[render] Download internal images

examples/nextjs/.gitignore

@@ -34,3 +34,6 @@ yarn-error.log*
 # typescript
 *.tsbuildinfo
 next-env.d.ts
+
+# Images downloaded from CMS
+public/cms/*

examples/nextjs/src/lib/fetchers.ts

@@ -1,18 +1,56 @@
 import { unstable_cache } from "next/cache"
 import { fetchBlocksChildren } from "@react-notion-cms/render"
 import { Client } from "@notionhq/client"
+import { statSync, writeFileSync } from "node:fs"
+import nextConfig from "../../next.config.mjs"
 
 const notionClient = new Client({
   auth: process.env.NOTION_ACCESS_TOKEN,
 })
 
 export const getCachedPageContent = unstable_cache(
   async (blockId: string) => {
-    return fetchBlocksChildren(notionClient, blockId)
+    return fetchBlocksChildren(
+      notionClient,
+      {
+        block_id: blockId,
+        page_size: 100,
+      },
+      {
+        resolveImageFn: downloadImageToPublicDir,
+      },
+    )
   },
   ["cms-page"],
   // {
   //   revalidate: false,
   //   tags: ["cms-page"],
   // },
 )
+
+const downloadImageToPublicDir = async (url: string, meta: { blockId: string; lastEditedTime: Date }) => {
+  const fileUrl = new URL(url)
+  const originalFileName = fileUrl.pathname.substring(fileUrl.pathname.lastIndexOf("/") + 1)
+  const newFileName = `public/cms/${meta.blockId}-${originalFileName}`
+
+  let savedLastEditedTime: Date | null = null
+  try {
+    const stat = statSync(newFileName)
+    savedLastEditedTime = stat.mtime
+  } catch (error) {
+    if (error instanceof Error && "code" in error && error.code === "ENOENT") {
+      console.debug(`${newFileName} not found`)
+    } else {
+      console.warn(`${newFileName}: ${error}`)
+    }
+  }
+  // Avoid download the file again and again
+  if (savedLastEditedTime === null || meta.lastEditedTime > savedLastEditedTime) {
+    const resp = await fetch(fileUrl)
+    const blob = await resp.blob()
+    writeFileSync(newFileName, new DataView(await blob.arrayBuffer()))
+    console.info(`downloaded image ${newFileName} of block ${meta.blockId}`)
+  }
+
+  return newFileName.replace("public", nextConfig.basePath ?? "")
+}

packages/render/src/fetch.ts

@@ -1,13 +1,24 @@
 import { type Client, isFullBlock, iteratePaginatedAPI } from "@notionhq/client"
 import { BlockObjectResponseWithChildren } from "./types"
+import { type ListBlockChildrenParameters } from "@notionhq/client/build/src/api-endpoints"
 
-export const fetchBlocksChildren = async (client: Client, blockId: string) => {
+type ResolveImageFn = (url: string, meta: { blockId: string; lastEditedTime: Date }) => Promise<string>
+
+export const fetchBlocksChildren = async (
+  client: Client,
+  firstPageArgs: ListBlockChildrenParameters,
+  options: { resolveImageFn?: ResolveImageFn },
+) => {
   const result: Array<BlockObjectResponseWithChildren> = []
-  for await (const block of iteratePaginatedAPI(client.blocks.children.list, {
-    block_id: blockId,
-    page_size: 100,
-  })) {
+  for await (const block of iteratePaginatedAPI(client.blocks.children.list, firstPageArgs)) {
     if (isFullBlock(block)) {
+      if (options.resolveImageFn && block.type === "image" && block.image.type === "file") {
+        block.image.file.url = await options.resolveImageFn(block.image.file.url, {
+          blockId: block.id,
+          lastEditedTime: new Date(block.last_edited_time),
+        })
+      }
+
       if (!block.has_children) {
         result.push(block)
         continue
@@ -19,10 +30,17 @@ export const fetchBlocksChildren = async (client: Client, blockId: string) => {
 
       result.push({
         ...block,
-        _children: await fetchBlocksChildren(client, childId),
+        _children: await fetchBlocksChildren(
+          client,
+          {
+            ...firstPageArgs,
+            block_id: childId,
+          },
+          options,
+        ),
       })
     }
   }
-  console.info(`fetched children from notion block ${blockId}`)
+  console.info(`fetched children from notion block ${firstPageArgs.block_id}`)
   return result
 }