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
|
use crate::default_emitter::DefaultEmitter;
use crate::offset::{Offset, Position};
use crate::reader::{IntoReader, Reader};
use crate::tokenizer::CdataAction;
use crate::{Emitter, Event, State, Tokenizer};
/// A naive HTML parser (**not** spec-compliant since it doesn't do tree construction).
///
/// * it **does not** correct [misnested tags]
///
/// * it **does not** recognize implicitly self-closing elements like
/// `<img>`, it will simply emit a start token
///
/// * it naively emits any CDATA sections as bogus comments, for example:
///
/// ```
/// # use html5tokenizer::{Error, NaiveParser, Tokenizer, Token};
/// let html = "<svg><![CDATA[I love SVG]]>";
/// let mut tokens = NaiveParser::new(html).flatten();
/// assert!(matches!(tokens.next().unwrap(), Token::StartTag(tag) if tag.name == "svg"));
/// assert!(matches!(tokens.next().unwrap(), Token::Error {error: Error::CdataInHtmlContent, ..}));
/// assert!(matches!(tokens.next().unwrap(), Token::Comment(_bogus_comment)));
/// ```
///
/// It has similar caveats to the [HTMLParser] from the Python standard library.
/// It should suffice for web scraping but you wouldn't use it to implement a browser.
///
/// [misnested tags]: https://html.spec.whatwg.org/multipage/parsing.html#an-introduction-to-error-handling-and-strange-cases-in-the-parser
/// [HTMLParser]: https://docs.python.org/3/library/html.parser.html
pub struct NaiveParser<R: Reader, O: Offset, E: Emitter<O>> {
tokenizer: Tokenizer<R, O, E>,
}
impl<R: Reader + Position<O>, O: Offset> NaiveParser<R, O, DefaultEmitter<O>> {
/// Constructs a new naive parser.
// TODO: add example for NaiveParser::new
pub fn new<'a>(reader: impl IntoReader<'a, Reader = R>) -> Self {
let mut tokenizer = Tokenizer::new(reader, DefaultEmitter::default());
tokenizer.naively_switch_state = true;
NaiveParser { tokenizer }
}
}
impl<R: Reader + Position<usize>> NaiveParser<R, usize, DefaultEmitter<usize>> {
/// Constructs a new naive parser with source code offsets and spans.
// TODO: add example for NaiveParser::new_with_spans
pub fn new_with_spans<'a>(reader: impl IntoReader<'a, Reader = R>) -> Self {
let mut tokenizer = Tokenizer::new(reader, DefaultEmitter::default());
tokenizer.naively_switch_state = true;
NaiveParser { tokenizer }
}
}
impl<R: Reader + Position<O>, O: Offset, E: Emitter<O>> NaiveParser<R, O, E> {
/// Constructs a new naive parser with a custom emitter.
// TODO: add example for NaiveParser::new_with_emitter
pub fn new_with_emitter<'a>(reader: impl IntoReader<'a, Reader = R>, emitter: E) -> Self {
let mut tokenizer = Tokenizer::new(reader, emitter);
tokenizer.naively_switch_state = true;
NaiveParser { tokenizer }
}
}
impl<R, O, E> Iterator for NaiveParser<R, O, E>
where
R: Reader + Position<O>,
O: Offset,
E: Emitter<O> + Iterator,
{
type Item = Result<E::Item, R::Error>;
fn next(&mut self) -> Option<Self::Item> {
loop {
let event = self.tokenizer.next()?;
match event {
Err(e) => return Some(Err(e)),
Ok(Event::Token(t)) => {
// A proper parser would follow the steps described under section '13.2.6 Tree construction'
// of the spec. Since this parser is naive, we directly return the token instead.
return Some(Ok(t));
}
Ok(Event::CdataOpen) => {
// Naively parse any CDATA sections as bogus comments.
self.tokenizer.handle_cdata_open(CdataAction::BogusComment)
}
}
}
}
}
pub(crate) fn naive_next_state(tag_name: &str) -> State {
// These transitions are defined in https://html.spec.whatwg.org/multipage/parsing.html#parsing-html-fragments.
// TODO: investigate what state logic Python's HTMLParser is using
match tag_name {
"title" | "textarea" => State::RcData,
"style" | "xmp" | "iframe" | "noembed" | "noframes" => State::RawText,
"script" => State::ScriptData,
"plaintext" => State::PlainText,
_other => State::Data,
}
}
|
