@nuxtjs/sitemap 8.3.0 → 8.3.2

package/dist/runtime/server/sitemap/urlset/normalise.js

@@ -144,13 +144,22 @@ export function normaliseEntry(_e, defaults, resolvers, cache) {
   return e;
 }
 const IS_VALID_W3C_DATE = [
-  /(\d{4}-[01]\d-[0-3]\dT[0-2]\d:[0-5]\d:[0-5]\d\.\d+([+-][0-2]\d:[0-5]\d|Z))|(\d{4}-[01]\d-[0-3]\dT[0-2]\d:[0-5]\d:[0-5]\d([+-][0-2]\d:[0-5]\d|Z))|(\d{4}-[01]\d-[0-3]\dT[0-2]\d:[0-5]\d([+-][0-2]\d:[0-5]\d|Z))/,
+  /^\d{4}-[01]\d-[0-3]\dT[0-2]\d:[0-5]\d:[0-5]\d\.\d+([+-][0-2]\d:[0-5]\d|Z)$/,
+  /^\d{4}-[01]\d-[0-3]\dT[0-2]\d:[0-5]\d:[0-5]\d([+-][0-2]\d:[0-5]\d|Z)$/,
+  /^\d{4}-[01]\d-[0-3]\dT[0-2]\d:[0-5]\d([+-][0-2]\d:[0-5]\d|Z)$/,
   /^\d{4}-[01]\d-[0-3]\d$/,
   /^\d{4}-[01]\d$/,
   /^\d{4}$/
 ];
 export function isValidW3CDate(d) {
-  return IS_VALID_W3C_DATE.some((r) => r.test(d));
+  if (!IS_VALID_W3C_DATE.some((r) => r.test(d)))
+    return false;
+  const [year, month, day] = d.slice(0, 10).split("-").map(Number);
+  if (month !== void 0 && (month < 1 || month > 12))
+    return false;
+  if (day !== void 0 && (day < 1 || day > new Date(year, month, 0).getDate()))
+    return false;
+  return true;
 }
 export function normaliseDate(d) {
   if (typeof d === "string") {

package/dist/runtime/server/sitemap/urlset/sources.js

@@ -1,8 +1,39 @@
-import { parseSitemapXml } from "@nuxtjs/sitemap/utils";
 import { defu } from "defu";
 import { getRequestHost } from "h3";
+import { collectSitemap } from "sitemapd/parse";
 import { parseURL } from "ufo";
 import { logger } from "../../../utils-pure.js";
+const changeFrequencies = /* @__PURE__ */ new Set([
+  "always",
+  "hourly",
+  "daily",
+  "weekly",
+  "monthly",
+  "yearly",
+  "never"
+]);
+function readerEntryToSitemapInput(entry) {
+  const priority = entry.priority === void 0 ? void 0 : Number.parseFloat(entry.priority);
+  const changefreq = entry.changefreq && changeFrequencies.has(entry.changefreq) ? entry.changefreq : void 0;
+  return {
+    loc: entry.loc,
+    ...entry.lastmod ? { lastmod: entry.lastmod } : {},
+    ...changefreq ? { changefreq } : {},
+    ...priority !== void 0 && Number.isFinite(priority) ? { priority } : {},
+    ...entry.extensions?.alternatives ? { alternatives: entry.extensions.alternatives.map(({ hreflang, href }) => ({ hreflang, href })) } : {},
+    ...entry.extensions?.images ? {
+      images: entry.extensions.images.map((image) => ({
+        loc: image.loc,
+        ...image.caption ? { caption: image.caption } : {},
+        ...image.geoLocation ? { geo_location: image.geoLocation } : {},
+        ...image.title ? { title: image.title } : {},
+        ...image.license ? { license: image.license } : {}
+      }))
+    } : {},
+    ...entry.extensions?.videos ? { videos: entry.extensions.videos } : {},
+    ...entry.extensions?.news ? { news: entry.extensions.news } : {}
+  };
+}
 export function normalizeSourceInput(source) {
   if (typeof source === "string") {
     return { context: { name: "hook" }, fetch: source };
@@ -48,7 +79,9 @@ export async function fetchDataSource(input, event) {
   const abortRequestTimeout = setTimeout(() => timeoutController.abort(), timeout);
   try {
     let isMaybeErrorResponse = false;
-    const isXmlRequest = parseURL(url).pathname.endsWith(".xml");
+    const pathname = parseURL(url).pathname.toLowerCase();
+    const isGzUrl = pathname.endsWith(".gz");
+    const isXmlRequest = pathname.endsWith(".xml") || isGzUrl;
     const mergedHeaders = defu(
       options?.headers,
       {
@@ -58,7 +91,10 @@ export async function fetchDataSource(input, event) {
     );
     const fetchOptions = {
       ...options,
-      responseType: isXmlRequest ? "text" : "json",
+      // Fetch XML sources as raw bytes so we can detect and decompress a gzip body
+      // (either a `.gz` URL, or a server that serves gzip without Content-Encoding)
+      // before it's mangled by a UTF-8 text decode.
+      responseType: isXmlRequest ? "arrayBuffer" : "json",
       signal: timeoutController.signal,
       headers: mergedHeaders,
       // Use ofetch's built-in retry for external sources
@@ -84,11 +120,16 @@ export async function fetchDataSource(input, event) {
       };
     }
     let urls = [];
-    if (typeof res === "object") {
+    if (isXmlRequest) {
+      const bytes = res instanceof Uint8Array ? res : new Uint8Array(res);
+      const result = await collectSitemap(bytes);
+      if (result._tag !== "document")
+        throw new Error(result.issues.map((issue) => issue.message).join("; ") || "Invalid sitemap document");
+      if (result.document._tag !== "urlset")
+        throw new Error("Sitemap URL source must be a URL set, not a sitemap index");
+      urls = result.document.entries.map(readerEntryToSitemapInput);
+    } else if (typeof res === "object") {
       urls = res.urls || res;
-    } else if (typeof res === "string" && isXmlRequest) {
-      const result = await parseSitemapXml(res);
-      urls = result.urls;
     }
     return {
       ...input,

package/dist/runtime/utils-pure.js

@@ -1,13 +1,8 @@
-import { createConsola } from "consola";
 import { createDefu } from "defu";
-import { createFilter } from "nuxtseo-shared/utils";
+import { createFilter, createModuleLogger } from "nuxtseo-shared/utils";
 import { parseURL, withoutBase } from "ufo";
 export { createFilter } from "nuxtseo-shared/utils";
-export const logger = createConsola({
-  defaults: {
-    tag: "@nuxt/sitemap"
-  }
-});
+export const logger = createModuleLogger("@nuxt/sitemap");
 const XML_ENTITIES = {
   "&": "&",
   "<": "&lt;",

package/dist/utils.d.mts

@@ -1,14 +1,11 @@
-import { SitemapUrl, SitemapUrlInput } from '../dist/runtime/types.js';
+import { SitemapUrlInput, SitemapUrl } from '../dist/runtime/types.js';
 export * from '../dist/runtime/types.js';
+import { SitemapInput } from 'sitemapd/parse';
+export { CollectSitemapResult, ParseSitemapOptions, SitemapCompleteness, SitemapDocument, SitemapDocumentKind, SitemapExtensions, SitemapFormat, SitemapInput, SitemapIssue, SitemapIssueCode, SitemapParseEvent, SitemapReference, SitemapUrlRecord, collectSitemap, parseSitemap } from 'sitemapd/parse';
+export { SitemapDocumentLoadResult, SitemapDocumentLoader, SitemapLoadFailureCode, SitemapLoadRequest, SitemapLoadSource, SitemapReadOptions, SitemapReadResult, SitemapReader, SitemapReaderOptions, SitemapTargetAuthorization, SitemapTargetAuthorizer, SitemapWalkDocument, SitemapWalkDocumentVisitor, SitemapWalkFailure, SitemapWalkNonRetainedResult, SitemapWalkOptions, SitemapWalkPartialReason, SitemapWalkResult, SitemapWalkRetainedResult, createSitemapReader } from 'sitemapd';
+export { FetchDocumentLoaderOptions, SitemapFetch, createFetchDocumentLoader } from 'sitemapd/fetch';
 
-declare function parseHtmlExtractSitemapMeta(html: string, options?: {
-    images?: boolean;
-    videos?: boolean;
-    lastmod?: boolean;
-    alternatives?: boolean;
-    resolveUrl?: (s: string) => string;
-}): Partial<SitemapUrl> | null;
-
+/** @deprecated Use `SitemapIssue` with `parseSitemap` or `collectSitemap`. */
 interface SitemapWarning {
     type: 'validation';
     message: string;
@@ -18,16 +15,31 @@ interface SitemapWarning {
         value?: unknown;
     };
 }
+/** @deprecated Use the tagged `CollectSitemapResult` returned by `collectSitemap`. */
 interface SitemapParseResult {
     urls: SitemapUrlInput[];
     warnings: SitemapWarning[];
 }
+/** @deprecated Use `SitemapReference`. */
+interface SitemapIndexEntry {
+    loc: string;
+    lastmod?: string;
+}
+/** @deprecated Use the tagged `CollectSitemapResult` returned by `collectSitemap`. */
+interface SitemapIndexParseResult {
+    entries: SitemapIndexEntry[];
+    warnings: SitemapWarning[];
+}
+/** @deprecated Use `SitemapInput`. */
 type SitemapXmlChunk = string | Uint8Array;
-type SitemapXmlInput = SitemapXmlChunk | Iterable<SitemapXmlChunk> | AsyncIterable<SitemapXmlChunk> | ReadableStream<SitemapXmlChunk>;
+/** @deprecated Use `SitemapInput`. */
+type SitemapXmlInput = SitemapInput;
+/** @deprecated Use `ParseSitemapOptions`. */
 interface SitemapStreamOptions {
     maxEntryBytes?: number;
     maxBufferBytes?: number;
 }
+/** @deprecated Use `SitemapParseEvent`. */
 type SitemapXmlStreamEvent = {
     _tag: 'url';
     url: SitemapUrlInput;
@@ -35,6 +47,7 @@ type SitemapXmlStreamEvent = {
     _tag: 'warning';
     warning: SitemapWarning;
 };
+/** @deprecated Use `SitemapParseEvent`. */
 type SitemapIndexStreamEvent = {
     _tag: 'sitemap';
     sitemap: SitemapIndexEntry;
@@ -42,26 +55,52 @@ type SitemapIndexStreamEvent = {
     _tag: 'warning';
     warning: SitemapWarning;
 };
+/** @deprecated Use `SitemapDocumentKind`. */
 type SitemapKind = 'urlset' | 'index';
+/** @deprecated Use `SitemapParseEvent`. */
 type SitemapStreamEvent = {
     _tag: 'kind';
     kind: SitemapKind;
 } | SitemapXmlStreamEvent | SitemapIndexStreamEvent;
+/**
+ * @deprecated Use `parseSitemap` from `@nuxtjs/sitemap/utils`. Canonical
+ * streams emit `document`, `url`, `sitemap`, `issue`, and terminal `end`
+ * events. URL and sitemap payloads use `entry`.
+ */
 declare function parseSitemapStream(input: SitemapXmlInput, options?: SitemapStreamOptions): AsyncGenerator<SitemapStreamEvent>;
+/**
+ * @deprecated Use `parseSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * events whose document kind is `urlset`.
+ */
 declare function parseSitemapXmlStream(input: SitemapXmlInput, options?: SitemapStreamOptions): AsyncGenerator<SitemapXmlStreamEvent>;
+/**
+ * @deprecated Use `parseSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * events whose document kind is `index`.
+ */
 declare function parseSitemapIndexStream(input: SitemapXmlInput, options?: SitemapStreamOptions): AsyncGenerator<SitemapIndexStreamEvent>;
+/**
+ * @deprecated Use `collectSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * its tagged result.
+ */
 declare function parseSitemapXml(xml: string): Promise<SitemapParseResult>;
-
-interface SitemapIndexEntry {
-    loc: string;
-    lastmod?: string;
-}
-interface SitemapIndexParseResult {
-    entries: SitemapIndexEntry[];
-    warnings: SitemapWarning[];
-}
+/**
+ * @deprecated Use `collectSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * an `index` document result.
+ */
 declare function parseSitemapIndex(xml: string): Promise<SitemapIndexParseResult>;
+/**
+ * @deprecated Use `collectSitemap` from `@nuxtjs/sitemap/utils` and inspect
+ * the tagged document result.
+ */
 declare function isSitemapIndex(xml: string): boolean;
 
+declare function parseHtmlExtractSitemapMeta(html: string, options?: {
+    images?: boolean;
+    videos?: boolean;
+    lastmod?: boolean;
+    alternatives?: boolean;
+    resolveUrl?: (s: string) => string;
+}): Partial<SitemapUrl> | null;
+
 export { isSitemapIndex, parseHtmlExtractSitemapMeta, parseSitemapIndex, parseSitemapIndexStream, parseSitemapStream, parseSitemapXml, parseSitemapXmlStream };
 export type { SitemapIndexEntry, SitemapIndexParseResult, SitemapIndexStreamEvent, SitemapKind, SitemapParseResult, SitemapStreamEvent, SitemapStreamOptions, SitemapWarning, SitemapXmlChunk, SitemapXmlInput, SitemapXmlStreamEvent };

package/dist/utils.d.ts

@@ -1,14 +1,11 @@
-import { SitemapUrl, SitemapUrlInput } from '../dist/runtime/types.js';
+import { SitemapUrlInput, SitemapUrl } from '../dist/runtime/types.js';
 export * from '../dist/runtime/types.js';
+import { SitemapInput } from 'sitemapd/parse';
+export { CollectSitemapResult, ParseSitemapOptions, SitemapCompleteness, SitemapDocument, SitemapDocumentKind, SitemapExtensions, SitemapFormat, SitemapInput, SitemapIssue, SitemapIssueCode, SitemapParseEvent, SitemapReference, SitemapUrlRecord, collectSitemap, parseSitemap } from 'sitemapd/parse';
+export { SitemapDocumentLoadResult, SitemapDocumentLoader, SitemapLoadFailureCode, SitemapLoadRequest, SitemapLoadSource, SitemapReadOptions, SitemapReadResult, SitemapReader, SitemapReaderOptions, SitemapTargetAuthorization, SitemapTargetAuthorizer, SitemapWalkDocument, SitemapWalkDocumentVisitor, SitemapWalkFailure, SitemapWalkNonRetainedResult, SitemapWalkOptions, SitemapWalkPartialReason, SitemapWalkResult, SitemapWalkRetainedResult, createSitemapReader } from 'sitemapd';
+export { FetchDocumentLoaderOptions, SitemapFetch, createFetchDocumentLoader } from 'sitemapd/fetch';
 
-declare function parseHtmlExtractSitemapMeta(html: string, options?: {
-    images?: boolean;
-    videos?: boolean;
-    lastmod?: boolean;
-    alternatives?: boolean;
-    resolveUrl?: (s: string) => string;
-}): Partial<SitemapUrl> | null;
-
+/** @deprecated Use `SitemapIssue` with `parseSitemap` or `collectSitemap`. */
 interface SitemapWarning {
     type: 'validation';
     message: string;
@@ -18,16 +15,31 @@ interface SitemapWarning {
         value?: unknown;
     };
 }
+/** @deprecated Use the tagged `CollectSitemapResult` returned by `collectSitemap`. */
 interface SitemapParseResult {
     urls: SitemapUrlInput[];
     warnings: SitemapWarning[];
 }
+/** @deprecated Use `SitemapReference`. */
+interface SitemapIndexEntry {
+    loc: string;
+    lastmod?: string;
+}
+/** @deprecated Use the tagged `CollectSitemapResult` returned by `collectSitemap`. */
+interface SitemapIndexParseResult {
+    entries: SitemapIndexEntry[];
+    warnings: SitemapWarning[];
+}
+/** @deprecated Use `SitemapInput`. */
 type SitemapXmlChunk = string | Uint8Array;
-type SitemapXmlInput = SitemapXmlChunk | Iterable<SitemapXmlChunk> | AsyncIterable<SitemapXmlChunk> | ReadableStream<SitemapXmlChunk>;
+/** @deprecated Use `SitemapInput`. */
+type SitemapXmlInput = SitemapInput;
+/** @deprecated Use `ParseSitemapOptions`. */
 interface SitemapStreamOptions {
     maxEntryBytes?: number;
     maxBufferBytes?: number;
 }
+/** @deprecated Use `SitemapParseEvent`. */
 type SitemapXmlStreamEvent = {
     _tag: 'url';
     url: SitemapUrlInput;
@@ -35,6 +47,7 @@ type SitemapXmlStreamEvent = {
     _tag: 'warning';
     warning: SitemapWarning;
 };
+/** @deprecated Use `SitemapParseEvent`. */
 type SitemapIndexStreamEvent = {
     _tag: 'sitemap';
     sitemap: SitemapIndexEntry;
@@ -42,26 +55,52 @@ type SitemapIndexStreamEvent = {
     _tag: 'warning';
     warning: SitemapWarning;
 };
+/** @deprecated Use `SitemapDocumentKind`. */
 type SitemapKind = 'urlset' | 'index';
+/** @deprecated Use `SitemapParseEvent`. */
 type SitemapStreamEvent = {
     _tag: 'kind';
     kind: SitemapKind;
 } | SitemapXmlStreamEvent | SitemapIndexStreamEvent;
+/**
+ * @deprecated Use `parseSitemap` from `@nuxtjs/sitemap/utils`. Canonical
+ * streams emit `document`, `url`, `sitemap`, `issue`, and terminal `end`
+ * events. URL and sitemap payloads use `entry`.
+ */
 declare function parseSitemapStream(input: SitemapXmlInput, options?: SitemapStreamOptions): AsyncGenerator<SitemapStreamEvent>;
+/**
+ * @deprecated Use `parseSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * events whose document kind is `urlset`.
+ */
 declare function parseSitemapXmlStream(input: SitemapXmlInput, options?: SitemapStreamOptions): AsyncGenerator<SitemapXmlStreamEvent>;
+/**
+ * @deprecated Use `parseSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * events whose document kind is `index`.
+ */
 declare function parseSitemapIndexStream(input: SitemapXmlInput, options?: SitemapStreamOptions): AsyncGenerator<SitemapIndexStreamEvent>;
+/**
+ * @deprecated Use `collectSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * its tagged result.
+ */
 declare function parseSitemapXml(xml: string): Promise<SitemapParseResult>;
-
-interface SitemapIndexEntry {
-    loc: string;
-    lastmod?: string;
-}
-interface SitemapIndexParseResult {
-    entries: SitemapIndexEntry[];
-    warnings: SitemapWarning[];
-}
+/**
+ * @deprecated Use `collectSitemap` from `@nuxtjs/sitemap/utils` and handle
+ * an `index` document result.
+ */
 declare function parseSitemapIndex(xml: string): Promise<SitemapIndexParseResult>;
+/**
+ * @deprecated Use `collectSitemap` from `@nuxtjs/sitemap/utils` and inspect
+ * the tagged document result.
+ */
 declare function isSitemapIndex(xml: string): boolean;
 
+declare function parseHtmlExtractSitemapMeta(html: string, options?: {
+    images?: boolean;
+    videos?: boolean;
+    lastmod?: boolean;
+    alternatives?: boolean;
+    resolveUrl?: (s: string) => string;
+}): Partial<SitemapUrl> | null;
+
 export { isSitemapIndex, parseHtmlExtractSitemapMeta, parseSitemapIndex, parseSitemapIndexStream, parseSitemapStream, parseSitemapXml, parseSitemapXmlStream };
 export type { SitemapIndexEntry, SitemapIndexParseResult, SitemapIndexStreamEvent, SitemapKind, SitemapParseResult, SitemapStreamEvent, SitemapStreamOptions, SitemapWarning, SitemapXmlChunk, SitemapXmlInput, SitemapXmlStreamEvent };