Create Embeddings for Canvas

4 files changed, 169 insertions, 6 deletions

diff --git a/apps/web/app/(canvas)/lib/createEmbeds.ts b/apps/web/app/(canvas)/lib/createEmbeds.ts
index 322e697e..53d81533 100644
--- a/apps/web/app/(canvas)/lib/createEmbeds.ts
+++ b/apps/web/app/(canvas)/lib/createEmbeds.ts
@@ -50,10 +50,18 @@ export default async function createEmbedsFromUrl({url, point, sources, editor}:
           type: "url",
           url,
         });
-        const fetchWebsite = await (await fetch(`https://unfurl-bookmark.pruthvirajthinks.workers.dev/?url=${url}`)).json()
-        if (fetchWebsite.title) bookmarkAsset.props.title = fetchWebsite.title;
-        if (fetchWebsite.image) bookmarkAsset.props.image = fetchWebsite.image;
-        if (fetchWebsite.description) bookmarkAsset.props.description = fetchWebsite.description;
+        const fetchWebsite: {
+          title?: string;
+          image?: string;
+          description?: string;
+        } = await (await fetch(`/api/unfirlsite?website=${url}`, {
+          method: "POST"
+        })).json()
+        if (bookmarkAsset){
+          if (fetchWebsite.title) bookmarkAsset.props.title = fetchWebsite.title;
+          if (fetchWebsite.image) bookmarkAsset.props.image = fetchWebsite.image;
+          if (fetchWebsite.description) bookmarkAsset.props.description = fetchWebsite.description;
+        }
         if (!bookmarkAsset) throw Error("Could not create an asset");
         asset = bookmarkAsset;
       } catch (e) {
diff --git a/apps/web/app/api/editorai/route.ts b/apps/web/app/api/editorai/route.ts
new file mode 100644
index 00000000..6ee0aed2
--- /dev/null
+++ b/apps/web/app/api/editorai/route.ts
@@ -0,0 +1,20 @@
+import type { NextRequest } from "next/server";
+import { ensureAuth } from "../ensureAuth";
+
+export const runtime = "edge";
+
+export async function POST(request: NextRequest) {
+  const d = await ensureAuth(request);
+  if (!d) {
+    return new Response("Unauthorized", { status: 401 });
+  }
+  const res : {context: string, request: string} = await request.json()
+
+  try {
+    const response = await fetch(`${process.env.BACKEND_BASE_URL}/api/editorai?context=${res.context}&request=${res.request}`);
+    const result = await response.json();
+    return new Response(JSON.stringify(result));
+  } catch (error) {
+    return new Response(`Error, ${error}`)
+  }
+}
\ No newline at end of file
diff --git a/apps/web/app/api/unfirlsite/route.ts b/apps/web/app/api/unfirlsite/route.ts
new file mode 100644
index 00000000..4b8b4858
--- /dev/null
+++ b/apps/web/app/api/unfirlsite/route.ts
@@ -0,0 +1,134 @@
+import { load } from 'cheerio'
+import { AwsClient } from "aws4fetch";
+
+import type { NextRequest } from "next/server";
+import { ensureAuth } from "../ensureAuth";
+
+export const runtime = "edge";
+
+const r2 = new AwsClient({
+  accessKeyId: process.env.R2_ACCESS_KEY_ID,
+  secretAccessKey: process.env.R2_SECRET_ACCESS_KEY,
+});
+
+
+export async function POST(request: NextRequest) {
+  
+  const d = await ensureAuth(request);
+  if (!d) {
+    return new Response("Unauthorized", { status: 401 });
+  }
+
+  if (
+    !process.env.R2_ACCESS_KEY_ID ||
+    !process.env.R2_ACCOUNT_ID ||
+    !process.env.R2_SECRET_ACCESS_KEY ||
+    !process.env.R2_BUCKET_NAME
+  ) {
+    return new Response(
+      "Missing one or more R2 env variables: R2_ENDPOINT, R2_ACCESS_ID, R2_SECRET_KEY, R2_BUCKET_NAME. To get them, go to the R2 console, create and paste keys in a `.dev.vars` file in the root of this project.",
+      { status: 500 },
+    );
+  }
+
+  const website = new URL(request.url).searchParams.get("website");
+  
+  if (!website) {
+    return new Response("Missing website", { status: 400 });
+  }
+
+  const salt = () => Math.floor(Math.random() * 11);
+  const encodeWebsite = `${encodeURIComponent(website)}${salt()}`;
+
+  try {
+    // this returns the og image, description and title of website 
+    const response = await unfurl(website);
+
+    if (!response.image){
+      return new Response(JSON.stringify(response))
+    }
+
+    const imageUrl = await process.env.DEV_IMAGES.get(encodeWebsite)
+    if (imageUrl){
+      return new Response(JSON.stringify({
+        image: imageUrl,
+        title: response.title,
+        description: response.description,
+      }))
+    }
+
+    const res = await fetch(`${response.image}`)
+    const image = await res.blob();
+
+    const url = new URL(
+      `https://${process.env.R2_BUCKET_NAME}.${process.env.R2_ACCOUNT_ID}.r2.cloudflarestorage.com`
+    );
+
+    url.pathname = encodeWebsite;
+    url.searchParams.set("X-Amz-Expires", "3600");
+
+    const signedPuturl = await r2.sign(
+      new Request(url, {
+        method: "PUT",
+      }),
+      {
+        aws: { signQuery: true },
+      }
+    );
+    await fetch(signedPuturl.url, {
+      method: 'PUT',
+      body: image,
+    });
+
+    await process.env.DEV_IMAGES.put(encodeWebsite, `${process.env.R2_PUBLIC_BUCKET_ADDRESS}/${encodeWebsite}`)
+
+    return new Response(JSON.stringify({
+      image: `${process.env.R2_PUBLIC_BUCKET_ADDRESS}/${encodeWebsite}`,
+      title: response.title,
+      description: response.description,
+    }));
+
+  } catch (error) {
+    console.log(error)
+    return new Response(JSON.stringify({
+      status: 500,
+      error: error,
+    }))
+  }
+  }
+
+export async function unfurl(url: string) {
+	const response = await fetch(url)
+	if (response.status >= 400) {
+		throw new Error(`Error fetching url: ${response.status}`)
+	}
+	const contentType = response.headers.get('content-type')
+	if (!contentType?.includes('text/html')) {
+		throw new Error(`Content-type not right: ${contentType}`)
+	}
+
+	const content = await response.text()
+	const $ = load(content)
+
+	const og: { [key: string]: string | undefined } = {}
+	const twitter: { [key: string]: string | undefined } = {}
+
+  // @ts-ignore, it just works so why care of type safety if someone has better way go ahead
+	$('meta[property^=og:]').each((_, el) => (og[$(el).attr('property')!] = $(el).attr('content')))
+  // @ts-ignore
+	$('meta[name^=twitter:]').each((_, el) => (twitter[$(el).attr('name')!] = $(el).attr('content')))
+
+	const title = og['og:title'] ?? twitter['twitter:title'] ?? $('title').text() ?? undefined
+	const description =
+		og['og:description'] ??
+		twitter['twitter:description'] ??
+		$('meta[name="description"]').attr('content') ??
+		undefined
+	const image = og['og:image:secure_url'] ?? og['og:image'] ?? twitter['twitter:image'] ?? undefined
+
+	return {
+		title,
+		description,
+		image,
+	}
+}
diff --git a/apps/web/cf-env.d.ts b/apps/web/cf-env.d.ts
index e98c36cf..be5c991a 100644
--- a/apps/web/cf-env.d.ts
+++ b/apps/web/cf-env.d.ts
@@ -5,8 +5,9 @@ declare global {
       GOOGLE_CLIENT_SECRET: string;
       AUTH_SECRET: string;
       R2_ENDPOINT: string;
-      R2_ACCESS_ID: string;
-      R2_SECRET_KEY: string;
+      R2_ACCESS_KEY_ID: string;
+      R2_SECRET_ACCESS_KEY: string;
+      R2_PUBLIC_BUCKET_ADDRESS: string;
       R2_BUCKET_NAME: string;
       BACKEND_SECURITY_KEY: string;
       BACKEND_BASE_URL: string;