import TshetUinh from "tshet-uinh";
import * as Examples from "tshet-uinh-examples";

function tokensFromAnnotatedText(text, derivePron) {
  const tokens = [];
  const re = /([\s\S])\(([^)]+)\)/g;
  let lastIndex = 0;
  for (let m; (m = re.exec(text)); ) {
    const start = m.index;
    if (start > lastIndex) {
      tokens.push({ type: "raw", text: text.slice(lastIndex, start) });
    }
    const ch = m[1];
    const desc = m[2];
    let pron = "";
    try {
      const pos = TshetUinh.音韻地位.from描述(desc);
      pron = derivePron(pos, ch) ?? "";
    } catch {
      // 保底：不中断整段推导
      pron = "";
    }
    tokens.push({ type: "annotated", ch, desc, pron });
    lastIndex = start + m[0].length;
  }
  if (lastIndex < text.length) tokens.push({ type: "raw", text: text.slice(lastIndex) });
  return tokens;
}

function choosePositionByChar(ch) {
  const entries = TshetUinh.資料.query字頭(ch);
  if (!entries || !entries.length) return null;
  // 尽量贴近原站：优先选「廣韻」代表条目
  const guangyun = entries.find(e => e?.來源?.文獻 === "廣韻");
  return (guangyun || entries[0]).音韻地位 || null;
}

function tokensFromCharText(text, derivePron) {
  const tokens = [];
  let rawBuf = "";

  const flushRaw = () => {
    if (!rawBuf) return;
    tokens.push({ type: "raw", text: rawBuf });
    rawBuf = "";
  };

  for (const ch of Array.from(text)) {
    const pos = choosePositionByChar(ch);
    if (!pos) {
      rawBuf += ch;
      continue;
    }
    flushRaw();
    let pron = "";
    try {
      pron = derivePron(pos, ch) ?? "";
    } catch {
      pron = "";
    }
    tokens.push({ type: "annotated", ch, desc: pos.描述, pron });
  }
  flushRaw();
  return tokens;
}

function getDeriver(scheme) {
  return getDeriverWithOptions(scheme, undefined);
}

function getDeriverWithOptions(scheme, options) {
  if (scheme === "tupa") return Examples.tupa(options ?? {});
  if (scheme === "high_tang") return Examples.high_tang(options ?? {});
  if (scheme === "unt") return Examples.unt(options ?? {});
  if (scheme === "unt_legacy") return Examples.unt_legacy(options ?? {});
  if (scheme === "putonghua") return Examples.putonghua(options ?? {});
  if (scheme === "gwongzau") return Examples.gwongzau(options ?? {});
  if (scheme === "zaonhe") return Examples.zaonhe(options ?? {});
  throw new Error(`Unsupported scheme: ${scheme}`);
}

export function derive({ scheme = "tupa", text = "", options = undefined } = {}) {
  const s = scheme.toString();
  const t = text.toString();
  if (s === "tupa_high_tang") {
    const deriveTupa = getDeriverWithOptions("tupa", undefined);
    const deriveHighTang = getDeriverWithOptions("high_tang", undefined);
    const tokens = /[\s\S]\([^)]+\)/.test(t)
      ? // 兼容旧格式：括号内地位直接推导（两套方案）
        (() => {
          const base = tokensFromAnnotatedText(t, () => "");
          return base.map(tok => {
            if (tok.type !== "annotated") return tok;
            const pos = TshetUinh.音韻地位.from描述(tok.desc);
            const tupa = (deriveTupa(pos, tok.ch) ?? "").trim();
            const ipa = (deriveHighTang(pos, tok.ch) ?? "").trim();
            return { ...tok, pron: `${tupa} /${ipa}/` };
          });
        })()
      : // 新格式：按字头查資料推导（两套方案）
        (() => {
          const tokens = [];
          let rawBuf = "";
          const flushRaw = () => {
            if (!rawBuf) return;
            tokens.push({ type: "raw", text: rawBuf });
            rawBuf = "";
          };
          for (const ch of Array.from(t)) {
            const pos = choosePositionByChar(ch);
            if (!pos) {
              rawBuf += ch;
              continue;
            }
            flushRaw();
            const tupa = (deriveTupa(pos, ch) ?? "").trim();
            const ipa = (deriveHighTang(pos, ch) ?? "").trim();
            tokens.push({ type: "annotated", ch, desc: pos.描述, pron: `${tupa} /${ipa}/` });
          }
          flushRaw();
          return tokens;
        })();
    return { scheme: s, tokens };
  }

  const derivePron = getDeriverWithOptions(s, options);
  // 兼容两种输入：
  // 1) 旧格式：風(幫三C東平) —— 直接用括号内地位推导
  // 2) 新格式：只输入汉字 —— 按字头查資料推导
  const tokens = /[\s\S]\([^)]+\)/.test(t) ? tokensFromAnnotatedText(t, derivePron) : tokensFromCharText(t, derivePron);
  return { scheme: s, tokens };
}