@sonordev/contracts 2.12.0 → 2.15.0

package/dist/llms.mjs

@@ -1,3 +1,622 @@
+// src/contracts/error-page.ts
+var ERROR_PAGE_MAX_WORDS = 120;
+var PREFIXED_MAX_WORDS = 400;
+var SHORT_ONLY_MAX_WORDS = 60;
+var START_STATE_MAX_WORDS = 40;
+var HEAD_CHARS = 300;
+var DESCRIBE_WINDOW = 45;
+var NAMED_ENTITIES = {
+  amp: "&",
+  quot: '"',
+  apos: "'",
+  nbsp: " ",
+  lt: "<",
+  gt: ">",
+  rsquo: "'",
+  lsquo: "'",
+  ldquo: '"',
+  ndash: "\u2013",
+  mdash: "\u2014",
+  hellip: "\u2026"
+};
+function decodeEntities(text) {
+  let out = text;
+  for (let pass = 0; pass < 2 && out.indexOf("&") !== -1; pass++) {
+    out = out.replace(/&(?:#x([0-9a-f]{1,6})|#(\d{1,7})|([a-z]{2,8}));/gi, (whole, hex, dec, name) => {
+      if (name) {
+        const key = name.toLowerCase();
+        return Object.prototype.hasOwnProperty.call(NAMED_ENTITIES, key) ? NAMED_ENTITIES[key] : whole;
+      }
+      const code = hex ? parseInt(hex, 16) : parseInt(dec, 10);
+      return code <= 1114111 ? String.fromCodePoint(code) : whole;
+    });
+  }
+  return out;
+}
+function tidy(text) {
+  return decodeEntities(text).replace(/[‘’‛ʼ]/g, "'").replace(/\s+/g, " ").trim();
+}
+function textOf(text) {
+  const oc = tidy(text);
+  const lc = oc.toLowerCase();
+  return { lc, oc: lc.length === oc.length ? oc : lc };
+}
+function slice(text, from, to) {
+  return { lc: text.lc.slice(from, to), oc: text.oc.slice(from, to) };
+}
+function words(text) {
+  return text.split(/\s+/).filter(Boolean).length;
+}
+function wordsAcrossElements(text) {
+  return words(text.replace(/([a-z0-9.,;:!?)"'’])([A-Z])/g, "$1 $2"));
+}
+var NOT_A_QUESTION = String.raw`(?!\s*\?|\s+(?:in|on|for|at|with|under|from|during|to)\b[^.!?]{0,50}\?)`;
+var NOT_IN_YOUR_SIZE = String.raw`(?!\s+in\s+your\s+size\b)`;
+var RECORD_NOUN = String.raw`(?:page|project|post|article|product|listing|resource|file|deployment)`;
+var CANT_BE_FOUND = String.raw`\b(?:page|url|link|resource|route|file|deployment|site|website|app|project)\b[^.!?]{0,40}?(?:could not|couldn't|can't|cannot|can not) be found`;
+var WHICH_PAGE = String.raw`(?:\s+(?:you(?:'re| are| were)? (?:looking for|requested|asked for|(?:tried|were trying|are trying) to (?:reach|visit|access|find|open))|that you (?:requested|asked for)|at\s+\S[^.!?]{0,40}|on this (?:site|server|website)))?`;
+var DOESNT_EXIST = String.raw`(?:\b(?:this|that|the|requested)\s+(?:[a-z0-9'-]+\s+){0,2}(?:page|url|link|post|article|listing|product|resource|project|profile|record|entry|story|event|course|job|item|offer|video)|\bpage)${WHICH_PAGE}\s+(?:doesn't|does not)(?: (?:seem|appear))?(?: to)? exist(?!\s+(?:to|for|so)\s)`;
+var MAINTENANCE_QUALIFIER = String.raw`(?!\s+(?:on|from|every|each|this|next|between|during|at|after|before|once|twice|monthly|weekly|nightly|daily|contract|agreement|plan|program|warranty|coverage|with\s+(?:us|our)|for\s+(?:the|a|\d)|until(?!\s+(?:\d{1,2}(?::\d\d)?\s*(?:[ap]\.?m\b|o'clock)|\d{1,2}:\d\d|noon|midnight|tomorrow|further notice)))\b)`;
+var ERROR_PAGE_SIGNATURES = [
+  // ---- Not found: the framework defaults and the common custom wordings ----
+  new RegExp(RECORD_NOUN + " (?:was |is )?not found" + NOT_A_QUESTION + NOT_IN_YOUR_SIZE),
+  // "Can't find the page you need? Use the search box" is a help line; "We can't find that page." isn't.
+  /(?:can't|cannot|can not|couldn't|could not|unable to) (?:be )?(?:find|found|locate)(?: that| this| the)? page(?!\s*\?|\s+(?:you|that|which)\b[^.!?]{0,40}\?)/,
+  // "We can't seem to find the page", "It seems we can't find what you're looking for".
+  /(?:we|i) (?:can't|cannot|couldn't|could not|were unable to|are unable to|weren't able to)(?: seem to)? (?:find|locate) (?:that page|this page|the page|that one|that(?=[.!,?]| link)|what you(?:'re| are| were)? looking for)/,
+  // "can not" is two words in the Vercel default.
+  new RegExp(CANT_BE_FOUND),
+  // Next's own not-found.tsx example: "Not FoundCould not find requested resourceReturn Home".
+  /(?:could not|couldn't) find (?:the )?requested resource/,
+  // "doesn't exist" only with a page, link or record as the subject: a short
+  // marketing page can say "if a tool doesn't exist, we'll build it".
+  new RegExp(DOESNT_EXIST),
+  /(?:doesn't|does not) exist or (?:has|have) been (?:moved|removed|deleted)/,
+  /looking for (?:doesn't|does not) exist/,
+  // A page-ish subject, like "doesn't exist" above: a history page can say "the 1890 mill no longer exists".
+  new RegExp(String.raw`\b(?:this|that|the)\s+(?:[a-z0-9'-]+\s+){0,2}(?:page|url|link|project|post|article|product|listing|resource|content|item|offer|event|record|story|profile|entry|document|course|job|video)` + WHICH_PAGE + String.raw`\s+no longer exists`),
+  /(?:404: ?not[ _]found|code: ?not_found)/,
+  // "Error 404" and "404 Error" as a label ("404 Error | Acme") or leading into what a 404 says
+  // ("Error 404Page not found"). Not "a custom 404 error page" and not a name ("Error 404 Studio").
+  /(?:^|[.!?|:]\s*)(?:http |server )?(?:error 404|404 error)(?=\s*[-–—|:.!?,)]|\s*(?:page|this|that|the|we|oops|whoops|sorry|looks|uh|hmm|nothing|there|it|oh|please)\b)|(?:error 404|404 error)(?=\s*$|\s*[-–—|:]?\s*(?:not found|this (?:page|url)|that page)\b)/,
+  // "The page you requested was not found", "The page you were looking for has been removed".
+  /page you (?:requested|(?:were|are) looking for|tried to (?:reach|visit|access))[^.!?]{0,60}(?:not found|no longer|(?:has|have|had|might have|may have|may) been|has (?:drifted|wandered|vanished|disappeared|moved|left)|isn't|is not|doesn't|does not|could not|can't|cannot|was not)/,
+  // Apache, Flask and Django: "The requested URL was not found on this server."
+  /requested (?:url|uri|resource|file|page|document)\b[^.!?]{0,80}(?:not found|not be found|cannot be found)|(?:was|is) not found on (?:this|the) server/,
+  // Netlify, Render, Flask and Express answer with the bare words "Not Found".
+  /^not found(?=$|\s*[-–—|:.!]|the requested|request id)/,
+  // WordPress core and its default themes: "It looks like nothing was found at this location.", "Nothing Found".
+  /nothing was found at this location|^nothing found(?=$|[.,:;!?|-]|sorry|it looks|we couldn|nothing|oops|the page)/,
+  // Custom 404 copy that never says "404": "Oops! This page took a day off.", "You've hit a dead end.",
+  // "This page doesn't live here anymore." The opener needs the page to be the one that's gone ("Oops! This
+  // page is a work in progress" is a holding page), and the dead end needs to be a sentence of its own.
+  /^(?:oops|whoops|uh[- ]oh)[!,.…]*\s*(?:this|that) page (?:took|went|wandered|got lost|ran away|has (?:gone|wandered|moved|left|vanished|disappeared|been (?:removed|moved|deleted))|is (?:missing|gone|lost)|isn't (?:here|available|there)|(?:doesn't|does not|can't|cannot|couldn't) (?:exist|be found|seem)|seems? (?:to be )?(?:missing|lost|gone))/,
+  /(?:^|[.!?]\s*)(?:you've|you have|you) (?:hit|reached) a dead end(?=[.!?…]|$)/,
+  /(?:page|link|url)[^.!?]{0,40}(?:doesn't|does not|no longer|isn't|is not) (?:live |exist |work )?here(?: anymore)?(?![a-z]*\s+to\b)/,
+  // Gatsby's 404 route: "NOT FOUNDYou just hit a route that doesn't exist... the sadness."
+  /hit a route that doesn't exist/,
+  // Tomcat and IIS: "HTTP Status 404 – Not Found", "HTTP Error 403.14 - Forbidden", "HTTP Error 503. The service is unavailable."
+  /http (?:error|status) [45]\d\d(?:\.\d+)?\s*[–—-]\s*(?:not found|forbidden|unauthori[sz]ed|bad request|internal|method not allowed|service unavailable)|http error [45]\d\d(?:\.\d+)?\. the /,
+  // A status number with its reason, anywhere in the head (a nav can precede it):
+  // "404 | Not Found", "502 - Bad Gateway", "403: Forbidden", "400Bad Request", "419 | Page Expired".
+  // Not "401 unauthorized means your key is missing": a status explained is documentation.
+  /(?:^|\D)(?:(?:404|410)\s*[|:–—-]?\s*(?:not found|page not found|file not found|web site not found)|408[\s:|–—-]*request time-?out|50[0-4][\s:|–—-]*(?:internal (?:server )?error|server error|bad gateway|service (?:temporarily )?unavailable|gateway time-?out)|40[13][\s:|–—-]*(?:forbidden|unauthori[sz]ed)|400[\s:|–—-]*bad request|405[\s:|–—-]*method not allowed|413[\s:|–—-]*(?:request entity|payload|content) too large|419[\s:|–—-]*page expired|429[\s:|–—-]*too many)(?!\s+(?:(?:requests?|errors?|responses?|status(?: codes?)?|codes?|messages?|pages?)\s+)?(?:means?|is|are|indicates?|occurs?|happens?|when|if|returns?|tells?|shows?|appears?|can|should|will|occurred)\b)/,
+  // Netlify's bare "Not Found - Request ID: ..." and Express's "Cannot GET /about", even with a nav in front.
+  /not found\s*[-–—]\s*request id|\bcannot (?:get|post|put|patch|delete) \/[\w\-./?=&%]*(?:$|\s)/,
+  // IIS's classic "Server Error 404 - File or directory not found." and its "had its name changed" sentence.
+  /file or directory not found|removed,? had its name changed/,
+  // This page has been removed / moved: its own sentence, not "removed from our menu".
+  /(?:^|[.!?]\s*)this page (?:has been |has |was |is )?(?:permanently )?(?:removed|deleted|moved)(?=[.!,;]|$|\s+(?:or|and)\b|\s*(?:return|go|back|click|please|visit|head)\b)/,
+  // ---- Server and client errors ----
+  // Not "Application error monitoring for Node teams".
+  /^application error(?!\s+(?:monitoring|tracking|reporting|handling|management|detection|analytics|logs?|alerts?|rates?|budgets?|codes?))/,
+  /(?:client|server)-side exception has occurred/,
+  // The hosts spell a status as a code too: "500: INTERNAL_SERVER_ERROR", "504: GATEWAY_TIMEOUT".
+  /internal[ _]server[ _]error/,
+  new RegExp(String.raw`(?:bad[ _]gateway|gateway[ _]time-?out)` + NOT_A_QUESTION),
+  // An exception's message rendered as the page: "Error: Failed to fetch data", "TypeError: x is undefined".
+  /^(?:(?:type|reference|syntax|range)error: |error: (?:failed|cannot|can't|unable|invalid|network|request|fetch|something|unexpected|no |not ))/,
+  // "Something went wrong with your insurance claim? ..." is a headline about a claim, not an error state.
+  new RegExp(String.raw`something(?: went|'s gone| has gone| had gone) wrong(?!\s+(?:with|during|on|in|at)\s+(?:your|my)\b)` + NOT_A_QUESTION),
+  /an (?:unexpected )?error (?:has )?occurred/,
+  /this (?:page|site|website) (?:isn't working|can't be reached)/,
+  // Next.js 16 error shells: "This page couldn't load. A server error occurred. Reload to try again."
+  // and the 403 / 401 pages ("This page could not be accessed.").
+  /this page couldn't load|a server error (?:has )?occurred|reload to try again(?=[.,!]|$)|this page could not be accessed/,
+  // React Router and Remix, Spring Boot, and the "unexpected error" boundaries.
+  /unexpected application error|whitelabel error page|(?:there was|encountered) an unexpected error/,
+  // Django and ASP.NET: "Server Error (500)", "Server Error in '/' Application", "Not Found (404)".
+  /server error (?:\(500\)|in '?\/)|(?:bad request|forbidden|not found) \((?:400|403|404)\)/,
+  // Next's runtime messages when a chunk or a hydration fails and the message is the page.
+  /chunkloaderror|loading chunk [\w/.-]+ failed|hydration failed because the|there was an error while hydrating/,
+  // Generated error.tsx boundaries: "Oops, we hit a snag.", "Something unexpected happened", "Unable to load page".
+  new RegExp(
+    String.raw`(?:^|[.!?]\s*)(?:(?:oops|whoops|uh[- ]oh)[!,.…]*\s*)?we hit a snag(?!\s+(?:with|in|on|during|when|while|at)\b)|(?:^|[.!?]\s*)something unexpected happened|(?:^|[.!?]\s*)(?:failed|unable) to load (?:the |this )?(?:page|content|data|resource)` + NOT_A_QUESTION + String.raw`|^error loading (?:the |this )?(?:page|content|data)`
+  ),
+  // "We're having trouble loading this page", "We couldn't load this page".
+  /(?:we're|we are|i'm) having (?:some )?(?:trouble|difficulty|difficulties|problems?) (?:loading|displaying|showing|retrieving) (?:this|that|the) (?:page|content)|we (?:couldn't|could not|can't|cannot) load (?:this|that|the) page/,
+  // WordPress core's fatal-error and database screens.
+  /there has been a critical error on this (?:website|site)|error establishing a database connection|(?:the|this) (?:site|website) is experiencing technical difficulties/,
+  // Apache's stock pages: 503, 400, 401, 408, 410 and 414 each in its own words.
+  /the server is temporarily unable to service your request|your browser sent a request that this server could not understand|this server could not verify that you are authorized|server timeout waiting for the http request|(?:was|is) no longer available on this server|request-uri too large/,
+  // ---- Unavailable and maintenance ----
+  // A qualifier ("in Hawaii", "until April") makes it content about a service, not a dead page.
+  /(?:this|the) (?:page|site|website|service) (?:is|are) (?:temporarily|currently) unavailable(?!\s+(?:in|on|for|at|during|to|until|from|outside|after|before|through)\b)/,
+  /^(?:service (?:temporarily )?unavailable|(?:temporarily|currently) unavailable(?!:\s))(?!\s+(?:in|on|for|at|during|to|until|from|outside)\b)/,
+  // The site or the page is the one under maintenance. Not "is under maintenance contract", not "down for
+  // maintenance on Fridays": a day, a time or a contract makes it content.
+  new RegExp(String.raw`(?:site|website|page|store|shop|app|portal|server|we're|we are)(?: is| are)?(?: currently| temporarily| now)*,? (?:under|down for) (?:scheduled |routine |some |emergency )?maintenance` + MAINTENANCE_QUALIFIER),
+  // "Our website is currently undergoing scheduled maintenance", "Briefly unavailable for scheduled maintenance",
+  // "We are offline for a short while". Not "Our rental fleet is undergoing routine maintenance": the subject is the site.
+  new RegExp(String.raw`(?:(?:site|website|page|store|shop|app|portal|server|we're|we are)(?: is| are)?(?: currently| temporarily| now)*|briefly|temporarily|currently|now) (?:undergoing|(?:offline|unavailable|down)(?: temporarily)? for) (?:scheduled |routine |some |emergency )?maintenance` + MAINTENANCE_QUALIFIER),
+  /(?:we're|we are|site is|website is) offline for an? (?:short|little|brief) (?:while|time|period)/,
+  /^be right back\.?$/,
+  // "We'll be back soon." ending its sentence, or leading into the next one. Not "We should be back soon with a replacement
+  // part", "back soon after our holiday break": what follows decides.
+  /(?:we'll|we will|we should|it should) (?:be )?(?:right )?back (?:online )?(?:soon|shortly)(?=[.!;]|$|,?\s*(?:our|the|this|thank|thanks|please|we)\b|\s+(?:once|after)\s+(?:a|the|our)\s+(?:(?:short|brief|quick|scheduled|planned)\s+)*(?:maintenance|upgrade|update|outage|downtime|migration|work)\b)/,
+  // An apology for the downtime makes what follows an outage notice, whatever it promises.
+  /sorry for the (?:downtime|inconvenience|interruption|outage)\W+(?:we'll|we will|we should|we are working)/,
+  // ---- Walls that stand in for the page ----
+  new RegExp(String.raw`(?:403 forbidden|access denied)(?!\s+(?:is|was|means|errors?|messages?|pages?)\b)` + NOT_A_QUESTION),
+  // A permission wall is addressed to the visitor and about THIS page or resource, and opens the text or follows its own
+  // heading ("Sorry, you are not allowed to access this page.", "ForbiddenYou are not authorized to access this resource.").
+  // Not a person or a file ("our team is not allowed to access this page of your file"), not a gallery ("you don't have
+  // permission to view the gallery"), and not a sentence further in ("Members only. You do not have permission to ...").
+  /(?:^|sorry|oops|error|forbidden|unauthori[sz]ed|access denied|40[13])\W*you(?:'re| are| aren't) (?:not |un)?(?:allowed|authori[sz]ed|permitted) to (?:access|view|see) (?:this|that) (?:page|resource|site|website|area|section|content)/,
+  /(?:^|sorry|oops|error|forbidden|unauthori[sz]ed|access denied|40[13])\W*you (?:don't|do not) have permission to (?:access|view) (?:this (?:page|resource|site|website|area|section|content)|\S+ on this server)/,
+  /^(?:forbidden|unauthori[sz]ed)(?=$|\s*[-–—|:.!]|you(?:'re| are)\b|please\b|this (?:page|resource))/,
+  /(?:checking your browser before accessing|enable javascript and cookies to continue(?!\s+\w)|verif(?:y|ying) you are (?:a )?human)/
+];
+var LEADING_STATUS = /^(?:(?:http )?error:? ?)?(?:400|401|402|403|405|408|410|413|414|419|429|500|501|502|503|504)\W{0,3}(?:(?:service|server|temporarily|internal|gateway|bad|proxy|request|too|access|page|file|resource|connection|http|status|method|payload|entity|payment|authentication)\s+){0,3}(?:error|forbidden|unauthori[sz]ed|gone|unavailable|bad gateway|time-?out|timed out|not found|not allowed|too many|too large|bad request|page expired|payment required|not implemented)/;
+var LEADING_404 = /^404(?![\d,])(?![\s.\-–—)]*\d)/;
+var ADDRESS_AFTER_404 = /^404\s+(?:[a-z0-9.'-]+\s+){0,3}(?:street|st|avenue|ave|road|rd|boulevard|blvd|drive|dr|lane|ln|way|court|ct|highway|hwy|pike|parkway|pkwy|place|pl|suite|ste|unit|apt)\b/;
+var SHORT_ONLY_SIGNATURES = [
+  // Opening the text, and not conditional: "Please refresh the page and try again." is the boundary's own line;
+  // "Clear your cache, then refresh the page and try again" is a tip, and "...try again if the calendar doesn't load" is one.
+  /^(?:please )?(?:refresh|reload)(?: the| this)? (?:page )?(?:and|or) try again(?!\s+(?:if|when|unless|whether|once|after|later|next|the next)\b)/
+];
+var SHORT_START_SIGNATURES = [new RegExp(String.raw`^(?:[a-z0-9'&-]+\s+){0,2}[a-z0-9'&-]+ not found` + NOT_A_QUESTION + NOT_IN_YOUR_SIZE)];
+var GLUED_404 = /(?:^|[\s|])404(?=[a-z])(?!th\b)/g;
+var DELIMITED_404 = /(?:^|\s)404\s*[|:–—]\s*[a-z]/g;
+var NUMBER_PREFIX = /(?:suite|ste|room|rm|unit|apt|apartment|route|rt|hwy|highway|building|bldg|floor|#)\.?\s*$/;
+var ABOUT_THE_ERROR = String.raw`(?!\W*(?:errors?\b|explained|guide|what (?:it|this|that|is|are)|how to|how do|why|means|fix(?:es|ing)?\b|causes|vs\b|\?))`;
+var PREFIX_SIGNATURES = [
+  new RegExp(String.raw`(?:page|resource|post|article|product|listing) (?:was |is )?not found`),
+  new RegExp(CANT_BE_FOUND),
+  /(?:we|i) (?:can't|cannot|couldn't|could not)(?: seem to)? (?:find|locate) (?:that page|this page|the page|what you(?:'re| are| were)? looking for)/,
+  new RegExp(DOESNT_EXIST),
+  /something went wrong|(?:client|server)-side exception|internal server error|this page couldn't load/
+].map((pattern) => new RegExp("(?:" + pattern.source + ")" + ABOUT_THE_ERROR));
+var PREFIXED_LEADING_404 = new RegExp(LEADING_404.source + ABOUT_THE_ERROR);
+var PREFIXED_LEADING_STATUS = new RegExp(LEADING_STATUS.source + ABOUT_THE_ERROR);
+var PREFIX_SEPARATOR = " \u2014 ";
+var PREFIX_MIN_CHARS = 60;
+var PREFIX_MAX_CHARS = 400;
+var SKIP_LINK = /^(?:skip to (?:main )?content|skip to main|skip navigation)\s*/;
+var PREFIX_WINDOW_CHARS = 90;
+var DESCRIBERS = String.raw`if|when|whether|unless|how to|how do|why|what is|what's|fix(?:ing|es)?|handle|handling|resolve|resolving|troubleshoot(?:ing)?|diagnos(?:e|ing)|watch(?:es|ing)? for|encounter(?:ing)?|history of`;
+var LOWERCASE_DESCRIBERS = String.raw`about|monitor(?:ing)?|alert(?:s|ing)? (?:on|for)`;
+var DESCRIBING_CONTEXT = new RegExp(
+  String.raw`(?:\b(?:(${DESCRIBERS})|(${LOWERCASE_DESCRIBERS}))(?:\s+(?:you|your|the|a|an|their|our|my|see|get|are|is|seeing|getting|receive|receiving|hit|to|every|any|each|all|\d{3})){0,3}\s*["'“]?|["“])$`
+);
+var INNER_DESCRIBER = /\b(?:if|when|whether|unless|how to)\b/;
+var SENTENCE_OPENING = /[.!?:;|—–"“(]\s*$/;
+var VISIBILITY_SUBJECT = /\b(?:you|your|we|our|they|their|customers?|clients?|people|visitors?|buyers?|patients?|business(?:es)?|brands?|compan(?:y|ies))\s+$/;
+var FORM_FIELDS = /\b(?:name|e-?mail|phone|message|subscribe|submit)\b/;
+function globalOf(patterns) {
+  return patterns.map((pattern) => new RegExp(pattern.source, "g"));
+}
+var SIGNATURES = globalOf(ERROR_PAGE_SIGNATURES);
+var SHORT_ONLY_SIGS = globalOf(SHORT_ONLY_SIGNATURES);
+var SHORT_START_SIGS = globalOf(SHORT_START_SIGNATURES);
+var PREFIX_SIGS = globalOf(PREFIX_SIGNATURES);
+var TITLE_VERB = /\b(?:fix|handle|resolve|troubleshoot|diagnose)\b/;
+function describerIsProse(orig, start, describer, context) {
+  const word = orig.slice(start, start + describer.length);
+  if (word === describer || word === describer.charAt(0).toUpperCase() + describer.slice(1)) return true;
+  return start === 0 || SENTENCE_OPENING.test(orig.slice(Math.max(0, start - 3), start)) || TITLE_VERB.test(context);
+}
+function isDescribed(head, hit) {
+  const from = Math.max(0, hit.index - DESCRIBE_WINDOW);
+  const before = head.lc.slice(from, hit.index);
+  const context = DESCRIBING_CONTEXT.exec(before);
+  if (context) {
+    const [, describer, lowercaseDescriber] = context;
+    const at = from + context.index;
+    if (describer !== void 0) {
+      if (describerIsProse(head.oc, at, describer, context[0])) return true;
+    } else if (lowercaseDescriber !== void 0) {
+      if (head.oc.slice(at, at + lowercaseDescriber.length) === lowercaseDescriber) return true;
+    } else {
+      return true;
+    }
+  }
+  const inner = INNER_DESCRIBER.exec(hit[0]);
+  if (inner && head.oc.slice(hit.index + inner.index, hit.index + inner.index + inner[0].length) === inner[0]) return true;
+  return /be found$/.test(hit[0]) && VISIBILITY_SUBJECT.test(before) || /you are (?:a )?human$/.test(hit[0]) && FORM_FIELDS.test(head.lc);
+}
+function showsSignature(head, patterns, inFirstSentence = false) {
+  return patterns.some((pattern) => {
+    pattern.lastIndex = 0;
+    for (let hit = pattern.exec(head.lc); hit; hit = pattern.exec(head.lc)) {
+      if (inFirstSentence && /[.!?]/.test(head.lc.slice(0, hit.index))) return false;
+      if (!isDescribed(head, hit)) return true;
+      pattern.lastIndex = hit.index + 1;
+    }
+    return false;
+  });
+}
+function showsLoose404(head) {
+  for (const pattern of [GLUED_404, DELIMITED_404]) {
+    pattern.lastIndex = 0;
+    for (let hit = pattern.exec(head); hit; hit = pattern.exec(head)) {
+      if (!NUMBER_PREFIX.test(head.slice(Math.max(0, hit.index - 14), hit.index))) return true;
+    }
+  }
+  return false;
+}
+function afterPrefix(text) {
+  let start = 0;
+  const cut = text.lc.indexOf(PREFIX_SEPARATOR);
+  if (cut >= PREFIX_MIN_CHARS && cut < PREFIX_MAX_CHARS) start = cut + PREFIX_SEPARATOR.length;
+  const skip = SKIP_LINK.exec(text.lc.slice(start));
+  if (skip) start += skip[0].length;
+  return start === 0 ? text : slice(text, start);
+}
+function opensWith404(head) {
+  return LEADING_404.test(head) && !ADDRESS_AFTER_404.test(head);
+}
+function showsErrorState(head, count) {
+  return opensWith404(head.lc) || LEADING_STATUS.test(head.lc) || showsSignature(head, SIGNATURES) || count <= START_STATE_MAX_WORDS && showsSignature(head, SHORT_START_SIGS) || count <= SHORT_ONLY_MAX_WORDS && (showsSignature(head, SHORT_ONLY_SIGS) || showsLoose404(head.lc));
+}
+function looksLikeErrorPage(text, wordCount2) {
+  if (!text || typeof text !== "string") return false;
+  const trimmed = text.trim();
+  if (!trimmed) return false;
+  const own = wordsAcrossElements(trimmed);
+  const count = own;
+  if (count > PREFIXED_MAX_WORDS) return false;
+  const whole = textOf(trimmed);
+  const rest = afterPrefix(whole);
+  if (own > ERROR_PAGE_MAX_WORDS) {
+    const lead = rest.lc.slice(0, HEAD_CHARS);
+    return PREFIXED_LEADING_404.test(lead) && !ADDRESS_AFTER_404.test(lead) || PREFIXED_LEADING_STATUS.test(lead) || showsSignature(slice(rest, 0, PREFIX_WINDOW_CHARS), PREFIX_SIGS, true);
+  }
+  if (count > ERROR_PAGE_MAX_WORDS) return false;
+  return showsErrorState(slice(whole, 0, HEAD_CHARS), count) || // A long description pushes the page's own words past the first 300 characters.
+  rest !== whole && showsErrorState(slice(rest, 0, HEAD_CHARS), words(rest.lc));
+}
+function usablePageText(text, wordCount2) {
+  if (!text || typeof text !== "string" || !text.trim()) return null;
+  return looksLikeErrorPage(text) ? null : text;
+}
+
+// src/contracts/seo-meta.ts
+var TITLE_SEPARATOR = " | ";
+function stripTitleSuffix(title) {
+  const t = title.trim();
+  const i = t.lastIndexOf(TITLE_SEPARATOR);
+  return i > 0 ? t.slice(0, i).trim() : t;
+}
+
+// src/contracts/page-titles.ts
+var PIPE_RUN = /\|(?:\s*\|)+/g;
+var EDGE_PIPES = /^[\s|]+|[\s|]+$/g;
+function titleHead(title) {
+  if (typeof title !== "string") return "";
+  const tidy2 = title.replace(/\s+/g, " ").replace(PIPE_RUN, " | ").replace(EDGE_PIPES, "").replace(/\s+/g, " ");
+  return tidy2 ? stripTitleSuffix(tidy2) : "";
+}
+var ACRONYMS = /* @__PURE__ */ new Map([
+  ["ac", "AC"],
+  ["ai", "AI"],
+  ["aeo", "AEO"],
+  ["api", "API"],
+  ["cms", "CMS"],
+  ["crm", "CRM"],
+  ["faq", "FAQ"],
+  ["faqs", "FAQs"],
+  ["fsbo", "FSBO"],
+  ["hvac", "HVAC"],
+  ["jl", "JL"],
+  ["llc", "LLC"],
+  ["llp", "LLP"],
+  ["mls", "MLS"],
+  ["pdf", "PDF"],
+  ["pllc", "PLLC"],
+  ["ppc", "PPC"],
+  ["roi", "ROI"],
+  ["seo", "SEO"],
+  ["ui", "UI"],
+  ["usa", "USA"],
+  ["ux", "UX"]
+]);
+var STATE_CODES = new Set(
+  "al ak az ar ca co ct de dc fl ga hi id il in ia ks ky la me md ma mi mn ms mo mt ne nv nh nj nm ny nc nd oh ok or pa ri sc sd tn tx ut vt va wa wv wi wy".split(" ")
+);
+var WORD_STATE_CODES = /* @__PURE__ */ new Set([
+  "in",
+  "or",
+  "me",
+  "hi",
+  "id",
+  "la",
+  "ma",
+  "pa",
+  "al",
+  "de",
+  "co"
+]);
+var TRAILING_STATE_CODES = /* @__PURE__ */ new Set(["mt"]);
+function isStateCode(lower, isLast) {
+  if (!STATE_CODES.has(lower) || WORD_STATE_CODES.has(lower)) return false;
+  return TRAILING_STATE_CODES.has(lower) ? isLast : true;
+}
+var SMALL_WORDS = /* @__PURE__ */ new Set([
+  "a",
+  "an",
+  "and",
+  "at",
+  "by",
+  "for",
+  "in",
+  "of",
+  "on",
+  "or",
+  "the",
+  "to",
+  "vs",
+  "with"
+]);
+var PAGE_EXTENSION = /\.(?:html?|php|aspx)$/i;
+var URL_ORIGIN = /^[a-z][a-z0-9+.-]*:\/\/[^/\\]*/i;
+var PERCENT_RUN = /(?:%[0-9a-f]{2})+/gi;
+function decodeEscapes(value) {
+  return value.replace(PERCENT_RUN, (run) => {
+    try {
+      return decodeURIComponent(run);
+    } catch {
+      return run;
+    }
+  });
+}
+function capitalizeFirst(word) {
+  const code = word.codePointAt(0);
+  if (code === void 0) return word;
+  const first = String.fromCodePoint(code);
+  return first.toUpperCase() + word.slice(first.length);
+}
+function labelWord(token, isFirst, isLast) {
+  const lower = token.toLowerCase();
+  const acronym = ACRONYMS.get(lower);
+  if (acronym) return acronym;
+  if (isStateCode(lower, isLast)) return lower.toUpperCase();
+  if (/\p{N}/u.test(token)) return /^\p{L}/u.test(token) ? capitalizeFirst(token) : token;
+  if (!isFirst && SMALL_WORDS.has(lower)) return lower;
+  return capitalizeFirst(lower);
+}
+function slugLabel(input) {
+  if (typeof input !== "string") return "";
+  const path = input.split(/[?#]/, 1)[0].replace(URL_ORIGIN, "");
+  const segment = path.split(/[/\\]/).filter((s) => s.trim() !== "").pop();
+  if (!segment) return "";
+  const name = decodeEscapes(segment).replace(PAGE_EXTENSION, "");
+  const tokens = name.split(/[-_+\s]+/).filter(Boolean);
+  return tokens.map((token, i) => labelWord(token, i === 0, i === tokens.length - 1)).join(" ");
+}
+var PLACEHOLDER_TOKEN = /^(?::[A-Za-z_][\w-]*[?*+]?|\[\[\.\.\.[A-Za-z_][\w-]*\]\]|\[(?:\.\.\.)?[A-Za-z_][\w-]*\])$/;
+function isRoutePlaceholder(value) {
+  if (typeof value !== "string") return false;
+  return value.split("/").some((segment) => PLACEHOLDER_TOKEN.test(segment.trim()));
+}
+var ABSOLUTE_URL = /^[a-z][a-z0-9+.-]*:\/\//i;
+var HOST_NAME = /^(?:[a-z0-9-]+\.)+[a-z]{2,}(?::\d{1,5})?$/i;
+function withoutOuterSlashes(value) {
+  let start = 0;
+  let end = value.length;
+  while (start < end && value[start] === "/") start++;
+  while (end > start && value[end - 1] === "/") end--;
+  return value.slice(start, end);
+}
+function isHostPath(value) {
+  const slash = value.indexOf("/");
+  if (slash < 0 || /\s/.test(value)) return false;
+  return HOST_NAME.test(value.slice(0, slash));
+}
+function isPageAddress(value, path) {
+  if (typeof value !== "string") return false;
+  const given = value.trim();
+  if (!given) return false;
+  if (given.startsWith("/") || ABSOLUTE_URL.test(given) || isHostPath(given) || isRoutePlaceholder(given)) return true;
+  if (typeof path !== "string") return false;
+  const own = withoutOuterSlashes(path.trim());
+  return withoutOuterSlashes(given) === own;
+}
+function pipePieces(title) {
+  return title.split("|").map((piece) => piece.trim()).filter(Boolean);
+}
+function isPageName(title, path) {
+  if (typeof title !== "string") return false;
+  const pieces = pipePieces(title);
+  return pieces.length > 0 && pieces.every((piece) => !isPageAddress(piece, path));
+}
+
+// src/contracts/sentences.ts
+var ABBREVIATIONS = /* @__PURE__ */ new Set([
+  "mr",
+  "mrs",
+  "ms",
+  "dr",
+  "prof",
+  "st",
+  "jr",
+  "sr",
+  "vs",
+  "inc",
+  "co",
+  "corp",
+  "ltd",
+  "llc",
+  "approx",
+  "dept",
+  "jan",
+  "feb",
+  "mar",
+  "apr",
+  "jun",
+  "jul",
+  "aug",
+  "sep",
+  "sept",
+  "oct",
+  "nov",
+  "dec",
+  "gen",
+  "sen",
+  "gov",
+  "lt",
+  "col",
+  "capt",
+  "sgt",
+  "cpl",
+  "maj",
+  "adm",
+  "pres",
+  "atty",
+  "bros",
+  "mfg",
+  "assoc",
+  "univ",
+  "bldg",
+  "ste",
+  "ave",
+  "blvd",
+  "rd",
+  "ln",
+  "hwy",
+  "pkwy",
+  "pl"
+]);
+var CAPITALISED_ABBREVIATIONS = /* @__PURE__ */ new Set([
+  "Mon",
+  "Tue",
+  "Tues",
+  "Wed",
+  "Thu",
+  "Thur",
+  "Thurs",
+  "Fri",
+  "Sat",
+  "Sun",
+  "Mt",
+  "Ft",
+  "Ct",
+  "Rep",
+  "Rev",
+  "Hon"
+]);
+var BEFORE_NUMBER = /* @__PURE__ */ new Set([
+  "no",
+  "nos",
+  "fig",
+  "figs",
+  "vol",
+  "vols",
+  "ch",
+  "sec",
+  "pp",
+  "cf",
+  "ca",
+  "est",
+  "ext",
+  "tel",
+  "ph",
+  "ste",
+  "apt",
+  "bldg",
+  "rm",
+  "hwy",
+  "rt",
+  "rte"
+]);
+var DOTTED_SHORT = /^\d*(?:[A-Za-z]{1,2}\.)+[A-Za-z]{1,2}$/;
+var DOTTED_CAPS = /^(?:[A-Z][A-Za-z]*\.)+[A-Z][A-Za-z]*$/;
+var ABBREVIATION_RANGE = /[A-Za-z]\.[-–—/]/;
+var LIST_MARKER_ALONE = /^\s*(?:(?:step|tip|reason|rule|point|item|option)\s+)?\d{1,3}\.$/i;
+var LIST_MARKER_AFTER_COLON = /:\s+\d{1,3}\.$/;
+var LIST_MARKER_REACH = 24;
+var OPENERS = /* @__PURE__ */ new Set(['"', "'", "\u201C", "\u2018", "(", "["]);
+var CLOSERS = /* @__PURE__ */ new Set(['"', "'", "\u201D", "\u2019", ")", "]"]);
+var SPACE = /\s/;
+var LETTER = /\p{L}/u;
+var DIGIT = /[0-9]/;
+var SENTENCE_START = /[A-Z0-9]/;
+var isStop = (c) => c === "." || c === "!" || c === "?";
+var isSpace = (c) => SPACE.test(c);
+function startsSentence(source, at) {
+  const c = OPENERS.has(source[at]) ? source[at + 1] : source[at];
+  return c !== void 0 && SENTENCE_START.test(c);
+}
+function digitFollows(source, at) {
+  const c = OPENERS.has(source[at]) ? source[at + 1] : source[at];
+  return c !== void 0 && DIGIT.test(c);
+}
+function isDottedAbbreviation(bare) {
+  const compound = Math.max(bare.lastIndexOf("-"), bare.lastIndexOf("\u2013"), bare.lastIndexOf("\u2014"), bare.lastIndexOf("/"));
+  const last = bare.slice(compound + 1);
+  return DOTTED_SHORT.test(last) || DOTTED_CAPS.test(last) || ABBREVIATION_RANGE.test(bare);
+}
+function endsWithAbbreviation(source, end, digitNext) {
+  let from = end;
+  while (from > 0 && !isSpace(source[from - 1])) from--;
+  let to = end;
+  while (from < to && OPENERS.has(source[from])) from++;
+  while (to > from && CLOSERS.has(source[to - 1])) to--;
+  while (to > from && source[to - 1] === ".") to--;
+  const bare = source.slice(from, to);
+  const lower = bare.toLowerCase();
+  return isDottedAbbreviation(bare) || /^[A-Za-z]$/.test(bare) || ABBREVIATIONS.has(lower) || CAPITALISED_ABBREVIATIONS.has(bare) || digitNext && BEFORE_NUMBER.has(lower);
+}
+function endsWithListMarker(source, start, end) {
+  const from = Math.max(start, end - LIST_MARKER_REACH);
+  const tail = source.slice(from, end);
+  return from === start && LIST_MARKER_ALONE.test(tail) || LIST_MARKER_AFTER_COLON.test(tail);
+}
+function splitSentences(text) {
+  const source = (text ?? "").trim();
+  if (!source) return [];
+  const sentences = [];
+  let start = 0;
+  let scanned = 0;
+  let sawLetter = false;
+  let i = 0;
+  while (i < source.length) {
+    if (!isStop(source[i])) {
+      i++;
+      continue;
+    }
+    let endsStrong = false;
+    while (i < source.length && isStop(source[i])) {
+      if (source[i] !== ".") endsStrong = true;
+      i++;
+    }
+    while (i < source.length && CLOSERS.has(source[i])) i++;
+    let next = i;
+    while (next < source.length && isSpace(source[next])) next++;
+    if (next === i || !startsSentence(source, next)) continue;
+    if (!sawLetter) sawLetter = LETTER.test(source.slice(scanned, i));
+    scanned = i;
+    if (!sawLetter) continue;
+    if (!endsStrong && (endsWithAbbreviation(source, i, digitFollows(source, next)) || endsWithListMarker(source, start, i))) continue;
+    sentences.push(source.slice(start, i).trim());
+    start = i;
+    sawLetter = false;
+  }
+  const rest = source.slice(start).trim();
+  if (rest) sentences.push(rest);
+  return sentences;
+}
+
 // src/contracts/llms.ts
 var LLM_GEO_CONTRACT_VERSION = 1;
 var LLMS_PUBLIC_SUMMARY_MAX_LENGTH = 400;
@@ -18,8 +637,8 @@ var MANAGED_LLM_SCHEMA_KNOWN_KEYS = /* @__PURE__ */ new Set([
 ]);
 var SENTENCE_SHARE = 0.4;
 var ELLIPSIS = "\u2026";
-var CLOSERS = /* @__PURE__ */ new Set(['"', "'", ")", "]", "\u201D", "\u2019", "\xBB"]);
-var ABBREVIATIONS = /* @__PURE__ */ new Set([
+var CLOSERS2 = /* @__PURE__ */ new Set(['"', "'", ")", "]", "\u201D", "\u2019", "\xBB"]);
+var ABBREVIATIONS2 = /* @__PURE__ */ new Set([
   "mr",
   "mrs",
   "ms",
@@ -56,7 +675,7 @@ function clipAtWordBoundary(text, max) {
 function endsSentence(flat, end) {
   if (flat[end] !== " ") return false;
   let i = end - 1;
-  while (i > 0 && CLOSERS.has(flat[i])) i--;
+  while (i > 0 && CLOSERS2.has(flat[i])) i--;
   const terminator = flat[i];
   if (terminator === "!" || terminator === "?") return true;
   return terminator === "." && !isAbbreviation(flat, i);
@@ -66,16 +685,43 @@ function isAbbreviation(flat, dot) {
   while (start > 0 && flat[start - 1] !== " ") start--;
   const token = flat.slice(start, dot).replace(/^[("'[\u201c\u2018]+/, "").toLowerCase();
   if (token.includes(".") || /^[a-z]$/.test(token)) return true;
-  return ABBREVIATIONS.has(token);
+  return ABBREVIATIONS2.has(token);
+}
+var PHONE_SOURCE = String.raw`(?:\+\d{1,3}[\s.-]?)?(?:\(\d{3}\)\s?|\b\d{3}[\s.-])\d{3}[\s.-]\d{4}\b` + String.raw`|\+\d[\d\s().-]{8,}\d` + String.raw`|\b\d{10}\b`;
+var EMAIL_SOURCE = String.raw`[A-Za-z0-9._%+-]+@[A-Za-z0-9-]+(?:\.[A-Za-z0-9-]+)*\.[A-Za-z]{2,}`;
+var CONTACT_DETAIL = new RegExp(`${EMAIL_SOURCE}|${PHONE_SOURCE}`);
+var CONTACT_LEAD_IN = /\b(?:call|text|phone|telephone|tel|dial|e-?mail|contact|reach|whatsapp)\b[^.!?]{0,40}$/i;
+var HANGING_END = /(?:[\s,;:\u2013\u2014-]+|\s(?:and|or|at|on|by|via|to|from|with|using))$/i;
+function stripContactDetails(text) {
+  if (!CONTACT_DETAIL.test(text)) return text;
+  const kept = [];
+  for (const sentence of splitSentences(text)) {
+    const hit = CONTACT_DETAIL.exec(sentence);
+    if (!hit) {
+      kept.push(sentence);
+      continue;
+    }
+    let before = sentence.slice(0, hit.index);
+    const lead = CONTACT_LEAD_IN.exec(before);
+    if (lead) before = before.slice(0, lead.index);
+    for (let pass = 0; pass < 4 && HANGING_END.test(before); pass++) before = before.replace(HANGING_END, "");
+    before = before.trim();
+    if (before.split(/\s+/).length >= 4) kept.push(`${before}.`);
+  }
+  return kept.join(" ");
+}
+var PAGE_FILLER = /^(?:this|the)\s+(?:[\w'\u2019-]+\s+){0,2}?(?:page|url|post|article|resource|guide|section|listing|case study)\s+(?:also\s+)?(?:covers|presents|describes|features|explains|provides|outlines|highlights|offers|details|shows|introduces|lists|includes|compares|answers|walks through|breaks down)\s+/i;
+function stripPageFiller(note) {
+  const rest = note.replace(PAGE_FILLER, "");
+  if (rest === note || rest.length < 20) return note;
+  return `${rest.charAt(0).toUpperCase()}${rest.slice(1)}`;
 }
 function sanitizeLlmsPublicSummary(raw) {
   if (raw === null || raw === void 0) return null;
   if (typeof raw !== "string") return null;
-  let s = raw.trim().replace(/\s+/g, " ");
-  s = s.replace(/[\[\]\r\n]/g, " ").replace(/\s+/g, " ").trim();
+  let s = raw.replace(/[\[\]\r\n]/g, " ").replace(/\s+/g, " ").trim();
+  s = stripPageFiller(stripContactDetails(s));
   s = clipAtWordBoundary(s, LLMS_PUBLIC_SUMMARY_MAX_LENGTH);
-  if (/\b[\w.%+-]+@[\w.-]+\.[A-Za-z]{2,}\b/.test(s)) return null;
-  if (/\b\+?\d[\d\s().-]{8,}\b/.test(s)) return null;
   if (s.length < 8) return null;
   return s;
 }
@@ -141,5 +787,339 @@ function pickManagedLlmSchemaForJsonLd(schema) {
   }
   return Object.keys(out).length ? out : null;
 }
+var LLMS_DESCRIPTION_MAX_LENGTH = 500;
+var HARD_JOIN = "";
+var SOFT_JOIN = "";
+var SCRAPE_READ_LIMIT = 6e3;
+var FORM_LABEL = String.raw`(?:first name|last name|full name|your name|name|e-?mail(?: address)?|your e-?mail|phone(?: number)?|mobile(?: number)?|message|your message|subject|company(?: name)?|address|city|state|zip(?: code)?|comments?|how can we help(?: you)?\??|preferred (?:contact method|date|time)|service (?:needed|requested)|submit|send(?: message)?|send us a message|get (?:a )?(?:free )?quote|request (?:a )?(?:free )?quote|required)`;
+var FORM_RUN = new RegExp(String.raw`(?:\b${FORM_LABEL}\b[\s:*.]*){3,}`, "gi");
+var STARRED_LABEL = new RegExp(String.raw`\b${FORM_LABEL}\b\s*\*`, "gi");
+var NAV_WORD = String.raw`(?:Home|About(?: Us)?|Services|Contact(?: Us)?|Blog|Pricing|Resources|Work|Portfolio|Careers|Login|Log In|Sign In|Sign Up|FAQs?|Products|Gallery|Menu|Search|Cart|Shop|Reviews|Locations|Our Work|Our Services|Get a Quote|Book Now|Request a Quote)`;
+var NAV_RUN = new RegExp(String.raw`(?:\b${NAV_WORD}\b[\s|\u2022\u00b7/>-]*){4,}`, "g");
+var SKIP_AND_TOGGLE = /\b(?:skip to (?:main )?content|(?:toggle|open|close) (?:the )?(?:main )?(?:navigation|menu))\b/gi;
+var COOKIE_BANNER = /\bcookies?\b.*\b(?:accept|consent|agree|settings|preferences|policy)\b|\b(?:accept|consent|agree)\b.*\bcookies?\b/i;
+var CTA_BUTTON = String.raw`(?:book (?:a )?(?:free )?(?:demo|call|consultation)|read (?:the )?docs|read more|learn more|get started|get (?:a )?free (?:quote|audit|estimate)|call (?:us )?(?:now|today)|see (?:our )?work|view all|view (?:the )?demo|sign up|watch (?:the )?(?:video|demo))`;
+var CHROME_PIECE = new RegExp(String.raw`^(?:${FORM_LABEL}|${NAV_WORD}|${CTA_BUTTON}|skip to (?:main )?content)\s*\*?\.?$`, "i");
+var EYEBROW_MARK = /[\u00b7\u2022|\u203a\u00bb]/;
+function wordCount(text) {
+  const trimmed = text.trim();
+  return trimmed ? trimmed.split(/\s+/).length : 0;
+}
+function camelCaseTerms(texts) {
+  const found = /* @__PURE__ */ new Set();
+  for (const text of texts) {
+    if (typeof text !== "string") continue;
+    for (const token of text.split(/\s+/)) {
+      const bare = token.replace(/^[^A-Za-z0-9]+|[^A-Za-z0-9]+$/g, "");
+      if (/[a-z][A-Z]/.test(bare)) found.add(bare);
+    }
+  }
+  return [...found];
+}
+var JOIN_WORDS = /* @__PURE__ */ new Set(["first", "last", "name", "email", "phone", "message", "subject", "company", "address", "city", "state", "zip", "submit", "send"]);
+function isElementJoin(left, right) {
+  if (/^[a-z]$/.test(left)) return false;
+  if (JOIN_WORDS.has(left.toLowerCase()) || JOIN_WORDS.has(right.toLowerCase())) return true;
+  if (/^[a-z]/.test(left)) return left.length >= 3;
+  return left.length >= 8 || right.length >= 8;
+}
+function markGluedJoins(token, known) {
+  if (token.length < 5) return token;
+  if (known.has(token.replace(/^[^A-Za-z0-9]+|[^A-Za-z0-9]+$/g, ""))) return token;
+  return token.replace(/([.!?*:)\]])(?=[A-Z][a-z]{2,})/g, `$1${HARD_JOIN}`).replace(/([A-Z]{2,})(?=[A-Z][a-z]{2,})/g, `$1${SOFT_JOIN}`).replace(
+    /([A-Z]?[a-z]+)(?=([A-Z][a-z]+))/g,
+    (match, left, right) => isElementJoin(left, right) ? `${left}${HARD_JOIN}` : match
+  );
+}
+function isPageChrome(piece) {
+  if (CHROME_PIECE.test(piece)) return true;
+  if (piece.endsWith("*") && wordCount(piece) <= 4) return true;
+  return EYEBROW_MARK.test(piece) && piece.length <= 70 && wordCount(piece) <= 9;
+}
+function cleanScrapedPageText(raw, options = {}) {
+  if (typeof raw !== "string" || !raw.trim()) return "";
+  const known = new Set(options.knownTerms ?? []);
+  const marked = raw.slice(0, SCRAPE_READ_LIMIT).replace(/[\u00a0\u200b\u2060\ufeff\r]/g, " ").replace(/\n+/g, HARD_JOIN).replace(/\S+/g, (token) => markGluedJoins(token, known)).replace(new RegExp(SOFT_JOIN, "g"), " ");
+  const pieces = marked.split(HARD_JOIN).map((piece) => piece.replace(/\s+/g, " ").trim()).filter((piece) => piece && !isPageChrome(piece));
+  let joined = "";
+  for (const piece of pieces) {
+    if (!joined) {
+      joined = piece;
+      continue;
+    }
+    const clause = joined.slice(joined.search(/[^.!?]*$/));
+    const open = !/[.!?:;,"')\]]$/.test(joined);
+    joined += open && wordCount(clause) >= 3 && /^[A-Z]/.test(piece) ? `. ${piece}` : ` ${piece}`;
+  }
+  const text = joined.replace(SKIP_AND_TOGGLE, " ").replace(FORM_RUN, " ").replace(STARRED_LABEL, " ").replace(NAV_RUN, " ").replace(/\s+/g, " ").trim();
+  const cleaned = splitSentences(text).filter((sentence) => !COOKIE_BANNER.test(sentence)).join(" ").replace(/\s+([.,;:!?])/g, "$1").replace(/([.!?])(?:\s*[.!?])+/g, "$1").trim();
+  return cleaned.length >= 20 && wordCount(cleaned) >= 3 ? cleaned : "";
+}
+function plainNote(raw) {
+  return typeof raw === "string" ? clipAtWordBoundary(raw.replace(/[\[\]\r\n]/g, " "), LLMS_DESCRIPTION_MAX_LENGTH) : "";
+}
+function sameWords(a, b) {
+  const norm = (s) => s.toLowerCase().replace(/[^\p{L}\p{N}]+/gu, " ").trim();
+  return norm(a) === norm(b);
+}
+function isDescriptive(note, title) {
+  if (note.length < 15 || wordCount(note) < 3) return false;
+  if (title && sameWords(note, title)) return false;
+  return !/^(?:undefined|null|n\/a|none|lorem ipsum)\b/i.test(note);
+}
+function pickLlmsPageDescription(input) {
+  const title = plainNote(input.title);
+  const withheld = Boolean(input.copyWithheld);
+  const known = camelCaseTerms([
+    input.title,
+    input.llms_public_summary,
+    input.managed_meta_description,
+    input.meta_description,
+    input.llm_schema_description
+  ]);
+  const meta = plainNote(input.meta_description);
+  const candidates = [
+    ["summary", () => withheld ? "" : sanitizeLlmsPublicSummary(input.llms_public_summary) ?? ""],
+    ["managed_meta", () => withheld ? "" : plainNote(input.managed_meta_description)],
+    ["meta", () => withheld || looksLikeErrorPage(meta) ? "" : meta],
+    ["schema", () => withheld ? "" : plainNote(stripPageFiller(input.llm_schema_description ?? ""))],
+    [
+      "scrape",
+      () => plainNote(stripContactDetails(cleanScrapedPageText(usablePageText(input.content_text), { knownTerms: known })))
+    ]
+  ];
+  for (const [source, note] of candidates) {
+    const text = note();
+    if (isDescriptive(text, title)) return { text, source };
+  }
+  return null;
+}
+function resolveLlmsPageDescription(input) {
+  return pickLlmsPageDescription(input)?.text ?? null;
+}
+function llmsPageListNote(page, publicSummaryOnly = false) {
+  const summary = sanitizeLlmsPublicSummary(page.llms_public_summary);
+  if (summary) return summary;
+  if (publicSummaryOnly) return void 0;
+  return plainNote(page.description) || void 0;
+}
+function pagePathSegments(path) {
+  return path.replace(/^https?:\/\/[^/]+/i, "").split(/[?#]/, 1)[0].split("/").filter(Boolean);
+}
+var UTILITY_SLUGS = /* @__PURE__ */ new Set([
+  "thank-you",
+  "thanks",
+  "thankyou",
+  "confirmation",
+  "order-confirmation",
+  "cart",
+  "checkout",
+  "login",
+  "log-in",
+  "signin",
+  "sign-in",
+  "register",
+  "search",
+  "404",
+  "500",
+  "not-found",
+  "offline",
+  "preview",
+  "draft",
+  "unsubscribe",
+  "wp-admin",
+  "wp-login",
+  "admin"
+]);
+function isLlmsListablePage(page) {
+  if (typeof page.http_status === "number" && page.http_status >= 400) return false;
+  const robots = (page.managed_robots_meta || page.robots_meta || "").toLowerCase();
+  if (/\bnoindex\b|\bnone\b/.test(robots)) return false;
+  if (!robots && page.has_noindex === true) return false;
+  const segments = pagePathSegments(page.path);
+  return !segments.some((segment) => UTILITY_SLUGS.has(segment.toLowerCase()));
+}
+function llmsPageTitle(page, homeName) {
+  for (const candidate of [page.managed_title, page.title, page.h1]) {
+    const head = titleHead(candidate);
+    if (head && isPageName(head, page.path) && !looksLikeErrorPage(head)) {
+      return clipAtWordBoundary(head, 120);
+    }
+  }
+  if (pagePathSegments(page.path).length === 0) return homeName?.trim() || "Home";
+  return slugLabel(page.path) || page.path;
+}
+function selectLlmsPages(pages, cap) {
+  if (cap <= 0) return [];
+  if (pages.length <= cap) return pages;
+  const clusters = /* @__PURE__ */ new Map();
+  for (const row of pages) {
+    const segment = row.path.replace(/^\//, "").split("/")[0] || "";
+    const cluster = `/${segment}`;
+    const bucket = clusters.get(cluster);
+    if (bucket) bucket.push(row);
+    else clusters.set(cluster, [row]);
+  }
+  const rank = ([cluster]) => {
+    const segment = cluster.slice(1);
+    if (!segment) return -1;
+    return LEGAL_SLUGS.has(segment.toLowerCase()) ? SECTION_VALUE_ORDER.length + 1 : valueRank(segment);
+  };
+  const selected = [];
+  const buckets = [...clusters.entries()].sort((a, b) => rank(a) - rank(b)).map(([, bucket]) => bucket);
+  for (let depth = 0; selected.length < cap; depth++) {
+    let placedAny = false;
+    for (const bucket of buckets) {
+      if (depth >= bucket.length) continue;
+      selected.push(bucket[depth]);
+      placedAny = true;
+      if (selected.length >= cap) break;
+    }
+    if (!placedAny) break;
+  }
+  const order = new Map(pages.map((row, index) => [row.path, index]));
+  selected.sort((a, b) => (order.get(a.path) ?? 0) - (order.get(b.path) ?? 0));
+  return selected;
+}
+var SECTION_VALUE_ORDER = [
+  "services",
+  "service",
+  "what-we-do",
+  "solutions",
+  "products",
+  "product",
+  "features",
+  "shop",
+  "locations",
+  "location",
+  "service-area",
+  "service-areas",
+  "areas-served",
+  "industries",
+  "work",
+  "portfolio",
+  "case-studies",
+  "projects",
+  "reviews",
+  "testimonials",
+  "about",
+  "about-us",
+  "team",
+  "pricing",
+  "plans",
+  "resources",
+  "guides",
+  "insights",
+  "blog",
+  "articles",
+  "news",
+  "faq",
+  "faqs",
+  "contact",
+  "contact-us"
+];
+var LEGAL_SLUGS = /* @__PURE__ */ new Set([
+  "privacy",
+  "privacy-policy",
+  "terms",
+  "terms-of-service",
+  "terms-and-conditions",
+  "terms-of-use",
+  "cookies",
+  "cookie-policy",
+  "accessibility",
+  "disclaimer",
+  "legal",
+  "sitemap",
+  "dmca",
+  "refund-policy",
+  "returns",
+  "shipping-policy"
+]);
+var FLAT_LIST_MAX = 12;
+function valueRank(segment) {
+  const at = SECTION_VALUE_ORDER.indexOf(segment.toLowerCase());
+  return at === -1 ? SECTION_VALUE_ORDER.length : at;
+}
+function groupLlmsPages(pages) {
+  const index = new Map(pages.map((page, i) => [page, i]));
+  const main = [];
+  const legal = [];
+  const sections = /* @__PURE__ */ new Map();
+  for (const page of pages) {
+    const segments = pagePathSegments(page.path);
+    if (segments.some((segment) => LEGAL_SLUGS.has(segment.toLowerCase())) && segments.length <= 2) legal.push(page);
+    else if (segments.length <= 1) main.push(page);
+    else {
+      const key = segments[0].toLowerCase();
+      const bucket = sections.get(key);
+      if (bucket) bucket.push(page);
+      else sections.set(key, [page]);
+    }
+  }
+  const byOriginal = (a, b) => (index.get(a) ?? 0) - (index.get(b) ?? 0);
+  const mainRank = (page) => {
+    const segments = pagePathSegments(page.path);
+    return segments.length === 0 ? -1 : valueRank(segments[0]);
+  };
+  main.sort((a, b) => mainRank(a) - mainRank(b) || byOriginal(a, b));
+  const grouped = [];
+  const strays = [];
+  for (const [key, bucket] of sections) {
+    if (bucket.length === 1) strays.push(bucket[0]);
+    else grouped.push([key, bucket.sort(byOriginal)]);
+  }
+  grouped.sort(([ka, a], [kb, b]) => valueRank(ka) - valueRank(kb) || b.length - a.length || ka.localeCompare(kb));
+  strays.sort((a, b) => valueRank(pagePathSegments(a.path)[0]) - valueRank(pagePathSegments(b.path)[0]) || byOriginal(a, b));
+  if (pages.length <= FLAT_LIST_MAX || grouped.length === 0) {
+    return [{ heading: null, pages: [...main, ...strays, ...grouped.flatMap(([, bucket]) => bucket), ...legal] }];
+  }
+  const groups = [];
+  if (main.length > 0) groups.push({ heading: "Key pages", pages: main });
+  for (const [key, bucket] of grouped) groups.push({ heading: slugLabel(key) || key, pages: bucket });
+  if (strays.length > 0) groups.push({ heading: "More pages", pages: strays });
+  if (legal.length > 0) groups.push({ heading: "Legal", pages: legal });
+  return groups;
+}
+function llmsBlockquoteSummary(business) {
+  const tagline = business.tagline?.replace(/\s+/g, " ").trim();
+  if (tagline) return tagline;
+  const first = splitSentences(business.description)[0];
+  return clipAtWordBoundary(first || business.name || "Website", 300);
+}
+function squash(value) {
+  return typeof value === "string" ? value.replace(/\s+/g, " ").trim() : "";
+}
+function hoursLine(hours) {
+  if (typeof hours === "string") return squash(hours);
+  if (!hours || typeof hours !== "object" || Array.isArray(hours)) return "";
+  return Object.entries(hours).map(([day, value]) => typeof value === "string" && squash(value) ? `${squash(day)} ${squash(value)}` : "").filter(Boolean).join("; ");
+}
+function llmsContactLines(contact) {
+  if (!contact) return [];
+  const lines = [];
+  const phone = squash(contact.phone);
+  const email = squash(contact.email);
+  if (phone) lines.push(`- **Phone:** ${phone}`);
+  if (email) lines.push(`- **Email:** ${email}`);
+  const regionAndPostal = [squash(contact.state), squash(contact.postal_code)].filter(Boolean).join(" ");
+  const country = squash(contact.country);
+  const address = [
+    squash(contact.address),
+    squash(contact.city),
+    regionAndPostal,
+    /^(?:us|usa|united states(?: of america)?)$/i.test(country) ? "" : country
+  ].filter(Boolean);
+  if (address.length > 0) lines.push(`- **Address:** ${address.join(", ")}`);
+  const hours = hoursLine(contact.hours);
+  if (hours) lines.push(`- **Hours:** ${hours}`);
+  return lines;
+}
+function llmsAbsoluteUrl(baseUrl, pathOrUrl) {
+  if (/^[a-z][a-z0-9+.-]*:/i.test(pathOrUrl)) return pathOrUrl;
+  const base = (baseUrl ?? "").trim().replace(/\/+$/, "");
+  if (!base) return pathOrUrl;
+  return `${base}${pathOrUrl.startsWith("/") ? "" : "/"}${pathOrUrl}`;
+}
 
-export { LLMS_DISCLAIMER_MAX_LENGTH, LLMS_PUBLIC_SUMMARY_MAX_LENGTH, LLM_GEO_CONTRACT_VERSION, MANAGED_LLM_SCHEMA_KNOWN_KEYS, clipAtWordBoundary, isEmittableReviewNode, pickManagedLlmSchemaForJsonLd, sanitizeLlmsDisclaimerLine, sanitizeLlmsPublicSummary, sanitizePrimaryLanguageTag };
+export { LLMS_DESCRIPTION_MAX_LENGTH, LLMS_DISCLAIMER_MAX_LENGTH, LLMS_PUBLIC_SUMMARY_MAX_LENGTH, LLM_GEO_CONTRACT_VERSION, MANAGED_LLM_SCHEMA_KNOWN_KEYS, cleanScrapedPageText, clipAtWordBoundary, groupLlmsPages, isEmittableReviewNode, isLlmsListablePage, llmsAbsoluteUrl, llmsBlockquoteSummary, llmsContactLines, llmsPageListNote, llmsPageTitle, pickLlmsPageDescription, pickManagedLlmSchemaForJsonLd, resolveLlmsPageDescription, sanitizeLlmsDisclaimerLine, sanitizeLlmsPublicSummary, sanitizePrimaryLanguageTag, selectLlmsPages, stripContactDetails, stripPageFiller };