depth, f, " tag_name: {:?}" , child.tag_name()); |
272 | print_into_iter("attributes" , child.attributes(), depth + 1, f)?; |
273 | print_into_iter("namespaces" , child.namespaces(), depth + 1, f)?; |
274 | |
275 | if child.has_children() { |
276 | writeln_indented!(depth, f, " children: [" ); |
277 | print_children(child, depth + 2, f)?; |
278 | writeln_indented!(depth, f, " ]" ); |
279 | } |
280 | |
281 | writeln_indented!(depth, f, " }}" ); |
282 | } else { |
283 | writeln_indented!(depth, f, " {:?}" , child); |
284 | } |
285 | } |
286 | |
287 | Ok(()) |
288 | } |
289 | |
290 | writeln!(f, "Document [" )?; |
291 | print_children(self.root(), 1, f)?; |
292 | writeln!(f, "]" )?; |
293 | |
294 | Ok(()) |
295 | } |
296 | } |
297 | |
298 | /// A list of supported node types. |
299 | #[derive (Clone, Copy, PartialEq, Eq, Debug)] |
300 | pub enum NodeType { |
301 | /// The root node of the `Document`. |
302 | Root, |
303 | /// An element node. |
304 | /// |
305 | /// Only an element can have a tag name and attributes. |
306 | Element, |
307 | /// A processing instruction. |
308 | PI, |
309 | /// A comment node. |
310 | Comment, |
311 | /// A text node. |
312 | Text, |
313 | } |
314 | |
315 | /// A processing instruction. |
316 | #[derive (Clone, Copy, PartialEq, Eq, Debug)] |
317 | #[allow (missing_docs)] |
318 | pub struct PI<'input> { |
319 | pub target: &'input str, |
320 | pub value: Option<&'input str>, |
321 | } |
322 | |
323 | /// A short range. |
324 | /// |
325 | /// Just like Range, but only for `u32` and copyable. |
326 | #[derive (Clone, Copy, Debug)] |
327 | struct ShortRange { |
328 | start: u32, |
329 | end: u32, |
330 | } |
331 | |
332 | impl From<Range<usize>> for ShortRange { |
333 | #[inline ] |
334 | fn from(range: Range<usize>) -> Self { |
335 | debug_assert!(range.start <= core::u32::MAX as usize); |
336 | debug_assert!(range.end <= core::u32::MAX as usize); |
337 | ShortRange::new(range.start as u32, range.end as u32) |
338 | } |
339 | } |
340 | |
341 | impl ShortRange { |
342 | #[inline ] |
343 | fn new(start: u32, end: u32) -> Self { |
344 | ShortRange { start, end } |
345 | } |
346 | |
347 | #[inline ] |
348 | fn to_urange(self) -> Range<usize> { |
349 | self.start as usize..self.end as usize |
350 | } |
351 | } |
352 | |
353 | /// A node ID stored as `u32`. |
354 | /// |
355 | /// An index into a `Tree`-internal `Vec`. |
356 | /// |
357 | /// Note that this value should be used with care since `roxmltree` doesn't |
358 | /// check that `NodeId` actually belongs to a selected `Document`. |
359 | /// So you can end up in a situation, when `NodeId` produced by one `Document` |
360 | /// is used to select a node in another `Document`. |
361 | #[derive (Clone, Copy, PartialEq, Eq, Hash, Debug)] |
362 | pub struct NodeId(NonZeroU32); |
363 | |
364 | impl NodeId { |
365 | /// Construct a new `NodeId` from a `u32`. |
366 | #[inline ] |
367 | pub fn new(id: u32) -> Self { |
368 | debug_assert!(id < core::u32::MAX); |
369 | |
370 | // We are using `NonZeroU32` to reduce overhead of `Option<NodeId>`. |
371 | NodeId(NonZeroU32::new(id + 1).unwrap()) |
372 | } |
373 | |
374 | /// Returns the `u32` representation of the `NodeId`. |
375 | #[inline ] |
376 | pub fn get(self) -> u32 { |
377 | self.0.get() - 1 |
378 | } |
379 | |
380 | /// Returns the `usize` representation of the `NodeId`. |
381 | #[inline ] |
382 | pub fn get_usize(self) -> usize { |
383 | self.get() as usize |
384 | } |
385 | } |
386 | |
387 | impl From<u32> for NodeId { |
388 | #[inline ] |
389 | fn from(id: u32) -> Self { |
390 | NodeId::new(id) |
391 | } |
392 | } |
393 | |
394 | impl From<usize> for NodeId { |
395 | #[inline ] |
396 | fn from(id: usize) -> Self { |
397 | // We already checked that `id` is limited by u32::MAX. |
398 | debug_assert!(id <= core::u32::MAX as usize); |
399 | NodeId::new(id as u32) |
400 | } |
401 | } |
402 | |
403 | #[derive (Debug)] |
404 | enum NodeKind<'input> { |
405 | Root, |
406 | Element { |
407 | tag_name: ExpandedNameIndexed<'input>, |
408 | attributes: ShortRange, |
409 | namespaces: ShortRange, |
410 | }, |
411 | PI(PI<'input>), |
412 | Comment(StringStorage<'input>), |
413 | Text(StringStorage<'input>), |
414 | } |
415 | |
416 | #[derive (Debug)] |
417 | struct NodeData<'input> { |
418 | parent: Option<NodeId>, |
419 | prev_sibling: Option<NodeId>, |
420 | next_subtree: Option<NodeId>, |
421 | last_child: Option<NodeId>, |
422 | kind: NodeKind<'input>, |
423 | #[cfg (feature = "positions" )] |
424 | range: Range<usize>, |
425 | } |
426 | |
427 | #[cfg (target_has_atomic = "ptr" )] |
428 | type OwnedSharedString = alloc::sync::Arc<str>; |
429 | |
430 | #[cfg (not(target_has_atomic = "ptr" ))] |
431 | type OwnedSharedString = alloc::rc::Rc<str>; |
432 | |
433 | /// A string storage. |
434 | /// |
435 | /// Used by text nodes and attributes values. |
436 | /// |
437 | /// We try our best not to allocate strings, referencing the input string as much as possible. |
438 | /// But in some cases post-processing is necessary and we have to allocate them. |
439 | /// |
440 | /// All owned, allocated strings are stored as `Arc<str>` or as `Rc<str>` on targets |
441 | /// were `Arc` isn't available. |
442 | /// And unlike `Cow<&str>`, `StringStorage` is immutable and can be cheaply cloned. |
443 | #[derive (Clone, Eq, Debug)] |
444 | pub enum StringStorage<'input> { |
445 | /// A raw slice of the input string. |
446 | Borrowed(&'input str), |
447 | |
448 | /// A reference-counted string. |
449 | Owned(OwnedSharedString), |
450 | } |
451 | |
452 | impl StringStorage<'_> { |
453 | /// Creates a new owned string from `&str` or `String`. |
454 | pub fn new_owned<T: Into<OwnedSharedString>>(s: T) -> Self { |
455 | StringStorage::Owned(s.into()) |
456 | } |
457 | |
458 | /// Returns a string slice. |
459 | pub fn as_str(&self) -> &str { |
460 | match self { |
461 | StringStorage::Borrowed(s: &&str) => s, |
462 | StringStorage::Owned(s: &Arc) => s, |
463 | } |
464 | } |
465 | } |
466 | |
467 | impl PartialEq for StringStorage<'_> { |
468 | fn eq(&self, other: &Self) -> bool { |
469 | self.as_str() == other.as_str() |
470 | } |
471 | } |
472 | |
473 | impl core::fmt::Display for StringStorage<'_> { |
474 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
475 | write!(f, " {}" , self.as_str()) |
476 | } |
477 | } |
478 | |
479 | impl core::ops::Deref for StringStorage<'_> { |
480 | type Target = str; |
481 | |
482 | fn deref(&self) -> &Self::Target { |
483 | self.as_str() |
484 | } |
485 | } |
486 | |
487 | #[derive (Clone, Debug)] |
488 | struct AttributeData<'input> { |
489 | name: ExpandedNameIndexed<'input>, |
490 | value: StringStorage<'input>, |
491 | #[cfg (feature = "positions" )] |
492 | pos: usize, |
493 | } |
494 | |
495 | /// An attribute. |
496 | #[derive (Copy, Clone)] |
497 | pub struct Attribute<'a, 'input: 'a> { |
498 | doc: &'a Document<'input>, |
499 | data: &'a AttributeData<'input>, |
500 | } |
501 | |
502 | impl<'a, 'input> Attribute<'a, 'input> { |
503 | /// Returns attribute's namespace URI. |
504 | /// |
505 | /// # Examples |
506 | /// |
507 | /// ``` |
508 | /// let doc = roxmltree::Document::parse( |
509 | /// "<e xmlns:n='http://www.w3.org' a='b' n:a='c'/>" |
510 | /// ).unwrap(); |
511 | /// |
512 | /// assert_eq!(doc.root_element().attributes().nth(0).unwrap().namespace(), None); |
513 | /// assert_eq!(doc.root_element().attributes().nth(1).unwrap().namespace(), Some("http://www.w3.org" )); |
514 | /// ``` |
515 | #[inline ] |
516 | pub fn namespace(&self) -> Option<&'a str> { |
517 | self.data.name.namespace(self.doc).map(Namespace::uri) |
518 | } |
519 | |
520 | /// Returns attribute's name. |
521 | /// |
522 | /// # Examples |
523 | /// |
524 | /// ``` |
525 | /// let doc = roxmltree::Document::parse( |
526 | /// "<e xmlns:n='http://www.w3.org' a='b' n:a='c'/>" |
527 | /// ).unwrap(); |
528 | /// |
529 | /// assert_eq!(doc.root_element().attributes().nth(0).unwrap().name(), "a" ); |
530 | /// assert_eq!(doc.root_element().attributes().nth(1).unwrap().name(), "a" ); |
531 | /// ``` |
532 | #[inline ] |
533 | pub fn name(&self) -> &'input str { |
534 | self.data.name.local_name |
535 | } |
536 | |
537 | /// Returns attribute's value. |
538 | /// |
539 | /// # Examples |
540 | /// |
541 | /// ``` |
542 | /// let doc = roxmltree::Document::parse( |
543 | /// "<e xmlns:n='http://www.w3.org' a='b' n:a='c'/>" |
544 | /// ).unwrap(); |
545 | /// |
546 | /// assert_eq!(doc.root_element().attributes().nth(0).unwrap().value(), "b" ); |
547 | /// assert_eq!(doc.root_element().attributes().nth(1).unwrap().value(), "c" ); |
548 | /// ``` |
549 | #[inline ] |
550 | pub fn value(&self) -> &'a str { |
551 | &self.data.value |
552 | } |
553 | |
554 | /// Returns attribute's value storage. |
555 | /// |
556 | /// Useful when you need a more low-level access to an allocated string. |
557 | #[inline ] |
558 | pub fn value_storage(&self) -> &StringStorage<'input> { |
559 | &self.data.value |
560 | } |
561 | |
562 | /// Returns attribute's position in bytes in the original document. |
563 | /// |
564 | /// You can calculate a human-readable text position via [Document::text_pos_at]. |
565 | /// |
566 | /// ```text |
567 | /// <e attr='value'/> |
568 | /// ^ |
569 | /// ``` |
570 | /// |
571 | /// [Document::text_pos_at]: struct.Document.html#method.text_pos_at |
572 | #[cfg (feature = "positions" )] |
573 | #[inline ] |
574 | pub fn position(&self) -> usize { |
575 | self.data.pos |
576 | } |
577 | } |
578 | |
579 | impl PartialEq for Attribute<'_, '_> { |
580 | #[inline ] |
581 | fn eq(&self, other: &Attribute<'_, '_>) -> bool { |
582 | self.data.name.as_expanded_name(self.doc) == other.data.name.as_expanded_name(other.doc) |
583 | && self.data.value == other.data.value |
584 | } |
585 | } |
586 | |
587 | impl fmt::Debug for Attribute<'_, '_> { |
588 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
589 | write!( |
590 | f, |
591 | "Attribute {{ name: {:?}, value: {:?} }}" , |
592 | self.data.name.as_expanded_name(self.doc), |
593 | self.data.value |
594 | ) |
595 | } |
596 | } |
597 | |
598 | /// A namespace. |
599 | /// |
600 | /// Contains URI and *prefix* pair. |
601 | #[derive (Clone, PartialEq, Eq, Debug)] |
602 | pub struct Namespace<'input> { |
603 | name: Option<&'input str>, |
604 | uri: StringStorage<'input>, |
605 | } |
606 | |
607 | impl<'input> Namespace<'input> { |
608 | /// Returns namespace name/prefix. |
609 | /// |
610 | /// # Examples |
611 | /// |
612 | /// ``` |
613 | /// let doc = roxmltree::Document::parse( |
614 | /// "<e xmlns:n='http://www.w3.org'/>" |
615 | /// ).unwrap(); |
616 | /// |
617 | /// assert_eq!(doc.root_element().namespaces().nth(0).unwrap().name(), Some("n" )); |
618 | /// ``` |
619 | /// |
620 | /// ``` |
621 | /// let doc = roxmltree::Document::parse( |
622 | /// "<e xmlns='http://www.w3.org'/>" |
623 | /// ).unwrap(); |
624 | /// |
625 | /// assert_eq!(doc.root_element().namespaces().nth(0).unwrap().name(), None); |
626 | /// ``` |
627 | #[inline ] |
628 | pub fn name(&self) -> Option<&'input str> { |
629 | self.name |
630 | } |
631 | |
632 | /// Returns namespace URI. |
633 | /// |
634 | /// # Examples |
635 | /// |
636 | /// ``` |
637 | /// let doc = roxmltree::Document::parse( |
638 | /// "<e xmlns:n='http://www.w3.org'/>" |
639 | /// ).unwrap(); |
640 | /// |
641 | /// assert_eq!(doc.root_element().namespaces().nth(0).unwrap().uri(), "http://www.w3.org" ); |
642 | /// ``` |
643 | #[inline ] |
644 | pub fn uri(&self) -> &str { |
645 | self.uri.as_ref() |
646 | } |
647 | } |
648 | |
649 | #[derive (Default)] |
650 | struct Namespaces<'input> { |
651 | // Deduplicated namespace values used throughout the document |
652 | values: Vec<Namespace<'input>>, |
653 | // Indices into the above in tree order as the document is parsed |
654 | tree_order: Vec<NamespaceIdx>, |
655 | // Indices into the above sorted by value used for deduplication |
656 | sorted_order: Vec<NamespaceIdx>, |
657 | } |
658 | |
659 | impl<'input> Namespaces<'input> { |
660 | fn push_ns( |
661 | &mut self, |
662 | name: Option<&'input str>, |
663 | uri: StringStorage<'input>, |
664 | ) -> Result<(), Error> { |
665 | debug_assert_ne!(name, Some("" )); |
666 | |
667 | let idx = match self.sorted_order.binary_search_by(|idx| { |
668 | let value = &self.values[idx.0 as usize]; |
669 | |
670 | (value.name, value.uri.as_ref()).cmp(&(name, uri.as_str())) |
671 | }) { |
672 | Ok(sorted_idx) => self.sorted_order[sorted_idx], |
673 | Err(sorted_idx) => { |
674 | if self.values.len() > core::u16::MAX as usize { |
675 | return Err(Error::NamespacesLimitReached); |
676 | } |
677 | let idx = NamespaceIdx(self.values.len() as u16); |
678 | self.values.push(Namespace { name, uri }); |
679 | self.sorted_order.insert(sorted_idx, idx); |
680 | idx |
681 | } |
682 | }; |
683 | |
684 | self.tree_order.push(idx); |
685 | |
686 | Ok(()) |
687 | } |
688 | |
689 | #[inline ] |
690 | fn push_ref(&mut self, tree_idx: usize) { |
691 | let idx = self.tree_order[tree_idx]; |
692 | self.tree_order.push(idx); |
693 | } |
694 | |
695 | #[inline ] |
696 | fn exists(&self, start: usize, prefix: Option<&str>) -> bool { |
697 | self.tree_order[start..] |
698 | .iter() |
699 | .any(|idx| self.values[idx.0 as usize].name == prefix) |
700 | } |
701 | |
702 | fn shrink_to_fit(&mut self) { |
703 | self.values.shrink_to_fit(); |
704 | self.tree_order.shrink_to_fit(); |
705 | self.sorted_order.shrink_to_fit(); |
706 | } |
707 | |
708 | #[inline ] |
709 | fn get(&self, idx: NamespaceIdx) -> &Namespace<'input> { |
710 | &self.values[idx.0 as usize] |
711 | } |
712 | } |
713 | |
714 | #[derive (Clone, Copy, Debug)] |
715 | #[repr (transparent)] |
716 | struct NamespaceIdx(u16); |
717 | |
718 | #[derive (Clone, Copy, Debug)] |
719 | struct ExpandedNameIndexed<'input> { |
720 | namespace_idx: Option<NamespaceIdx>, |
721 | local_name: &'input str, |
722 | } |
723 | |
724 | impl<'input> ExpandedNameIndexed<'input> { |
725 | #[inline ] |
726 | fn namespace<'a>(&self, doc: &'a Document<'input>) -> Option<&'a Namespace<'input>> { |
727 | self.namespace_idx.map(|idx: NamespaceIdx| doc.namespaces.get(idx)) |
728 | } |
729 | |
730 | #[inline ] |
731 | fn as_expanded_name<'a>(&self, doc: &'a Document<'input>) -> ExpandedName<'a, 'input> { |
732 | ExpandedName { |
733 | uri: self.namespace(doc).map(Namespace::uri), |
734 | name: self.local_name, |
735 | } |
736 | } |
737 | } |
738 | |
739 | /// An expanded name. |
740 | /// |
741 | /// Contains an namespace URI and name pair. |
742 | #[derive (Clone, Copy, PartialEq, Eq)] |
743 | pub struct ExpandedName<'a, 'b> { |
744 | uri: Option<&'a str>, |
745 | name: &'b str, |
746 | } |
747 | |
748 | impl<'a, 'b> ExpandedName<'a, 'b> { |
749 | /// Returns a namespace URI. |
750 | /// |
751 | /// # Examples |
752 | /// |
753 | /// ``` |
754 | /// let doc = roxmltree::Document::parse("<e xmlns='http://www.w3.org'/>" ).unwrap(); |
755 | /// |
756 | /// assert_eq!(doc.root_element().tag_name().namespace(), Some("http://www.w3.org" )); |
757 | /// ``` |
758 | #[inline ] |
759 | pub fn namespace(&self) -> Option<&'a str> { |
760 | self.uri |
761 | } |
762 | |
763 | /// Returns a local name. |
764 | /// |
765 | /// # Examples |
766 | /// |
767 | /// ``` |
768 | /// let doc = roxmltree::Document::parse("<e/>" ).unwrap(); |
769 | /// |
770 | /// assert_eq!(doc.root_element().tag_name().name(), "e" ); |
771 | /// ``` |
772 | #[inline ] |
773 | pub fn name(&self) -> &'b str { |
774 | self.name |
775 | } |
776 | } |
777 | |
778 | impl ExpandedName<'static, 'static> { |
779 | /// Create a new instance from static data. |
780 | /// |
781 | /// # Example |
782 | /// |
783 | /// ```rust |
784 | /// use roxmltree::ExpandedName; |
785 | /// const DAV_HREF: ExpandedName = |
786 | /// ExpandedName::from_static("urn:ietf:params:xml:ns:caldav:" , "calendar-data" ); |
787 | /// ``` |
788 | pub const fn from_static(uri: &'static str, name: &'static str) -> Self { |
789 | Self { |
790 | uri: Some(uri), |
791 | name, |
792 | } |
793 | } |
794 | } |
795 | |
796 | impl fmt::Debug for ExpandedName<'_, '_> { |
797 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
798 | match self.namespace() { |
799 | Some(ns: &str) => write!(f, " {{{}}}{}" , ns, self.name), |
800 | None => write!(f, " {}" , self.name), |
801 | } |
802 | } |
803 | } |
804 | |
805 | impl<'a, 'b> From<&'b str> for ExpandedName<'a, 'b> { |
806 | #[inline ] |
807 | fn from(v: &'b str) -> Self { |
808 | ExpandedName { uri: None, name: v } |
809 | } |
810 | } |
811 | |
812 | impl<'a, 'b> From<(&'a str, &'b str)> for ExpandedName<'a, 'b> { |
813 | #[inline ] |
814 | fn from(v: (&'a str, &'b str)) -> Self { |
815 | ExpandedName { |
816 | uri: Some(v.0), |
817 | name: v.1, |
818 | } |
819 | } |
820 | } |
821 | |
822 | /// A node in a document. |
823 | /// |
824 | /// # Document Order |
825 | /// |
826 | /// The implementation of the `Ord` traits for `Node` is based on the concept of *document-order*. |
827 | /// In layman's terms, document-order is the order in which one would see each element if |
828 | /// one opened a document in a text editor or web browser and scrolled down. |
829 | /// Document-order convention is followed in XPath, CSS Counters, and DOM selectors API |
830 | /// to ensure consistent results from selection. |
831 | /// One difference in `roxmltree` is that there is the notion of more than one document |
832 | /// in existence at a time. While Nodes within the same document are in document-order, |
833 | /// Nodes in different documents will be grouped together, but not in any particular |
834 | /// order. |
835 | /// |
836 | /// As an example, if we have a Document `a` with Nodes `[a0, a1, a2]` and a |
837 | /// Document `b` with Nodes `[b0, b1]`, these Nodes in order could be either |
838 | /// `[a0, a1, a2, b0, b1]` or `[b0, b1, a0, a1, a2]` and roxmltree makes no |
839 | /// guarantee which it will be. |
840 | /// |
841 | /// Document-order is defined here in the |
842 | /// [W3C XPath Recommendation](https://www.w3.org/TR/xpath-3/#id-document-order) |
843 | /// The use of document-order in DOM Selectors is described here in the |
844 | /// [W3C Selectors API Level 1](https://www.w3.org/TR/selectors-api/#the-apis) |
845 | #[derive (Clone, Copy)] |
846 | pub struct Node<'a, 'input: 'a> { |
847 | /// Node's ID. |
848 | id: NodeId, |
849 | |
850 | /// The tree containing the node. |
851 | doc: &'a Document<'input>, |
852 | |
853 | /// Node's data. |
854 | d: &'a NodeData<'input>, |
855 | } |
856 | |
857 | impl Eq for Node<'_, '_> {} |
858 | |
859 | impl PartialEq for Node<'_, '_> { |
860 | #[inline ] |
861 | fn eq(&self, other: &Self) -> bool { |
862 | (self.id, self.doc as *const _) == (other.id, other.doc as *const _) |
863 | } |
864 | } |
865 | |
866 | impl PartialOrd for Node<'_, '_> { |
867 | fn partial_cmp(&self, other: &Self) -> Option<Ordering> { |
868 | Some(self.cmp(other)) |
869 | } |
870 | } |
871 | |
872 | impl Ord for Node<'_, '_> { |
873 | fn cmp(&self, other: &Self) -> Ordering { |
874 | (self.id.0, self.doc as *const _).cmp(&(other.id.0, other.doc as *const _)) |
875 | } |
876 | } |
877 | |
878 | impl Hash for Node<'_, '_> { |
879 | fn hash<H: Hasher>(&self, state: &mut H) { |
880 | self.id.0.hash(state); |
881 | (self.doc as *const Document).hash(state); |
882 | (self.d as *const NodeData).hash(state); |
883 | } |
884 | } |
885 | |
886 | impl<'a, 'input: 'a> Node<'a, 'input> { |
887 | /// Returns node's type. |
888 | #[inline ] |
889 | pub fn node_type(&self) -> NodeType { |
890 | match self.d.kind { |
891 | NodeKind::Root => NodeType::Root, |
892 | NodeKind::Element { .. } => NodeType::Element, |
893 | NodeKind::PI { .. } => NodeType::PI, |
894 | NodeKind::Comment(_) => NodeType::Comment, |
895 | NodeKind::Text(_) => NodeType::Text, |
896 | } |
897 | } |
898 | |
899 | /// Checks that node is a root node. |
900 | #[inline ] |
901 | pub fn is_root(&self) -> bool { |
902 | self.node_type() == NodeType::Root |
903 | } |
904 | |
905 | /// Checks that node is an element node. |
906 | #[inline ] |
907 | pub fn is_element(&self) -> bool { |
908 | self.node_type() == NodeType::Element |
909 | } |
910 | |
911 | /// Checks that node is a processing instruction node. |
912 | #[inline ] |
913 | pub fn is_pi(&self) -> bool { |
914 | self.node_type() == NodeType::PI |
915 | } |
916 | |
917 | /// Checks that node is a comment node. |
918 | #[inline ] |
919 | pub fn is_comment(&self) -> bool { |
920 | self.node_type() == NodeType::Comment |
921 | } |
922 | |
923 | /// Checks that node is a text node. |
924 | #[inline ] |
925 | pub fn is_text(&self) -> bool { |
926 | self.node_type() == NodeType::Text |
927 | } |
928 | |
929 | /// Returns node's document. |
930 | #[inline ] |
931 | pub fn document(&self) -> &'a Document<'input> { |
932 | self.doc |
933 | } |
934 | |
935 | /// Returns node's tag name. |
936 | /// |
937 | /// Returns an empty name with no namespace if the current node is not an element. |
938 | /// |
939 | /// # Examples |
940 | /// |
941 | /// ``` |
942 | /// let doc = roxmltree::Document::parse("<e xmlns='http://www.w3.org'/>" ).unwrap(); |
943 | /// |
944 | /// assert_eq!(doc.root_element().tag_name().namespace(), Some("http://www.w3.org" )); |
945 | /// assert_eq!(doc.root_element().tag_name().name(), "e" ); |
946 | /// ``` |
947 | #[inline ] |
948 | pub fn tag_name(&self) -> ExpandedName<'a, 'input> { |
949 | match self.d.kind { |
950 | NodeKind::Element { ref tag_name, .. } => tag_name.as_expanded_name(self.doc), |
951 | _ => "" .into(), |
952 | } |
953 | } |
954 | |
955 | /// Checks that node has a specified tag name. |
956 | /// |
957 | /// # Examples |
958 | /// |
959 | /// ``` |
960 | /// let doc = roxmltree::Document::parse("<e xmlns='http://www.w3.org'/>" ).unwrap(); |
961 | /// |
962 | /// assert!(doc.root_element().has_tag_name("e" )); |
963 | /// assert!(doc.root_element().has_tag_name(("http://www.w3.org" , "e" ))); |
964 | /// |
965 | /// assert!(!doc.root_element().has_tag_name("b" )); |
966 | /// assert!(!doc.root_element().has_tag_name(("http://www.w4.org" , "e" ))); |
967 | /// ``` |
968 | pub fn has_tag_name<'n, 'm, N>(&self, name: N) -> bool |
969 | where |
970 | N: Into<ExpandedName<'n, 'm>>, |
971 | { |
972 | let name = name.into(); |
973 | |
974 | match self.d.kind { |
975 | NodeKind::Element { ref tag_name, .. } => match name.namespace() { |
976 | Some(_) => tag_name.as_expanded_name(self.doc) == name, |
977 | None => tag_name.local_name == name.name, |
978 | }, |
979 | _ => false, |
980 | } |
981 | } |
982 | |
983 | /// Returns node's default namespace URI. |
984 | /// |
985 | /// # Examples |
986 | /// |
987 | /// ``` |
988 | /// let doc = roxmltree::Document::parse("<e xmlns='http://www.w3.org'/>" ).unwrap(); |
989 | /// |
990 | /// assert_eq!(doc.root_element().default_namespace(), Some("http://www.w3.org" )); |
991 | /// ``` |
992 | /// |
993 | /// ``` |
994 | /// let doc = roxmltree::Document::parse("<e xmlns:n='http://www.w3.org'/>" ).unwrap(); |
995 | /// |
996 | /// assert_eq!(doc.root_element().default_namespace(), None); |
997 | /// ``` |
998 | pub fn default_namespace(&self) -> Option<&'a str> { |
999 | self.namespaces() |
1000 | .find(|ns| ns.name.is_none()) |
1001 | .map(|v| v.uri.as_ref()) |
1002 | } |
1003 | |
1004 | /// Returns a prefix for a given namespace URI. |
1005 | /// |
1006 | /// # Examples |
1007 | /// |
1008 | /// ``` |
1009 | /// let doc = roxmltree::Document::parse("<e xmlns:n='http://www.w3.org'/>" ).unwrap(); |
1010 | /// |
1011 | /// assert_eq!(doc.root_element().lookup_prefix("http://www.w3.org" ), Some("n" )); |
1012 | /// ``` |
1013 | /// |
1014 | /// ``` |
1015 | /// let doc = roxmltree::Document::parse("<e xmlns:n=''/>" ).unwrap(); |
1016 | /// |
1017 | /// assert_eq!(doc.root_element().lookup_prefix("" ), Some("n" )); |
1018 | /// ``` |
1019 | pub fn lookup_prefix(&self, uri: &str) -> Option<&'input str> { |
1020 | if uri == NS_XML_URI { |
1021 | return Some(NS_XML_PREFIX); |
1022 | } |
1023 | |
1024 | self.namespaces() |
1025 | .find(|ns| &*ns.uri == uri) |
1026 | .map(|v| v.name) |
1027 | .unwrap_or(None) |
1028 | } |
1029 | |
1030 | /// Returns an URI for a given prefix. |
1031 | /// |
1032 | /// # Examples |
1033 | /// |
1034 | /// ``` |
1035 | /// let doc = roxmltree::Document::parse("<e xmlns:n='http://www.w3.org'/>" ).unwrap(); |
1036 | /// |
1037 | /// assert_eq!(doc.root_element().lookup_namespace_uri(Some("n" )), Some("http://www.w3.org" )); |
1038 | /// ``` |
1039 | /// |
1040 | /// ``` |
1041 | /// let doc = roxmltree::Document::parse("<e xmlns='http://www.w3.org'/>" ).unwrap(); |
1042 | /// |
1043 | /// assert_eq!(doc.root_element().lookup_namespace_uri(None), Some("http://www.w3.org" )); |
1044 | /// ``` |
1045 | pub fn lookup_namespace_uri(&self, prefix: Option<&str>) -> Option<&'a str> { |
1046 | self.namespaces() |
1047 | .find(|ns| ns.name == prefix) |
1048 | .map(|v| v.uri.as_ref()) |
1049 | } |
1050 | |
1051 | /// Returns element's attribute value. |
1052 | /// |
1053 | /// # Examples |
1054 | /// |
1055 | /// ``` |
1056 | /// let doc = roxmltree::Document::parse("<e a='b'/>" ).unwrap(); |
1057 | /// |
1058 | /// assert_eq!(doc.root_element().attribute("a" ), Some("b" )); |
1059 | /// ``` |
1060 | /// |
1061 | /// ``` |
1062 | /// let doc = roxmltree::Document::parse( |
1063 | /// "<e xmlns:n='http://www.w3.org' a='b' n:a='c'/>" |
1064 | /// ).unwrap(); |
1065 | /// |
1066 | /// assert_eq!(doc.root_element().attribute("a" ), Some("b" )); |
1067 | /// assert_eq!(doc.root_element().attribute(("http://www.w3.org" , "a" )), Some("c" )); |
1068 | /// ``` |
1069 | pub fn attribute<'n, 'm, N>(&self, name: N) -> Option<&'a str> |
1070 | where |
1071 | N: Into<ExpandedName<'n, 'm>>, |
1072 | { |
1073 | let name = name.into(); |
1074 | self.attributes() |
1075 | .find(|a| a.data.name.as_expanded_name(self.doc) == name) |
1076 | .map(|a| a.value()) |
1077 | } |
1078 | |
1079 | /// Returns element's attribute object. |
1080 | /// |
1081 | /// The same as [`attribute()`], but returns the `Attribute` itself instead of a value string. |
1082 | /// |
1083 | /// [`attribute()`]: struct.Node.html#method.attribute |
1084 | pub fn attribute_node<'n, 'm, N>(&self, name: N) -> Option<Attribute<'a, 'input>> |
1085 | where |
1086 | N: Into<ExpandedName<'n, 'm>>, |
1087 | { |
1088 | let name = name.into(); |
1089 | self.attributes() |
1090 | .find(|a| a.data.name.as_expanded_name(self.doc) == name) |
1091 | } |
1092 | |
1093 | /// Checks that element has a specified attribute. |
1094 | /// |
1095 | /// # Examples |
1096 | /// |
1097 | /// ``` |
1098 | /// let doc = roxmltree::Document::parse( |
1099 | /// "<e xmlns:n='http://www.w3.org' a='b' n:a='c'/>" |
1100 | /// ).unwrap(); |
1101 | /// |
1102 | /// assert!(doc.root_element().has_attribute("a" )); |
1103 | /// assert!(doc.root_element().has_attribute(("http://www.w3.org" , "a" ))); |
1104 | /// |
1105 | /// assert!(!doc.root_element().has_attribute("b" )); |
1106 | /// assert!(!doc.root_element().has_attribute(("http://www.w4.org" , "a" ))); |
1107 | /// ``` |
1108 | pub fn has_attribute<'n, 'm, N>(&self, name: N) -> bool |
1109 | where |
1110 | N: Into<ExpandedName<'n, 'm>>, |
1111 | { |
1112 | let name = name.into(); |
1113 | self.attributes() |
1114 | .any(|a| a.data.name.as_expanded_name(self.doc) == name) |
1115 | } |
1116 | |
1117 | /// Returns element's attributes. |
1118 | /// |
1119 | /// # Examples |
1120 | /// |
1121 | /// ``` |
1122 | /// let doc = roxmltree::Document::parse( |
1123 | /// "<e xmlns:n='http://www.w3.org' a='b' n:a='c'/>" |
1124 | /// ).unwrap(); |
1125 | /// |
1126 | /// assert_eq!(doc.root_element().attributes().len(), 2); |
1127 | /// ``` |
1128 | #[inline ] |
1129 | pub fn attributes(&self) -> Attributes<'a, 'input> { |
1130 | Attributes::new(self) |
1131 | } |
1132 | |
1133 | /// Returns element's namespaces. |
1134 | /// |
1135 | /// # Examples |
1136 | /// |
1137 | /// ``` |
1138 | /// let doc = roxmltree::Document::parse( |
1139 | /// "<e xmlns:n='http://www.w3.org'/>" |
1140 | /// ).unwrap(); |
1141 | /// |
1142 | /// assert_eq!(doc.root_element().namespaces().len(), 1); |
1143 | /// ``` |
1144 | #[inline ] |
1145 | pub fn namespaces(&self) -> NamespaceIter<'a, 'input> { |
1146 | let namespaces = match self.d.kind { |
1147 | NodeKind::Element { ref namespaces, .. } => { |
1148 | &self.doc.namespaces.tree_order[namespaces.to_urange()] |
1149 | } |
1150 | _ => &[], |
1151 | }; |
1152 | |
1153 | NamespaceIter { |
1154 | doc: self.doc, |
1155 | namespaces: namespaces.iter(), |
1156 | } |
1157 | } |
1158 | |
1159 | /// Returns node's text. |
1160 | /// |
1161 | /// - for an element will return a first text child |
1162 | /// - for a comment will return a self text |
1163 | /// - for a text node will return a self text |
1164 | /// |
1165 | /// # Examples |
1166 | /// |
1167 | /// ``` |
1168 | /// let doc = roxmltree::Document::parse("\ |
1169 | /// <p> |
1170 | /// text |
1171 | /// </p> |
1172 | /// " ).unwrap(); |
1173 | /// |
1174 | /// assert_eq!(doc.root_element().text(), |
1175 | /// Some(" \n text \n" )); |
1176 | /// assert_eq!(doc.root_element().first_child().unwrap().text(), |
1177 | /// Some(" \n text \n" )); |
1178 | /// ``` |
1179 | /// |
1180 | /// ``` |
1181 | /// let doc = roxmltree::Document::parse("<!-- comment --><e/>" ).unwrap(); |
1182 | /// |
1183 | /// assert_eq!(doc.root().first_child().unwrap().text(), Some(" comment " )); |
1184 | /// ``` |
1185 | #[inline ] |
1186 | pub fn text(&self) -> Option<&'a str> { |
1187 | self.text_storage().map(|s| s.as_str()) |
1188 | } |
1189 | |
1190 | /// Returns node's text storage. |
1191 | /// |
1192 | /// Useful when you need a more low-level access to an allocated string. |
1193 | pub fn text_storage(&self) -> Option<&'a StringStorage<'input>> { |
1194 | match self.d.kind { |
1195 | NodeKind::Element { .. } => match self.first_child() { |
1196 | Some(child) if child.is_text() => match self.doc.nodes[child.id.get_usize()].kind { |
1197 | NodeKind::Text(ref text) => Some(text), |
1198 | _ => None, |
1199 | }, |
1200 | _ => None, |
1201 | }, |
1202 | NodeKind::Comment(ref text) => Some(text), |
1203 | NodeKind::Text(ref text) => Some(text), |
1204 | _ => None, |
1205 | } |
1206 | } |
1207 | |
1208 | /// Returns element's tail text. |
1209 | /// |
1210 | /// # Examples |
1211 | /// |
1212 | /// ``` |
1213 | /// let doc = roxmltree::Document::parse("\ |
1214 | /// <root> |
1215 | /// text1 |
1216 | /// <p/> |
1217 | /// text2 |
1218 | /// </root> |
1219 | /// " ).unwrap(); |
1220 | /// |
1221 | /// let p = doc.descendants().find(|n| n.has_tag_name("p" )).unwrap(); |
1222 | /// assert_eq!(p.tail(), Some(" \n text2 \n" )); |
1223 | /// ``` |
1224 | #[inline ] |
1225 | pub fn tail(&self) -> Option<&'a str> { |
1226 | self.tail_storage().map(|s| s.as_str()) |
1227 | } |
1228 | |
1229 | /// Returns element's tail text storage. |
1230 | /// |
1231 | /// Useful when you need a more low-level access to an allocated string. |
1232 | pub fn tail_storage(&self) -> Option<&'a StringStorage<'input>> { |
1233 | if !self.is_element() { |
1234 | return None; |
1235 | } |
1236 | |
1237 | match self.next_sibling().map(|n| n.id) { |
1238 | Some(id) => match self.doc.nodes[id.get_usize()].kind { |
1239 | NodeKind::Text(ref text) => Some(text), |
1240 | _ => None, |
1241 | }, |
1242 | None => None, |
1243 | } |
1244 | } |
1245 | |
1246 | /// Returns node as Processing Instruction. |
1247 | #[inline ] |
1248 | pub fn pi(&self) -> Option<PI<'input>> { |
1249 | match self.d.kind { |
1250 | NodeKind::PI(pi) => Some(pi), |
1251 | _ => None, |
1252 | } |
1253 | } |
1254 | |
1255 | /// Returns the parent of this node. |
1256 | #[inline ] |
1257 | pub fn parent(&self) -> Option<Self> { |
1258 | self.d.parent.map(|id| self.doc.get_node(id).unwrap()) |
1259 | } |
1260 | |
1261 | /// Returns the parent element of this node. |
1262 | pub fn parent_element(&self) -> Option<Self> { |
1263 | self.ancestors().skip(1).find(|n| n.is_element()) |
1264 | } |
1265 | |
1266 | /// Returns the previous sibling of this node. |
1267 | #[inline ] |
1268 | pub fn prev_sibling(&self) -> Option<Self> { |
1269 | self.d.prev_sibling.map(|id| self.doc.get_node(id).unwrap()) |
1270 | } |
1271 | |
1272 | /// Returns the previous sibling element of this node. |
1273 | pub fn prev_sibling_element(&self) -> Option<Self> { |
1274 | self.prev_siblings().skip(1).find(|n| n.is_element()) |
1275 | } |
1276 | |
1277 | /// Returns the next sibling of this node. |
1278 | #[inline ] |
1279 | pub fn next_sibling(&self) -> Option<Self> { |
1280 | self.d |
1281 | .next_subtree |
1282 | .map(|id| self.doc.get_node(id).unwrap()) |
1283 | .and_then(|node| { |
1284 | let possibly_self = node |
1285 | .d |
1286 | .prev_sibling |
1287 | .expect("next_subtree will always have a previous sibling" ); |
1288 | if possibly_self == self.id { |
1289 | Some(node) |
1290 | } else { |
1291 | None |
1292 | } |
1293 | }) |
1294 | } |
1295 | |
1296 | /// Returns the next sibling element of this node. |
1297 | pub fn next_sibling_element(&self) -> Option<Self> { |
1298 | self.next_siblings().skip(1).find(|n| n.is_element()) |
1299 | } |
1300 | |
1301 | /// Returns the first child of this node. |
1302 | #[inline ] |
1303 | pub fn first_child(&self) -> Option<Self> { |
1304 | self.d |
1305 | .last_child |
1306 | .map(|_| self.doc.get_node(NodeId::new(self.id.get() + 1)).unwrap()) |
1307 | } |
1308 | |
1309 | /// Returns the first element child of this node. |
1310 | pub fn first_element_child(&self) -> Option<Self> { |
1311 | self.children().find(|n| n.is_element()) |
1312 | } |
1313 | |
1314 | /// Returns the last child of this node. |
1315 | #[inline ] |
1316 | pub fn last_child(&self) -> Option<Self> { |
1317 | self.d.last_child.map(|id| self.doc.get_node(id).unwrap()) |
1318 | } |
1319 | |
1320 | /// Returns the last element child of this node. |
1321 | pub fn last_element_child(&self) -> Option<Self> { |
1322 | self.children().filter(|n| n.is_element()).last() |
1323 | } |
1324 | |
1325 | /// Returns true if this node has siblings. |
1326 | #[inline ] |
1327 | pub fn has_siblings(&self) -> bool { |
1328 | self.d.prev_sibling.is_some() || self.next_sibling().is_some() |
1329 | } |
1330 | |
1331 | /// Returns true if this node has children. |
1332 | #[inline ] |
1333 | pub fn has_children(&self) -> bool { |
1334 | self.d.last_child.is_some() |
1335 | } |
1336 | |
1337 | /// Returns an iterator over ancestor nodes starting at this node. |
1338 | #[inline ] |
1339 | pub fn ancestors(&self) -> AxisIter<'a, 'input> { |
1340 | AxisIter { |
1341 | node: Some(*self), |
1342 | next: Node::parent, |
1343 | } |
1344 | } |
1345 | |
1346 | /// Returns an iterator over previous sibling nodes starting at this node. |
1347 | #[inline ] |
1348 | pub fn prev_siblings(&self) -> AxisIter<'a, 'input> { |
1349 | AxisIter { |
1350 | node: Some(*self), |
1351 | next: Node::prev_sibling, |
1352 | } |
1353 | } |
1354 | |
1355 | /// Returns an iterator over next sibling nodes starting at this node. |
1356 | #[inline ] |
1357 | pub fn next_siblings(&self) -> AxisIter<'a, 'input> { |
1358 | AxisIter { |
1359 | node: Some(*self), |
1360 | next: Node::next_sibling, |
1361 | } |
1362 | } |
1363 | |
1364 | /// Returns an iterator over first children nodes starting at this node. |
1365 | #[inline ] |
1366 | pub fn first_children(&self) -> AxisIter<'a, 'input> { |
1367 | AxisIter { |
1368 | node: Some(*self), |
1369 | next: Node::first_child, |
1370 | } |
1371 | } |
1372 | |
1373 | /// Returns an iterator over last children nodes starting at this node. |
1374 | #[inline ] |
1375 | pub fn last_children(&self) -> AxisIter<'a, 'input> { |
1376 | AxisIter { |
1377 | node: Some(*self), |
1378 | next: Node::last_child, |
1379 | } |
1380 | } |
1381 | |
1382 | /// Returns an iterator over children nodes. |
1383 | #[inline ] |
1384 | pub fn children(&self) -> Children<'a, 'input> { |
1385 | Children { |
1386 | front: self.first_child(), |
1387 | back: self.last_child(), |
1388 | } |
1389 | } |
1390 | |
1391 | /// Returns an iterator over this node and its descendants. |
1392 | #[inline ] |
1393 | pub fn descendants(&self) -> Descendants<'a, 'input> { |
1394 | Descendants::new(*self) |
1395 | } |
1396 | |
1397 | /// Returns node's range in bytes in the original document. |
1398 | #[cfg (feature = "positions" )] |
1399 | #[inline ] |
1400 | pub fn range(&self) -> Range<usize> { |
1401 | self.d.range.clone() |
1402 | } |
1403 | |
1404 | /// Returns node's NodeId |
1405 | #[inline ] |
1406 | pub fn id(&self) -> NodeId { |
1407 | self.id |
1408 | } |
1409 | } |
1410 | |
1411 | impl<'a, 'input: 'a> fmt::Debug for Node<'a, 'input> { |
1412 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
1413 | match self.d.kind { |
1414 | NodeKind::Root => write!(f, "Root" ), |
1415 | NodeKind::Element { .. } => { |
1416 | write!( |
1417 | f, |
1418 | "Element {{ tag_name: {:?}, attributes: {:?}, namespaces: {:?} }}" , |
1419 | self.tag_name(), |
1420 | self.attributes(), |
1421 | self.namespaces() |
1422 | ) |
1423 | } |
1424 | NodeKind::PI(pi: PI<'_>) => { |
1425 | write!(f, "PI {{ target: {:?}, value: {:?} }}" , pi.target, pi.value) |
1426 | } |
1427 | NodeKind::Comment(ref text: &StringStorage<'_>) => write!(f, "Comment( {:?})" , text.as_str()), |
1428 | NodeKind::Text(ref text: &StringStorage<'_>) => write!(f, "Text( {:?})" , text.as_str()), |
1429 | } |
1430 | } |
1431 | } |
1432 | |
1433 | /// Iterator over a node's attributes |
1434 | #[derive (Clone)] |
1435 | pub struct Attributes<'a, 'input> { |
1436 | doc: &'a Document<'input>, |
1437 | attrs: core::slice::Iter<'a, AttributeData<'input>>, |
1438 | } |
1439 | |
1440 | impl<'a, 'input> Attributes<'a, 'input> { |
1441 | #[inline ] |
1442 | fn new(node: &Node<'a, 'input>) -> Attributes<'a, 'input> { |
1443 | let attrs: &[AttributeData<'_>] = match node.d.kind { |
1444 | NodeKind::Element { ref attributes: &ShortRange, .. } => { |
1445 | &node.doc.attributes[attributes.to_urange()] |
1446 | } |
1447 | _ => &[], |
1448 | }; |
1449 | Attributes { |
1450 | doc: node.doc, |
1451 | attrs: attrs.iter(), |
1452 | } |
1453 | } |
1454 | } |
1455 | |
1456 | impl<'a, 'input> Iterator for Attributes<'a, 'input> { |
1457 | type Item = Attribute<'a, 'input>; |
1458 | |
1459 | #[inline ] |
1460 | fn next(&mut self) -> Option<Self::Item> { |
1461 | self.attrs.next().map(|attr: &AttributeData<'_>| Attribute { |
1462 | doc: self.doc, |
1463 | data: attr, |
1464 | }) |
1465 | } |
1466 | |
1467 | #[inline ] |
1468 | fn nth(&mut self, n: usize) -> Option<Self::Item> { |
1469 | self.attrs.nth(n).map(|attr: &AttributeData<'_>| Attribute { |
1470 | doc: self.doc, |
1471 | data: attr, |
1472 | }) |
1473 | } |
1474 | |
1475 | #[inline ] |
1476 | fn size_hint(&self) -> (usize, Option<usize>) { |
1477 | self.attrs.size_hint() |
1478 | } |
1479 | } |
1480 | |
1481 | impl<'a, 'input> DoubleEndedIterator for Attributes<'a, 'input> { |
1482 | #[inline ] |
1483 | fn next_back(&mut self) -> Option<Self::Item> { |
1484 | self.attrs.next_back().map(|attr: &AttributeData<'_>| Attribute { |
1485 | doc: self.doc, |
1486 | data: attr, |
1487 | }) |
1488 | } |
1489 | } |
1490 | |
1491 | impl ExactSizeIterator for Attributes<'_, '_> {} |
1492 | |
1493 | impl fmt::Debug for Attributes<'_, '_> { |
1494 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
1495 | f&mut DebugStruct<'_, '_>.debug_struct("Attributes" ) |
1496 | .field(name:"attrs" , &self.attrs) |
1497 | .finish() |
1498 | } |
1499 | } |
1500 | |
1501 | /// Iterator over specified axis. |
1502 | #[derive (Clone)] |
1503 | pub struct AxisIter<'a, 'input: 'a> { |
1504 | node: Option<Node<'a, 'input>>, |
1505 | next: fn(&Node<'a, 'input>) -> Option<Node<'a, 'input>>, |
1506 | } |
1507 | |
1508 | impl<'a, 'input: 'a> Iterator for AxisIter<'a, 'input> { |
1509 | type Item = Node<'a, 'input>; |
1510 | |
1511 | #[inline ] |
1512 | fn next(&mut self) -> Option<Self::Item> { |
1513 | let node: Option> = self.node.take(); |
1514 | self.node = node.as_ref().and_then(self.next); |
1515 | node |
1516 | } |
1517 | } |
1518 | |
1519 | impl fmt::Debug for AxisIter<'_, '_> { |
1520 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
1521 | f&mut DebugStruct<'_, '_>.debug_struct("AxisIter" ) |
1522 | .field("node" , &self.node) |
1523 | .field(name:"next" , &"fn()" ) |
1524 | .finish() |
1525 | } |
1526 | } |
1527 | |
1528 | /// Iterator over children. |
1529 | #[derive (Clone, Debug)] |
1530 | pub struct Children<'a, 'input: 'a> { |
1531 | front: Option<Node<'a, 'input>>, |
1532 | back: Option<Node<'a, 'input>>, |
1533 | } |
1534 | |
1535 | impl<'a, 'input: 'a> Iterator for Children<'a, 'input> { |
1536 | type Item = Node<'a, 'input>; |
1537 | |
1538 | #[inline ] |
1539 | fn next(&mut self) -> Option<Self::Item> { |
1540 | if self.front == self.back { |
1541 | let node: Option> = self.front.take(); |
1542 | self.back = None; |
1543 | node |
1544 | } else { |
1545 | let node: Option> = self.front.take(); |
1546 | self.front = node.as_ref().and_then(Node::next_sibling); |
1547 | node |
1548 | } |
1549 | } |
1550 | } |
1551 | |
1552 | impl<'a, 'input: 'a> DoubleEndedIterator for Children<'a, 'input> { |
1553 | #[inline ] |
1554 | fn next_back(&mut self) -> Option<Self::Item> { |
1555 | if self.back == self.front { |
1556 | let node: Option> = self.back.take(); |
1557 | self.front = None; |
1558 | node |
1559 | } else { |
1560 | let node: Option> = self.back.take(); |
1561 | self.back = node.as_ref().and_then(Node::prev_sibling); |
1562 | node |
1563 | } |
1564 | } |
1565 | } |
1566 | |
1567 | /// Iterator over a node and its descendants. |
1568 | #[derive (Clone)] |
1569 | pub struct Descendants<'a, 'input> { |
1570 | doc: &'a Document<'input>, |
1571 | nodes: core::iter::Enumerate<core::slice::Iter<'a, NodeData<'input>>>, |
1572 | from: usize, |
1573 | } |
1574 | |
1575 | impl<'a, 'input> Descendants<'a, 'input> { |
1576 | #[inline ] |
1577 | fn new(start: Node<'a, 'input>) -> Self { |
1578 | let from: usize = start.id.get_usize(); |
1579 | |
1580 | let until: usize = start |
1581 | .d |
1582 | .next_subtree |
1583 | .map(NodeId::get_usize) |
1584 | .unwrap_or(default:start.doc.nodes.len()); |
1585 | |
1586 | let nodes: impl Iterator = start.doc.nodes[from..until].iter().enumerate(); |
1587 | |
1588 | Self { |
1589 | doc: start.doc, |
1590 | nodes, |
1591 | from, |
1592 | } |
1593 | } |
1594 | } |
1595 | |
1596 | impl<'a, 'input> Iterator for Descendants<'a, 'input> { |
1597 | type Item = Node<'a, 'input>; |
1598 | |
1599 | #[inline ] |
1600 | fn next(&mut self) -> Option<Self::Item> { |
1601 | self.nodes.next().map(|(idx, data)| Node { |
1602 | id: NodeId::from(self.from + idx), |
1603 | d: data, |
1604 | doc: self.doc, |
1605 | }) |
1606 | } |
1607 | |
1608 | #[inline ] |
1609 | fn nth(&mut self, n: usize) -> Option<Self::Item> { |
1610 | self.nodes.nth(n).map(|(idx, data)| Node { |
1611 | id: NodeId::from(self.from + idx), |
1612 | d: data, |
1613 | doc: self.doc, |
1614 | }) |
1615 | } |
1616 | |
1617 | #[inline ] |
1618 | fn size_hint(&self) -> (usize, Option<usize>) { |
1619 | self.nodes.size_hint() |
1620 | } |
1621 | } |
1622 | |
1623 | impl<'a, 'input> DoubleEndedIterator for Descendants<'a, 'input> { |
1624 | #[inline ] |
1625 | fn next_back(&mut self) -> Option<Self::Item> { |
1626 | self.nodes.next_back().map(|(idx: usize, data: &NodeData<'_>)| Node { |
1627 | id: NodeId::from(self.from + idx), |
1628 | d: data, |
1629 | doc: self.doc, |
1630 | }) |
1631 | } |
1632 | } |
1633 | |
1634 | impl ExactSizeIterator for Descendants<'_, '_> {} |
1635 | |
1636 | impl fmt::Debug for Descendants<'_, '_> { |
1637 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
1638 | f&mut DebugStruct<'_, '_>.debug_struct("Descendants" ) |
1639 | .field("nodes" , &self.nodes) |
1640 | .field(name:"from" , &self.from) |
1641 | .finish() |
1642 | } |
1643 | } |
1644 | |
1645 | /// Iterator over the namespaces attached to a node. |
1646 | #[derive (Clone)] |
1647 | pub struct NamespaceIter<'a, 'input> { |
1648 | doc: &'a Document<'input>, |
1649 | namespaces: core::slice::Iter<'a, NamespaceIdx>, |
1650 | } |
1651 | |
1652 | impl<'a, 'input> Iterator for NamespaceIter<'a, 'input> { |
1653 | type Item = &'a Namespace<'input>; |
1654 | |
1655 | #[inline ] |
1656 | fn next(&mut self) -> Option<Self::Item> { |
1657 | self.namespaces |
1658 | .next() |
1659 | .map(|idx: &NamespaceIdx| self.doc.namespaces.get(*idx)) |
1660 | } |
1661 | |
1662 | #[inline ] |
1663 | fn nth(&mut self, n: usize) -> Option<Self::Item> { |
1664 | self.namespaces |
1665 | .nth(n) |
1666 | .map(|idx: &NamespaceIdx| self.doc.namespaces.get(*idx)) |
1667 | } |
1668 | |
1669 | #[inline ] |
1670 | fn size_hint(&self) -> (usize, Option<usize>) { |
1671 | self.namespaces.size_hint() |
1672 | } |
1673 | } |
1674 | |
1675 | impl<'a, 'input> DoubleEndedIterator for NamespaceIter<'a, 'input> { |
1676 | #[inline ] |
1677 | fn next_back(&mut self) -> Option<Self::Item> { |
1678 | self.namespaces |
1679 | .next() |
1680 | .map(|idx: &NamespaceIdx| self.doc.namespaces.get(*idx)) |
1681 | } |
1682 | } |
1683 | |
1684 | impl ExactSizeIterator for NamespaceIter<'_, '_> {} |
1685 | |
1686 | impl fmt::Debug for NamespaceIter<'_, '_> { |
1687 | fn fmt(&self, f: &mut fmt::Formatter) -> Result<(), fmt::Error> { |
1688 | f&mut DebugStruct<'_, '_>.debug_struct("NamespaceIter" ) |
1689 | .field(name:"namespaces" , &self.namespaces) |
1690 | .finish() |
1691 | } |
1692 | } |
1693 | |