Chinese words

use chinese_segment.qweruse chinese_cut, chinese_search_terms from chinese_segment      // Quillo 0.6: only these

jieba-rs = "0.11"pinyin = "0.11"

// A search index for Chinese product titles, and a contact list sorted and searched by pinyin.////   quillo run chinese_segment.qwe      then open http://localhost:8080/wordsuse "chinese_segment.qwer"route GET /words ? { text: str = "我来到北京清华大学" } = { words: chinese_cut(text)?, index: chinese_search_terms(text)?, tags: chinese_tag(text)? }let contacts = ["张伟", "王芳", "李娜", "刘洋"]route GET /contacts = contacts    | map((c) => { name: c, pinyin: chinese_pinyin(c, "tone")?, sort: chinese_pinyin(c, "plain")?, initials: chinese_initials(c)? })    | sort_by((c) => c.sort)route GET /pinyin ? { text: str } = { tone: chinese_pinyin(text, "tone")?, number: chinese_pinyin(text, "number")? }

//! Chinese text split into words for search indexes and word counts (Chinese has no spaces), and written in pinyin, with tones or initials, for sorting and searching names.//! version: 1.0.0//!//! [dependencies]//! jieba-rs = "0.11"//! pinyin = "0.11"use jieba_rs::Jieba;use pinyin::ToPinyin;use serde_json::{json, Value};use std::sync::OnceLock;const MAX: usize = 1_000_000;fn jieba() -> &'static Jieba {    static J: OnceLock<Jieba> = OnceLock::new();    J.get_or_init(Jieba::new)}fn check(text: &str) -> Result<(), String> {    if text.len() > MAX { Err("text up to 1 MB".into()) } else { Ok(()) }}fn is_word(w: &str) -> bool {    w.chars().any(|c| c.is_alphanumeric())}/// The words of `text`, in order: "我来到北京清华大学" is ["我", "来到", "北京", "清华大学"]. Spaces and punctuation are left out; words in other scripts come through whole.pub fn chinese_cut(text: &str) -> Result<Vec<String>, String> {    check(text)?;    Ok(jieba().cut(text, true).into_iter().map(|t| t.word).filter(|w| is_word(w)).map(String::from).collect())}/// The terms to index `text` under for search: its words and the shorter words inside long ones ("清华大学" gives "清华", "大学" and "清华大学" too), each once.pub fn chinese_search_terms(text: &str) -> Result<Vec<String>, String> {    check(text)?;    let mut seen = std::collections::HashSet::new();    Ok(jieba().cut_for_search(text, true).into_iter().map(|t| t.word).filter(|w| is_word(w) && seen.insert(w.to_string())).map(String::from).collect())}/// `text` in pinyin, a syllable for each character, spaces between: `style` is "tone" ("zhōng guó"), "number" ("zhong1 guo2") or "plain" ("zhong guo"). Other characters stay as they are.pub fn chinese_pinyin(text: &str, style: &str) -> Result<String, String> {    check(text)?;    let mut parts: Vec<String> = Vec::new();    let mut other = String::new();    for (c, p) in text.chars().zip(text.to_pinyin()) {        match p {            Some(p) => {                if !other.trim().is_empty() {                    parts.push(other.trim().to_string());                }                other.clear();                parts.push(match style {                    "plain" => p.plain(),                    "number" => p.with_tone_num_end(),                    "tone" | "" => p.with_tone(),                    o => return Err(format!("style is \"tone\", \"number\" or \"plain\", not {:?}", o)),                }.to_string());            }            None => other.push(c),        }    }    if !other.trim().is_empty() {        parts.push(other.trim().to_string());    }    Ok(parts.join(" "))}/// The first letters of the pinyin of each Chinese character, in lower case: "张伟" is "zw", for finding contacts by their initials.pub fn chinese_initials(text: &str) -> Result<String, String> {    check(text)?;    Ok(text.to_pinyin().flatten().map(|p| p.first_letter()).collect())}/// The words of `text` with their part of speech, `[{ word, tag }]`: "n" noun, "v" verb, "ns" place name, "nr" person's name, "a" adjective... (the ICTCLAS tags).pub fn chinese_tag(text: &str) -> Result<Vec<Value>, String> {    check(text)?;    Ok(jieba().tag(text, true).into_iter().filter(|t| is_word(t.word)).map(|t| json!({ "word": t.word, "tag": t.tag })).collect())}