The platform uses two separate SharePoint sites accessed through the Microsoft Graph API:
- CMS site — source of truth for categories, products, and page content. Read at build time by the sync script.
- CRM site — stores price request submissions, comments, and attachment metadata. Written at runtime by Nitro server routes.
Both sites require an Azure App registration with the Sites.ReadWrite.All and Files.ReadWrite.All Microsoft Graph application permissions granted by a tenant administrator.
CMS site
The CMS site holds three SharePoint Lists:
| List | Env var | Purpose |
|---|
| Categories | CMS_CATEGORIES_LIST_ID | Product category tree, slugs, SEO fields, tabs |
| Products | CMS_PRODUCTS_LIST_ID | Individual product records with pricing and attributes |
| Assets | CMS_ASSETS_LIST_ID | Media asset metadata |
Sync script
scripts/sync-sharepoint-to-cms.mjs fetches all published items from the CMS Lists and writes two files:
cms/catalog.json — full catalog with categories and productscms/routes.json — flat list of URL paths for Nuxt pre-rendering
This script is automatically called by npm run build and npm run generate. For local development, run it once after cloning and again whenever SharePoint content changes.
The script uses @azure/identity ClientSecretCredential and @microsoft/microsoft-graph-client to authenticate:
import { ClientSecretCredential } from "@azure/identity"
import { Client } from "@microsoft/microsoft-graph-client"
const credential = new ClientSecretCredential(TENANT_ID, CLIENT_ID, CLIENT_SECRET)
const graph = Client.init({
authProvider: async (done) => {
const token = await credential.getToken("https://graph.microsoft.com/.default")
done(null, token.token)
},
})
$top=999 and retries throttled requests (HTTP 429/503) with exponential backoff up to 6 attempts.
Category data shape
Each category in cms/catalog.json includes:
{
id: string
slug: string // URL segment, e.g. "gran-formato"
path: string // full path, e.g. "/categorias/gran-formato"
parent?: string // parent slug for nested categories
type: "categoria" | "subcategoria"
title: string
nav: string // short nav label
order: number
isPublished: boolean
tabs: TabItem[] // content tabs (from TabsJson or parsed BodyMd)
image: { src, width, height, alt }
seo: { metaTitle, metaDescription, canonical, schema, keywords }
faqs: FaqItem[]
}
CRM site
The CRM site stores operational data created at runtime:
| List | Env var | Purpose |
|---|
| Price requests | CRM_PRICE_REQUESTS_LIST_ID | One item per quote submission |
| Comments | CRM_PRICE_REQUESTS_COMMENTS_LIST_ID | Internal notes on requests |
| Attachments library | CRM_ATTACHMENTS_LIBRARY_LIST_ID | Document library for uploaded files |
createPriceRequest service writes to the price requests list and optionally uploads a file to the attachments Drive via Graph.
ISR caching strategy
Content served from SharePoint is cached at two levels:
// nuxt.config.ts
routeRules: {
"/categorias/**": { isr: 600 }, // pages regenerate every 10 minutes
"/api/categorias/**": { swr: 300 }, // API responses stale-while-revalidate 5 min
"/productos/**": { isr: 600 },
"/api/productos/**": { swr: 300 },
}
Environment variables
# Azure authentication
AZURE_TENANT_ID=
AZURE_CLIENT_ID=
AZURE_CLIENT_SECRET=
# CMS site
CMS_SITE_ID=
CMS_SITE_HOSTNAME=reprodisseny.sharepoint.com
CMS_SITE_PATH=
CMS_CATEGORIES_LIST_ID=
CMS_PRODUCTS_LIST_ID=
CMS_ASSETS_LIST_ID=
# CRM site
CRM_SITE_ID=
CRM_SITE_HOSTNAME=reprodisseny.sharepoint.com
CRM_SITE_PATH=/sites/portal
CRM_PRICE_REQUESTS_LIST_ID=
CRM_PRICE_REQUESTS_COMMENTS_LIST_ID=
# Sync script (.env.imports)
TENANT_ID=
CLIENT_ID=
CLIENT_SECRET=
SHAREPOINT_SITE_ID=
SP_LIST_CATEGORIES_ID=
SP_LIST_PRODUCTS_ID=
