Filenames are urlencoded when import from Notion

[and7ey]

Here is an example how some file names are imported:

Here is an example of full filename:
%D0%A0%D1%83%D0%BA%D0%BE%D0%B2%D0%BE%D0%B4%D1%81%D1%82%D0%B2%D0%BE_%D0%BF%D0%BE_%D0%BF%D0%BE%D0%B4%D0%BA%D0%BB%D1%8E%D1%87%D0%B5%D0%BD%D0%B8%D1%8E_DK103M(W)(-4G)

[tgrosinger]

Would it be possible for you to provide us with a Notion export zip file that we could use to help test and resolve this issue? Thank you!

[and7ey]

@tgrosinger I can not share the export zip since it contains confidential info.
But I can give you more details.

Here is the screenshot from Notion exported html:

the file is referenced somewhere here I think:
<figure id="8d1a494d-0a00-4958-b879-876fe051cf94"><div class="source"><a href="%D0%94%D0%BE%D0%BC%D0%BE%D1%84%D0%BE%D0%BD%20Beward%2016aaf973752a4cc89313e48964bc76a0/%25D0%25A0%25D1%2583%25D0%25BA%25D0%25BE%25D0%25B2%25D0%25BE%25D0%25B4%25D1%2581%25D1%2582%25D0%25B2%25D0%25BE_%25D0%25BF%25D0%25BE_%25D0%25BF%25D0%25BE%25D0%25B4%25D0%25BA%25D0%25BB%25D1%258E%25D1%2587%25D0%25B5%25D0%25BD%25D0%25B8%25D1%258E_DK103M(W)(-4G).pdf">Руководство по подключению DK103M(W)(-4G).pdf</a></div></figure>

The filename is actually URL encoded:

[felciabatta]

Hi @and7ey, was Руководство по подключению DK103M(W)(-4G).pdf the original filename of the PDF stored in your Notion workspace?

If so, then this means Notion is responsible for encoding the filename on export, and perhaps we can decode this with Importer...

[and7ey]

The decoded filename looks Ok

[felciabatta]

I just tested this in my Notion workspace and can confirm the same thing happens when exporting.

Will have a look in decoding back into the original characters — the challenge will be updating this across the Vault for all references to the file.

[felciabatta]

The filename is actually URL encoded:

It's actually even worse than that; this is the component of the href referring to the filename itself (not the full path):

%25D0%25A0%25D1%2583%25D0%25BA%25D0%25BE%25D0%25B2%25D0%25BE%25D0%25B4%25D1%2581%25D1%2582%25D0%25B2%25D0%25BE_%25D0%25BF%25D0%25BE_%25D0%25BF%25D0%25BE%25D0%25B4%25D0%25BA%25D0%25BB%25D1%258E%25D1%2587%25D0%25B5%25D0%25BD%25D0%25B8%25D1%258E_DK103M(W)(-4G).pdf

and compare this to the URL-encoded filename:

%D0%A0%D1%83%D0%BA%D0%BE%D0%B2%D0%BE%D0%B4%D1%81%D1%82%D0%B2%D0%BE_%D0%BF%D0%BE_%D0%BF%D0%BE%D0%B4%D0%BA%D0%BB%D1%8E%D1%87%D0%B5%D0%BD%D0%B8%D1%8E_DK103M(W)(-4G).pdf

They're not the same — the href is a double-encoded version of the original filename:

Руководство по подключению DK103M(W)(-4G).pdf — but not quite — when you decode the href you get:

Руководство_по_подключению_DK103M(W)(-4G).pdf — so it also replaces spaces with underscores!

This leads me to believe this is actually a bug with Notion's export utility: the encoding should only be required for the href, not the actual file itself — it makes no sense to double-encode like this. 😭

[and7ey]

What if we follow the MS Excel approach when importing CSV file - it shows the file content, suggests the encoding and asks if user would like to change it. What if we do the same? Single encoding, double encoding?

[felciabatta]

I think it's probably simpler to default to decoding the filenames — we know how Notion encodes them so we can just apply the same decoder across the board.

For filenames that aren't encoded in the first place, applying the URL decoder should leave them unmodified anyway — although in theory someone could intentionally have %XX strings in their filename...