@bagelink/blox 1.18.3 → 1.18.7

package/dist/{validate-output-3fNPagvj.js → validate-output-B8pmE7hf.js}

@@ -3,7 +3,7 @@ import fs, { readFile } from "node:fs/promises";
 import path, { resolve } from "node:path";
 import process$1 from "node:process";
 import { pathToFileURL } from "node:url";
-import { r as resolvePageTitle, p as primaryPageContext, c as contextPageTitle } from "./pageTitle-B1mtDJKn.js";
+import { r as resolvePageTitle, p as primaryPageContext, c as contextPageTitle } from "./localeStrategy-hXlNof1E.js";
 class BloxBuildResponseError extends Error {
   constructor(status, url, statusText) {
     super(`Build API ${status} ${statusText}: ${url}`);
@@ -72,6 +72,24 @@ async function bloxBuildJson(input, init = {}, options = {}) {
   if (response.status === 204) return null;
   return response.json();
 }
+function buildPageSlugMap(localizedBundles, defaultLocale) {
+  var _a;
+  const map = {};
+  const defaultPages = ((_a = localizedBundles.find((result) => result.locale === defaultLocale)) == null ? void 0 : _a.bundle.pages) ?? [];
+  const canonicalById = new Map(
+    defaultPages.flatMap((page) => page.id && page.slug && !page.slug.includes(":") ? [[page.id, page.slug]] : [])
+  );
+  for (const { locale, bundle } of localizedBundles) {
+    for (const page of bundle.pages) {
+      if (!page.slug || page.slug.includes(":") || page.status && page.status !== "published") continue;
+      const canonical = page.id ? canonicalById.get(page.id) : void 0;
+      if (!canonical) continue;
+      const entry = map[canonical] ?? (map[canonical] = { [defaultLocale]: canonical });
+      entry[locale] = page.slug;
+    }
+  }
+  return map;
+}
 function localizePath(slug, locale, defaultLocale) {
   if (locale === defaultLocale) return slug === "" ? "/" : slug;
   const s = slug === "" || slug === "/" ? "" : slug;
@@ -118,7 +136,7 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
         return [{
           from_path: fromPath,
           to_path: localizePath(page.slug, locale, defaultLocale2),
-          status_code: 302
+          status_code: 301
         }];
       });
     });
@@ -132,6 +150,7 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
       website: website2,
       paths: paths2,
       pageSlugs: [...new Set(localizedBundles.flatMap((result) => result.bundle.pages.flatMap((page) => page.slug ? [localizePath(page.slug, result.locale, defaultLocale2)] : [])))],
+      pageSlugMap: buildPageSlugMap(localizedBundles, defaultLocale2),
       redirects: redirects2,
       collections: extractCollectionRefs(localizedBundles.flatMap((result) => result.bundle.pages)),
       collectionItems
@@ -175,7 +194,7 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
       return [{
         from_path: fromPath,
         to_path: localizePath(page.slug, locale, defaultLocale),
-        status_code: 302
+        status_code: 301
       }];
     });
   });
@@ -191,7 +210,18 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
   const pages = localizedPages.flatMap((result) => result.pages);
   const collections = extractCollectionRefs(pages);
   const pageSlugs = [...new Set(localizedPages.flatMap((result) => result.pages.flatMap((page) => page.slug ? [localizePath(page.slug, result.locale, defaultLocale)] : [])))];
-  return { websiteId: site.id, website, paths, pageSlugs, redirects, collections };
+  return {
+    websiteId: site.id,
+    website,
+    paths,
+    pageSlugs,
+    pageSlugMap: buildPageSlugMap(
+      localizedPages.map((result) => ({ locale: result.locale, bundle: { pages: result.pages } })),
+      defaultLocale
+    ),
+    redirects,
+    collections
+  };
 }
 async function expandRoutes(apiBase, pages, requestOptions, locale, collectItems) {
   const routes = [];
@@ -992,6 +1022,7 @@ async function prerender({
   website = null,
   websiteId = "",
   redirects = [],
+  pageSlugMap = {},
   llm = {}
 } = {}) {
   var _a, _b, _c, _d, _e, _f;
@@ -1099,6 +1130,7 @@ async function prerender({
       template,
       website,
       websiteId,
+      pageSlugMap,
       resolveCache,
       llm
     });
@@ -1242,7 +1274,7 @@ ${existing}`,
     totalDiscovered: seen.size
   };
 }
-function normalizePath(p) {
+function normalizePath$1(p) {
   if (!p) return "/";
   try {
     if (p.startsWith("http://") || p.startsWith("https://")) {
@@ -1259,7 +1291,7 @@ function normalizePath(p) {
   return p || "/";
 }
 function enqueue(p, queue, seen, isExcluded) {
-  const np = normalizePath(p);
+  const np = normalizePath$1(p);
   if (seen.has(np)) return;
   if (isExcluded(np)) return;
   seen.add(np);
@@ -1268,7 +1300,7 @@ function enqueue(p, queue, seen, isExcluded) {
 function makeExcludeChecker(excludePaths) {
   const rules = Array.isArray(excludePaths) ? excludePaths : [excludePaths];
   return (p) => {
-    const np = normalizePath(p);
+    const np = normalizePath$1(p);
     for (const r of rules) {
       if (typeof r === "string" && r === "") continue;
       if (typeof r === "function") {
@@ -1276,7 +1308,7 @@ function makeExcludeChecker(excludePaths) {
       } else if (r instanceof RegExp) {
         if (r.test(np)) return true;
       } else if (typeof r === "string") {
-        const prefix = normalizePath(r);
+        const prefix = normalizePath$1(r);
         if (np === prefix || np.startsWith(`${prefix}/`)) return true;
       }
     }
@@ -1514,6 +1546,60 @@ async function validateInternalLinks(outDir) {
 ${failures.map((item) => `  - ${item}`).join("\n")}`);
   }
 }
+async function validateNoRedirectingLinks(outDir) {
+  const root = path.resolve(outDir);
+  const redirectSources = await readRedirectSources(root, "301", "302", "307", "308");
+  if (!redirectSources.size) return;
+  const failures = [];
+  for (const file of await collectHtmlFiles(root)) {
+    const name = path.relative(root, file);
+    const html = await fs.readFile(file, "utf8");
+    for (const href of internalHrefs(html)) {
+      const [withoutSuffix] = href.split(/[?#]/);
+      const normalized = normalizePath(withoutSuffix || "/");
+      if (redirectSources.has(normalized)) {
+        failures.push(`${name}: links to redirecting URL "${href}"`);
+      }
+    }
+  }
+  if (failures.length) {
+    throw new Error(
+      `Invalid Blox SSG output:
+${failures.slice(0, 25).map((item) => `  - ${item}`).join("\n")}`
+    );
+  }
+}
+async function readRedirectSources(root, ...statuses) {
+  const sources = /* @__PURE__ */ new Set();
+  let content = "";
+  try {
+    content = await fs.readFile(path.join(root, "_redirects"), "utf8");
+  } catch {
+    return sources;
+  }
+  for (const line of content.split("\n")) {
+    const trimmed = line.trim();
+    if (!trimmed || trimmed.startsWith("#")) continue;
+    const parts = trimmed.split(/\s+/);
+    if (parts.length < 3) continue;
+    const status = parts[2];
+    if (!statuses.includes(status)) continue;
+    sources.add(normalizePath(parts[0]));
+  }
+  return sources;
+}
+function normalizePath(value) {
+  const clean = value.split(/[?#]/)[0] || "/";
+  return clean !== "/" ? clean.replace(/\/+$/, "") : "/";
+}
+function* internalHrefs(html) {
+  const attr = /(?:<a\b[^>]*\shref=["']([^"']+)["'])|(?:data-to=["']([^"']+)["'])/gi;
+  let match;
+  while (match = attr.exec(html)) {
+    const value = match[1] ?? match[2] ?? "";
+    if (value.startsWith("/") && !value.startsWith("//")) yield value;
+  }
+}
 const anchorHref = /<a\b[^>]*\shref=["']([^"']*)["']/gi;
 const ABSOLUTE_OR_SPECIAL = /^(?:[a-z][a-z0-9+.-]*:|\/\/|\/|#|\?|$)/i;
 async function readSiteNoindex(outDir) {
@@ -1654,33 +1740,34 @@ async function validateSitemap(outDir, content) {
   return failures;
 }
 export {
-  splitSitemapPathsByLocale as A,
+  resolveManifestProvenance as A,
   BloxBuildResponseError as B,
-  llmArtifacts as C,
+  splitSitemapPathsByLocale as C,
+  llmArtifacts as D,
   prerender as a,
   buildCollectionResolver as b,
   validateInternalLinks as c,
-  buildPageHead as d,
-  primaryContextEntry as e,
+  validateNoRedirectingLinks as d,
+  buildPageHead as e,
   fetchCmsSiteData as f,
   generateSitemapArtifacts as g,
-  buildJsonLd as h,
-  extractPageText as i,
-  bloxBuildJson as j,
-  bloxBuildRequest as k,
-  buildSiteHead as l,
-  generateLlmsFullTxt as m,
-  generateLlmsTxt as n,
-  generateNetlifyRedirects as o,
+  primaryContextEntry as h,
+  buildJsonLd as i,
+  extractPageText as j,
+  bloxBuildJson as k,
+  bloxBuildRequest as l,
+  buildSiteHead as m,
+  generateLlmsFullTxt as n,
+  generateLlmsTxt as o,
   polyfillBloxSsgGlobals as p,
-  generateRobotsTxt as q,
-  generateSitemapIndexXml as r,
-  generateSitemapXml as s,
-  getBloxBuildRequestOptions as t,
+  generateNetlifyRedirects as q,
+  generateRobotsTxt as r,
+  generateSitemapIndexXml as s,
+  generateSitemapXml as t,
   uploadBloxManifest as u,
   validateSsgOutput as v,
-  getSiteIconFlags as w,
-  localeSitemapPath as x,
-  mergeLlmParts as y,
-  resolveManifestProvenance as z
+  getBloxBuildRequestOptions as w,
+  getSiteIconFlags as x,
+  localeSitemapPath as y,
+  mergeLlmParts as z
 };

package/dist/{validate-output-CrhXrC1l.cjs → validate-output-CbIQrCwa.cjs}

@@ -26,7 +26,7 @@ const fs = require("node:fs/promises");
 const path = require("node:path");
 const process$1 = require("node:process");
 const node_url = require("node:url");
-const pageTitle = require("./pageTitle-wobGhtiZ.cjs");
+const localeStrategy = require("./localeStrategy-OwRIn35C.cjs");
 class BloxBuildResponseError extends Error {
   constructor(status, url, statusText) {
     super(`Build API ${status} ${statusText}: ${url}`);
@@ -95,6 +95,24 @@ async function bloxBuildJson(input, init = {}, options = {}) {
   if (response.status === 204) return null;
   return response.json();
 }
+function buildPageSlugMap(localizedBundles, defaultLocale) {
+  var _a;
+  const map = {};
+  const defaultPages = ((_a = localizedBundles.find((result) => result.locale === defaultLocale)) == null ? void 0 : _a.bundle.pages) ?? [];
+  const canonicalById = new Map(
+    defaultPages.flatMap((page) => page.id && page.slug && !page.slug.includes(":") ? [[page.id, page.slug]] : [])
+  );
+  for (const { locale, bundle } of localizedBundles) {
+    for (const page of bundle.pages) {
+      if (!page.slug || page.slug.includes(":") || page.status && page.status !== "published") continue;
+      const canonical = page.id ? canonicalById.get(page.id) : void 0;
+      if (!canonical) continue;
+      const entry = map[canonical] ?? (map[canonical] = { [defaultLocale]: canonical });
+      entry[locale] = page.slug;
+    }
+  }
+  return map;
+}
 function localizePath(slug, locale, defaultLocale) {
   if (locale === defaultLocale) return slug === "" ? "/" : slug;
   const s = slug === "" || slug === "/" ? "" : slug;
@@ -141,7 +159,7 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
         return [{
           from_path: fromPath,
           to_path: localizePath(page.slug, locale, defaultLocale2),
-          status_code: 302
+          status_code: 301
         }];
       });
     });
@@ -155,6 +173,7 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
       website: website2,
       paths: paths2,
       pageSlugs: [...new Set(localizedBundles.flatMap((result) => result.bundle.pages.flatMap((page) => page.slug ? [localizePath(page.slug, result.locale, defaultLocale2)] : [])))],
+      pageSlugMap: buildPageSlugMap(localizedBundles, defaultLocale2),
       redirects: redirects2,
       collections: extractCollectionRefs(localizedBundles.flatMap((result) => result.bundle.pages)),
       collectionItems
@@ -198,7 +217,7 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
       return [{
         from_path: fromPath,
         to_path: localizePath(page.slug, locale, defaultLocale),
-        status_code: 302
+        status_code: 301
       }];
     });
   });
@@ -214,7 +233,18 @@ async function fetchCmsSiteData(apiBase, websiteName, requestOptions = {}) {
   const pages = localizedPages.flatMap((result) => result.pages);
   const collections = extractCollectionRefs(pages);
   const pageSlugs = [...new Set(localizedPages.flatMap((result) => result.pages.flatMap((page) => page.slug ? [localizePath(page.slug, result.locale, defaultLocale)] : [])))];
-  return { websiteId: site.id, website, paths, pageSlugs, redirects, collections };
+  return {
+    websiteId: site.id,
+    website,
+    paths,
+    pageSlugs,
+    pageSlugMap: buildPageSlugMap(
+      localizedPages.map((result) => ({ locale: result.locale, bundle: { pages: result.pages } })),
+      defaultLocale
+    ),
+    redirects,
+    collections
+  };
 }
 async function expandRoutes(apiBase, pages, requestOptions, locale, collectItems) {
   const routes = [];
@@ -444,7 +474,7 @@ function buildPageHead(options) {
   const page = resolvedData == null ? void 0 : resolvedData.page;
   const ctx = extractPrimaryContext(resolvedData == null ? void 0 : resolvedData.contexts);
   const siteName = seo.site_name || void 0;
-  const title = pageTitle.resolvePageTitle(resolvedData ?? {}, {
+  const title = localeStrategy.resolvePageTitle(resolvedData ?? {}, {
     override: override == null ? void 0 : override.title,
     siteName,
     defaultTitle: seo.default_meta_title
@@ -738,20 +768,20 @@ function getSiteIconFlags(website) {
 function extractPrimaryContext(contexts) {
   var _a;
   const empty = { title: null, description: null, image: null };
-  const ctx = (_a = pageTitle.primaryPageContext(contexts)) == null ? void 0 : _a.ctx;
+  const ctx = (_a = localeStrategy.primaryPageContext(contexts)) == null ? void 0 : _a.ctx;
   if (!ctx) return empty;
   const str2 = (key) => {
     const v = ctx[key];
     return typeof v === "string" && v.trim() ? v.trim() : null;
   };
   return {
-    title: pageTitle.contextPageTitle(ctx),
+    title: localeStrategy.contextPageTitle(ctx),
     description: str2("meta_description") || str2("excerpt") || str2("description") || str2("blurb") || str2("summary"),
     image: str2("og_image") || str2("cover_image_url") || str2("image_url") || str2("image")
   };
 }
 function primaryContextEntry(contexts) {
-  return pageTitle.primaryPageContext(contexts);
+  return localeStrategy.primaryPageContext(contexts);
 }
 function esc(s) {
   return s.replace(/&/g, "&amp;").replace(/"/g, "&quot;").replace(/</g, "&lt;").replace(/>/g, "&gt;");
@@ -1015,6 +1045,7 @@ async function prerender({
   website = null,
   websiteId = "",
   redirects = [],
+  pageSlugMap = {},
   llm = {}
 } = {}) {
   var _a, _b, _c, _d, _e, _f;
@@ -1122,6 +1153,7 @@ async function prerender({
       template,
       website,
       websiteId,
+      pageSlugMap,
       resolveCache,
       llm
     });
@@ -1265,7 +1297,7 @@ ${existing}`,
     totalDiscovered: seen.size
   };
 }
-function normalizePath(p) {
+function normalizePath$1(p) {
   if (!p) return "/";
   try {
     if (p.startsWith("http://") || p.startsWith("https://")) {
@@ -1282,7 +1314,7 @@ function normalizePath(p) {
   return p || "/";
 }
 function enqueue(p, queue, seen, isExcluded) {
-  const np = normalizePath(p);
+  const np = normalizePath$1(p);
   if (seen.has(np)) return;
   if (isExcluded(np)) return;
   seen.add(np);
@@ -1291,7 +1323,7 @@ function enqueue(p, queue, seen, isExcluded) {
 function makeExcludeChecker(excludePaths) {
   const rules = Array.isArray(excludePaths) ? excludePaths : [excludePaths];
   return (p) => {
-    const np = normalizePath(p);
+    const np = normalizePath$1(p);
     for (const r of rules) {
       if (typeof r === "string" && r === "") continue;
       if (typeof r === "function") {
@@ -1299,7 +1331,7 @@ function makeExcludeChecker(excludePaths) {
       } else if (r instanceof RegExp) {
         if (r.test(np)) return true;
       } else if (typeof r === "string") {
-        const prefix = normalizePath(r);
+        const prefix = normalizePath$1(r);
         if (np === prefix || np.startsWith(`${prefix}/`)) return true;
       }
     }
@@ -1537,6 +1569,60 @@ async function validateInternalLinks(outDir) {
 ${failures.map((item) => `  - ${item}`).join("\n")}`);
   }
 }
+async function validateNoRedirectingLinks(outDir) {
+  const root = path.resolve(outDir);
+  const redirectSources = await readRedirectSources(root, "301", "302", "307", "308");
+  if (!redirectSources.size) return;
+  const failures = [];
+  for (const file of await collectHtmlFiles(root)) {
+    const name = path.relative(root, file);
+    const html = await fs.readFile(file, "utf8");
+    for (const href of internalHrefs(html)) {
+      const [withoutSuffix] = href.split(/[?#]/);
+      const normalized = normalizePath(withoutSuffix || "/");
+      if (redirectSources.has(normalized)) {
+        failures.push(`${name}: links to redirecting URL "${href}"`);
+      }
+    }
+  }
+  if (failures.length) {
+    throw new Error(
+      `Invalid Blox SSG output:
+${failures.slice(0, 25).map((item) => `  - ${item}`).join("\n")}`
+    );
+  }
+}
+async function readRedirectSources(root, ...statuses) {
+  const sources = /* @__PURE__ */ new Set();
+  let content = "";
+  try {
+    content = await fs.readFile(path.join(root, "_redirects"), "utf8");
+  } catch {
+    return sources;
+  }
+  for (const line of content.split("\n")) {
+    const trimmed = line.trim();
+    if (!trimmed || trimmed.startsWith("#")) continue;
+    const parts = trimmed.split(/\s+/);
+    if (parts.length < 3) continue;
+    const status = parts[2];
+    if (!statuses.includes(status)) continue;
+    sources.add(normalizePath(parts[0]));
+  }
+  return sources;
+}
+function normalizePath(value) {
+  const clean = value.split(/[?#]/)[0] || "/";
+  return clean !== "/" ? clean.replace(/\/+$/, "") : "/";
+}
+function* internalHrefs(html) {
+  const attr = /(?:<a\b[^>]*\shref=["']([^"']+)["'])|(?:data-to=["']([^"']+)["'])/gi;
+  let match;
+  while (match = attr.exec(html)) {
+    const value = match[1] ?? match[2] ?? "";
+    if (value.startsWith("/") && !value.startsWith("//")) yield value;
+  }
+}
 const anchorHref = /<a\b[^>]*\shref=["']([^"']*)["']/gi;
 const ABSOLUTE_OR_SPECIAL = /^(?:[a-z][a-z0-9+.-]*:|\/\/|\/|#|\?|$)/i;
 async function readSiteNoindex(outDir) {
@@ -1704,4 +1790,5 @@ exports.resolveManifestProvenance = resolveManifestProvenance;
 exports.splitSitemapPathsByLocale = splitSitemapPathsByLocale;
 exports.uploadBloxManifest = uploadBloxManifest;
 exports.validateInternalLinks = validateInternalLinks;
+exports.validateNoRedirectingLinks = validateNoRedirectingLinks;
 exports.validateSsgOutput = validateSsgOutput;