1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
use crate::Result;
use pest::{iterators::Pair, iterators::Pairs, Parser};
use serde::Serialize;
use std::default::Default;
use crate::error::Error;
use crate::grammar::Grammar;
use crate::Rule;
pub mod element;
pub mod formatting;
pub mod node;
pub mod span;
use crate::dom::span::SourceSpan;
use element::{Element, ElementVariant};
use node::Node;
/// Document, DocumentFragment or Empty
#[derive(Debug, Clone, PartialEq, Serialize)]
#[serde(rename_all = "camelCase")]
pub enum DomVariant {
/// This means that the parsed html had the representation of an html document. The doctype is optional but a document should only have one root node with the name of html.
/// Example:
/// ```text
/// <!doctype html>
/// <html>
/// <head></head>
/// <body>
/// <h1>Hello world</h1>
/// </body>
/// </html>
/// ```
Document,
/// A document fragment means that the parsed html did not have the representation of a document. A fragment can have multiple root children of any name except html, body or head.
/// Example:
/// ```text
/// <h1>Hello world</h1>
/// ```
DocumentFragment,
/// An empty dom means that the input was empty
Empty,
}
/// **The main struct** & the result of the parsed html
#[derive(Debug, Clone, Serialize, PartialEq)]
#[serde(rename_all = "camelCase")]
pub struct Dom {
/// The type of the tree that was parsed
pub tree_type: DomVariant,
/// All of the root children in the tree
#[serde(skip_serializing_if = "Vec::is_empty")]
pub children: Vec<Node>,
/// A collection of all errors during parsing
#[serde(skip_serializing)]
pub errors: Vec<String>,
}
impl Default for Dom {
fn default() -> Self {
Self {
tree_type: DomVariant::Empty,
children: vec![],
errors: vec![],
}
}
}
impl Dom {
pub fn parse(input: &str) -> Result<Self> {
let pairs = match Grammar::parse(Rule::html, input) {
Ok(pairs) => pairs,
Err(error) => return formatting::error_msg(error),
};
Self::build_dom(pairs)
}
pub fn to_json(&self) -> Result<String> {
Ok(serde_json::to_string(self)?)
}
pub fn to_json_pretty(&self) -> Result<String> {
Ok(serde_json::to_string_pretty(self)?)
}
fn build_dom(pairs: Pairs<Rule>) -> Result<Self> {
let mut dom = Self::default();
// NOTE: The logic is roughly as follows:
// 1) A document containing nothing but comments is DomVariant::Empty even though it will have
// children in this first pass. We fix this in the next section. This allows us to use
// DomVariant::Empty to indicate "we haven't decided the type yet".
// 2) If the type is DomVariant::Empty _so far_, then it can be changed to DomVariant::Document
// or DomVariant::DocumentFragment. DomVariant is only selected in this stage if we see a
// DOCTYPE tag. Comments do not change the type.
// 3) If the type is non-empty, we don't re-set the type. We do look for conflicts between
// the type and the tokens in the next stage.
for pair in pairs {
match pair.as_rule() {
// A <!DOCTYPE> tag means a full-fledged document. Note that because of the way
// the grammar is written, we will only get this token if the <!DOCTYPE> occurs
// before any other tag; otherwise it will be parsed as a custom tag.
Rule::doctype => {
if dom.tree_type == DomVariant::Empty {
dom.tree_type = DomVariant::Document;
}
}
// If we see an element, build the sub-tree and add it as a child. If we don't
// have a document type yet (i.e. "empty"), select DocumentFragment
Rule::node_element => match Self::build_node_element(pair, &mut dom) {
Ok(el) => {
if let Some(node) = el {
if dom.tree_type == DomVariant::Empty {
dom.tree_type = DomVariant::DocumentFragment;
};
dom.children.push(node);
}
}
Err(error) => {
dom.errors.push(format!("{}", error));
}
},
// Similar to an element, we add it as a child and select DocumentFragment if we
// don't already have a document type.
Rule::node_text => {
if dom.tree_type == DomVariant::Empty {
dom.tree_type = DomVariant::DocumentFragment;
}
let text = pair.as_str().to_string();
if !text.is_empty() {
dom.children.push(Node::Text(text));
}
}
// Store comments as a child, but it doesn't affect the document type selection
// until the next phase (validation).
Rule::node_comment => {
dom.children
.push(Node::Comment(pair.into_inner().as_str().to_string()));
}
// Ignore 'end of input', which then allows the catch-all unreachable!() arm to
// function properly.
Rule::EOI => (),
// This should be unreachable, due to the way the grammar is written
_ => unreachable!("[build dom] unknown rule: {:?}", pair.as_rule()),
};
}
// Implement some checks on the generated dom's data and initial type. The type may be
// modified in this section.
match dom.tree_type {
// A DomVariant::Empty can only have comments. Anything else is an error.
DomVariant::Empty => {
for node in &dom.children {
if let Node::Comment(_) = node {
// An "empty" document, but it has comments - this is where we cleanup the
// earlier assumption that a document with only comments is "empty".
// Really, it is a "fragment".
dom.tree_type = DomVariant::DocumentFragment
} else {
// Anything else (i.e. Text() or Element() ) can't happen at the top level;
// if we had seen one, we would have set the document type above
unreachable!("[build dom] empty document with an Element {:?}", node)
}
}
}
// A DomVariant::Document can only have comments and an <HTML> node at the top level.
// Only one <HTML> tag is permitted.
DomVariant::Document => {
if dom
.children
.iter()
.filter(|x| match x {
Node::Element(el) if el.name.to_lowercase() == "html" => true,
_ => false,
})
.count()
> 1
{
return Err(Error::Parsing(format!("Document with multiple HTML tags",)));
}
}
// A DomVariant::DocumentFragment should not have <HEAD>, or <BODY> tags at the
// top-level. If we find an <HTML> tag, then we consider this a Document instead (if
// it comes before any other elements, and if there is only one <HTML> tag).
DomVariant::DocumentFragment => {
let mut seen_html = false;
let mut seen_elements = false;
for node in &dom.children {
match node {
// Nodes other than <HTML> - reject <HEAD> and <BODY>
Node::Element(ref el) if el.name.clone().to_lowercase() != "html" => {
if el.name == "head" || el.name == "body" {
return Err(Error::Parsing(format!(
"A document fragment should not include {}",
el.name
)));
}
seen_elements = true;
}
// <HTML> Nodes - one (before any other elements) is okay
Node::Element(ref el) if el.name.clone().to_lowercase() == "html" => {
if seen_html || seen_elements {
return Err(Error::Parsing(format!(
"A document fragment should not include {}",
el.name
)));
};
// A fragment with just an <HTML> tag is a document
dom.tree_type = DomVariant::Document;
seen_html = true;
}
// Comment() and Text() nodes are permitted at the top-level of a
// DocumentFragment
_ => (),
}
}
}
}
// The result is the validated tree
Ok(dom)
}
fn build_node_element(pair: Pair<Rule>, dom: &mut Dom) -> Result<Option<Node>> {
let source_span = {
let pair_span = pair.as_span();
let (start_line, start_column) = pair_span.start_pos().line_col();
let (end_line, end_column) = pair_span.end_pos().line_col();
SourceSpan::new(
String::from(pair_span.as_str()),
start_line,
end_line,
start_column,
end_column,
)
};
let mut element = Element {
source_span,
..Element::default()
};
for pair in pair.into_inner() {
match pair.as_rule() {
Rule::node_element | Rule::el_raw_text => {
match Self::build_node_element(pair, dom) {
Ok(el) => {
if let Some(child_element) = el {
element.children.push(child_element)
}
}
Err(error) => {
dom.errors.push(format!("{}", error));
}
}
}
Rule::node_text | Rule::el_raw_text_content => {
let text = pair.as_str().to_string();
if !text.is_empty() {
element.children.push(Node::Text(text));
}
}
Rule::node_comment => {
element
.children
.push(Node::Comment(pair.into_inner().as_str().to_string()));
}
// TODO: To enable some kind of validation we should probably align this with
// https://html.spec.whatwg.org/multipage/syntax.html#elements-2
// Also see element variants
Rule::el_name | Rule::el_void_name | Rule::el_raw_text_name => {
element.name = pair.as_str().to_string();
}
Rule::attr => match Self::build_attribute(pair.into_inner()) {
Ok((attr_key, attr_value)) => {
match attr_key.as_str() {
"id" => element.id = attr_value,
"class" => {
if let Some(classes) = attr_value {
let classes = classes.split_whitespace().collect::<Vec<_>>();
for class in classes {
element.classes.push(class.to_string());
}
}
}
_ => {
element.attributes.insert(attr_key, attr_value);
}
};
}
Err(error) => {
dom.errors.push(format!("{}", error));
}
},
Rule::el_normal_end | Rule::el_raw_text_end => {
element.variant = ElementVariant::Normal;
break;
}
Rule::el_dangling => (),
Rule::EOI => (),
_ => {
return Err(Error::Parsing(format!(
"Failed to create element at rule: {:?}",
pair.as_rule()
)))
}
}
}
if element.name != "" {
Ok(Some(Node::Element(element)))
} else {
Ok(None)
}
}
fn build_attribute(pairs: Pairs<Rule>) -> Result<(String, Option<String>)> {
let mut attribute = ("".to_string(), None);
for pair in pairs {
match pair.as_rule() {
Rule::attr_key => {
attribute.0 = pair.as_str().trim().to_string();
}
Rule::attr_non_quoted => {
attribute.1 = Some(pair.as_str().trim().to_string());
}
Rule::attr_quoted => {
let inner_pair = pair
.into_inner()
.into_iter()
.next()
.expect("attribute value");
match inner_pair.as_rule() {
Rule::attr_value => attribute.1 = Some(inner_pair.as_str().to_string()),
_ => {
return Err(Error::Parsing(format!(
"Failed to parse attr value: {:?}",
inner_pair.as_rule()
)))
}
}
}
_ => {
return Err(Error::Parsing(format!(
"Failed to parse attr: {:?}",
pair.as_rule()
)))
}
}
}
Ok(attribute)
}
}