This commit is contained in:
2026-07-01 14:13:32 -05:00
commit d69955ea39
10 changed files with 2145 additions and 0 deletions
+13
View File
@@ -0,0 +1,13 @@
[package]
name = "mmaschedule_scraper"
version = "0.1.0"
edition = "2024"
[dependencies]
anyhow = "1.0.103"
scraper = "0.27.0"
serde = { version = "1.0.228", features = ["derive"] }
serde_json = "1.0.150"
time_rs = { package = "time", version = "0.3.51", features = ["formatting", "parsing", "serde"] }
wreq = "5.3.0"
wreq-util = "2.2.6"
+97
View File
@@ -0,0 +1,97 @@
use anyhow::Result;
use scraper::{Html, Selector};
use wreq::Client;
use wreq_util::Emulation;
mod one;
use crate::one::ONE;
pub struct Event {
pub url: String,
pub slug: String,
pub name: String,
pub location: String,
pub organization: String,
pub image: String,
pub date: i64,
pub fights: Vec<Fight>,
}
pub struct Fight {
pub weight: String,
pub fighter_a: Fighter,
pub fighter_b: Fighter,
}
pub struct Fighter {
pub name: String,
pub image: String,
pub country: String,
pub link: String,
}
pub struct Tapology {
pub name: String,
pub url: String,
}
trait ClientScraper {
async fn get(&self, url: &str) -> Result<Html>;
}
trait EventScraper {
async fn scrape<C: ClientScraper>(client: &C) -> Result<Vec<Event>>;
}
enum UFC {}
impl EventScraper for UFC {
async fn scrape<C: ClientScraper>(client: &C) -> Result<Vec<Event>> {
todo!()
}
}
pub async fn hello() -> wreq::Result<()> {
// Build a client
let client = Client::builder().emulation(Emulation::Firefox139).build()?;
// Use the API you're already familiar with
let resp = client.get("https://tls.peet.ws/api/all").send().await?;
println!("{}", resp.text().await?);
Ok(())
}
pub fn add(left: u64, right: u64) -> u64 {
left + right
}
fn yes() {
let html = r#"
<ul>
<li>Foo</li>
<li>Bar</li>
<li>Baz</li>
</ul>
"#;
let fragment = Html::parse_fragment(html);
let ul_selector = Selector::parse("ul").unwrap();
let li_selector = Selector::parse("li").unwrap();
let ul = fragment.select(&ul_selector).next().unwrap();
for element in ul.select(&li_selector) {
assert_eq!("li", element.value().name());
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn it_works() {
let result = add(2, 2);
assert_eq!(result, 4);
}
}
+87
View File
@@ -0,0 +1,87 @@
use anyhow::Result;
use scraper::{ElementRef, Selector, selectable::Selectable};
use serde::{Deserialize, Serialize};
use time_rs::OffsetDateTime;
use crate::{ClientScraper, Event, EventScraper, Fight, Fighter};
pub enum ONE {}
impl EventScraper for ONE {
async fn scrape<C: ClientScraper>(client: &C) -> Result<Vec<Event>> {
let page = client.get("https://www.onefc.com/events/").await?;
// should panic if selectors are invalid
let selector_event = Selector::parse("#upcoming-events-section .is-event").unwrap();
let selector_check = Selector::parse("a.smart-link").unwrap();
let selector_event_url = Selector::parse("a.is-image-zoom[href]").unwrap();
for event_fragment in page.select(&selector_event) {
do_thing(client, event_fragment).await;
if event_fragment
.select(&selector_check)
.next()
.map(|e| e.attr("href"))
.flatten()
.is_none()
{
continue;
}
let Some(url) = event_fragment
.select(&selector_event_url)
.next()
.map(|e| e.attr("href"))
.flatten()
else {
continue;
};
}
todo!()
}
}
#[derive(Deserialize)]
struct Location {
name: String,
}
#[derive(Deserialize)]
#[serde(rename_all = "camelCase")]
struct EventData {
name: String,
image: Vec<String>,
location: Location,
#[serde(with = "time_rs::serde::rfc3339")]
start_date: OffsetDateTime,
}
async fn do_thing<C: ClientScraper>(client: &C, elem: ElementRef<'_>) -> Option<Vec<Event>> {
let selector_check = Selector::parse("a.smart-link").unwrap();
let selector_event_url = Selector::parse("a.is-image-zoom[href]").unwrap();
let selector_data = Selector::parse("#site-main script[type=\"application/ld+json\"]").unwrap();
let mut events = Vec::new();
elem.select(&selector_check)
.next()
.map(|e| e.attr("href"))
.flatten()?;
let url = elem
.select(&selector_event_url)
.next()
.map(|e| e.attr("href"))
.flatten()?;
let page = client.get(url).await.ok()?;
let data_elem = page.select(&selector_data).next()?;
let data_str: String = data_elem.text().collect();
let data: EventData = serde_json::from_str(&data_str).ok()?;
data.start_date.unix_timestamp();
Some(events)
}