oolatoocs/src/utils.rs

use html_escape::decode_html_entities;
use megalodon::entities::status::Tag;
use regex::Regex;
use std::error::Error;

pub fn strip_everything(content: &str, tags: &Vec<Tag>) -> Result<String, Box<dyn Error>> {
    let mut res = strip_html_tags(&content.replace("</p><p>", "\n\n").replace("<br />", "\n"));

    strip_mastodon_tags(&mut res, tags).unwrap();

    res = res.trim_end_matches('\n').trim_end_matches(' ').to_string();
    res = decode_html_entities(&res).to_string();

    Ok(res)
}

fn strip_mastodon_tags(content: &mut String, tags: &Vec<Tag>) -> Result<(), Box<dyn Error>> {
    for tag in tags {
        let re = Regex::new(&format!("(?i)(#{} ?)", &tag.name))?;
        *content = re.replace(content, "").to_string();
    }

    Ok(())
}

fn strip_html_tags(input: &str) -> String {
    let mut data = String::new();
    let mut inside = false;

    for c in input.chars() {
        if c == '<' {
            inside = true;
            continue;
        }
        if c == '>' {
            inside = false;
            continue;
        }
        if !inside {
            data.push(c);
        }
    }

    data
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_strip_mastodon_tags() {
        let tags = vec![
            Tag {
                name: "putaclic".to_string(),
                url: "https://m.nintendojo.fr/tags/putaclic".to_string(),
            },
            Tag {
                name: "tamerelol".to_string(),
                url: "https://m.nintendojo.fr/tags/tamerelol".to_string(),
            },
            Tag {
                name: "JeFaisNawakEnCamelCase".to_string(),
                url: "https://m.nintendojo.fr/tags/jefaisnawakencamelcase".to_string(),
            },
        ];

        let mut content =
            "C’est super ça ! #putaclic #TAMERELOL #JeFaisNawakEnCamelCase".to_string();
        let sample = "C’est super ça ! ".to_string();

        strip_mastodon_tags(&mut content, &tags).unwrap();

        assert_eq!(content, sample);
    }

    #[test]
    fn test_strip_everything() {
        let content = "<p>Ce soir à 21h, c&#39;est le Dojobar ! Au programme ce soir, une rétrospective sur la série Mario &amp; Luigi.<br />Comme d&#39;hab, le Twitch sera ici : <a href=\"https://twitch.tv/nintendojofr\" target=\"_blank\" rel=\"nofollow noopener noreferrer\" translate=\"no\"><span class=\"invisible\">https://</span><span class=\"\">twitch.tv/nintendojofr</span><span class=\"invisible\"></span></a><br />Ou juste l&#39;audio là : <a href=\"https://nintendojo.fr/dojobar\" target=\"_blank\" rel=\"nofollow noopener noreferrer\" translate=\"no\"><span class=\"invisible\">https://</span><span class=\"\">nintendojo.fr/dojobar</span><span class=\"invisible\"></span></a><br />A toute !</p>";
        let expected_result = "Ce soir à 21h, c'est le Dojobar ! Au programme ce soir, une rétrospective sur la série Mario & Luigi.\nComme d'hab, le Twitch sera ici : https://twitch.tv/nintendojofr\nOu juste l'audio là : https://nintendojo.fr/dojobar\nA toute !".to_string();
        let result = strip_everything(content, &vec![]).unwrap();

        assert_eq!(result, expected_result);
    }
}