Results

• Embed GitHub

Doc 2 notion flow

• I add a line of text in any doc synced with my one drive
• I run the script manually
• saves the date/time when started to run in a variable runStartTimestamp
• get last time the script started to run
• const lastRunTimestamp = ./lastRunTimestamp.txt
• gets the files names in the folders to be scanned
• list of folders to be scanned will be a hardcoded array in V1
Copy

[{
	"file_path": "<file_path>",
	"files": []
}]

• Save the remaining file names as objects in the file_path array item as

[{
	"file_path": "<file_path>",
	"files": [{
		"name": "<file_name>",
		"todos": []
	}]
}]

• get stats for every file name in the list and save in the obj as:

[{
	"file_path": "<file_path>",
	"files": [{
		"name": "<file_name>",
		"last_modified": yyyy-MM-dd'T'HH:mm:ss.SSSZ
		"todos": []
	}]
}]

• filter file names that have end with .docx and stat.mtime > lastTimeScriptRan
• for each file name resulting from the filter
• get the the word/comments.xml content and saves in a variable commentsContent
• unzips the .docx file to a local directory ./tmp using bash command unzip
• reads the content of the word/comments.xml file using fs.readFileSync
• extract the TODOs with pattern below from file content
• TODO <task name> by <due date YYYY-MM-dd> [(tracked)]
• Convert todo strings to list of objs by parsing fields from string
Copy

[{
	"original_text": "TODO <task name> by <due date YYYY-MM-dd> [(tracked)]",
  "desc": "<task name>",
	"due_date": "YYYY-MM-dd",
	"track_status": false | true
	"error_msg": ""
}]

• filter untracked todos where track_status = false
• for each untracked todo in the list
• create task in notion
• Use the logic in this tutorial https://developers.notion.com/docs/create-a-notion-integration
• POC code implemented in notion_poc.js
• field mapping
• desc > name of the DB item
• due date > Date prioritized
• file_name > Src file name
• file_path > Src file path
• if success
• update text in the note to indicate with (tracked) to indicate that the todo has been tracked
• update in commentsContent with the content of comments.xml using a regex to identify the text of the original todo and replace with:
• text + ' (tracked)'
• update note.track_status = true
• if fail
• update note.error_msg = <error message>
• Append list of untracked todos (that now might be tracked) objs to corresponding file obj in file.todos as
Copy

[{
	"file_path": "<file_path>",
	"files": [{
		"name": "<file_name>",
		"last_modify_dt": yyyy-MM-dd'T'HH:mm:ss.SSSZ
		"todos": [{
			"original_text": "TODO <task name> by <due date YYYY-MM-dd> [(tracked)]",
			"desc": "<task name>",
			"due_date": "YYYY-MM-dd",
			"track_status": false
			"error_msg": ""
		}]
	}]
}]

• update the content of word/comments.xml file in ./tmp with the value of commentsContent using fs.writeFileSync
• Zip the contents of ./tmp in a .docx file with the same name as the original file and write it in the original path in the OneDrive folder
• saves execution report
• saves in local file in ./.logs/<lastTimeScriptRan>.txt with lastTimeScriptRan in same format as 2023-07-25T23:18:07.000Z and in UTC
• Execution report should be the final json used through the execution generated by JSON.stringify(report, null, 2)
Copy

[{
	"file_path": "<file_path>",
	"files": [{
		"name": "<file_name>",
		"todos": [{
			"desc": "<task name>",
			"due_date": "YYYY-MM-dd",
			"track_status": false
			"error_msg": ""
		}]
	}]
}]

• saves runStartTimestamp value in ./lastTimeScriptRan.txt

Insights

• Had to change the approach from using the content of the document.xml to comments.xml instead because content in document.xml was weird sometimes. Ex: TODO create script by Jul 26 would appear as below
• I thought this was hard to predict and the contents of the comments seemed more standardized

<w:p w14:paraId="1921F443" w14:textId="1511460D" w:rsidR="00015415"
    w:rsidRDefault="00015415" w:rsidP="00015415">
    <w:pPr>
        <w:pStyle w:val="ListParagraph" />
        <w:numPr>
            <w:ilvl w:val="0" />
            <w:numId w:val="2" />
        </w:numPr>
    </w:pPr>
    <w:commentRangeStart w:id="0" />
    <w:r w:rsidRPr="00937622">
        <w:t xml:space="preserve">TODO create </w:t>
    </w:r>
    <w:r w:rsidR="00937622" w:rsidRPr="00937622">
        <w:t>script by Jul 2</w:t>
    </w:r>
    <w:r w:rsidR="00937622">
        <w:t>6</w:t>
    </w:r>
    <w:commentRangeEnd w:id="0" />
    <w:r w:rsidR="00416F57">
        <w:rPr>
            <w:rStyle w:val="CommentReference" />
        </w:rPr>
        <w:commentReference w:id="0" />
    </w:r>
</w:p>

• Note for future versions, by using the comments it’s possible to mark the comment as “resolved” by changing the prop done of the commentsExtended.xml file like below. This will not be implemented in V1, but seems like a nice feature for the Notion to doc flow:

<w15:commentEx w15:paraId="58C6F26B" w15:done="1" />

• When I have the doc open in my desktop and the script runs updating a comment, the content of the doc is not updated automatically. If I close the doc and reopen without changing anything, the comment appears updated.
• If I update the doc after it was updated by the script and save it, than the script changes are overwritten. If I then close and reopen the doc, my changes are there, but the script changes are not

Notion to doc flow

TODO in next version