misc-mirrors/article_scraper
1
0
Fork 0
mirror of https://gitlab.com/news-flash/article_scraper.git synced 2025-07-08 08:30:00 +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

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"))
}
}