webarc/src/archive.rs
2024-12-29 16:51:34 +01:00

123 lines
3.3 KiB
Rust

use std::{fmt::format, fs::read_to_string, path::{Path, PathBuf}};
pub struct WebsiteArchive {
pub dir: PathBuf
}
pub struct Domain {
pub name: String,
dir: PathBuf
}
impl Domain {
pub fn new(name: &str, dir: PathBuf) -> Self {
std::fs::create_dir_all(&dir).unwrap();
Self { name: name.to_string(), dir }
}
pub fn path(&self, path: &str) -> Document {
Document::new(&self.name, path, self.dir.parent().unwrap().to_path_buf())
}
}
pub struct Document {
pub domain: String,
pub path: String,
base_dir: PathBuf
}
impl Document {
pub fn new(domain: &str, path: &str, base_dir: PathBuf) -> Self {
Self { domain: domain.to_string(), path: path.to_string(), base_dir }
}
pub fn render_local(&self, version: Option<String>) -> String {
let mut file_path = self.base_dir.join(&self.domain);
for p in self.path.split('/') {
file_path = file_path.join(p);
}
let latest_version = if let Some(version) = version {
format!("index_{version}.html")
} else {
let versions = Self::versions(&file_path);
versions.first().cloned().unwrap()
};
file_path = file_path.join(latest_version);
// TODO : Replace links with local ones
return std::fs::read_to_string(file_path).unwrap();
}
pub fn versions(path: &PathBuf) -> Vec<String> {
let mut version_list = Vec::new();
if let Ok(entries) = std::fs::read_dir(path) {
for entry in entries {
if let Ok(entry) = entry {
if let Some(file_name) = entry.file_name().to_str() {
version_list.push(file_name.to_string());
}
}
}
}
version_list
}
}
impl WebsiteArchive {
pub fn new(dir: &str) -> Self {
Self { dir: PathBuf::from(dir) }
}
pub fn get_domain(&self, domain: &str) -> Domain {
Domain::new(domain, self.dir.join(domain))
}
/// Archive a URL
pub fn archive_url(&self, url: &str) {
let parsed_url = url::Url::parse(url).unwrap();
let domain = parsed_url.domain().unwrap().trim_start_matches("www");
let path = parsed_url.path();
let mut folder_name = self.dir.join(&domain);
for paths in path.split('/') {
if !paths.is_empty() {
folder_name = folder_name.join(paths);
}
}
std::fs::create_dir_all(&folder_name).unwrap();
let timestamp = chrono::Utc::now().format("%Y-%m-%d").to_string();
let filename = folder_name.join(&format!("index_{timestamp}.html"));
run_command(&vec![
"monolith",
"-I",
"-o",
filename.to_str().unwrap(),
&format!("https://{}/{}", domain, path)
]);
}
}
// full text search
// add new sites?
// transparent auto page downloading
// redownload after threshold
fn run_command(cmd: &[&str]) {
let mut cmd_setup = std::process::Command::new(cmd[0].clone());
let cmd_setup = cmd_setup.args(cmd.into_iter().skip(1).collect::<Vec<_>>());
let child = cmd_setup.spawn().unwrap();
let status = child.wait_with_output().unwrap();
assert!(status.status.success());
}