whoareyou/src/probe/vem_ringde.rs

use std::str;

// use log::debug;
use scraper::{Html, Selector};

use crate::context::Context;
use crate::probe::{Entry, Probe};

fn from_html(document: &str) -> Result<Entry, ()> {
    let html = Html::parse_document(document);

    let mut messages = Vec::new();
    let history = Vec::new();
    let comments = Vec::new();

    let selector = Selector::parse("#toporganisations li").unwrap();

    for element in html.select(&selector) {
        let message = element
            .text()
            .map(str::trim)
            .filter(|s| !s.is_empty())
            .collect::<Vec<_>>()
            .join(" ");

        messages.push(message);
    }

    Ok(Entry {
        messages,
        history,
        comments,
    })
}

pub struct VemRingde;

impl Probe for VemRingde {
    fn uri(&self, number: &str) -> String {
        format!("http://vemringde.se/?q={}", number)
    }

    fn search(&mut self, ctx: &mut Context, number: &str) -> Result<(), ()> {
        let body = if let Some(cache) = ctx.cache_get("vem_ringde", &number) {
            String::from_utf8(cache.data).unwrap()
        } else {
            let body = reqwest::get(&self.uri(number)).unwrap().text().unwrap();

            ctx.cache_set("vem_ringde", &number, body.as_bytes())
                .expect("wut?! why not?!");

            body
        };

        match from_html(&body) {
            Ok(entry) => {
                println!("vemringde.se:");
                print!("{}", entry);

                Ok(())
            }
            Err(_) => Err(()),
        }
    }
}

#[cfg(test)]
mod tests {
    use pretty_assertions::assert_eq;

    use super::*;

    #[test]
    fn test_0104754350() {
        let document = include_str!("../../fixtures/vemringde/0104754350.html");

        let expected = Entry {
            messages: vec!["Folksam (5 samtal)".to_string()],
            history: vec![],
            comments: vec![],
        };

        assert_eq!(from_html(&document), Ok(expected));
    }

    #[test]
    fn test_0313908905() {
        let document = include_str!("../../fixtures/vemringde/0313908905.html");

        let expected = Entry {
            messages: vec![],
            history: vec![],
            comments: vec![],
        };

        assert_eq!(from_html(&document), Ok(expected));
    }

    #[test]
    fn test_0702269893() {
        let document = include_str!("../../fixtures/vemringde/0702269893.html");

        let expected = Entry {
            messages: vec![],
            history: vec![],
            comments: vec![],
        };

        assert_eq!(from_html(&document), Ok(expected));
    }

    #[test]
    fn test_0726443387() {
        let document = include_str!("../../fixtures/vemringde/0726443387.html");

        let expected = Entry {
            messages: vec![],
            history: vec![],
            comments: vec![],
        };

        assert_eq!(from_html(&document), Ok(expected));
    }

    #[test]
    fn test_0751793426() {
        let document = include_str!("../../fixtures/vemringde/0751793426.html");

        let expected = Entry {
            messages: vec![],
            history: vec![],
            comments: vec![],
        };

        assert_eq!(from_html(&document), Ok(expected));
    }

    #[test]
    fn test_0751793483() {
        let document = include_str!("../../fixtures/vemringde/0751793483.html");

        let expected = Entry {
            messages: vec![],
            history: vec![],
            comments: vec![],
        };

        assert_eq!(from_html(&document), Ok(expected));
    }

    #[test]
    fn test_0751793499() {
        let document = include_str!("../../fixtures/vemringde/0751793499.html");

        let expected = Entry {
            messages: vec![],
            history: vec![],
            comments: vec![],
        };

        assert_eq!(from_html(&document), Ok(expected));
    }
}