misc-mirrors/article_scraper
1
0
Fork 0
mirror of https://gitlab.com/news-flash/article_scraper.git synced 2025-07-07 16:15:32 +02:00
Code Releases Activity

start refactor & fingerprints

Browse source
This commit is contained in:
Jan Lukas Gernert 2022-10-08 23:09:00 +02:00
parent 29df3aa698
commit 273ddd832c
16 changed files with 944 additions and 848 deletions
Download patch file Download diff file Expand all files Collapse all files

1
Cargo.toml
View file

@ -20,3 +20,4 @@ base64 = "0.13"
image = "0.24"
log = "0.4"
rust-embed="6.4"
once_cell = "1.15"

15
src/article.rs
View file

@ -1,8 +1,7 @@
use crate::error::{ScraperError, ScraperErrorKind};
use chrono::{DateTime, Utc};
use failure::ResultExt;
use std::io::Write;
use std::io::{Error, ErrorKind, Write};
use std::path::PathBuf;
use std::fs::File;
use url::Url;
pub struct Article {
@ -14,7 +13,7 @@ pub struct Article {
}
impl Article {
pub fn save_html(&self, path: &PathBuf) -> Result<(), ScraperError> {
pub fn save_html(&self, path: &PathBuf) -> Result<(), Error> {
if let Some(ref html) = self.html {
if let Ok(()) = std::fs::create_dir_all(&path) {
let mut file_name = match self.title.clone() {
@ -23,14 +22,12 @@ impl Article {
};
file_name.push_str(".html");
let path = path.join(file_name);
let mut html_file = std::fs::File::create(&path).context(ScraperErrorKind::IO)?;
html_file
.write_all(html.as_bytes())
.context(ScraperErrorKind::IO)?;
let mut html_file = File::create(&path)?;
html_file.write_all(html.as_bytes())?;
return Ok(());
}
}
Err(ScraperErrorKind::Unknown.into())
Err(Error::new(ErrorKind::NotFound, "Article does not contain HTML"))
}
}

14
src/error.rs
View file

@ -8,20 +8,6 @@ pub struct ScraperError {
#[derive(Copy, Clone, Eq, PartialEq, Debug, Fail)]
pub enum ScraperErrorKind {
#[fail(display = "libXml Error")]
Xml,
#[fail(display = "No content found")]
Scrape,
#[fail(display = "Url Error")]
Url,
#[fail(display = "Http request failed")]
Http,
#[fail(display = "Config Error")]
Config,
#[fail(display = "IO Error")]
IO,
#[fail(display = "Content-type suggest no html")]
ContentType,
#[fail(display = "Unknown Error")]
Unknown,
}

4
src/config/config_collection.rs → src/full_text_parser/config/config_collection.rs
View file

@ -66,8 +66,4 @@ impl ConfigCollection {
self.embedded_entries.get(key)
}
}
pub fn contains_config(&self, key: &str) -> bool {
self.user_entries.contains_key(key) || self.embedded_entries.contains_key(key)
}
}

0
src/config/config_entry.rs → src/full_text_parser/config/config_entry.rs
View file

0
src/config/error.rs → src/full_text_parser/config/error.rs
View file

0
src/config/macros.rs → src/full_text_parser/config/macros.rs
View file

0
src/config/mod.rs → src/full_text_parser/config/mod.rs
View file

71
src/full_text_parser/error.rs Normal file
View file

@ -0,0 +1,71 @@
use failure::{Backtrace, Context, Error, Fail};
use std::fmt;
#[derive(Debug)]
pub struct FullTextParserError {
inner: Context<FullTextParserErrorKind>,
}
#[derive(Copy, Clone, Eq, PartialEq, Debug, Fail)]
pub enum FullTextParserErrorKind {
#[fail(display = "libXml Error")]
Xml,
#[fail(display = "No content found")]
Scrape,
#[fail(display = "Url Error")]
Url,
#[fail(display = "Http request failed")]
Http,
#[fail(display = "Config Error")]
Config,
#[fail(display = "IO Error")]
IO,
#[fail(display = "Content-type suggest no html")]
ContentType,
#[fail(display = "Unknown Error")]
Unknown,
}
impl Fail for FullTextParserError {
fn cause(&self) -> Option<&dyn Fail> {
self.inner.cause()
}
fn backtrace(&self) -> Option<&Backtrace> {
self.inner.backtrace()
}
}
impl fmt::Display for FullTextParserError {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
fmt::Display::fmt(&self.inner, f)
}
}
impl FullTextParserError {
pub fn kind(&self) -> FullTextParserErrorKind {
*self.inner.get_context()
}
}
impl From<FullTextParserErrorKind> for FullTextParserError {
fn from(kind: FullTextParserErrorKind) -> FullTextParserError {
FullTextParserError {
inner: Context::new(kind),
}
}
}
impl From<Context<FullTextParserErrorKind>> for FullTextParserError {
fn from(inner: Context<FullTextParserErrorKind>) -> FullTextParserError {
FullTextParserError { inner }
}
}
impl From<Error> for FullTextParserError {
fn from(_: Error) -> FullTextParserError {
FullTextParserError {
inner: Context::new(FullTextParserErrorKind::Unknown),
}
}
}

47
src/full_text_parser/fingerprints.rs Normal file
View file

@ -0,0 +1,47 @@
use once_cell::sync::Lazy;
use regex::Regex;
use reqwest::Url;
use std::collections::HashMap;
static FINGERPRINT_REGEXES: Lazy<HashMap<&'static str, Regex>> = Lazy::new(|| {
let mut m = HashMap::with_capacity(4);
m.insert(
"fingerprint.blogspot.com",
regex::Regex::new(
r#"/\<meta\s*content=([\'"])blogger([\'"])\s*name=([\'"])generator([\'"])/i"#,
)
.expect("failed to build static regex"),
);
m.insert(
"fingerprint.blogspot.com",
regex::Regex::new(
r#"/\<meta\s*name=([\'"])generator([\'"])\s*content=([\'"])Blogger([\'"])/i"#,
)
.expect("failed to build static regex"),
);
m.insert(
"fingerprint.wordpress.com",
regex::Regex::new(r#"/\<meta\s*name=([\'"])generator([\'"])\s*content=([\'"])WordPress/i"#)
.expect("failed to build static regex"),
);
m.insert(
"fingerprint.ippen.media",
regex::Regex::new(r#"/\<div\s*class=([\'"])id-SiteBEEPWrap([\'"])\>/i"#)
.expect("failed to build static regex"),
);
m
});
pub struct Fingerprints;
impl Fingerprints {
pub fn detect(html: &str) -> Option<Url> {
for (url, regex) in FINGERPRINT_REGEXES.iter() {
if regex.captures(html).is_some() {
return Some(Url::parse(url).expect("failed to parse static url"));
}
}
None
}
}

768
src/full_text_parser/mod.rs Normal file
View file

@ -0,0 +1,768 @@
pub mod error;
pub mod config;
mod fingerprints;
#[cfg(test)]
mod tests;
use self::error::{FullTextParserError, FullTextParserErrorKind};
use crate::article::Article;
use self::config::{ConfigCollection, ConfigEntry};
use chrono::DateTime;
use encoding_rs::Encoding;
use failure::ResultExt;
use fingerprints::Fingerprints;
use libxml::parser::Parser;
use libxml::tree::{Document, Node, SaveOptions};
use libxml::xpath::Context;
use log::{debug, error, info, warn};
use reqwest::header::HeaderMap;
use reqwest::Client;
use std::path::Path;
use std::str::FromStr;
use crate::util::Util;
pub struct FullTextParser {
config_files: ConfigCollection,
}
impl FullTextParser {
pub async fn new(config_path: Option<&Path>) -> Self {
let config_files = ConfigCollection::parse(config_path).await;
Self {
config_files,
}
}
pub async fn parse(
&self,
url: &url::Url,
client: &Client,
) -> Result<Article, FullTextParserError> {
info!("Scraping article: '{}'", url.as_str());
// check if we have a config for the url
let config = self.get_grabber_config(url);
let global_config = self
.config_files
.get("global.txt")
.ok_or(FullTextParserErrorKind::Config)?;
let headers = Util::generate_headers(config, global_config)?;
let response = client
.head(url.clone())
.headers(headers)
.send()
.await
.map_err(|err| {
error!("Failed head request to: '{}' - '{}'", url.as_str(), err);
err
})
.context(FullTextParserErrorKind::Http)?;
// check if url redirects and we need to pick up the new url
let url = if let Some(new_url) = Util::check_redirect(&response, url) {
debug!("Url '{}' redirects to '{}'", url.as_str(), new_url.as_str());
new_url
} else {
url.clone()
};
// check if we are dealing with text/html
if !Util::check_content_type(&response)? {
return Err(FullTextParserErrorKind::ContentType.into());
}
let mut article = Article {
title: None,
author: None,
url: url.clone(),
date: None,
html: None,
};
let mut document = Document::new().map_err(|()| FullTextParserErrorKind::Xml)?;
let mut root = Node::new("article", None, &document).map_err(|()| FullTextParserErrorKind::Xml)?;
document.set_root_element(&root);
Self::generate_head(&mut root, &document)?;
self.parse_pages(&mut article, &url, &mut root, config, global_config, client)
.await?;
let context = Context::new(&document).map_err(|()| {
error!("Failed to create xpath context for extracted article");
FullTextParserErrorKind::Xml
})?;
if let Err(error) = Self::prevent_self_closing_tags(&context) {
error!("Preventing self closing tags failed - '{}'", error);
return Err(error);
}
// serialize content
let options = SaveOptions {
format: false,
no_declaration: false,
no_empty_tags: true,
no_xhtml: false,
xhtml: false,
as_xml: false,
as_html: true,
non_significant_whitespace: false,
};
let html = document.to_string_with_options(options);
article.html = Some(html);
Ok(article)
}
async fn parse_pages(
&self,
article: &mut Article,
url: &url::Url,
root: &mut Node,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
client: &Client,
) -> Result<(), FullTextParserError> {
let headers = Util::generate_headers(config, global_config)?;
let html = Self::download(url, client, headers).await?;
// see if
let config = if config.is_none() {
if let Some(url) = Fingerprints::detect(&html) {
self.get_grabber_config(&url)
} else {
config
}
} else {
config
};
let mut document = Self::parse_html(html, config, global_config)?;
let mut xpath_ctx = Self::get_xpath_ctx(&document)?;
// check for single page link
let rule = Util::select_rule(
config.and_then(|c| c.single_page_link.as_deref()),
global_config.single_page_link.as_deref(),
);
if let Some(xpath_single_page_link) = rule {
debug!(
"Single page link xpath specified in config '{}'",
xpath_single_page_link
);
if let Some(single_page_url) = Util::find_page_url(&xpath_ctx, xpath_single_page_link) {
// parse again with single page url
debug!("Single page link found '{}'", single_page_url);
return self
.parse_single_page(
article,
&single_page_url,
root,
config,
global_config,
client,
)
.await;
}
}
Self::extract_metadata(&xpath_ctx, config, global_config, article);
Self::strip_junk(&xpath_ctx, config, global_config, url);
Self::extract_body(&xpath_ctx, root, config, global_config)?;
while let Some(url) = self.check_for_next_page(&xpath_ctx, config, global_config) {
let headers = Util::generate_headers(config, global_config)?;
let html = Self::download(&url, client, headers).await?;
document = Self::parse_html(html, config, global_config)?;
xpath_ctx = Self::get_xpath_ctx(&document)?;
Self::strip_junk(&xpath_ctx, config, global_config, &url);
Self::extract_body(&xpath_ctx, root, config, global_config)?;
}
Ok(())
}
fn parse_html(
html: String,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
) -> Result<Document, FullTextParserError> {
// replace matches in raw html
let mut html = html;
if let Some(config) = config {
for replace in &config.replace {
html = html.replace(&replace.to_replace, &replace.replace_with);
}
}
for replace in &global_config.replace {
html = html.replace(&replace.to_replace, &replace.replace_with);
}
// parse html
let parser = Parser::default_html();
Ok(parser.parse_string(html.as_str()).map_err(|err| {
error!("Parsing HTML failed for downloaded HTML {:?}", err);
FullTextParserErrorKind::Xml
})?)
}
fn get_xpath_ctx(doc: &Document) -> Result<Context, FullTextParserError> {
Ok(Context::new(doc).map_err(|()| {
error!("Creating xpath context failed for downloaded HTML");
FullTextParserErrorKind::Xml
})?)
}
async fn parse_single_page(
&self,
article: &mut Article,
url: &url::Url,
root: &mut Node,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
client: &Client,
) -> Result<(), FullTextParserError> {
let headers = Util::generate_headers(config, global_config)?;
let html = Self::download(url, client, headers).await?;
let document = Self::parse_html(html, config, global_config)?;
let xpath_ctx = Self::get_xpath_ctx(&document)?;
Self::extract_metadata(&xpath_ctx, config, global_config, article);
Self::strip_junk(&xpath_ctx, config, global_config, url);
Self::extract_body(&xpath_ctx, root, config, global_config)?;
Ok(())
}
async fn download(
url: &url::Url,
client: &Client,
headers: HeaderMap,
) -> Result<String, FullTextParserError> {
let response = client
.get(url.as_str())
.headers(headers)
.send()
.await
.map_err(|err| {
error!(
"Downloading HTML failed: GET '{}' - '{}'",
url.as_str(),
err
);
err
})
.context(FullTextParserErrorKind::Http)?;
if response.status().is_success() {
let headers = response.headers().clone();
let text = response.text().await.context(FullTextParserErrorKind::Http)?;
{
if let Some(decoded_html) =
Self::decode_html(&text, Self::get_encoding_from_html(&text))
{
return Ok(decoded_html);
}
if let Some(decoded_html) =
Self::decode_html(&text, Self::get_encoding_from_http_header(&headers))
{
return Ok(decoded_html);
}
}
warn!("No encoding of HTML detected - assuming utf-8");
return Ok(text);
}
Err(FullTextParserErrorKind::Http.into())
}
fn get_encoding_from_http_header(headers: &reqwest::header::HeaderMap) -> Option<&str> {
if let Some(content_type) = headers.get(reqwest::header::CONTENT_TYPE) {
if let Ok(content_type) = content_type.to_str() {
let regex =
regex::Regex::new(r#"charset=([^"']+)"#).expect("Failed to parse regex");
if let Some(captures) = regex.captures(content_type) {
if let Some(regex_match) = captures.get(1) {
return Some(regex_match.as_str());
}
}
}
}
None
}
fn get_encoding_from_html(html: &str) -> Option<&str> {
let regex =
regex::Regex::new(r#"<meta.*?charset=([^"']+)"#).expect("Failed to parse regex");
if let Some(captures) = regex.captures(html) {
if let Some(regex_match) = captures.get(1) {
return Some(regex_match.as_str());
}
}
None
}
fn decode_html(html: &str, encoding: Option<&str>) -> Option<String> {
if let Some(encoding) = encoding {
if let Some(encoding) = Encoding::for_label(encoding.as_bytes()) {
let (decoded_html, _, invalid_chars) = encoding.decode(html.as_bytes());
if !invalid_chars {
return Some(decoded_html.into_owned());
}
}
warn!("Could not decode HTML. Encoding: '{}'", encoding);
}
None
}
fn get_host_name(url: &url::Url) -> Result<String, FullTextParserError> {
match url.host_str() {
Some(name) => {
let mut name = name;
if name.starts_with("www.") && name.len() > 4 {
name = &name[4..]
}
Ok(name.into())
}
None => {
error!("Getting config failed due to bad Url");
Err(FullTextParserErrorKind::Config.into())
}
}
}
fn get_grabber_config(&self, url: &url::Url) -> Option<&ConfigEntry> {
let conf = Self::get_host_name(url)
.ok()
.map(|url| url + ".txt")
.and_then(|name| self.config_files.get(&name));
if conf.is_none() {
log::warn!("No config found for url '{}'", url);
}
conf
}
fn fix_lazy_images(
context: &Context,
class: &str,
property_url: &str,
) -> Result<(), FullTextParserError> {
let xpath = &format!("//img[contains(@class, '{}')]", class);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if let Some(correct_url) = node.get_property(property_url) {
if node.set_property("src", &correct_url).is_err() {
return Err(FullTextParserErrorKind::Xml.into());
}
}
}
Ok(())
}
fn fix_iframe_size(context: &Context, site_name: &str) -> Result<(), FullTextParserError> {
let xpath = &format!("//iframe[contains(@src, '{}')]", site_name);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if let Some(mut parent) = node.get_parent() {
if let Ok(mut video_wrapper) = parent.new_child(None, "div") {
if let Ok(()) = video_wrapper.set_property("class", "videoWrapper") {
if let Ok(()) = node.set_property("width", "100%") {
if let Ok(()) = node.set_property("height", "100%") {
node.unlink();
video_wrapper.add_child(&mut node).map_err(|_| {
error!("Failed to add iframe as child of video wrapper <div>");
FullTextParserErrorKind::Xml
})?;
}
}
}
}
error!("Failed to add video wrapper <div> as parent of iframe");
return Err(FullTextParserErrorKind::Xml.into());
}
error!("Failed to get parent of iframe");
// return Err(ScraperErrorKind::Xml.into());
}
Ok(())
}
fn remove_attribute(
context: &Context,
tag: Option<&str>,
attribute: &str,
) -> Result<(), FullTextParserError> {
let xpath_tag = tag.unwrap_or("*");
let xpath = &format!("//{}[@{}]", xpath_tag, attribute);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.remove_property(attribute).is_err() {
return Err(FullTextParserErrorKind::Xml.into());
}
}
Ok(())
}
fn add_attribute(
context: &Context,
tag: Option<&str>,
attribute: &str,
value: &str,
) -> Result<(), FullTextParserError> {
let xpath_tag = tag.unwrap_or("*");
let xpath = &format!("//{}", xpath_tag);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.set_attribute(attribute, value).is_err() {
return Err(FullTextParserErrorKind::Xml.into());
}
}
Ok(())
}
fn get_attribute(
context: &Context,
xpath: &str,
attribute: &str,
) -> Result<String, FullTextParserError> {
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for node in node_vec {
if let Some(value) = node.get_attribute(attribute) {
return Ok(value);
}
}
Err(FullTextParserErrorKind::Xml.into())
}
fn repair_urls(
context: &Context,
xpath: &str,
attribute: &str,
article_url: &url::Url,
) -> Result<(), FullTextParserError> {
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if let Some(val) = node.get_attribute(attribute) {
if let Err(url::ParseError::RelativeUrlWithoutBase) = url::Url::parse(&val) {
if let Ok(fixed_url) = Self::complete_url(article_url, &val) {
if node.set_attribute(attribute, fixed_url.as_str()).is_err() {
return Err(FullTextParserErrorKind::Scrape.into());
}
}
}
}
}
Ok(())
}
fn complete_url(
article_url: &url::Url,
incomplete_url: &str,
) -> Result<url::Url, FullTextParserError> {
let mut completed_url = article_url.scheme().to_owned();
completed_url.push(':');
if !incomplete_url.starts_with("//") {
match article_url.host() {
Some(url::Host::Domain(host)) => {
completed_url.push_str("//");
completed_url.push_str(host);
}
_ => return Err(FullTextParserErrorKind::Scrape.into()),
};
}
if !completed_url.ends_with('/') && !incomplete_url.starts_with('/') {
completed_url.push('/');
}
completed_url.push_str(incomplete_url);
let url = url::Url::parse(&completed_url).context(FullTextParserErrorKind::Url)?;
Ok(url)
}
fn strip_junk(
context: &Context,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
url: &url::Url,
) {
// strip specified xpath
if let Some(config) = config {
for xpath_strip in &config.xpath_strip {
let _ = Util::strip_node(context, xpath_strip);
}
}
for xpath_strip in &global_config.xpath_strip {
let _ = Util::strip_node(context, xpath_strip);
}
// strip everything with specified 'id' or 'class'
if let Some(config) = config {
for xpaht_strip_class in &config.strip_id_or_class {
let _ = Util::strip_id_or_class(context, xpaht_strip_class);
}
}
for xpaht_strip_class in &global_config.strip_id_or_class {
let _ = Util::strip_id_or_class(context, xpaht_strip_class);
}
// strip any <img> element where @src attribute contains this substring
if let Some(config) = config {
for xpath_strip_img_src in &config.strip_image_src {
let _ = Util::strip_node(
context,
&format!("//img[contains(@src,'{}')]", xpath_strip_img_src),
);
}
}
for xpath_strip_img_src in &global_config.strip_image_src {
let _ = Util::strip_node(
context,
&format!("//img[contains(@src,'{}')]", xpath_strip_img_src),
);
}
let _ = Self::fix_lazy_images(context, "lazyload", "data-src");
let _ = Self::fix_iframe_size(context, "youtube.com");
let _ = Self::remove_attribute(context, None, "style");
let _ = Self::remove_attribute(context, Some("a"), "onclick");
let _ = Self::remove_attribute(context, Some("img"), "srcset");
let _ = Self::remove_attribute(context, Some("img"), "sizes");
let _ = Self::add_attribute(context, Some("a"), "target", "_blank");
let _ = Self::repair_urls(context, "//img", "src", url);
let _ = Self::repair_urls(context, "//a", "src", url);
let _ = Self::repair_urls(context, "//a", "href", url);
let _ = Self::repair_urls(context, "//object", "data", url);
let _ = Self::repair_urls(context, "//iframe", "src", url);
// strip elements using Readability.com and Instapaper.com ignore class names
// .entry-unrelated and .instapaper_ignore
// See http://blog.instapaper.com/post/730281947
let _ = Util::strip_node(
context,
"//*[contains(@class,' entry-unrelated ') or contains(@class,' instapaper_ignore ')]",
);
// strip elements that contain style="display: none;"
let _ = Util::strip_node(context, "//*[contains(@style,'display:none')]");
// strip all comments
let _ = Util::strip_node(context, "//comment()");
// strip all empty url-tags <a/>
let _ = Util::strip_node(context, "//a[not(node())]");
// strip all external css and fonts
let _ = Util::strip_node(context, "//*[@type='text/css']");
}
fn extract_metadata(
context: &Context,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
article: &mut Article,
) {
// try to get title
if let Some(config) = config {
for xpath_title in &config.xpath_title {
if let Ok(title) = Util::extract_value_merge(context, xpath_title) {
debug!("Article title: '{}'", title);
article.title = Some(title);
break;
}
}
}
if article.title.is_none() {
for xpath_title in &global_config.xpath_title {
if let Ok(title) = Util::extract_value_merge(context, xpath_title) {
debug!("Article title: '{}'", title);
article.title = Some(title);
break;
}
}
}
// try to get the author
if let Some(config) = config {
for xpath_author in &config.xpath_author {
if let Ok(author) = Util::extract_value(context, xpath_author) {
debug!("Article author: '{}'", author);
article.author = Some(author);
break;
}
}
}
if article.author.is_none() {
for xpath_author in &global_config.xpath_author {
if let Ok(author) = Util::extract_value(context, xpath_author) {
debug!("Article author: '{}'", author);
article.author = Some(author);
break;
}
}
}
// try to get the date
if let Some(config) = config {
for xpath_date in &config.xpath_date {
if let Ok(date_string) = Util::extract_value(context, xpath_date) {
debug!("Article date: '{}'", date_string);
if let Ok(date) = DateTime::from_str(&date_string) {
article.date = Some(date);
break;
} else {
warn!("Parsing the date string '{}' failed", date_string);
}
}
}
}
if article.date.is_none() {
for xpath_date in &global_config.xpath_date {
if let Ok(date_string) = Util::extract_value(context, xpath_date) {
debug!("Article date: '{}'", date_string);
if let Ok(date) = DateTime::from_str(&date_string) {
article.date = Some(date);
break;
} else {
warn!("Parsing the date string '{}' failed", date_string);
}
}
}
}
}
fn extract_body(
context: &Context,
root: &mut Node,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
) -> Result<(), FullTextParserError> {
let mut found_something = false;
if let Some(config) = config {
for xpath_body in &config.xpath_body {
found_something = Self::extract_body_single(context, root, xpath_body)?;
}
}
if !found_something {
for xpath_body in &global_config.xpath_body {
found_something = Self::extract_body_single(context, root, xpath_body)?;
}
}
if !found_something {
log::error!("no body found");
return Err(FullTextParserErrorKind::Scrape.into());
}
Ok(())
}
fn extract_body_single(
context: &Context,
root: &mut Node,
xpath: &str,
) -> Result<bool, FullTextParserError> {
let mut found_something = false;
{
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.get_property("style").is_some() && node.remove_property("style").is_err() {
return Err(FullTextParserErrorKind::Xml.into());
}
node.unlink();
if root.add_child(&mut node).is_ok() {
found_something = true;
} else {
error!("Failed to add body to prepared document");
return Err(FullTextParserErrorKind::Xml.into());
}
}
}
Ok(found_something)
}
fn check_for_next_page(
&self,
context: &Context,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
) -> Option<url::Url> {
if let Some(config) = config {
if let Some(next_page_xpath) = config.next_page_link.as_deref() {
if let Ok(next_page_string) = Self::get_attribute(context, next_page_xpath, "href")
{
if let Ok(next_page_url) = url::Url::parse(&next_page_string) {
return Some(next_page_url);
}
}
}
} else if let Some(next_page_xpath) = global_config.next_page_link.as_deref() {
if let Ok(next_page_string) = Self::get_attribute(context, next_page_xpath, "href") {
if let Ok(next_page_url) = url::Url::parse(&next_page_string) {
return Some(next_page_url);
}
}
}
// last page reached
None
}
fn generate_head(root: &mut Node, document: &Document) -> Result<(), FullTextParserError> {
if let Ok(mut head_node) = Node::new("head", None, document) {
if let Ok(()) = root.add_prev_sibling(&mut head_node) {
if let Ok(mut meta) = head_node.new_child(None, "meta") {
if meta.set_property("charset", "utf-8").is_ok() {
return Ok(());
}
}
}
}
Err(FullTextParserErrorKind::Xml.into())
}
fn prevent_self_closing_tags(context: &Context) -> Result<(), FullTextParserError> {
// search document for empty tags and add a empty text node as child
// this prevents libxml from self closing non void elements such as iframe
let xpath = "//*[not(node())]";
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.get_name() == "meta" {
continue;
}
let _ = node.add_text_child(None, "empty", "");
}
Ok(())
}
}

14
src/tests.rs → src/full_text_parser/tests.rs
View file

@ -1,4 +1,4 @@
use crate::*;
use super::FullTextParser;
use reqwest::Client;
use std::path::PathBuf;
@ -7,8 +7,8 @@ async fn golem() {
let out_path = PathBuf::from(r"./test_output");
let url = url::Url::parse("https://www.golem.de/news/http-error-418-fehlercode-ich-bin-eine-teekanne-darf-bleiben-1708-129460.html").unwrap();
let grabber = ArticleScraper::new(None).await;
let article = grabber.parse(&url, true, &Client::new()).await.unwrap();
let grabber = FullTextParser::new(None).await;
let article = grabber.parse(&url, &Client::new()).await.unwrap();
article.save_html(&out_path).unwrap();
assert_eq!(
@ -27,8 +27,8 @@ async fn phoronix() {
url::Url::parse("http://www.phoronix.com/scan.php?page=article&item=amazon_ec2_bare&num=1")
.unwrap();
let grabber = ArticleScraper::new(None).await;
let article = grabber.parse(&url, false, &Client::new()).await.unwrap();
let grabber = FullTextParser::new(None).await;
let article = grabber.parse(&url, &Client::new()).await.unwrap();
article.save_html(&out_path).unwrap();
assert_eq!(
@ -44,8 +44,8 @@ async fn youtube() {
let out_path = PathBuf::from(r"./test_output");
let url = url::Url::parse("https://www.youtube.com/watch?v=8KjaIumu-jI").unwrap();
let grabber = ArticleScraper::new(None).await;
let article = grabber.parse(&url, false, &Client::new()).await.unwrap();
let grabber = FullTextParser::new(None).await;
let article = grabber.parse(&url, &Client::new()).await.unwrap();
article.save_html(&out_path).unwrap();
assert_eq!(

8
src/images/error.rs
View file

@ -1,4 +1,4 @@
use super::super::ScraperErrorKind;
use crate::full_text_parser::error::FullTextParserErrorKind;
use failure::{Backtrace, Context, Error, Fail};
use std::fmt;
@ -67,10 +67,10 @@ impl From<Context<ImageDownloadErrorKind>> for ImageDownloadError {
}
}
impl From<ScraperErrorKind> for ImageDownloadError {
fn from(kind: ScraperErrorKind) -> ImageDownloadError {
impl From<FullTextParserErrorKind> for ImageDownloadError {
fn from(kind: FullTextParserErrorKind) -> ImageDownloadError {
let kind = match kind {
ScraperErrorKind::Xml => ImageDownloadErrorKind::HtmlParse,
FullTextParserErrorKind::Xml => ImageDownloadErrorKind::HtmlParse,
_ => ImageDownloadErrorKind::Unknown,
};

782
src/lib.rs
View file

@ -1,41 +1,26 @@
mod article;
mod config;
mod error;
pub mod images;
mod article;
mod full_text_parser;
mod util;
mod youtube;
mod error;
#[cfg(test)]
mod tests;
use self::error::{ScraperError, ScraperErrorKind};
use crate::article::Article;
use crate::config::{ConfigCollection, ConfigEntry};
use crate::images::ImageDownloader;
use chrono::DateTime;
use encoding_rs::Encoding;
use failure::ResultExt;
use libxml::parser::Parser;
use libxml::tree::{Document, Node, SaveOptions};
use libxml::xpath::Context;
use log::{debug, error, info, warn};
use reqwest::header::HeaderMap;
use reqwest::Client;
use std::path::Path;
use std::str::FromStr;
use util::Util;
use article::Article;
use full_text_parser::FullTextParser;
use error::{ScraperError, ScraperErrorKind};
use images::ImageDownloader;
use reqwest::Client;
pub struct ArticleScraper {
pub image_downloader: ImageDownloader,
config_files: ConfigCollection,
full_text_parser: FullTextParser,
image_downloader: ImageDownloader,
}
impl ArticleScraper {
pub async fn new(config_path: Option<&Path>) -> Self {
let config_files = ConfigCollection::parse(config_path).await;
ArticleScraper {
pub async fn new(user_configs: Option<&Path>) -> Self {
Self {
full_text_parser: FullTextParser::new(user_configs).await,
image_downloader: ImageDownloader::new((2048, 2048)),
config_files,
}
}
@ -45,742 +30,19 @@ impl ArticleScraper {
download_images: bool,
client: &Client,
) -> Result<Article, ScraperError> {
info!("Scraping article: '{}'", url.as_str());
// custom youtube handling, but prefer config if exists
if !self.config_files.contains_config("youtube.com.txt") {
if let Some(article) = youtube::Youtube::handle(url) {
return Ok(article);
}
}
// check if we have a config for the url
let config = self.get_grabber_config(url);
let global_config = self
.config_files
.get("global.txt")
.ok_or(ScraperErrorKind::Config)?;
let headers = Util::generate_headers(config, global_config)?;
let response = client
.head(url.clone())
.headers(headers)
.send()
.await
.map_err(|err| {
error!("Failed head request to: '{}' - '{}'", url.as_str(), err);
err
})
.context(ScraperErrorKind::Http)?;
// check if url redirects and we need to pick up the new url
let url = if let Some(new_url) = Util::check_redirect(&response, url) {
debug!("Url '{}' redirects to '{}'", url.as_str(), new_url.as_str());
new_url
} else {
url.clone()
};
// check if we are dealing with text/html
if !Util::check_content_type(&response)? {
return Err(ScraperErrorKind::ContentType.into());
}
let mut article = Article {
title: None,
author: None,
url: url.clone(),
date: None,
html: None,
};
let mut document = Document::new().map_err(|()| ScraperErrorKind::Xml)?;
let mut root = Node::new("article", None, &document).map_err(|()| ScraperErrorKind::Xml)?;
document.set_root_element(&root);
ArticleScraper::generate_head(&mut root, &document)?;
self.parse_pages(&mut article, &url, &mut root, config, global_config, client)
.await?;
let context = Context::new(&document).map_err(|()| {
error!("Failed to create xpath context for extracted article");
ScraperErrorKind::Xml
})?;
if let Err(error) = ArticleScraper::prevent_self_closing_tags(&context) {
error!("Preventing self closing tags failed - '{}'", error);
return Err(error);
}
let res = self.full_text_parser.parse(url, client).await;
if download_images {
if let Err(error) = self
.image_downloader
.download_images_from_context(&context, client)
.await
{
error!("Downloading images failed: '{}'", error);
}
}
// serialize content
let options = SaveOptions {
format: false,
no_declaration: false,
no_empty_tags: true,
no_xhtml: false,
xhtml: false,
as_xml: false,
as_html: true,
non_significant_whitespace: false,
};
let html = document.to_string_with_options(options);
article.html = Some(html);
Ok(article)
}
async fn parse_pages(
&self,
article: &mut Article,
url: &url::Url,
root: &mut Node,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
client: &Client,
) -> Result<(), ScraperError> {
let headers = Util::generate_headers(config, global_config)?;
let html = ArticleScraper::download(url, client, headers).await?;
let mut document = Self::parse_html(html, config, global_config)?;
let mut xpath_ctx = Self::get_xpath_ctx(&document)?;
// check for single page link
let rule = Util::select_rule(
config.and_then(|c| c.single_page_link.as_deref()),
global_config.single_page_link.as_deref(),
);
if let Some(xpath_single_page_link) = rule {
debug!(
"Single page link xpath specified in config '{}'",
xpath_single_page_link
);
if let Some(single_page_url) = Util::find_page_url(&xpath_ctx, xpath_single_page_link) {
// parse again with single page url
debug!("Single page link found '{}'", single_page_url);
return self
.parse_single_page(
article,
&single_page_url,
root,
config,
global_config,
client,
)
.await;
}
}
ArticleScraper::extract_metadata(&xpath_ctx, config, global_config, article);
ArticleScraper::strip_junk(&xpath_ctx, config, global_config, url);
ArticleScraper::extract_body(&xpath_ctx, root, config, global_config)?;
while let Some(url) = self.check_for_next_page(&xpath_ctx, config, global_config) {
let headers = Util::generate_headers(config, global_config)?;
let html = ArticleScraper::download(&url, client, headers).await?;
document = Self::parse_html(html, config, global_config)?;
xpath_ctx = Self::get_xpath_ctx(&document)?;
ArticleScraper::strip_junk(&xpath_ctx, config, global_config, &url);
ArticleScraper::extract_body(&xpath_ctx, root, config, global_config)?;
}
Ok(())
}
fn parse_html(
html: String,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
) -> Result<Document, ScraperError> {
// replace matches in raw html
let mut html = html;
if let Some(config) = config {
for replace in &config.replace {
html = html.replace(&replace.to_replace, &replace.replace_with);
}
}
for replace in &global_config.replace {
html = html.replace(&replace.to_replace, &replace.replace_with);
}
// parse html
let parser = Parser::default_html();
Ok(parser.parse_string(html.as_str()).map_err(|err| {
error!("Parsing HTML failed for downloaded HTML {:?}", err);
ScraperErrorKind::Xml
})?)
}
fn get_xpath_ctx(doc: &Document) -> Result<Context, ScraperError> {
Ok(Context::new(doc).map_err(|()| {
error!("Creating xpath context failed for downloaded HTML");
ScraperErrorKind::Xml
})?)
}
async fn parse_single_page(
&self,
article: &mut Article,
url: &url::Url,
root: &mut Node,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
client: &Client,
) -> Result<(), ScraperError> {
let headers = Util::generate_headers(config, global_config)?;
let html = ArticleScraper::download(url, client, headers).await?;
let document = Self::parse_html(html, config, global_config)?;
let xpath_ctx = Self::get_xpath_ctx(&document)?;
ArticleScraper::extract_metadata(&xpath_ctx, config, global_config, article);
ArticleScraper::strip_junk(&xpath_ctx, config, global_config, url);
ArticleScraper::extract_body(&xpath_ctx, root, config, global_config)?;
Ok(())
}
async fn download(
url: &url::Url,
client: &Client,
headers: HeaderMap,
) -> Result<String, ScraperError> {
let response = client
.get(url.as_str())
.headers(headers)
.send()
.await
.map_err(|err| {
error!(
"Downloading HTML failed: GET '{}' - '{}'",
url.as_str(),
err
);
err
})
.context(ScraperErrorKind::Http)?;
if response.status().is_success() {
let headers = response.headers().clone();
let text = response.text().await.context(ScraperErrorKind::Http)?;
{
if let Some(decoded_html) = ArticleScraper::decode_html(
&text,
ArticleScraper::get_encoding_from_html(&text),
) {
return Ok(decoded_html);
}
if let Some(decoded_html) = ArticleScraper::decode_html(
&text,
ArticleScraper::get_encoding_from_http_header(&headers),
) {
return Ok(decoded_html);
}
}
warn!("No encoding of HTML detected - assuming utf-8");
return Ok(text);
}
Err(ScraperErrorKind::Http.into())
}
fn get_encoding_from_http_header(headers: &reqwest::header::HeaderMap) -> Option<&str> {
if let Some(content_type) = headers.get(reqwest::header::CONTENT_TYPE) {
if let Ok(content_type) = content_type.to_str() {
let regex =
regex::Regex::new(r#"charset=([^"']+)"#).expect("Failed to parse regex");
if let Some(captures) = regex.captures(content_type) {
if let Some(regex_match) = captures.get(1) {
return Some(regex_match.as_str());
}
}
}
}
None
}
fn get_encoding_from_html(html: &str) -> Option<&str> {
let regex =
regex::Regex::new(r#"<meta.*?charset=([^"']+)"#).expect("Failed to parse regex");
if let Some(captures) = regex.captures(html) {
if let Some(regex_match) = captures.get(1) {
return Some(regex_match.as_str());
}
}
None
}
fn decode_html(html: &str, encoding: Option<&str>) -> Option<String> {
if let Some(encoding) = encoding {
if let Some(encoding) = Encoding::for_label(encoding.as_bytes()) {
let (decoded_html, _, invalid_chars) = encoding.decode(html.as_bytes());
if !invalid_chars {
return Some(decoded_html.into_owned());
}
}
warn!("Could not decode HTML. Encoding: '{}'", encoding);
}
None
}
fn get_host_name(url: &url::Url) -> Result<String, ScraperError> {
match url.host_str() {
Some(name) => {
let mut name = name;
if name.starts_with("www.") && name.len() > 4 {
name = &name[4..]
}
Ok(name.into())
}
None => {
error!("Getting config failed due to bad Url");
Err(ScraperErrorKind::Config.into())
}
}
}
fn get_grabber_config(&self, url: &url::Url) -> Option<&ConfigEntry> {
let conf = Self::get_host_name(url)
.ok()
.map(|url| url + ".txt")
.and_then(|name| self.config_files.get(&name));
if conf.is_none() {
log::warn!("No config found for url '{}'", url);
}
conf
}
fn fix_lazy_images(
context: &Context,
class: &str,
property_url: &str,
) -> Result<(), ScraperError> {
let xpath = &format!("//img[contains(@class, '{}')]", class);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if let Some(correct_url) = node.get_property(property_url) {
if node.set_property("src", &correct_url).is_err() {
return Err(ScraperErrorKind::Xml.into());
}
}
}
Ok(())
}
fn fix_iframe_size(context: &Context, site_name: &str) -> Result<(), ScraperError> {
let xpath = &format!("//iframe[contains(@src, '{}')]", site_name);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if let Some(mut parent) = node.get_parent() {
if let Ok(mut video_wrapper) = parent.new_child(None, "div") {
if let Ok(()) = video_wrapper.set_property("class", "videoWrapper") {
if let Ok(()) = node.set_property("width", "100%") {
if let Ok(()) = node.set_property("height", "100%") {
node.unlink();
video_wrapper.add_child(&mut node).map_err(|_| {
error!("Failed to add iframe as child of video wrapper <div>");
ScraperErrorKind::Xml
})?;
}
}
}
}
error!("Failed to add video wrapper <div> as parent of iframe");
return Err(ScraperErrorKind::Xml.into());
}
error!("Failed to get parent of iframe");
// return Err(ScraperErrorKind::Xml.into());
}
Ok(())
}
fn remove_attribute(
context: &Context,
tag: Option<&str>,
attribute: &str,
) -> Result<(), ScraperError> {
let xpath_tag = tag.unwrap_or("*");
let xpath = &format!("//{}[@{}]", xpath_tag, attribute);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.remove_property(attribute).is_err() {
return Err(ScraperErrorKind::Xml.into());
}
}
Ok(())
}
fn add_attribute(
context: &Context,
tag: Option<&str>,
attribute: &str,
value: &str,
) -> Result<(), ScraperError> {
let xpath_tag = tag.unwrap_or("*");
let xpath = &format!("//{}", xpath_tag);
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.set_attribute(attribute, value).is_err() {
return Err(ScraperErrorKind::Xml.into());
}
}
Ok(())
}
fn get_attribute(
context: &Context,
xpath: &str,
attribute: &str,
) -> Result<String, ScraperError> {
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for node in node_vec {
if let Some(value) = node.get_attribute(attribute) {
return Ok(value);
}
}
Err(ScraperErrorKind::Xml.into())
}
fn repair_urls(
context: &Context,
xpath: &str,
attribute: &str,
article_url: &url::Url,
) -> Result<(), ScraperError> {
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if let Some(val) = node.get_attribute(attribute) {
if let Err(url::ParseError::RelativeUrlWithoutBase) = url::Url::parse(&val) {
if let Ok(fixed_url) = ArticleScraper::complete_url(article_url, &val) {
if node.set_attribute(attribute, fixed_url.as_str()).is_err() {
return Err(ScraperErrorKind::Scrape.into());
}
}
}
}
}
Ok(())
}
fn complete_url(
article_url: &url::Url,
incomplete_url: &str,
) -> Result<url::Url, ScraperError> {
let mut completed_url = article_url.scheme().to_owned();
completed_url.push(':');
if !incomplete_url.starts_with("//") {
match article_url.host() {
Some(url::Host::Domain(host)) => {
completed_url.push_str("//");
completed_url.push_str(host);
}
_ => return Err(ScraperErrorKind::Scrape.into()),
};
}
if !completed_url.ends_with('/') && !incomplete_url.starts_with('/') {
completed_url.push('/');
}
completed_url.push_str(incomplete_url);
let url = url::Url::parse(&completed_url).context(ScraperErrorKind::Url)?;
Ok(url)
}
fn strip_junk(
context: &Context,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
url: &url::Url,
) {
// strip specified xpath
if let Some(config) = config {
for xpath_strip in &config.xpath_strip {
let _ = Util::strip_node(context, xpath_strip);
}
}
for xpath_strip in &global_config.xpath_strip {
let _ = Util::strip_node(context, xpath_strip);
}
// strip everything with specified 'id' or 'class'
if let Some(config) = config {
for xpaht_strip_class in &config.strip_id_or_class {
let _ = Util::strip_id_or_class(context, xpaht_strip_class);
}
}
for xpaht_strip_class in &global_config.strip_id_or_class {
let _ = Util::strip_id_or_class(context, xpaht_strip_class);
}
// strip any <img> element where @src attribute contains this substring
if let Some(config) = config {
for xpath_strip_img_src in &config.strip_image_src {
let _ = Util::strip_node(
context,
&format!("//img[contains(@src,'{}')]", xpath_strip_img_src),
);
}
}
for xpath_strip_img_src in &global_config.strip_image_src {
let _ = Util::strip_node(
context,
&format!("//img[contains(@src,'{}')]", xpath_strip_img_src),
);
}
let _ = ArticleScraper::fix_lazy_images(context, "lazyload", "data-src");
let _ = ArticleScraper::fix_iframe_size(context, "youtube.com");
let _ = ArticleScraper::remove_attribute(context, None, "style");
let _ = ArticleScraper::remove_attribute(context, Some("a"), "onclick");
let _ = ArticleScraper::remove_attribute(context, Some("img"), "srcset");
let _ = ArticleScraper::remove_attribute(context, Some("img"), "sizes");
let _ = ArticleScraper::add_attribute(context, Some("a"), "target", "_blank");
let _ = ArticleScraper::repair_urls(context, "//img", "src", url);
let _ = ArticleScraper::repair_urls(context, "//a", "src", url);
let _ = ArticleScraper::repair_urls(context, "//a", "href", url);
let _ = ArticleScraper::repair_urls(context, "//object", "data", url);
let _ = ArticleScraper::repair_urls(context, "//iframe", "src", url);
// strip elements using Readability.com and Instapaper.com ignore class names
// .entry-unrelated and .instapaper_ignore
// See http://blog.instapaper.com/post/730281947
let _ = Util::strip_node(
context,
"//*[contains(@class,' entry-unrelated ') or contains(@class,' instapaper_ignore ')]",
);
// strip elements that contain style="display: none;"
let _ = Util::strip_node(context, "//*[contains(@style,'display:none')]");
// strip all comments
let _ = Util::strip_node(context, "//comment()");
// strip all empty url-tags <a/>
let _ = Util::strip_node(context, "//a[not(node())]");
// strip all external css and fonts
let _ = Util::strip_node(context, "//*[@type='text/css']");
}
fn extract_metadata(
context: &Context,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
article: &mut Article,
) {
// try to get title
if let Some(config) = config {
for xpath_title in &config.xpath_title {
if let Ok(title) = Util::extract_value_merge(context, xpath_title) {
debug!("Article title: '{}'", title);
article.title = Some(title);
break;
}
}
}
if article.title.is_none() {
for xpath_title in &global_config.xpath_title {
if let Ok(title) = Util::extract_value_merge(context, xpath_title) {
debug!("Article title: '{}'", title);
article.title = Some(title);
break;
}
}
}
// try to get the author
if let Some(config) = config {
for xpath_author in &config.xpath_author {
if let Ok(author) = Util::extract_value(context, xpath_author) {
debug!("Article author: '{}'", author);
article.author = Some(author);
break;
}
}
}
if article.author.is_none() {
for xpath_author in &global_config.xpath_author {
if let Ok(author) = Util::extract_value(context, xpath_author) {
debug!("Article author: '{}'", author);
article.author = Some(author);
break;
}
}
}
// try to get the date
if let Some(config) = config {
for xpath_date in &config.xpath_date {
if let Ok(date_string) = Util::extract_value(context, xpath_date) {
debug!("Article date: '{}'", date_string);
if let Ok(date) = DateTime::from_str(&date_string) {
article.date = Some(date);
break;
} else {
warn!("Parsing the date string '{}' failed", date_string);
}
}
}
}
if article.date.is_none() {
for xpath_date in &global_config.xpath_date {
if let Ok(date_string) = Util::extract_value(context, xpath_date) {
debug!("Article date: '{}'", date_string);
if let Ok(date) = DateTime::from_str(&date_string) {
article.date = Some(date);
break;
} else {
warn!("Parsing the date string '{}' failed", date_string);
}
}
}
}
}
fn extract_body(
context: &Context,
root: &mut Node,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
) -> Result<(), ScraperError> {
let mut found_something = false;
if let Some(config) = config {
for xpath_body in &config.xpath_body {
found_something = ArticleScraper::extract_body_single(context, root, xpath_body)?;
}
}
if !found_something {
for xpath_body in &global_config.xpath_body {
found_something = ArticleScraper::extract_body_single(context, root, xpath_body)?;
}
}
if !found_something {
log::error!("no body found");
return Err(ScraperErrorKind::Scrape.into());
}
Ok(())
}
fn extract_body_single(
context: &Context,
root: &mut Node,
xpath: &str,
) -> Result<bool, ScraperError> {
let mut found_something = false;
{
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.get_property("style").is_some() && node.remove_property("style").is_err() {
return Err(ScraperErrorKind::Xml.into());
}
node.unlink();
if root.add_child(&mut node).is_ok() {
found_something = true;
} else {
error!("Failed to add body to prepared document");
return Err(ScraperErrorKind::Xml.into());
}
}
}
Ok(found_something)
}
fn check_for_next_page(
&self,
context: &Context,
config: Option<&ConfigEntry>,
global_config: &ConfigEntry,
) -> Option<url::Url> {
if let Some(config) = config {
if let Some(next_page_xpath) = config.next_page_link.as_deref() {
if let Ok(next_page_string) =
ArticleScraper::get_attribute(context, next_page_xpath, "href")
{
if let Ok(next_page_url) = url::Url::parse(&next_page_string) {
return Some(next_page_url);
}
}
}
} else if let Some(next_page_xpath) = global_config.next_page_link.as_deref() {
if let Ok(next_page_string) =
ArticleScraper::get_attribute(context, next_page_xpath, "href")
{
if let Ok(next_page_url) = url::Url::parse(&next_page_string) {
return Some(next_page_url);
}
}
}
// last page reached
None
}
fn generate_head(root: &mut Node, document: &Document) -> Result<(), ScraperError> {
if let Ok(mut head_node) = Node::new("head", None, document) {
if let Ok(()) = root.add_prev_sibling(&mut head_node) {
if let Ok(mut meta) = head_node.new_child(None, "meta") {
if meta.set_property("charset", "utf-8").is_ok() {
return Ok(());
}
}
}
}
Err(ScraperErrorKind::Xml.into())
}
fn prevent_self_closing_tags(context: &Context) -> Result<(), ScraperError> {
// search document for empty tags and add a empty text node as child
// this prevents libxml from self closing non void elements such as iframe
let xpath = "//*[not(node())]";
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
for mut node in node_vec {
if node.get_name() == "meta" {
continue;
}
let _ = node.add_text_child(None, "empty", "");
// if let Err(error) = self
// .image_downloader
// .download_images_from_context(&context, client)
// .await
// {
// log::error!("Downloading images failed: '{}'", error);
// }
}
Ok(())
unimplemented!()
}
}

34
src/util.rs
View file

@ -6,9 +6,9 @@ use reqwest::{
};
use tokio::fs::DirEntry;
use crate::{
use crate::full_text_parser::{
config::ConfigEntry,
error::{ScraperError, ScraperErrorKind},
error::{FullTextParserError, FullTextParserErrorKind},
};
pub struct Util;
@ -49,28 +49,28 @@ impl Util {
pub fn generate_headers(
site_specific_rule: Option<&ConfigEntry>,
global_rule: &ConfigEntry,
) -> Result<HeaderMap, ScraperError> {
) -> Result<HeaderMap, FullTextParserError> {
let mut headers = HeaderMap::new();
if let Some(config) = site_specific_rule {
for header in &config.header {
let name = HeaderName::from_bytes(header.name.as_bytes())
.context(ScraperErrorKind::Config)?;
.context(FullTextParserErrorKind::Config)?;
let value = header
.value
.parse::<HeaderValue>()
.context(ScraperErrorKind::Config)?;
.context(FullTextParserErrorKind::Config)?;
headers.insert(name, value);
}
}
for header in &global_rule.header {
let name =
HeaderName::from_bytes(header.name.as_bytes()).context(ScraperErrorKind::Config)?;
HeaderName::from_bytes(header.name.as_bytes()).context(FullTextParserErrorKind::Config)?;
let value = header
.value
.parse::<HeaderValue>()
.context(ScraperErrorKind::Config)?;
.context(FullTextParserErrorKind::Config)?;
headers.insert(name, value);
}
@ -102,10 +102,10 @@ impl Util {
xpath_ctx: &Context,
xpath: &str,
thorw_if_empty: bool,
) -> Result<Vec<Node>, ScraperError> {
) -> Result<Vec<Node>, FullTextParserError> {
let res = xpath_ctx.evaluate(xpath).map_err(|()| {
log::debug!("Evaluation of xpath '{}' yielded no results", xpath);
ScraperErrorKind::Xml
FullTextParserErrorKind::Xml
})?;
let node_vec = res.get_nodes_as_vec();
@ -113,14 +113,14 @@ impl Util {
if node_vec.is_empty() {
log::debug!("Evaluation of xpath '{}' yielded no results", xpath);
if thorw_if_empty {
return Err(ScraperErrorKind::Xml.into());
return Err(FullTextParserErrorKind::Xml.into());
}
}
Ok(node_vec)
}
pub fn check_content_type(response: &Response) -> Result<bool, ScraperError> {
pub fn check_content_type(response: &Response) -> Result<bool, FullTextParserError> {
if response.status().is_success() {
if let Some(content_type) = response.headers().get(reqwest::header::CONTENT_TYPE) {
if let Ok(content_type) = content_type.to_str() {
@ -135,7 +135,7 @@ impl Util {
}
log::error!("Failed to determine content type");
Err(ScraperErrorKind::Http.into())
Err(FullTextParserErrorKind::Http.into())
}
pub fn check_redirect(response: &Response, original_url: &url::Url) -> Option<url::Url> {
@ -149,16 +149,16 @@ impl Util {
None
}
pub fn extract_value(context: &Context, xpath: &str) -> Result<String, ScraperError> {
pub fn extract_value(context: &Context, xpath: &str) -> Result<String, FullTextParserError> {
let node_vec = Util::evaluate_xpath(context, xpath, false)?;
if let Some(val) = node_vec.get(0) {
return Ok(val.get_content());
}
Err(ScraperErrorKind::Xml.into())
Err(FullTextParserErrorKind::Xml.into())
}
pub fn extract_value_merge(context: &Context, xpath: &str) -> Result<String, ScraperError> {
pub fn extract_value_merge(context: &Context, xpath: &str) -> Result<String, FullTextParserError> {
let node_vec = Util::evaluate_xpath(context, xpath, true)?;
let mut val = String::new();
for node in node_vec {
@ -174,7 +174,7 @@ impl Util {
Ok(val.trim().to_string())
}
pub fn strip_node(context: &Context, xpath: &str) -> Result<(), ScraperError> {
pub fn strip_node(context: &Context, xpath: &str) -> Result<(), FullTextParserError> {
let mut ancestor = xpath.to_string();
if ancestor.starts_with("//") {
ancestor = ancestor.chars().skip(2).collect();
@ -188,7 +188,7 @@ impl Util {
Ok(())
}
pub fn strip_id_or_class(context: &Context, id_or_class: &str) -> Result<(), ScraperError> {
pub fn strip_id_or_class(context: &Context, id_or_class: &str) -> Result<(), FullTextParserError> {
let xpath = &format!(
"//*[contains(@class, '{}') or contains(@id, '{}')]",
id_or_class, id_or_class

32
src/youtube.rs
View file

@ -1,32 +0,0 @@
use crate::article::Article;
use crate::ArticleScraper;
pub struct Youtube;
impl Youtube {
pub fn handle(url: &url::Url) -> Option<Article> {
let host_name = match ArticleScraper::get_host_name(url) {
Ok(host_name) => host_name,
Err(_) => return None,
};
if &host_name == "youtube.com" {
let regex =
regex::Regex::new(r#"youtube\.com/watch\?v=(.*)"#).expect("Failed to parse regex");
if let Some(captures) = regex.captures(url.as_str()) {
if let Some(video_id) = captures.get(1) {
let html = format!("<iframe width=\"650\" height=\"350\" frameborder=\"0\" src=\"https://www.youtube-nocookie.com/embed/{}\" allowfullscreen></iframe>", video_id.as_str());
return Some(Article {
title: None,
date: None,
author: None,
url: url.clone(),
html: Some(html),
});
}
}
}
None
}
}