#![doc(html_root_url = "http://wooya.me/tldextract-rs/tldextract/index.html")]
#![deny(missing_docs)]

//! # tldextract
//!
//! tldextract accurately separates the gTLD or ccTLD (generic or country code top-level domain)
//! from the registered domain and subdomains of a URL. For example,
//! say you want just the 'google' part of 'http://www.google.com'.
//!
//! Everybody gets this wrong. Splitting on the '.' and taking the last 2 elements goes a long way
//! only if you're thinking of simple e.g. .com domains. Think parsing http://forums.bbc.co.uk
//! for example: the naive splitting method above will give you 'co' as the domain and 'uk' as the TLD,
//! instead of 'bbc' and 'co.uk' respectively.
//!
//! tldextract on the other hand knows what all gTLDs and ccTLDs look like
//! by looking up the currently living ones according to the Public Suffix List.
//! So, given a URL, it knows its subdomain from its domain, and its domain from its country code.
//!
//! Thanks to [john-kurkowski](https://github.com/john-kurkowski),
//! this project is mainly inspired (Ok, stolen) by his [work](https://github.com/john-kurkowski/tldextract) in python

mod cache;
#[allow(missing_docs)]
pub mod errors;

pub use errors::{Result, TldExtractError};
use idna::punycode;
use std::collections::HashSet;
use url::{Host, Url};

/// The option for `TldExtractor`.
///
/// # Examples
///
/// ```
/// use tldextract::{TldExtractor, TldOption, TldResult};
///
/// let ext: TldExtractor = TldOption::default().cache_path(".tld_cache").build();
///
/// assert_eq!(ext.extract("https://m.facebook.com").unwrap(), TldResult::new("m", "facebook", "com"));
/// ```
#[derive(Default, Debug)]
pub struct TldOption {
    /// The path to file for storing tld cache
    cache_path: Option<String>,
    /// Whether to include private domains
    private_domains: bool,
    /// Should tldextract update local cache file if
    /// the cache is fetched from remote or from snapshot
    update_local: bool,
    /// When cannot finding valid suffix in PSL, should we naively
    /// treat the last piece of URL as the suffix and
    /// the last but one piece as the domain?
    naive_mode: bool,
}

impl TldOption {
    /// Set cache_path
    pub fn cache_path(mut self, path: &str) -> Self {
        self.cache_path = Some(path.into());
        self
    }

    /// Set private_domains
    pub fn private_domains(mut self, b: bool) -> Self {
        self.private_domains = b;
        self
    }

    /// Set update_local
    pub fn update_local(mut self, b: bool) -> Self {
        self.update_local = b;
        self
    }

    /// Set naive_mode
    pub fn naive_mode(mut self, b: bool) -> Self {
        self.naive_mode = b;
        self
    }

    /// Build TldExtractor
    pub fn build(self) -> TldExtractor {
        TldExtractor::new(self)
    }
}

/// The tld extractor, see TldOption for more docs.
#[derive(Debug)]
pub struct TldExtractor {
    tld_cache: HashSet<String>,
    naive_mode: bool,
}

impl TldExtractor {
    /// Constructs a new `TldExtractor`.
    ///
    /// see TldOption for more docs.
    pub fn new(option: TldOption) -> TldExtractor {
        let cache_path = option.cache_path.as_ref().map(|s| &s[..]);
        let tld_cache = cache::get_tld_cache(cache_path, option.private_domains);
        if option.update_local {
            let _ = cache::set_tld_cache(cache_path, &tld_cache);
        }
        TldExtractor {
            tld_cache,
            naive_mode: option.naive_mode,
        }
    }

    /// Extract (subdomain, domain, domain suffix) tuple from a given url or bare domain
    pub fn extract(&self, url: &str) -> Result<TldResult> {
        self._extract(url, None)
    }

    /// Extract (subdomain, domain, domain suffix) tuple from a given url or bare domain
    /// but override the universal naive_mode in TldExtractor.
    pub fn extract_naive(&self, url: &str) -> Result<TldResult> {
        self._extract(url, true)
    }

    fn _extract<O: Into<Option<bool>>>(&self, url: &str, naive: O) -> Result<TldResult> {
        if url.contains(':') {
            // : should only be in a URL, so parse as a URL

            let u = Url::parse(url)?;

            let host = u
                .host()
                .ok_or_else(|| TldExtractError::NoHostError(url.into()))?;

            match host {
                Host::Domain(host) => {
                    Ok(self.extract_triple(host, naive.into().unwrap_or(self.naive_mode)))
                }
                Host::Ipv4(ip) => Ok(TldResult {
                    domain: Some(ip.to_string()),
                    ..Default::default()
                }),
                Host::Ipv6(ip) => Ok(TldResult {
                    domain: Some(ip.to_string()),
                    ..Default::default()
                }),
            }
        } else {
            // no scheme, so assume we've just got a domain/subdomain, skip URL parsing
            Ok(self.extract_triple(url, naive.into().unwrap_or(self.naive_mode)))
        }
    }

    fn extract_triple(&self, host: &str, naive_mode: bool) -> TldResult {
        let segs: Vec<_> = host
            .split('.')
            .filter(|&s| !s.is_empty())
            .map(|seg| {
                if seg.starts_with("xn--") {
                    punycode::decode_to_string(seg.trim_start_matches("xn--")).unwrap_or(seg.into())
                } else {
                    seg.into()
                }
            })
            .collect();

        let mut suffix = None;
        let mut subdomain = None;
        let mut domain = None;
        for i in 0..segs.len() {
            let piece = segs[i..].join(".");
            let exception_piece = "!".to_string() + &piece;
            let wildcard_piece = "*.".to_string() + &segs[i + 1..].join(".");

            if self.tld_cache.get(&exception_piece).is_some() {
                continue;
            }

            if self
                .tld_cache
                .get(&piece)
                .or_else(|| self.tld_cache.get(&wildcard_piece))
                .is_some()
            {
                suffix = Some(piece);
                if i != 0 {
                    domain = Some(segs[i - 1].to_string());
                    subdomain = if segs[0..i - 1].is_empty() {
                        None
                    } else {
                        Some(segs[0..i - 1].join("."))
                    };
                }

                break;
            }
        }

        if let (None, None, None) = (subdomain.as_ref(), domain.as_ref(), suffix.as_ref()) {
            let mut iter = segs.into_iter().rev();
            if naive_mode {
                suffix = iter.next().map(|s| s.to_string());
            }
            domain = iter.next().map(|s| s.to_string());
            let maybe_subdomain = iter.collect::<Vec<_>>().join(".");
            subdomain = if maybe_subdomain.is_empty() {
                None
            } else {
                Some(maybe_subdomain)
            }
        }

        TldResult {
            suffix,
            subdomain,
            domain,
        }
    }
}

/// The Tld Result Type
///
/// E.g. "https://www.google.com" will be represent into
///
/// ```
/// use tldextract::TldResult;
///
/// TldResult { domain: Some("google".to_string()), subdomain: Some("www".to_string()), suffix: Some("com".to_string())};
/// ```
#[derive(Debug, Default, PartialEq, Eq)]
pub struct TldResult {
    /// The "google" part of "www.google.com"
    pub domain: Option<String>,
    /// The "www" part of "www.google.com"
    pub subdomain: Option<String>,
    /// The "com" part of "www.google.com"
    pub suffix: Option<String>,
}

impl TldResult {
    /// Create a new TldResult
    ///
    /// # Examples
    ///
    /// ```
    /// use tldextract::TldResult;
    /// assert_eq!(TldResult::new("www", "google", "com"),
    ///   TldResult {
    ///     domain: Some("google".to_string()),
    ///     subdomain: Some("www".to_string()),
    ///     suffix: Some("com".to_string())
    ///   });
    /// ```
    pub fn new<'a, O, P, Q>(subdomain: O, domain: P, suffix: Q) -> TldResult
    where
        O: Into<Option<&'a str>>,
        P: Into<Option<&'a str>>,
        Q: Into<Option<&'a str>>,
    {
        TldResult {
            domain: domain.into().map(|s| s.into()),
            subdomain: subdomain.into().map(|s| s.into()),
            suffix: suffix.into().map(|s| s.into()),
        }
    }
}