splitbee / notion-api-worker Goto Github PK

Currently the list is limited to 100 lines. In Notion's API it is possible to pass a cursor to get information from a specific part of the table

Calling whatever endpoint, like different page IDs, I always get a routing error:

{"error":"Route not found!","routes":["/v1/page/:pageId","/v1/table/:pageId","/v1/user/:pageId"]}

How about replacing the code with slice instead of deprecated substr in idToUuid function in /api/utils.ts?

I think we'd be better off matching Notion API's versioning for now.

If we have a major breaking change when Notion releases their public API, that's fine.. Maybe we should even make it explicit and have the URL prefix be /private/v3/?

Thanks for sharing this nice work and react-notion as well.

I'm trying to host it using Cloudflare, but it is a little bit difficult for me because I don't have an experience with it.

Could you please give me some information about how to host it?

There is a phenomenon where the content of notion gets cut off when it gets longer.

I tested the API with a large table and realized that only the first few items are returned. Is it possible to ask the API to fetch the next items?

Hey many thanks for working on this project. I came across the project, looking for a way to make a meaningful backup of our notion knowledge repo. However the preview is not functioning at the moment.

{"error":"Route not found!","routes":["/v1/page/:pageId","/v1/table/:pageId"," /v1/user/:pageId"]}

Sorry to bother! My Bad

They changed their API today. All loadPageChunk calls will fail as-is.

See fix here https://github.com/NotionX/react-notion-x/pull/39/files

Related: #27

Hi,

I'm getting a 400 error whenever I try to load an image that's stored as a file in a private table. I'm using my own worker with the correct NOTION_TOKEN because I can retrieve the info, but the image won't load. If I make the table publicly available then it loads just fine. Any ideas what may be the issue?

wrangler dev currently has a nasty bug that doesn't respond to ipv4 requests as outlined in this issue: cloudflare/wrangler-legacy#1198

Just a head's up.

I think it would be cool to automatically include the blockMap of each item inside a table (optionally).
Currently I have to manually do a request for each entry to receive it. Doing it on the server with caching sounds like a good idea.

Hello - I noticed while trying to read a private page on my Notion that most of my content looks okay, but when I used a toggle and placed content inside, the inside content returns JSON similar to this:

{
  "d53beb18-ae39-4422-ad5a-d23712d2d862": {
    "role": "none"
  }
}

Note that I'm passing the token as an environment variable rather than putting the AP in fetch as an Authorization Bearer. When the page becomes public the fetching of blocks work alright.

I think I've found a possible solution to this and am going to submit a PR shortly, but submitting an issue for the heads up.

After a recent update, many blocks are no longer returned for our pages. One page dropped 50-100 blocks from the end of the page.

In page bc7f910ef23b421487ca21b959c67c49, block 771f85de-f188-46e6-90d6-9b2dd814ba10 is listed in the content of the page, but the block content is not actually included.

We noticed that by deleting early blocks on the actual Notion page, missing blocks can be fetched by the API — in other words, blocks at the end of the page are more likely to be missing.

Hi,

Thank you for making this API 🙏. It has been really helpful.

Whenever I try to use the /page/<id> API on an empty page, it's throwing an error. Not sure why it's happening. Is there any way to get around this?

I get Access Denied message for image block. How to access the URL? My notion is public.

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/6b42051e-e390-4f74-a790-a9347a5adc65/1.jpg

This example from the readme is not working. When doing the request, I get "Error 1101 - Worker threw exception" from Cloudflare. Using Next.js, running npm run dev throws a FetchError for invalid json response body (Unexpected token < in JSON at position 0).

Is it me? Can you recreate?

On Safari, when making a request as documented with the Authorization header (with bearer token) I get the error:

Request header field Authorization is not allowed by Access-Control-Allow-Headers

This seems to indicate that this package cannot actually accept the Authorization header for private notion docs.
Is there a way around this?

Does this mean that the only way to use this package for private notion docs is to host our own worker?

The current implementation doesn't respect no-cache directives which makes debugging more difficult.

Specifically, we should disable caching if we encounter: pragma: no-cache, cache-control: no-store, or cache-control: no-cache directives.

Guys @transitive-bullshit @tobiaslins Could this worker also work with Coda.io tables?

Hi. Currently it seems that we can get only the part of date (YYYY-MM-DD), so I also want to get the part of time (HH:mm) if a property has

Hey, this could just be a user error, but I've tried both approaches and can't get private pages to return any BlockMap from the Splitbee Notion worker, or a personal worker that I spun up.

Personal Worker:

I created the notion worker from a local clone and uploaded it to CloudFlare:

And this is what comes back when I search a private page

This is what that said paid looks like in Notion:

I got the token from the cookie in notion it's key was "token_v2" so I'm pretty sure I grabbed the right thing.

Please let me know if I'm doing something wrong. Thanks.

Hi, this package works well for my use case (grabbing data from a Notion table, nice and simple) but it would be nice to have auth support via the token.

The notion-api-agent repo has some docs on how to get your token:
https://github.com/dragonman225/notionapi-agent/blob/master/documentation/get-token/get-token.md

I've tried to add support for it but sadly tiny-request-router doesn't support headers or query params so I'd like to get your opinion first, Is it something you would like to add? And do you have thoughts about changing to a different router option?

Something like POST /v1/search:

export interface SearchParams {
  query: string
  ancestorId: string
  limit?: number
}

export async function search(params: SearchParams) {
  const body = {
    type: 'BlocksInAncestor',
    source: 'quick_find_public',
    ancestorId: params.ancestorId,
    filters: {
      isDeletedOnly: false,
      excludeTemplates: true,
      isNavigableOnly: true,
      requireEditPermissions: false,
      ancestors: [],
      createdBy: [],
      editedBy: [],
      lastEditedTime: {},
      createdTime: {}
    },
    sort: 'Relevance',
    limit: params.limit || 20,
    query: params.query
  }

  return fetch('https://www.notion.so/api/v3/search', {
    method: 'POST',
    headers: {
      'content-type': 'application/json'
    },
    body: JSON.stringify(body)
  }).then((res) => res.json())
}

This is necessary to get around CORS from the client-side.

Maybe I'm missing something, but it looks like you can access private pages just with the user token. Is it possible to also use the workspace token? So you won't be able to access all the private pages, but just the pages from that particular workspace.
In the official Notion doc https://developers.notion.com/docs/authorization, there are these internal integrations, which are authorized by the workspace token.

Since Notion has launched their API in public beta - https://developers.notion.com/
Would be great if that’s integrated directly.

Here's a function which extracts the pageId in a consistent format (no dashes) and ignores any additional prefix (like my-page-about-e5a735e33baa458b988993b55a54ee54):

const pageIdRe = /\b([a-f0-9]{32})$/
const pageId2Re = /\b([a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12})$/

/** Extracts the notion page ID from a notion URL suffix */
export const parsePageId = (id: string) => {
  id = id.split('?')[0]
  const match = id.match(pageIdRe)

  if (match) {
    return match[1]
  }

  const match2 = id.match(pageId2Re)
  if (match2) {
    return match2[1].replace(/-/g, '')
  }
}

And some ava unit tests that will probably be useful:

import test from 'ava'
import * as notion from './notion'

const pageIdFixturesSuccess = [
  '267c0d1f1df8457f9b5c8f7efca16d83',
  'Twitter-Automation-Tool-267c0d1f1df8457f9b5c8f7efca16d83',
  'www.notion.so/saasifysh/Twitter-Automation-Tool-267c0d1f1df8457f9b5c8f7efca16d83',
  'www.notion.so/saasifysh/Twitter-Automation-Tool-267c0d1f1df8457f9b5c8f7efca16d83?foo=bar&bar=foo',
  'https://www.notion.so/saasifysh/Standalone-Notion-Hosting-717a3608b1874cc5bafb5b9680b53395',
  'Standalone-Notion-Hosting-717a3608b1874cc5bafb5b9680b53395',
  'Standalone-Notion-Hosting-717a3608b1874cc5bafb5b9680b53395?foo',
  '-717a3608b1874cc5bafb5b9680b53395',
  '717a3608b1874cc5bafb5b9680b53395',
  '717a3608b1874cc5bafb5b9680b53395?',
  'e5a735e3-3baa-458b-9889-93b55a54ee54',
  'fde5ac74-eea3-4527-8f00-4482710e1af3',
  'about-e5a735e3-3baa-458b-9889-93b55a54ee54',
  '.com/about-e5a735e3-3baa-458b-9889-93b55a54ee54',
  'About-d9ae0c6e7cad49a78e21d240cf2e3d04'
]

const pageIdFixturesFailure = [
  '717A3608b1874CC5bafb5b9680b53395',
  '717A36',
  '',
  'notion.so/saasifysh/Twitter-Automation-Tool-267c0d1f1df8457f9b5c8f7efca16d83abc',
  'a267c0d1f1df8457f9b5c8f7efca16d83',
  '267c0d1f1df8457f9b5c8f7efca16d83a',
  '267c0d1f1%f8457f9b5c8f7efca16d83',
  'Twitter-Automation-Tool',
  'fde5ac74-eea3-4527-8f00-4482710e1af'
]

test('notion.parsePageId success', (t) => {
  for (const id of pageIdFixturesSuccess) {
    const parsedId = notion.parsePageId(id)
    t.truthy(parsedId)
    t.snapshot(parsedId)
  }
})

test('notion.parsePageId failure', (t) => {
  for (const id of pageIdFixturesFailure) {
    const parsedId = notion.parsePageId(id)
    t.falsy(parsedId)
  }
})

I recommend using HTTP response fixtures from real notion API responses via nock and making sure that requests to this server always return the same result given the same inputs either via jest or ava snapshots.

Hey guys!

First of all, really love the project so keep that up. I was wondering if it's on the roadmap for a way of posting or adding data to pages or tables.

Curious if you folks have experimented with writing data back to Notion? What does the format of a block look like? Markdown, a JSON schema? Something else?

cache.put(request, response) uses the response's cache-control header to determine the object's TTL (and other standard headers).

Currently, we're passing responses with no cache-control which isn't really well-specified by Cloudflare as to what this actually means.

Experimentally, it appears that CF caches these objects for ~10 seconds, but we should really override this with a default value.

Note that for Notion2Site, I'm passing pragma: no-cache to disable caching within the worker because Vercel's SWR edge caching is all that's needed. Any additional caching of the page data just introduces additional lag before updates are reflected to their edge cache.

I have a use case where multiple users will be making use of a tool and so self-hosting and changing the NOTION_TOKEN as recommended wouldn't be a viable solution.

I've been adding the Authorization: Bearer <NOTION_TOKEN> header to my requests to private pages, but have noticed that because those pages are cached for a few seconds I if I then request the same URL without the authorisation header I can load private page data (tested from a separate device, so it's not just browser caching).

Either private pages shouldn't be cached, or their authorisations should be cached with them to prevent this.

Hello,

Let's see a screenshot of my notion page (available here)

And now, let's see a screenshot of my notion integration on my portfolio

As you can see, the content that are nested into the toggle heading aren't displaying. When I turn the toggle heading in classical heading, the nested content are successfully displaying.

Here is my code

NotionEmbed.jsx

import React, {useEffect, useState} from 'react'
import { useTranslation } from 'react-i18next'
import { NotionRenderer } from 'react-notion'
import "react-notion/src/styles.css"

export default function NotionEmbed(props) {
  const { t, i18n } = useTranslation();
  const { notionPageId } = props;
  let [notionContent, setNotionContent] = useState({});

    useEffect(() => {
        async function getStaticProps() {
            const data = await fetch(
                `https://notion-api.splitbee.io/v1/page/${notionPageId}`
            ).then(res => res.json());
            setNotionContent(data);
        }
        getStaticProps();
    }, []);

    useEffect(() => {
        console.log(notionContent)
    }, [notionContent]);

  return (
    <>
        <div className="notion-embed">
            <NotionRenderer blockMap={notionContent} fullPage={true} />
        </div>
    </>
  )
}

Courses.jsx (the component where the notion embed is used)

import './Courses.scss'
import React from 'react'
import { useTranslation } from 'react-i18next'
import NotionEmbed from "../NotionEmbed/NotionEmbed";

export default function Courses() {
  const { t, i18n } = useTranslation();
  return (
    <>
      <section id="section-courses">
        <h2 className="courses-title">
          {t('courses.MainTitle')}{' '}
            <span className="color-tertiary">{t('courses.TitleSpan')}</span>
        </h2>
        <div className="courses-container">
            <NotionEmbed notionPageId="1a9d25ee-18f8-4b92-98a3-d1a1e8149863" />
        </div>
      </section>
    </>
  )
}

And finally, here is my blockMap content :

pastebin

Could smbdy help me with this issue ? Thank you guys 💯

I am getting this error when I am using images from my notion doc

This is my notionId: 96defe8701444f0782cc9b428929c79a

I am rendering the page content using react-notion-x, those are working fine but the image URLs I am getting from the payload are not working

Would appreciate any help on this one 😊

I'm trying to deploy to cloudflare worker and after cloning and building the repository with the updated attributes, I get the following exception raised whilst trying to access the page;

{
  "outcome": "exception",
  "scriptName": "notion-cloudflare-worker",
  "exceptions": [
    {
      "name": "TypeError",
      "message": "Cannot read properties of undefined (reading 'block')",
      "timestamp": 1678203099458
    }
  ],
  "logs": [],
  "eventTimestamp": 1678203099274,
  "event": {
    "request": {
      "url": "https://notion-cloudflare-worker.XYZ.workers.dev/v1/page/REDACTED",
      "method": "GET",
      "headers": {
        "accept": "application/json, text/plain, */*",
        "accept-encoding": "gzip",
        "authorization": "REDACTED",
        "cf-connecting-ip": "86.8.247.150",
        "cf-ipcountry": "GB",
        "cf-ray": "7a43e2fa683c3867",
        "cf-visitor": "{\"scheme\":\"https\"}",
        "connection": "Keep-Alive",
        "host": "notion-cloudflare-worker.XYZ.workers.dev",
        "user-agent": "axios/0.21.1",
        "x-forwarded-proto": "https",
      },
      "cf": {
        "clientTcpRtt": 16,
        "longitude": "-2.55310",
        "latitude": "51.47710",
        "tlsCipher": "AEAD-AES128-GCM-SHA256",
        "continent": "EU",
        "asn": 5089,
        "country": "GB",
        "tlsClientAuth": {
          "certIssuerDNLegacy": "",
          "certIssuerSKI": "",
          "certSubjectDNRFC2253": "",
          "certSubjectDNLegacy": "",
          "certFingerprintSHA256": "",
          "certNotBefore": "",
          "certSKI": "",
          "certSerial": "",
          "certIssuerDN": "",
          "certVerified": "NONE",
          "certNotAfter": "",
          "certSubjectDN": "",
          "certPresented": "0",
          "certRevoked": "0",
          "certIssuerSerial": "",
          "certIssuerDNRFC2253": "",
          "certFingerprintSHA1": ""
        },
        "tlsVersion": "TLSv1.3",
        "colo": "LHR",
        "timezone": "Europe/London",
        "city": "Bristol",
        "edgeRequestKeepAliveStatus": 1,
        "requestPriority": "",
        "httpProtocol": "HTTP/1.1",
        "region": "England",
        "regionCode": "ENG",
        "asOrganization": "Virgin Media",
        "postalCode": "REDACTED"
      }
    },
    "response": {
      "status": 500
    }
  },
  "id": 12
}

I can't find a slug property in any page, but this code seems to suggest that there is, or at least at one point was:

https://github.com/noahsalvi/sotion/blob/ab2615b8157dffcde9fb96f4d2127660b7398268/src/sotion.ts#L94-L97

In addition, the sotion library barely works without the slug property on any page.

This looks very interesting... what are the benefits of using this over directly accessing Notion's loadPageChunk or getRecordValues endpoints, though? That'd be nice to have documented somewhere.

https://notion-api.splitbee.io/v1/page/76715632-d151-400d-8eb6-e7f1cc15ad6f

This API throw error and Idk what is the problem?

https://cuthanh.notion.site/Morphling-dApp-Tham-gia-Binance-Launchpad-76715632d151400d8eb6e7f1cc15ad6f