@griddo/cx 11.9.12-rc.0 → 11.9.13

package/exporter/{services → utils}/pages.ts

@@ -1,4 +1,3 @@
-import type { Core, Fields } from "@griddo/core";
 import type {
 	APIPageObject,
 	GriddoListPage,
@@ -8,10 +7,15 @@ import type {
 	MultiPageElement,
 	MultiPageElements,
 	PageAdditionalInfo,
-} from "../shared/types/pages";
-import type { TemplateWithReferenceField } from "../shared/types/templates";
+} from "../types/pages";
+import type { TemplateWithReferenceField } from "../types/templates";
+import type { Core, Fields } from "@griddo/core";
+
+import dotenv from "dotenv";
 
-import { formatImage } from "../core/images";
+import { formatImage } from "./images";
+
+dotenv.config();
 
 // Consts
 const DEFAULT_ITEMS_PER_PAGE_FOR_LIST_TEMPLATES = 25;
@@ -27,7 +31,10 @@ function getOpenGraph({
 	socialTitle,
 	socialDescription,
 	socialImage,
-}: Pick<Core.Page, "socialTitle" | "socialDescription" | "socialImage">): Core.Page["openGraph"] {
+}: Pick<
+	Core.Page,
+	"socialTitle" | "socialDescription" | "socialImage"
+>): Core.Page["openGraph"] {
 	return {
 		type: "website",
 		title: socialTitle,
@@ -42,7 +49,9 @@ function getOpenGraph({
  *
  * @param params A Page object
  */
-function getPageMetadata(params: Core.Page): GriddoPageObject["context"]["pageMetadata"] {
+function getPageMetadata(
+	params: Core.Page,
+): GriddoPageObject["context"]["pageMetadata"] {
 	const {
 		title,
 		metaTitle,
@@ -78,7 +87,9 @@ function getPageMetadata(params: Core.Page): GriddoPageObject["context"]["pageMe
 		index: isIndexed ? "index" : "noindex",
 		follow: follow ? "follow" : "nofollow",
 		translate: metasAdvancedList.includes("notranslate") ? "notranslate" : "",
-		metasAdvanced: metasAdvancedList.filter((item) => item !== "notranslate").join(),
+		metasAdvanced: metasAdvancedList
+			.filter((item) => item !== "notranslate")
+			.join(),
 		pageLanguages,
 		metaKeywords: metaKeywords
 			?.filter(Boolean)
@@ -122,9 +133,8 @@ async function createGriddoPageObject(
 		siteSlug,
 		socials,
 		theme,
-		navigations,
+		navigations: { header, footer },
 	} = additionalInfo;
-	const { header, footer } = navigations;
 
 	// Update page object
 	page.breadcrumb = breadcrumb;
@@ -197,7 +207,10 @@ async function createGriddoPageObject(
  * @param page A Griddo single page object.
  * @param additionalInfo Additional page info.
  */
-async function createGriddoSinglePage(page: GriddoSinglePage, additionalInfo: PageAdditionalInfo) {
+async function createGriddoSinglePage(
+	page: GriddoSinglePage,
+	additionalInfo: PageAdditionalInfo,
+) {
 	return createGriddoPageObject(page, additionalInfo);
 }
 
@@ -205,7 +218,14 @@ async function createGriddoSinglePage(page: GriddoSinglePage, additionalInfo: Pa
  * Create multiples pages from one page as list paginated pages
  */
 async function createGriddoListPages(
-	{ page, pages, isRoot = false, defaultLang, template, totalQueriedItems }: GriddoListPage,
+	{
+		page,
+		pages,
+		isRoot = false,
+		defaultLang,
+		template,
+		totalQueriedItems,
+	}: GriddoListPage,
 	additionalInfo: PageAdditionalInfo,
 ) {
 	const allPages = pages.map(async (dataInPage, idx) => {
@@ -213,8 +233,19 @@ async function createGriddoListPages(
 		const pageNumber = idx + 1;
 		const { domainUrl, compose } = page.fullPath;
 
+		// Crea un id como número negativo y añadiendo un sufijo para los
+		// listados estáticos. Esto es así para marcarlas y posteriormente
+		// borrarlas siempre en cada nuevo render desde el Adapter.
+		//
+		// id de página con mode:"list": 1546
+		// ids de las "sub-páginas": -15460, -15461, -1546n, (-)id(idx)
+		//
+		// @todo eliminar el concepto multipage de CX, debería ser algo core de
+		// Griddo itself: API/AX, que fuesen páginas con sus ids, etc..
+
 		const paginatedPage = {
 			...page,
+			id: Number.parseInt(`-${page.id}${idx}`),
 			fullPath: {
 				...page.fullPath,
 				// Add a page number (tailPageNumber) from page 2 onwards
@@ -256,11 +287,16 @@ async function createGriddoListPages(
  * @param page A Griddo Multipage object.
  * @param additionalInfo Additional page info.
  */
-function createGriddoMultiPages(page: GriddoMultiPage, additionalInfo: PageAdditionalInfo) {
+function createGriddoMultiPages(
+	page: GriddoMultiPage,
+	additionalInfo: PageAdditionalInfo,
+) {
 	const { multiPageElements: multiPageElementsBulk, ...cleanPage } = page;
 	// TODO: Use structuredClone() when node 18 is available.
 	// WARN!! Se ha probado que parse(stringify()) está tan optimizado que puede ser más rápido que structuredClone...
-	const multiPageElements: MultiPageElements = JSON.parse(JSON.stringify(multiPageElementsBulk));
+	const multiPageElements: MultiPageElements = JSON.parse(
+		JSON.stringify(multiPageElementsBulk),
+	);
 
 	// Si no hay un elemento sin slug, como mínimo hay que dibujar una página
 	// principal para el conjunto de páginas.
@@ -269,7 +305,7 @@ function createGriddoMultiPages(page: GriddoMultiPage, additionalInfo: PageAddit
 	}
 
 	// Creates each page based on `multiPageElements` from the schema.
-	const allPages = multiPageElements.map(async (pageElement) => {
+	const allPages = multiPageElements.map(async (pageElement, idx) => {
 		// TODO: Use structuredClone() when node 18 is available.
 		// WARN!! Se ha probado que parse(stringify()) está tan optimizado que puede ser más rápido que structuredClone...
 		const paginatedPage: APIPageObject = JSON.parse(JSON.stringify(cleanPage));
@@ -287,7 +323,9 @@ function createGriddoMultiPages(page: GriddoMultiPage, additionalInfo: PageAddit
 
 		const cleanSectionSlug = sectionSlug?.replace(/\//g, "");
 		const isRootSlug = sectionSlug === "/";
-		const rightSectionSlug = isRootSlug ? cleanSectionSlug : `${cleanSectionSlug}/`;
+		const rightSectionSlug = isRootSlug
+			? cleanSectionSlug
+			: `${cleanSectionSlug}/`;
 
 		const slash = compose.endsWith("/") ? "" : "/";
 		const newCompose = `${compose}${slash}${rightSectionSlug}`;
@@ -300,12 +338,24 @@ function createGriddoMultiPages(page: GriddoMultiPage, additionalInfo: PageAddit
 			paginatedPage.metaDescription = metaDescription;
 		}
 
+		// Crea un id como número negativo y añadiendo un sufijo para las multipages.
+		// Esto es así para marcarlas y posteriormente borrarlas siempre en cada
+		// nuevo render desde el Adapter.
+		//
+		// id de página con hasMultipageTrue: 1546
+		// ids de las "sub-páginas": -15460, -15461, -1546n, (-)id(idx)
+		//
+		// @todo eliminar el concepto multipage de CX, debería ser algo core de
+		// Griddo itself: API/AX, que fuesen páginas con sus ids, etc..
+		paginatedPage.id = Number.parseInt(`-${paginatedPage.id}${idx}`);
+
 		paginatedPage.fullUrl = `${fullUrl}/${rightSectionSlug}`;
 		paginatedPage.fullPath.compose = newCompose;
 		paginatedPage.slug = newCompose;
 		paginatedPage.template.activeSectionSlug = sectionSlug;
 		paginatedPage.template.activeSectionBase = fullUrl;
-		paginatedPage.metaTitle = metaTitle.trim() || title.trim() || paginatedPage.metaTitle;
+		paginatedPage.metaTitle =
+			metaTitle.trim() || title.trim() || paginatedPage.metaTitle;
 
 		return createGriddoPageObject(paginatedPage, additionalInfo);
 	});
@@ -318,7 +368,9 @@ function createGriddoMultiPages(page: GriddoMultiPage, additionalInfo: PageAddit
  *
  * @param page The page to get the multipage parts.
  */
-function getMultiPageElements(page: TemplateWithReferenceField): Promise<MultiPageElements> | null {
+function getMultiPageElements(
+	page: TemplateWithReferenceField,
+): Promise<MultiPageElements> | null {
 	const multiPageElements = new Promise((resolve) => {
 		// Recursive
 		const getMultiPageComponent = (
@@ -337,12 +389,13 @@ function getMultiPageElements(page: TemplateWithReferenceField): Promise<MultiPa
 				const currentComponent = template[key] as {
 					component: string;
 					hasGriddoMultiPage: boolean;
-					elements: Record<string, unknown>[];
+					elements: Array<Record<string, unknown>>;
 				};
-				const isValidComponent = currentComponent || typeof currentComponent === "object";
-				const hasGriddoMultiPageProp = JSON.stringify(currentComponent).includes(
-					'"hasGriddoMultiPage":true',
-				);
+				const isValidComponent =
+					currentComponent || typeof currentComponent === "object";
+				const hasGriddoMultiPageProp = JSON.stringify(
+					currentComponent,
+				).includes('"hasGriddoMultiPage":true');
 
 				if (!isValidComponent) {
 					continue;
@@ -392,7 +445,11 @@ function getMultiPageElements(page: TemplateWithReferenceField): Promise<MultiPa
  * getPage(3, ["a", "b", "c", "d", "e", "f", "g", "h"], 2)
  * // -> ["d", "e", "f"]
  */
-function getPage(itemsPerPage: number, items: Fields.QueriedDataItem[], page: number) {
+function getPage(
+	itemsPerPage: number,
+	items: Array<Fields.QueriedDataItem>,
+	page: number,
+) {
 	return items?.slice(itemsPerPage * (page - 1), itemsPerPage * page);
 }
 
@@ -406,11 +463,16 @@ function getPage(itemsPerPage: number, items: Fields.QueriedDataItem[], page: nu
  * getPageCluster(3, ["a", "b", "c", "d", "e", "f", "g", "h"])
  * // -> [["a", "b", "c"], ["d", "e", "f"], ["g", "h"]]
  */
-function getPageCluster(itemsPerPage: number, items: Fields.QueriedDataItem[]) {
+function getPageCluster(
+	itemsPerPage: number,
+	items: Array<Fields.QueriedDataItem>,
+) {
 	const totalPagesCount = Math.ceil(items.length / itemsPerPage) || 1;
 	const pageNumbers = Array.from({ length: totalPagesCount }, (_, i) => i + 1);
 
-	return pageNumbers?.map((pageNumber) => getPage(itemsPerPage, items, pageNumber));
+	return pageNumbers?.map((pageNumber) =>
+		getPage(itemsPerPage, items, pageNumber),
+	);
 }
 
 /**
@@ -420,7 +482,8 @@ function getPageCluster(itemsPerPage: number, items: Fields.QueriedDataItem[]) {
  */
 function getPaginatedPages(listTemplate: TemplateWithReferenceField) {
 	const items = listTemplate.queriedItems || [];
-	const itemsPerPage = listTemplate?.itemsPerPage || DEFAULT_ITEMS_PER_PAGE_FOR_LIST_TEMPLATES;
+	const itemsPerPage =
+		listTemplate?.itemsPerPage || DEFAULT_ITEMS_PER_PAGE_FOR_LIST_TEMPLATES;
 	const pageClusters = getPageCluster(itemsPerPage, items);
 
 	return pageClusters;
@@ -461,7 +524,9 @@ function addPageNumberToUrl(
 		return removeDuplicateTrailing(`${url}${endingSlash}`);
 	}
 
-	return removeDuplicateTrailing(`${url}${trailingSlash}${pageNumber}${endingSlash}`);
+	return removeDuplicateTrailing(
+		`${url}${trailingSlash}${pageNumber}${endingSlash}`,
+	);
 }
 
 /**

package/exporter/utils/render.ts

@@ -0,0 +1,71 @@
+import fs from "node:fs";
+import path from "node:path";
+
+import { envs } from "../constants";
+import { throwError } from "../errors";
+import { RenderUUIDError } from "../errors/errors-data";
+import { apiRegister } from "../registers";
+import { insertAlert } from "./alerts";
+import { getConfig } from "./core-utils";
+import { debugLog } from "./loggin";
+
+const config = getConfig();
+
+/**
+ * Creates a sentinel file with the current date and time.
+ * This file is used to track later if node_modules/@griddo/cx was cleaned by a
+ * npm install from a deploy.
+ */
+function createSentinelRenderFile() {
+	const { __cx } = config.paths();
+	const renderSentinelFile = path.join(__cx, ".render-sentinel");
+	fs.writeFileSync(renderSentinelFile, new Date().toISOString());
+}
+
+function deleteSentinelRenderFile() {
+	const { __cx } = config.paths();
+	const renderSentinelFile = path.join(__cx, ".render-sentinel");
+	fs.unlinkSync(renderSentinelFile);
+}
+
+function isValidRender() {
+	const { __cx } = config.paths();
+	const renderSentinelFile = path.join(__cx, ".render-sentinel");
+	if (!fs.existsSync(renderSentinelFile)) {
+		throwError(RenderUUIDError);
+	}
+}
+
+/**
+ * Send the default registers through Griddo alerts.
+ */
+async function sendGriddoDefaultAlerts() {
+	if (envs.GRIDDO_ALERT_FEATURE) {
+		const registerContent = apiRegister.get("API_RESPONSE_TOO_BIG");
+
+		if (registerContent.entries.length < 1) {
+			return;
+		}
+
+		debugLog("\nRender register report\n");
+		debugLog(registerContent);
+		debugLog();
+
+		await insertAlert({
+			description: `API response size is too large (${envs.GRIDDO_API_MAX_RESPONSE_SIZE}KB).`,
+			area: "Griddo",
+			level: "W",
+			fullData: {
+				output: registerContent,
+				date: new Date().toISOString(),
+			},
+		});
+	}
+}
+
+export {
+	createSentinelRenderFile,
+	deleteSentinelRenderFile,
+	isValidRender,
+	sendGriddoDefaultAlerts,
+};

package/exporter/utils/searches.ts

@@ -0,0 +1,156 @@
+import type { PostSearchInfoResponse } from "../types/api";
+import type {
+	AIEmbeddingsResponse,
+	PostSearchInfoProps,
+} from "../types/global";
+
+import fs from "node:fs";
+import path from "node:path";
+
+import { post } from "./api";
+import { getConfig } from "./core-utils";
+import { getBuildPagesFromCachedStore } from "./store";
+import { endpoints } from "../constants";
+
+/**
+ * Save in the BBDD the content of a page parsed without HTML tags.
+ *
+ * @param props Object with parts of the final page object to be saved in the BBDD.
+ */
+async function postSearchInfo(props: PostSearchInfoProps) {
+	const {
+		title,
+		description,
+		image,
+		pageId,
+		languageId,
+		siteId,
+		url,
+		content,
+		template,
+	} = props;
+
+	const response = await post<PostSearchInfoResponse>({
+		endpoint: endpoints.SEARCH,
+		body: {
+			title,
+			description,
+			image,
+			pageId,
+			languageId,
+			siteId,
+			url,
+			template,
+			content,
+		},
+		useApiCacheDir: false,
+	});
+
+	return response;
+}
+
+/**
+ * Remove the every cntent inside some HTML tags including them.
+ *
+ * @param content A string with the content.
+ */
+function prepareHTMLContentForSearch(content: string): string {
+	// 1. Remove script, style, and other unwanted block tags and their content.
+	// The regex looks for <tag...>...</tag> where tag is one of the specified ones.
+	const tagsToRemove = [
+		"meta",
+		"link",
+		"style",
+		"script",
+		"noscript",
+		"nav",
+		"header",
+		"footer",
+	];
+	const removeTagsRegex = new RegExp(
+		`<(${tagsToRemove.join("|")})\\b[^>]*>.*?<\\/\\1>`,
+		"gis",
+	);
+	let processedContent = content.replace(removeTagsRegex, "");
+
+	// 2. Strip all remaining HTML tags.
+	processedContent = processedContent.replace(/<[^>]+>/g, " ");
+
+	// 3. Normalize whitespace: replace multiple spaces/newlines with a single space and trim.
+	processedContent = processedContent.replace(/\s+/g, " ").trim();
+
+	return processedContent;
+}
+
+/**
+ * Function that search in the `/public` dir the content info of the pages and
+ * send it to the search table in the ddbb using the API.
+ */
+async function uploadRenderedSearchContentToAPI(
+	distDomainPath: string,
+	domain: string,
+) {
+	const config = getConfig();
+	const { __cache } = config.paths(domain);
+
+	// Extraemos el path de la carpeta store dentro de un dominio
+	const storeFolder = path.join(__cache, "store");
+
+	// Este caso sería que el dominio existe en la instancia pero no hay sites
+	// renderizardos bajo el mismo. Por lo que `store` no existe.
+	if (!fs.existsSync(storeFolder)) {
+		console.log(
+			`Skipping uploading content to the search endpoint for the domain ${domain}, it has not exported sites.`,
+		);
+		return;
+	}
+
+	// Obtenemos las páginas del store cacheado de un dominio
+	const storePages = getBuildPagesFromCachedStore(domain);
+
+	for await (const store of storePages) {
+		const {
+			context: { page, openGraph, pageMetadata },
+		} = store;
+
+		const { compose } = page.fullPath;
+
+		const contextPath = path.resolve(`${distDomainPath}/${compose}/index.html`);
+
+		const content = fs.readFileSync(contextPath).toString();
+
+		const pageObject: PostSearchInfoProps = {
+			siteId: page.site,
+			pageId: page.id,
+			// `pageMetadata.title` has already a fallback `metatitle ||
+			// pageTitle` so probably `title` never will be take the
+			// `openGraph?.title` value. Only when the `metatitle` and
+			// `pageTitle` are empty.
+			title: pageMetadata?.title || openGraph?.title,
+			languageId: page.language,
+			url: page.fullUrl,
+			template: page.template.templateType || page.templateId,
+			description: pageMetadata?.description || openGraph?.description,
+			image: openGraph.image,
+			content: prepareHTMLContentForSearch(content),
+		};
+
+		await postSearchInfo(pageObject);
+	}
+}
+
+async function startAIEmbeddings() {
+	try {
+		await post<AIEmbeddingsResponse>({
+			endpoint: endpoints.AI_EMBEDDINGS,
+			useApiCacheDir: false,
+		});
+	} catch (error) {
+		console.warn(
+			"There was an error with the ai embeddings",
+			(error as Error).message,
+		);
+	}
+}
+
+export { postSearchInfo, startAIEmbeddings, uploadRenderedSearchContentToAPI };