2 Commits

Author SHA1 Message Date
pyro 8228c51319 added ascii art 2026-05-18 13:43:43 -05:00
pyro 940673b381 made the healdess chromium window wait until the body and javascript
loaded before taking the screenshot.
2026-04-30 17:59:02 -05:00
2 changed files with 1753 additions and 31 deletions
Generated
+1625
View File
File diff suppressed because it is too large Load Diff
+113 -16
View File
@@ -2,7 +2,7 @@ use anyhow::{Context, Result};
use clap::Parser; use clap::Parser;
use headless_chrome::{Browser, LaunchOptions}; use headless_chrome::{Browser, LaunchOptions};
use rayon::prelude::*; use rayon::prelude::*;
use std::fs::{self, File}; use std::fs::{self, File, read_to_string};
use std::io::{BufRead, BufReader}; use std::io::{BufRead, BufReader};
use std::path::PathBuf; use std::path::PathBuf;
@@ -23,17 +23,103 @@ struct Args {
} }
fn main() -> Result<()> { fn main() -> Result<()> {
let args = Args::parse(); print!(
"
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣠⣴⣿⣿⣿⣿⣿⣆⠀⠀⠀⠀⣠⣿⣿⣿⣿⣿⣶⣦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣶⣶⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣼⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣧⣀⠀⠀⠀⣀⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣀⣴⣾⣿⣷⣶⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣶⣾⣿⣿⣦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣠⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣦⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠠⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣧⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⢀⣀⣀⣠⣴⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣦⡀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⢀⣴⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣦⡀⠀⠀⠀⠀
⠀⠀⢀⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣄⠀⠀⠀
⠀⠀⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡆⠀⠀
⠀⠀⠘⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠿⠟⠛⠛⠛⠛⠻⠿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡟⠀⠀
⠀⠀⠀⢸⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠟⠉⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠙⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡏⠀⠀⠀
⠀⠀⠀⣼⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡟⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣷⡀⠀⠀
⣠⣶⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡟⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠘⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣷⣦⣄
⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿
⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢘⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿
⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣷⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿
⠘⠻⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠿
⠀⠀⠀⢹⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣧⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠏⠁⠀
⠀⠀⠀⠀⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣷⣄⡀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣀⣴⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠂⠀⠀
⠀⠀⢀⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣶⣦⣤⣀⣀⣀⣠⣤⣴⣾⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣦⠀⠀
⠀⠀⠈⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡇⠀
⠀⠀⠀⠈⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡟⠀⠀
⠀⠀⠀⠀⠀⠻⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠏⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠈⠙⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠿⠋⠛⠋⠁⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠘⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡟⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠸⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠹⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠻⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠟⠉⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠉⠉⠀⠀⠉⠻⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠟⠁⠀⠀⠈⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢸⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡏⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠻⢿⣿⣿⣿⣿⣿⣿⡋⠀⠀⠀⢈⡻⣿⣿⣿⣿⠿⢿⡋⠀
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
|||||||||||||||||
______ __ __ _______ _______ _ _ ___ _______ __ _ _______ _______ _______
| _ | | | | || || || | _ | || | | || | | || || || |
| | || | | | || _____||_ _|| || || || | |_ _|| |_| || ___|| _____|| _____|
| |_||_ | |_| || |_____ | | | || | | | | || |___ | |_____ | |_____
| __ || ||_____ | | | | || | | | | _ || ___||_____ ||_____ |
| | | || | _____| | | | | _ || | | | | | | || |___ _____| | _____| |
|___| |_||_______||_______| |___| |__| |__||___| |___| |_| |__||_______||_______||_______|
"
);
let args = Args::parse();
if !args.output.exists() { if !args.output.exists() {
fs::create_dir_all(&args.output).context("Failed to create output directory")?; fs::create_dir_all(&args.output).context("Failed to create output directory")?;
} }
let mut log_path = args.output.clone();
log_path.push("log.txt");
let mut resume = None;
if log_path.exists() {
let log = read_to_string(log_path)?;
if let Some(last_line) = log.lines().last() {
resume = Some(last_line.to_string())
}
};
rayon::ThreadPoolBuilder::new() rayon::ThreadPoolBuilder::new()
.num_threads(args.workers) .num_threads(args.workers)
.build_global() .build_global()
.unwrap(); .unwrap();
let mut urls = Vec::new();
if resume.is_some() {
let url_resume = resume.clone().unwrap();
let url_text = read_to_string(&args.input)?;
let mut resume_met = false;
for line in url_text.lines() {
if !resume_met {
if line.trim().to_string() == url_resume {
resume_met = true;
}
} else {
urls.push(line.trim().to_string());
}
}
}
if resume.is_none() {
let file = File::open(&args.input).context("Failed to open input file")?; let file = File::open(&args.input).context("Failed to open input file")?;
let urls: Vec<String> = BufReader::new(file) let urls: Vec<String> = BufReader::new(file)
.lines() .lines()
@@ -41,40 +127,46 @@ fn main() -> Result<()> {
.map(|s| s.trim().to_string()) .map(|s| s.trim().to_string())
.filter(|s| !s.is_empty()) .filter(|s| !s.is_empty())
.collect(); .collect();
println!( println!(
"Starting capture of {} URLs using {} workers...", "Starting capture of {} URLs using {} workers...",
urls.len(), urls.len(),
args.workers args.workers
); );
urls.par_iter().enumerate().for_each(|(_index, url)| {
urls.par_iter().enumerate().for_each(|(index, url)| { if let Err(e) = capture_screenshot(url, &args) {
if let Err(e) = capture_screenshot(url, index, &args) {
eprintln!("Failed to process {}: {:?}", url, e); eprintln!("Failed to process {}: {:?}", url, e);
} }
}); });
println!("All tasks completed."); println!("All tasks completed.");
} else {
urls.par_iter().enumerate().for_each(|(_index, url)| {
if let Err(e) = capture_screenshot(url, &args) {
eprintln!("failed to process {}: {:?}", url, e);
}
});
}
Ok(()) Ok(())
} }
fn capture_screenshot(url: &str, args: &Args) -> Result<()> {
fn capture_screenshot(url: &str, index: usize, args: &Args) -> Result<()> {
let mut launch_options = LaunchOptions::default(); let mut launch_options = LaunchOptions::default();
let file_name = url.split("://").collect::<Vec<&str>>()[1].to_string();
if let Some(ref proxy_url) = args.proxy { if let Some(ref proxy_url) = args.proxy {
let proxy_arg = format!("--proxy-server={}", proxy_url); let proxy_arg = format!("--proxy-server={}", proxy_url);
launch_options launch_options
.args .args
.push(std::ffi::OsStr::new(Box::leak(proxy_arg.into_boxed_str()))); .push(std::ffi::OsStr::new(Box::leak(proxy_arg.into_boxed_str())));
} }
let browser = Browser::new(launch_options).context("Browser launch failed")?; let browser = Browser::new(launch_options).context("Browser launch failed")?;
let tab = browser.new_tab().context("Failed to open tab")?; let tab = browser.new_tab().context("Failed to open tab")?;
tab.navigate_to(url).context("Navigation failed")?; tab.navigate_to(url).context("Navigation failed")?;
tab.wait_until_navigated() tab.wait_until_navigated()
.context("Waiting for load failed")?; .context("Waiting for load failed")?;
tab.wait_for_element("body")?;
tab.evaluate("document.readyState === 'complete'", false)?;
let mut file_path = args.output.clone(); let mut file_path = args.output.clone();
file_path.push(format!("screenshot_{:03}.png", index + 1)); let mut log_path = args.output.clone();
log_path.push("log.txt");
file_path.push(format!("{}.png", file_name));
let png_data = tab let png_data = tab
.capture_screenshot( .capture_screenshot(
headless_chrome::protocol::cdp::Page::CaptureScreenshotFormatOption::Png, headless_chrome::protocol::cdp::Page::CaptureScreenshotFormatOption::Png,
@@ -84,8 +176,13 @@ fn capture_screenshot(url: &str, index: usize, args: &Args) -> Result<()> {
) )
.context("Screenshot capture failed")?; .context("Screenshot capture failed")?;
fs::write(&file_path, png_data).context("Failed to write file")?; fs::write(&file_path, png_data)
.context("Failed to write file")
.context(format!(
"failed to write png file! {}",
&file_path.display()
))?;
println!("Successfully captured: {}", url); println!("Successfully captured: {}", url);
fs::write(&log_path, format!("{}\n", url).as_bytes()).context("failed to write log file!")?;
Ok(()) Ok(())
} }