Change calls on replacing regexes to replace_all

Add `fix_relative_uris`, `clean_classes`, `clean_readability_attrs`
and `post_process_content`
This commit is contained in:
Kenneth Gitere 2020-10-21 16:21:53 +03:00
parent aacb442b7a
commit 350447d1c4
2 changed files with 353 additions and 7 deletions

View file

@ -1,4 +1,4 @@
use std::collections::{BTreeMap, HashMap};
use std::collections::{BTreeMap, HashMap, HashSet};
use std::str::FromStr;
use html5ever::{LocalName, Namespace, QualName};
@ -7,6 +7,7 @@ use kuchiki::{
traits::*,
NodeData, NodeRef,
};
use url::Url;
const SHARE_ELEMENT_THRESHOLD: usize = 500;
const READABILITY_SCORE: &'static str = "readability-score";
@ -68,13 +69,14 @@ impl Readability {
article_dir: None,
}
}
pub fn parse(&mut self) {
pub fn parse(&mut self, url: &str) {
self.unwrap_no_script_tags();
self.remove_scripts();
self.prep_document();
let meta_data = self.get_article_metadata();
self.article_title = meta_data.title.clone();
self.grab_article();
self.post_process_content(url);
}
/// Recursively check if node is image, or if node contains exactly only one image
@ -435,8 +437,8 @@ impl Readability {
let name_val = name_attr.unwrap();
if regexes::is_match_name_pattern(name_val) {
let name = name_val.to_lowercase();
let name = regexes::REPLACE_WHITESPACE_REGEX.replace(&name, "");
let name = regexes::REPLACE_DOT_REGEX.replace(&name, ":");
let name = regexes::REPLACE_WHITESPACE_REGEX.replace_all(&name, "");
let name = regexes::REPLACE_DOT_REGEX.replace_all(&name, ":");
values.insert(name.to_string(), content.trim().to_string());
}
}
@ -590,21 +592,162 @@ impl Readability {
}
}
cur_title = regexes::NORMALIZE_REGEX
.replace(cur_title.trim(), " ")
.replace_all(cur_title.trim(), " ")
.to_string();
let cur_word_count = word_count(&cur_title);
if cur_word_count <= 4
&& (!title_had_hierarchical_separators
|| cur_word_count
!= word_count(&regexes::REPLACE_MULTI_SEPARATOR_REGEX.replace(&orig_title, ""))
- 1)
!= word_count(
&regexes::REPLACE_MULTI_SEPARATOR_REGEX.replace_all(&orig_title, ""),
) - 1)
{
cur_title = orig_title;
}
cur_title
}
/// Removes the class="" attribute from every element in the given subtree, except those that
/// match CLASSES_TO_PRESERVE and the classesToPreserve array from the options object.
fn clean_classes(&mut self) {
// TODO: This should accessed from Self
let classes_to_preserve: HashSet<&str> = HashSet::new();
if let Some(article_node) = &mut self.article_node {
for elem in article_node.inclusive_descendants().elements() {
let mut elem_attrs = elem.attributes.borrow_mut();
if let Some(class_list) = elem_attrs.get_mut("class") {
let filtered_class: String = class_list
.split_whitespace()
.filter(|class| classes_to_preserve.contains(class))
.fold("".to_string(), |acc, x| acc + " " + x);
if filtered_class.is_empty() {
elem_attrs.remove("class");
} else {
*class_list = filtered_class;
}
}
}
}
}
/// Converts each <a> and <img> uri in the given element to an absolute URI, ignoring #ref URIs.
fn fix_relative_uris(&mut self, document_uri: &str) {
if let Some(article_node) = &mut self.article_node {
let document_uri =
Url::parse(document_uri).expect("Unable to parse the document's URI");
let base_uri = self
.root_node
.select("base")
.unwrap()
.filter(|node_ref| {
let node_attrs = node_ref.attributes.borrow();
node_attrs.contains("href")
})
.map(|node_ref| {
let node_attrs = node_ref.attributes.borrow();
Url::parse(node_attrs.get("href").unwrap()).unwrap()
})
.next()
.unwrap_or(document_uri.clone());
let to_absolute_uri = |uri_str: &str| -> String {
if base_uri == document_uri && uri_str.starts_with("#") {
return uri_str.to_string();
}
if let Ok(new_uri) = Url::parse(uri_str) {
if new_uri.has_host() {
return new_uri.to_string();
}
} else if let Ok(joined_uri) = base_uri.join(uri_str) {
return joined_uri.to_string();
}
uri_str.to_string()
};
let mut links = article_node.select("a").unwrap().filter(|a_ref| {
let link_attrs = a_ref.attributes.borrow();
link_attrs.contains("href")
});
let mut link = links.next();
while let Some(link_ref) = link {
link = links.next();
let mut link_attrs = link_ref.attributes.borrow_mut();
let href = link_attrs.get("href").map(|val| val.to_string()).unwrap();
if href.starts_with("javascript:") {
let link_node = link_ref.as_node();
if link_node.children().count() == 1
&& link_node
.first_child()
.map(|node_ref| node_ref.as_text().is_some())
.unwrap()
{
let text_node = NodeRef::new_text(link_node.text_contents());
link_node.insert_before(text_node);
link_node.detach();
} else {
let container = NodeRef::new_element(
QualName::new(None, Namespace::from(HTML_NS), LocalName::from("span")),
BTreeMap::new(),
);
let mut children = link_node.children();
let mut child = children.next();
while let Some(child_ref) = child {
child = children.next();
container.append(child_ref);
}
link_node.insert_before(container);
link_node.detach();
}
} else {
link_attrs.insert("href", to_absolute_uri(&href));
}
}
let media_nodes = article_node
.select("img, picture, figure, video, audio, source")
.unwrap();
for media_node in media_nodes {
let mut media_attrs = media_node.attributes.borrow_mut();
if let Some(src) = media_attrs.get_mut("src") {
*src = to_absolute_uri(&src);
}
if let Some(poster) = media_attrs.get_mut("poster") {
*poster = to_absolute_uri(&poster);
}
if let Some(srcset) = media_attrs.get_mut("srcset") {
let new_srcset = regexes::SRCSET_CAPTURE_REGEX.replace_all(
&srcset,
|captures: &regex::Captures| {
to_absolute_uri(&captures[1]) + &captures[2] + &captures[3]
},
);
*srcset = new_srcset.to_string();
}
}
}
}
/// Removes readability attributes from DOM nodes as they are not needed in the final article
fn clean_readability_attrs(&mut self) {
if let Some(article_node) = &mut self.article_node {
for node in article_node.inclusive_descendants().elements() {
let mut node_attrs = node.attributes.borrow_mut();
node_attrs.remove(READABILITY_SCORE);
node_attrs.remove("readability-data-table");
}
}
}
/// Run any post-process modifications to article content as necessary.
fn post_process_content(&mut self, url: &str) {
self.fix_relative_uris(url);
// TODO: Add flag check
self.clean_classes();
self.clean_readability_attrs();
}
/// Converts an inline CSS string to a [HashMap] of property and value(s)
fn inline_css_str_to_map(css_str: &str) -> HashMap<&str, &str> {
css_str
@ -3462,4 +3605,205 @@ characters. For that reason, this <p> tag could not be a byline because it's too
result.title = "A Longer Title".to_string();
assert_eq!(result, doc.get_article_metadata());
}
#[test]
fn test_fix_relative_uris() {
let html_str = r##"
<!DOCTYPE html>
<html>
<body>
<h1><a href="../home.html">Go back</a></h1>
<img id="ex-1" src="https://example.image.com/images/1.jpg" alt="Ex 1">
<img id="ex-2" src="https://example.image.com/images/2.jpg" alt="Ex 2">
<img id="ex-3" src="../images/2.jpg" alt="Ex 3">
<img id="ex-4" src="./images/1.jpg" alt="Ex 4">
<img id="ex-5" src="https://images.com/images/1.jpg" alt="Ex 5">
<p><a href="#ex-1">First image</a></p>
</body>
</html>
"##;
let mut doc = Readability::new(html_str);
doc.article_node = doc
.root_node
.select_first("body")
.ok()
.map(|node_ref| node_ref.as_node().clone());
doc.fix_relative_uris("https://example.image.com/blog/");
let node = doc.root_node.select_first("img#ex-1").unwrap();
let node_attrs = node.attributes.borrow();
assert_eq!(
Some("https://example.image.com/images/1.jpg"),
node_attrs.get("src")
);
let node = doc.root_node.select_first("img#ex-2").unwrap();
let node_attrs = node.attributes.borrow();
assert_eq!(
Some("https://example.image.com/images/2.jpg"),
node_attrs.get("src")
);
let node = doc.root_node.select_first("img#ex-3").unwrap();
let node_attrs = node.attributes.borrow();
assert_eq!(
Some("https://example.image.com/images/2.jpg"),
node_attrs.get("src")
);
let node = doc.root_node.select_first("img#ex-4").unwrap();
let node_attrs = node.attributes.borrow();
assert_eq!(
Some("https://example.image.com/blog/images/1.jpg"),
node_attrs.get("src")
);
let node = doc.root_node.select_first("img#ex-5").unwrap();
let node_attrs = node.attributes.borrow();
assert_eq!(
Some("https://images.com/images/1.jpg"),
node_attrs.get("src")
);
let node = doc.root_node.select_first("p a").unwrap();
let node_attrs = node.attributes.borrow();
assert_eq!(Some("#ex-1"), node_attrs.get("href"));
let node = doc.root_node.select_first("h1 a").unwrap();
let node_attrs = node.attributes.borrow();
assert_eq!(
Some("https://example.image.com/home.html"),
node_attrs.get("href")
);
}
#[test]
fn test_clean_classes() {
// TODO: This test will later be edited to ensure it checks to only remove certain classes
let html_str = r#"
<!DOCTYPE html>
<html>
<body>
<p class="a b c d">One</p>
<p class="b c d e">Two</p>
<div class="a b c div">Three</div>
<div class="b c d e">Four</div>
<ul class="a b c d">
<li class="a b c d">One</li>
<li class="b c d e">Two</li>
<li class="b c d e">Three</li>
</ul>
</body>
</html>
"#;
let mut doc = Readability::new(html_str);
doc.article_node = doc
.root_node
.select_first("body")
.ok()
.map(|node_ref| node_ref.as_node().clone());
doc.clean_classes();
assert_eq!(
true,
doc.root_node
.inclusive_descendants()
.elements()
.all(|node_elem| {
let node_attrs = node_elem.attributes.borrow();
!node_attrs.contains("class")
})
);
}
#[test]
fn test_clean_readability_attrs() {
let html_str = r#"
<!DOCTYPE html>
<html>
<body>
<div readability-score="0.921487">
<p readability-score="0.8102">Welcome to this awesome blog post. Only good content is here. No spam.</p>
<p readability-score="0.6004">Let's look at some statistics</p>
<table readability-score="0.719275" readability-data-table="true">
<caption>Monthly savings</caption>
<tr>
<th>Month</th>
<th>Savings</th>
</tr>
<tr>
<td>January</td>
<td>$100</td>
</tr>
<tr>
<td>February</td>
<td>$50</td>
</tr>
</table>
</div>
</body>
</html>
"#;
let mut doc = Readability::new(html_str);
doc.article_node = doc
.root_node
.select_first("body")
.ok()
.map(|node_ref| node_ref.as_node().clone());
doc.clean_readability_attrs();
assert_eq!(
true,
doc.root_node
.inclusive_descendants()
.elements()
.all(|node| {
let node_attrs = node.attributes.borrow();
node_attrs.map.len() == 0
})
);
}
#[test]
fn test_post_process_content() {
let html_str = r##"
<!DOCTYPE html>
<html>
<body>
<p class="a b c d">One</p>
<p class="b c d e">Two</p>
<div class="a b c div">Three</div>
<div class="b c d e">
<img src="./img.jpg" class="lazy">
</div>
<ul class="a b c d">
<li class="a b c d"><a href="#home">One</a></li>
<li class="b c d e">Two</li>
<li class="b c d e">Three</li>
</ul>
</body>
</html>
"##;
let mut doc = Readability::new(html_str);
doc.article_node = doc
.root_node
.select_first("body")
.ok()
.map(|node_ref| node_ref.as_node().clone());
doc.post_process_content("https://foo.blog/post/");
let has_class_attr = doc
.root_node
.inclusive_descendants()
.elements()
.any(|node_ref| {
let attrs = node_ref.attributes.borrow();
attrs.contains("class")
});
assert_eq!(false, has_class_attr);
let a_node = doc.root_node.select_first("a").unwrap();
let a_node_attrs = a_node.attributes.borrow();
assert_eq!(Some("#home"), a_node_attrs.get("href"));
let img_node = doc.root_node.select_first("img").unwrap();
let img_attrs = img_node.attributes.borrow();
assert_eq!(Some("https://foo.blog/post/img.jpg"), img_attrs.get("src"));
}
}

View file

@ -119,6 +119,8 @@ lazy_static! {
r"(?i)\s*(dc|dcterm|og|twitter)\s*:\s*(author|creator|description|title|site_name)\s*"
)
.unwrap();
pub static ref SRCSET_CAPTURE_REGEX: Regex =
Regex::new(r"(\S+)(\s+[\d.]+[xw])?(\s*(?:,|$))").unwrap();
pub static ref REPLACE_WHITESPACE_REGEX: Regex = Regex::new(r"\s").unwrap();
pub static ref REPLACE_DOT_REGEX: Regex = Regex::new(r"\.").unwrap();
pub static ref REPLACE_HTML_ESCAPE_REGEX: Regex =