Import Markdown pages (e.g Hugo) into Notion

import-markdown-files-into-notion.py

import os
import re
import json

from pprint import pprint
from subprocess import Popen, PIPE, STDOUT

from notion_client import Client

DATABASE_ID = os.getenv('DATABASE_ID') #NOTE: ID of your notion database you want to import into
NOTION_API_KEY = os.getenv('NOTION_API_KEY')

notion = Client(auth=NOTION_API_KEY)

def get_meta(name, content):
    return re.search(r'^' + name + r':(.*)$', content, re.MULTILINE).group(1).strip().strip('"').replace('\\', '')

# NOTE: You might want to extend this to more types (e.g `select` or `number`) in case you need to
class Notion:
    def title(content):
        return { "title": [{ "text": { "content": content } }] }

    def rich_text(content):
        return { "rich_text": [{ "text": { "content": content } }] }

    def date(content):
        return { "date": { "start": content } }

    def multi_select(array):
        return { "multi_select": list(map(lambda item: { "name": item }, array)) }

# NOTE: Make sure that your current working directory is close to your markdown articles
for article in os.listdir():
    if not os.path.isdir(f'./{article}'):
        continue

    # NOTE: My setup has article-name/index.md, so edit this to whatever your setup is
    file = open(f'./{article}/index.md', 'r')
    content = ''.join(file.readlines())
    file.close()


    title = get_meta('title', content)
    subtitle = get_meta('subtitle', content)
    date = re.sub(r'T.*$', '', get_meta('date', content))
    tags = [
        str.strip(string) for string in get_meta('tags', content) \
            .replace('[', '') \
            .replace(']', '') \
            .split(',') if string
    ]
    # NOTE: I had a custom property for cover image, you might not need this
    cover = re.search(r'^.*image:(.*)$', content, re.MULTILINE).group(1).strip().strip('"').replace('\\', '')

    if not 'http' in cover:
        cover = ''

    content = re.sub(r'^---(\n.+)+(?!---)', '', content, 1).strip()
  
    # NOTE: We use an npm/node library to convert our markdown into the JSON format we need. I could've ported it, but didn't feel like it.
    # NOTE: Update this to wherever your `node` binary lives and where you put the markdown-to-notion.js
    p = Popen(['/usr/local/bin/node', './markdown-to-notion.js'], stdout=PIPE, stdin=PIPE, stderr=PIPE)
    data = p.communicate(input=content.encode())[0]
    blocks = json.loads(data)

    notion.pages.create(
        **{
            "parent": {
                "type": "database_id",
                "database_id": DATABASE_ID
            },
            "cover": {
                "type": "external",
                "external": {
                    "url": cover
                }
            } if cover != "" else None,
            # NOTE: Make sure that these properties correspond in name and type with your database, edit otherwise
            "properties": { 
                "Title": Notion.title(title),
                "Subtitle": Notion.rich_text(subtitle),
                "Slug": Notion.rich_text(article),
                "Date": Notion.date(date),
                "Tags": Notion.multi_select(tags)
            },
            "children": blocks
        }
    )

markdown-to-notion.js

const {markdownToBlocks, markdownToRichText} = require('@tryfabric/martian');

var string = require('fs').readFileSync(0, 'utf-8');
console.log(JSON.stringify(markdownToBlocks(string), null, 4))