almeira-web/scripts/publish_cauca_sept17_to_date.js

339 lines
15 KiB
JavaScript

const axios = require('axios');
const cheerio = require('cheerio');
const fs = require('fs');
const path = require('path');
const os = require('os');
const { Client, Databases, Storage, ID, Query } = require('node-appwrite');
require('dotenv').config({ path: '.env.local' });
let InputFile;
try { InputFile = require('node-appwrite/file').InputFile; } catch (e) { InputFile = null; }
const { applyTablesDbShim } = require('../src/lib/appwrite-tables-shim');
const client = new Client()
.setEndpoint(process.env.NEXT_PUBLIC_APPWRITE_ENDPOINT)
.setProject(process.env.NEXT_PUBLIC_APPWRITE_PROJECT_ID)
.setKey(process.env.APPWRITE_API_KEY);
applyTablesDbShim(client);
const databases = new Databases(client);
const storage = new Storage(client);
const DB_ID = process.env.APPWRITE_DATABASE_ID;
const BUCKET_ID = process.env.NEXT_PUBLIC_APPWRITE_MEDIA_BUCKET_ID;
function slugify(text) {
return text
.toString()
.toLowerCase()
.normalize('NFD')
.replace(/[\u0300-\u036f]/g, '')
.replace(/[^a-z0-9 -]/g, '')
.replace(/\s+/g, '-')
.replace(/-+/g, '-')
.trim();
}
function cleanContentHtml(html) {
if (!html) return '';
const $ = cheerio.load(html, null, false);
$('style, script, link, iframe, object, embed, video, audio, source, noscript').remove();
const allowedTags = ['p', 'b', 'i', 'strong', 'em', 'ul', 'ol', 'li', 'br', 'a'];
function processNode(node) {
$(node).children().each((i, child) => {
processNode(child);
});
const tagName = node.tagName ? node.tagName.toLowerCase() : null;
if (tagName) {
const attribs = Object.keys(node.attribs || {});
attribs.forEach(attr => {
if (tagName !== 'a' || attr !== 'href') {
$(node).removeAttr(attr);
}
});
if (!allowedTags.includes(tagName)) {
$(node).replaceWith($(node).html() || '');
}
}
}
$.root().children().each((i, child) => {
processNode(child);
});
return $.html().trim();
}
async function getCollectionIds() {
const list = await databases.listCollections(DB_ID);
const map = {};
list.collections.forEach(c => { map[c.name.toLowerCase()] = c.$id; });
return map;
}
const sleep = ms => new Promise(res => setTimeout(res, ms));
async function run() {
console.log("===============================================================================");
console.log("Iniciando publicación de noticias de la Gobernación del Cauca (17 Sep a la fecha)");
console.log("===============================================================================");
try {
const collections = await getCollectionIds();
const postsCollectionId = collections['posts'] || 'posts';
const mediaCollectionId = collections['media'] || 'media';
const pubTypesCollectionId = collections['publication types'] || collections['publication_types'] || 'publication_types';
// 1. Obtener ID del tipo de publicación NOTICIA
const pubTypesRes = await databases.listDocuments(DB_ID, pubTypesCollectionId);
const typeMap = {};
pubTypesRes.documents.forEach(d => {
typeMap[d.codigo] = d.$id;
});
const typeId = typeMap['NOTICIA'];
if (!typeId) {
throw new Error("No se encontró el tipo de publicación NOTICIA en Appwrite.");
}
console.log(`Tipo de publicación NOTICIA ID: ${typeId}`);
// 2. Obtener lista de noticias desde la API SharePoint de la Gobernación
console.log("Consultando API SharePoint de cauca.gov.co...");
const listUrl = "https://cauca.gov.co/Prensa/SaladePrensa/_api/web/lists(guid'e1840969-e886-4c1e-8cde-7498ab33f10a')/items?$select=Id,Title,ArticleStartDate,Created,Modified,File/ServerRelativeUrl&$expand=File&$orderby=Created desc&$top=100";
const spRes = await axios.get(listUrl, {
headers: { 'Accept': 'application/json;odata=verbose', 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36' },
timeout: 15000
});
const rawItems = spRes.data.d?.results || [];
console.log(`Total artículos obtenidos de SharePoint: ${rawItems.length}`);
// Filtrar desde el 21 de agosto de 2026 (00:00:00) hasta la fecha
const targetStartDate = new Date('2026-08-21T00:00:00');
const filteredItems = rawItems.filter(it => {
const dateStr = it.ArticleStartDate || it.Created;
const d = new Date(dateStr);
return d >= targetStartDate;
});
console.log(`Artículos identificados desde el 17 de septiembre: ${filteredItems.length}`);
// Ordenar cronológicamente (los más antiguos primero para que los más nuevos queden en la parte superior del portal)
filteredItems.sort((a, b) => {
const da = new Date(a.ArticleStartDate || a.Created).getTime();
const db = new Date(b.ArticleStartDate || b.Created).getTime();
return da - db;
});
// 3. Consultar slugs existentes en Appwrite para evitar duplicados
console.log("Verificando publicaciones existentes en Appwrite...");
const existingSlugs = new Set();
let cursor = null;
while (true) {
const queries = [Query.limit(100), Query.select(['slug'])];
if (cursor) queries.push(Query.cursorAfter(cursor));
const existingRes = await databases.listDocuments(DB_ID, postsCollectionId, queries);
existingRes.documents.forEach(d => {
if (d.slug) existingSlugs.add(d.slug);
});
if (existingRes.documents.length < 100) break;
cursor = existingRes.documents[existingRes.documents.length - 1].$id;
}
console.log(`Slugs existentes en posts: ${existingSlugs.size}`);
let publishedCount = 0;
let skippedCount = 0;
let errorsCount = 0;
for (let i = 0; i < filteredItems.length; i++) {
const item = filteredItems[i];
const pubDate = new Date(item.ArticleStartDate || item.Created);
const rawTitle = (item.Title || '').trim();
const slug = `gobernacion-web-${slugify(rawTitle)}`.substring(0, 95);
console.log(`\n-------------------------------------------------------------------------------`);
console.log(`[${i + 1}/${filteredItems.length}] Noticia ID: ${item.Id} | Fecha: ${pubDate.toISOString().split('T')[0]}`);
console.log(`Título: "${rawTitle}"`);
console.log(`Slug: ${slug}`);
if (existingSlugs.has(slug)) {
console.log(`-> Ya existe en Appwrite con slug: ${slug}. Se omite duplicado.`);
skippedCount++;
continue;
}
const relativeUrl = item.File?.ServerRelativeUrl;
if (!relativeUrl) {
console.error("-> Error: No se encontró URL relativa para este artículo.");
errorsCount++;
continue;
}
const pageUrl = "https://cauca.gov.co" + encodeURI(decodeURI(relativeUrl));
console.log(`-> Obteniendo contenido de página: ${pageUrl}`);
try {
const pageRes = await axios.get(pageUrl, {
headers: { 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/115.0.0.0 Safari/537.36' },
timeout: 15000
});
const $ = cheerio.load(pageRes.data);
// A. Extraer imagen destacada
let imageUrl = null;
$('.ms-rtestate-field img').each((idx, el) => {
const src = $(el).attr('src');
if (src && !src.includes('logo') && !src.includes('icon')) {
imageUrl = src.startsWith('http') ? src : 'https://cauca.gov.co' + src;
return false;
}
});
if (!imageUrl) {
$('img').each((idx, el) => {
const src = $(el).attr('src');
if (src && src.includes('/Prensa/PublishingImages/')) {
imageUrl = src.startsWith('http') ? src : 'https://cauca.gov.co' + src;
return false;
}
});
}
if (!imageUrl) {
$('img').each((idx, el) => {
const src = $(el).attr('src');
if (src && !src.includes('logo') && !src.includes('icon') && !src.includes('menu') && src.length > 10) {
imageUrl = src.startsWith('http') ? src : 'https://cauca.gov.co' + src;
return false;
}
});
}
// B. Extraer contenido HTML
let contentHtml = null;
let maxLen = 0;
$('.ms-rtestate-field').each((idx, el) => {
const textLen = $(el).text().trim().length;
if (textLen > maxLen) {
maxLen = textLen;
contentHtml = $(el).html();
}
});
if (!contentHtml) {
contentHtml = $('.lg\\:container.mx-auto.flex.items-center.justify-center.my-\\[80px\\]').html();
}
if (!contentHtml || contentHtml.trim().length === 0) {
console.error("-> Error: Contenido de la noticia vacío.");
errorsCount++;
continue;
}
const cleanedContent = cleanContentHtml(contentHtml);
const plainText = cheerio.load(cleanedContent).text().replace(/\s+/g, ' ').trim();
let summary = plainText.substring(0, 150);
if (plainText.length > 150) summary += '...';
// C. Descargar y subir imagen destacada a Appwrite Storage y Media
let mediaId = null;
if (imageUrl) {
try {
console.log(`-> Descargando imagen: ${imageUrl}`);
const imgRes = await axios.get(encodeURI(decodeURI(imageUrl)), {
responseType: 'arraybuffer',
timeout: 15000,
headers: {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/115.0.0.0 Safari/537.36'
}
});
const buffer = Buffer.from(imgRes.data, 'binary');
const isHtml = buffer.toString('utf-8', 0, 50).trim().toLowerCase().startsWith('<!doctype') || buffer.toString('utf-8', 0, 20).trim().toLowerCase().startsWith('<html');
if (!isHtml && buffer.length > 500) {
const urlParts = imageUrl.split('/');
let filename = urlParts[urlParts.length - 1] || 'imagen.jpg';
filename = decodeURIComponent(filename).replace(/[^a-zA-Z0-9._-]/g, '_');
if (!filename.includes('.')) filename += '.jpg';
const tempFilePath = path.join(os.tmpdir(), `gob_${ID.unique()}_${filename}`);
fs.writeFileSync(tempFilePath, buffer);
let fileParam;
if (InputFile && InputFile.fromPath) {
fileParam = InputFile.fromPath(tempFilePath, filename);
} else {
fileParam = fs.createReadStream(tempFilePath);
}
console.log(`-> Subiendo imagen al bucket (${BUCKET_ID})...`);
const fileUploaded = await storage.createFile(BUCKET_ID, ID.unique(), fileParam);
console.log(`-> Registrando en colección media...`);
const mediaDoc = await databases.createDocument(DB_ID, mediaCollectionId, ID.unique(), {
file_id: fileUploaded.$id,
bucket_id: BUCKET_ID,
titulo: rawTitle.substring(0, 245),
tipo: 'imagen'
});
mediaId = mediaDoc.$id;
try { fs.unlinkSync(tempFilePath); } catch (e) {}
} else {
console.log("-> Aviso: La imagen obtenida no es un binario válido.");
}
} catch (imgErr) {
console.error(`-> Error al procesar imagen: ${imgErr.message}`);
}
}
// D. Crear publicación en Appwrite
const postDoc = {
titulo: rawTitle.substring(0, 245),
slug: slug,
tipo: 'noticia',
tipo_publicacion_id: typeId,
estado: 'publicado',
fecha_publicacion: pubDate.toISOString(),
contenido: cleanedContent,
resumen_redes: summary,
seo_descripcion: summary.substring(0, 155),
seo_titulo: rawTitle.substring(0, 58),
imagen_destacada_id: mediaId
};
console.log(`-> Guardando documento en Appwrite posts...`);
const createdPost = await databases.createDocument(DB_ID, postsCollectionId, ID.unique(), postDoc);
console.log(`✓ [ÉXITO] Publicada correctamente! Post ID: ${createdPost.$id}`);
existingSlugs.add(slug);
publishedCount++;
// Pausa prudente entre peticiones
await sleep(1200);
} catch (err) {
console.error(`-> Error publicando noticia: ${err.message}`);
errorsCount++;
}
}
console.log("\n===============================================================================");
console.log(`RESUMEN FINAL:`);
console.log(`Total identificadas: ${filteredItems.length}`);
console.log(`Publicadas exitosamente: ${publishedCount}`);
console.log(`Omitidas por ya existir: ${skippedCount}`);
console.log(`Errores: ${errorsCount}`);
console.log("===============================================================================");
} catch (e) {
console.error("Error fatal en proceso de publicación:", e.message);
}
}
run();