m4rkk · May 15, 2025 21:06 · m4rkk · May 15, 2025
diff --git a/md2n.sh b/md2n.sh
 #!/usr/bin/env bash
 #
 # md2n.sh — Move MarkDown document into notion.
 #
 # DESCRIPTION:
 #   Preserves code blocks and provides a copy button on each block.
 #   Yaml blocks are suspect, so patch them up with pandoc.
 #
 # DEPENDENCIES:
 #   - Pandoc - tried and true document converter tool.
 #   - fix-yaml.lua; this simple lua parser is expected at ~/bin/fix-yaml.lua,
 #  		it depends on pandoc to do the work once a block is identified.
 #   - Python - pushes to Notion.
 #	- API keys, Notion User Token V2, and parent page ID expected to be found
 #   	in a .env file in your ~/secrets/notion_keys.env.
 #   - ChatGPTExport or some other tool to generate the MarkDown file from threads.
 #
 # USAGE:
 #   md2n.sh <markdown-file> [page-title]
 #
 # AUTHOR:
 #   m4rkk
 #

 set -euo pipefail

 # Load environment variables from well-known locations
 CONFIG_DIR="${XDG_CONFIG_HOME:-$HOME/secrets}"
 for envfile in "$PWD/.env" "$CONFIG_DIR/notion_keys.env"; do
  if [ -f "$envfile" ]; then
    echo "Loading environment from $envfile"
    set -o allexport
    # shellcheck disable=SC1091
    source "$envfile"
    set +o allexport
    break
  fi
 done

 usage() {
  echo "Usage: $(basename "$0") <markdown-file> [page-title]"
  exit 1
 }

 if [ $# -lt 1 ] || [ $# -gt 2 ]; then
  usage
 fi

 SRC="$1"
 # Derive title from filename if not provided
 title_param="${2:-}"
 TITLE="${title_param:-$(basename "$SRC" .md)}"
 TMP="$(mktemp /tmp/fixed.XXXXXX).md"

 # Verify required environment variables
 : "${NOTION_TOKEN_V2:?Environment variable NOTION_TOKEN_V2 must be set (or defined in .env)}"
 : "${NOTION_PARENT_PAGE_ID:?Environment variable NOTION_PARENT_PAGE_ID must be set (or defined in .env)}"

 # 1) Fix YAML fences
 pandoc -f markdown -t gfm+yaml_metadata_block --lua-filter=$HOME/bin/fix-yaml.lua \
  -o "$TMP" "$SRC"

 # 2) Import into Notion with custom title
 PARENT_URL="https://www.notion.so/${NOTION_PARENT_PAGE_ID}"

 python3 - <<EOF
 import os
 from notion.client import NotionClient
 from notion.block import PageBlock
 from md2notion.upload import upload

 token = os.environ['NOTION_TOKEN_V2']
 parent_url = "$PARENT_URL"
 title = "$TITLE"
 file_path = "$TMP"

 client = NotionClient(token_v2=token)
 parent = client.get_block(parent_url)
 # Create a new child page with the given title
 new_page = parent.children.add_new(PageBlock, title=title)
 # Upload markdown content into that new page
 with open(file_path, "r", encoding="utf-8") as mdFile:
    upload(mdFile, new_page)
 EOF

 # Clean up
 echo -e "\nImported to Notion under parent ${NOTION_PARENT_PAGE_ID} with title \"${TITLE}\""
 rm "$TMP"
diff --git a/zfix-yaml.lua b/zfix-yaml.lua
 -- fix-yaml.lua
 -- Part of the md2n.sh solution to publish ChatGPT threads as Notion pages. See md2n.sh for details.
 -- Author: m4rkk 

 function CodeBlock(el)
  -- if it was a fenced block with no info-string…
  if #el.classes == 0 then
    -- and the first line looks like YAML key:value…
    local first = el.text:match("([^\n]+)")
    if first and first:match("^%s*[%w_%-]+%s*:%s*.+") then
      return pandoc.CodeBlock(el.text, {class="yaml"})
    end
  end
  return el
 end
	#!/usr/bin/env bash
	#
	# md2n.sh — Move MarkDown document into notion.
	#
	# DESCRIPTION:
	# Preserves code blocks and provides a copy button on each block.
	# Yaml blocks are suspect, so patch them up with pandoc.
	#
	# DEPENDENCIES:
	# - Pandoc - tried and true document converter tool.
	# - fix-yaml.lua; this simple lua parser is expected at ~/bin/fix-yaml.lua,
	# it depends on pandoc to do the work once a block is identified.
	# - Python - pushes to Notion.
	# - API keys, Notion User Token V2, and parent page ID expected to be found
	# in a .env file in your ~/secrets/notion_keys.env.
	# - ChatGPTExport or some other tool to generate the MarkDown file from threads.
	#
	# USAGE:
	# md2n.sh <markdown-file> [page-title]
	#
	# AUTHOR:
	# m4rkk
	#

	set -euo pipefail

	# Load environment variables from well-known locations
	CONFIG_DIR="${XDG_CONFIG_HOME:-$HOME/secrets}"
	for envfile in "$PWD/.env" "$CONFIG_DIR/notion_keys.env"; do
	if [ -f "$envfile" ]; then
	echo "Loading environment from $envfile"
	set -o allexport
	# shellcheck disable=SC1091
	source "$envfile"
	set +o allexport
	break
	fi
	done

	usage() {
	echo "Usage: $(basename "$0") <markdown-file> [page-title]"
	exit 1
	}

	if [ $# -lt 1 ] \|\| [ $# -gt 2 ]; then
	usage
	fi

	SRC="$1"
	# Derive title from filename if not provided
	title_param="${2:-}"
	TITLE="${title_param:-$(basename "$SRC" .md)}"
	TMP="$(mktemp /tmp/fixed.XXXXXX).md"

	# Verify required environment variables
	: "${NOTION_TOKEN_V2:?Environment variable NOTION_TOKEN_V2 must be set (or defined in .env)}"
	: "${NOTION_PARENT_PAGE_ID:?Environment variable NOTION_PARENT_PAGE_ID must be set (or defined in .env)}"

	# 1) Fix YAML fences
	pandoc -f markdown -t gfm+yaml_metadata_block --lua-filter=$HOME/bin/fix-yaml.lua \
	-o "$TMP" "$SRC"

	# 2) Import into Notion with custom title
	PARENT_URL="https://www.notion.so/${NOTION_PARENT_PAGE_ID}"

	python3 - <<EOF
	import os
	from notion.client import NotionClient
	from notion.block import PageBlock
	from md2notion.upload import upload

	token = os.environ['NOTION_TOKEN_V2']
	parent_url = "$PARENT_URL"
	title = "$TITLE"
	file_path = "$TMP"

	client = NotionClient(token_v2=token)
	parent = client.get_block(parent_url)
	# Create a new child page with the given title
	new_page = parent.children.add_new(PageBlock, title=title)
	# Upload markdown content into that new page
	with open(file_path, "r", encoding="utf-8") as mdFile:
	upload(mdFile, new_page)
	EOF

	# Clean up
	echo -e "\nImported to Notion under parent ${NOTION_PARENT_PAGE_ID} with title \"${TITLE}\""
	rm "$TMP"
	-- fix-yaml.lua
	-- Part of the md2n.sh solution to publish ChatGPT threads as Notion pages. See md2n.sh for details.
	-- Author: m4rkk

	function CodeBlock(el)
	-- if it was a fenced block with no info-string…
	if #el.classes == 0 then
	-- and the first line looks like YAML key:value…
	local first = el.text:match("([^\n]+)")
	if first and first:match("^%s[%w_%-]+%s:%s*.+") then
	return pandoc.CodeBlock(el.text, {class="yaml"})
	end
	end
	return el
	end