/* source-validator.v1.js
   Centralizes logic for deciding if a "Source" string should link to WOL.
   Loads publication codes from /static/data/wol-pub-codes.v1.json.

   Exposes:
     - SourceValidator.isWOLSource(text) -> boolean
     - SourceValidator.buildWOLSearchURL(text) -> string
*/
window.SourceValidator = (function () {
  // ---- Load publication codes (sync so callers can use API immediately) ----
  function loadPubCodesSync() {
    try {
      var xhr = new XMLHttpRequest();
      xhr.open("GET", "/static/data/wol-pub-codes.v1.json", false); // synchronous
      xhr.send(null);
      if (xhr.status >= 200 && xhr.status < 300) {
        var data = JSON.parse(xhr.responseText || "{}");
        if (data && Array.isArray(data.pub_codes)) {
          // de-duplicate and normalize to lowercase strings
          var uniq = Object.create(null), out = [];
          for (var i = 0; i < data.pub_codes.length; i++) {
            var c = String(data.pub_codes[i] || "").trim().toLowerCase();
            if (!c) continue;
            if (!uniq[c]) { uniq[c] = 1; out.push(c); }
          }
          return out;
        }
      }
    } catch (e) {
      // fall through to fallback
    }
    // Fallback (very small set) — only used if JSON cannot be loaded
    return ["w", "wp", "ws", "g", "rs"];
  }

  // Publications / codes loaded from JSON
  var PUB_CODES = loadPubCodesSync();

  // Choose the longest matching code at the start (so "ws" beats "w").
  var PUB_CODES_SORTED = PUB_CODES.slice().sort(function (a, b) { return b.length - a.length; });

  // Year validation rules (applies only if a year can be parsed from the source).
  // Watchtower (w/wp/ws) back to 1950; Awake (g) back to 1970.
  var YEAR_RULES = [
    { codes: ["w","wp","ws"], minYear: 1950 },
    { codes: ["g"],           minYear: 1970 }
  ];

  // Normalize helper
  function normalize(s) { return (s || "").trim().toLowerCase(); }

  function leadingCode(textLower) {
    for (var i = 0; i < PUB_CODES_SORTED.length; i++) {
      var code = PUB_CODES_SORTED[i];
      if (textLower.indexOf(code) === 0) return code;
    }
    return null;
  }

  // Try to extract a year that appears right after the leading code (allow spaces),
  // accepting either 4-digit (e.g., 1955, 2001) or 2-digit (e.g., 55, 95, 12) forms.
  function extractYearAfterCode(textLower, code) {
    var s = textLower.slice(code.length).trim();

    // 1) Look for a 4-digit year first (1800–2099)
    var m = s.match(/\b(1[89]\d{2}|20\d{2})\b/);
    if (m) return parseInt(m[1], 10);

    // 2) If not found, accept a 2-digit year at the *start* of the remainder
    //    (e.g., "w55 1/1", "w 95", "g70 1/22")
    m = s.match(/^\s*(\d{2})\b/);
    if (m) {
      var yy = parseInt(m[1], 10);
      if (code === "g")   return yy >= 70 ? (1900 + yy) : (2000 + yy);
      if (code === "w" || code === "wp" || code === "ws")
                          return yy >= 50 ? (1900 + yy) : (2000 + yy);
      return yy >= 70 ? (1900 + yy) : (2000 + yy);
    }
    // No recognizable year → don't enforce year limits
    return null;
  }

  function passesYearRuleIfPresent(textLower, code) {
    var rule = null;
    for (var i=0;i<YEAR_RULES.length;i++){
      if (YEAR_RULES[i].codes.indexOf(code) !== -1) { rule = YEAR_RULES[i]; break; }
    }
    if (!rule) return true; // no year rule for this pub

    var y = extractYearAfterCode(textLower, code);
    if (y == null) return true; // no year present → allow
    return y >= rule.minYear;
  }

  function isWOLSource(text) {
    var t = normalize(text);
    if (!t) return false;
    var code = leadingCode(t);
    if (!code) return false;
    // If starts with a known pub code, it’s WOL-capable — but enforce year rules where applicable.
    return passesYearRuleIfPresent(t, code);
  }

  function buildWOLSearchURL(text) {
    var q = encodeURIComponent(text || "");
    return "https://wol.jw.org/en/wol/l/r1/lp-e?q=" + q;
  }

  return { isWOLSource: isWOLSource, buildWOLSearchURL: buildWOLSearchURL };
})();