use std::fmt;
use std::fs;

use crate::config::*;

const fn build_lookup_table() -> [usize; 128] {
    let mut result = [0; 128];
    let mut i = 0;

    while i < NUM_KEYS {
        result[KEY_CHARS[i] as usize] = i;
        i += 1;
    }
    return result;
}
const CHAR_TO_INDEX: [usize; 128] = build_lookup_table();

const NUM_BIGRAMS: usize = NUM_KEYS * NUM_KEYS;
pub struct Corpus {
    bigram_count: [u32; NUM_BIGRAMS],
    character_count: [u32; 128],
    total_count: u32,
    index_threshold: u32,
}

fn pair_to_index(x: u8, y: u8) -> usize {
    let xi = CHAR_TO_INDEX[x as usize];
    let yi = CHAR_TO_INDEX[y as usize];
    xi + NUM_KEYS * yi
}

impl Corpus {
    pub fn get_character_count(&self, c: u8) -> u32 {
        self.character_count[c as usize]
    }

    pub fn get_bigram_count(&self, x: u8, y: u8) -> u32 {
        self.bigram_count[pair_to_index(x, y)]
    }

    pub fn get_index_threshold(&self) -> u32 {
        self.index_threshold
    }

    pub fn get_total_count(&self) -> u32 {
        self.total_count
    }

    pub fn load(path: &str) -> Result<Corpus, std::io::Error> {
        let contents = fs::read_to_string(path)?;

        let mut bigram_count = [0; NUM_BIGRAMS];
        let mut character_count = [0; 128];
        let mut total_count = 0;

        let mut last_char: Option<u8> = None;
        for c in contents.chars() {
            if let Some(c) = canonicalise(c) {
                if let Some(lc) = last_char {
                    // We don't count these anyway
                    if lc != c {
                        bigram_count[pair_to_index(c, lc)] += 1;
                        bigram_count[pair_to_index(lc, c)] += 1;
                    }
                }
                last_char = Some(c);
                character_count[c as usize] += 1;
                total_count += 1;
            } else {
                last_char = None;
            }
        }

        let index_threshold = (total_count as f32 * DESIRED_INDEX_USAGE_PERCENT) as u32;

        return Ok(Corpus {
            bigram_count,
            character_count,
            total_count,
            index_threshold,
        });
    }
}

fn dump_bigrams(corpus: &Corpus) -> Vec<(String, u32)> {
    let mut result: Vec<(String, u32)> = Vec::new();

    for (i, &x) in KEY_CHARS.iter().enumerate() {
        for &y in &KEY_CHARS[i..] {
            let mut pair = String::from(x);
            pair.push(y);
            result.push((pair, corpus.bigram_count[pair_to_index(x as u8, y as u8)]));
        }
    }

    result.sort_by(|(_, c1), (_, c2)| c1.cmp(c2).reverse());

    return result;
}

impl fmt::Display for Corpus {
    fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
        formatter.write_str(&format!(
            "Counted {} characters, top 5 bigrams {:?}",
            self.total_count,
            &dump_bigrams(self)[0..5]
        ))
    }
}