70 lines
1.7 KiB
Rust
70 lines
1.7 KiB
Rust
use reqwest::blocking::Client;
|
|
use scraper::{Html, Selector};
|
|
use serde::{Deserialize, Serialize};
|
|
use std::fs;
|
|
|
|
const BASE: &str = "https://letterboxd.com";
|
|
const LIST: &str = "https://letterboxd.com/mimuuu/list/2026/by/reverse/";
|
|
|
|
#[derive(Deserialize)]
|
|
struct FilmJson {
|
|
name: String,
|
|
released: Released,
|
|
directors: Vec<Person>,
|
|
}
|
|
|
|
#[derive(Deserialize)]
|
|
struct Released {
|
|
year: u16,
|
|
}
|
|
|
|
#[derive(Deserialize)]
|
|
struct Person {
|
|
name: String,
|
|
}
|
|
|
|
#[derive(Serialize)]
|
|
struct Film {
|
|
title: String,
|
|
year: u16,
|
|
director: String,
|
|
url: String,
|
|
}
|
|
|
|
fn main() -> Result<(), Box<dyn std::error::Error>> {
|
|
let client = Client::builder().user_agent("Mozilla/5.0").build()?;
|
|
|
|
// One request: list page (only to discover the 5 film endpoints)
|
|
let html = client.get(LIST).send()?.text()?;
|
|
let doc = Html::parse_document(&html);
|
|
let selector = Selector::parse("li.posteritem div.react-component")?;
|
|
|
|
let mut films = Vec::new();
|
|
|
|
for node in doc.select(&selector).take(5) {
|
|
let details = node.value().attr("data-details-endpoint").unwrap();
|
|
|
|
let slug = node.value().attr("data-item-link").unwrap();
|
|
|
|
// One request per film: JSON metadata
|
|
let film: FilmJson = client.get(format!("{BASE}{details}")).send()?.json()?;
|
|
|
|
films.push(Film {
|
|
title: film.name,
|
|
year: film.released.year,
|
|
director: film
|
|
.directors
|
|
.first()
|
|
.map(|d| d.name.clone())
|
|
.unwrap_or_default(),
|
|
url: format!("{BASE}{slug}"),
|
|
});
|
|
}
|
|
|
|
fs::create_dir_all("data")?;
|
|
fs::write("data/films.json", serde_json::to_string_pretty(&films)?)?;
|
|
|
|
println!("Saved {} films", films.len());
|
|
Ok(())
|
|
}
|