Pairing status visibility

[mscavnicky]

Thanks a great extension! I've been testing it out for a while now.

Is there a way to see notebook pairing status somewhere? It would be great to see somewhere whether a notebook is already paired (or not).

I have a lot of notebooks, so it's easy to get lost in them.

[caenrigen]

Hi, thanks :)

I am trying to keep the extension code and its executions as minimal as possible.
Checking if a text file or a notebook file is paired would require invoking jupytext most likely, which adds some extra overhead.

But besides that I have no idea if there is a good and reliable way to do this.

[mwouts]

Hi @caenrigen , I can try to make it easier for you to identify which notebooks are paired. Right now you can use jupytext --paired-paths but currently (v1.17.2) it does not load the configuration.

How do you get this information currently? With this content in jupytext.toml: formats="ipynb,py:percent", my .py file should have been paired but the extension said

onDidSaveTextDocument,zil8: Read paired formats for '/Users/marc/GitHub/jupytext/demo/vscode-ext/huge_notebook.py': ''
onDidSaveTextDocument,zil8: No paired formats in '/Users/marc/GitHub/jupytext/demo/vscode-ext/huge_notebook.py', skipping sync

[caenrigen]

Hi @mwouts

I think at some point I saw that --paired-paths seemed to have some unresolved bugs on your issue tracker and I decided to not use it.

How do you get this information currently?

It happens over here

vscode-jupytext-sync/src/jupytext.ts

Lines 320 to 329 in 7d9f999

	const pairedFormats = await readPairedFormats(document.uri.fsPath, logPrefix)
	// This fixes bug when default formats are set inside a config file e.g.
	// project.toml:
	// ```
	// [tool.jupytext]
	// formats = "ipynb,py:percent"
	// ```
	// and `jupytext` uses them right away even thought we never paired the file.
	// In such cases when saving any file matching supported extensions, would get
	// paired unintentionally.

And inside the readPairedFormats I use call jupytext.read via python:

vscode-jupytext-sync/src/jupytext.ts

Lines 440 to 441 in 7d9f999

	export async function readPairedFormats(filePath: string, logPrefix: string = "") {
	const py = `import jupytext; print(jupytext.read('${filePath}').metadata.get('jupytext', {}).get('formats', ''))`

You can understand why it is implemented like this over here: #2 (comment)

There might be solutions that cover both cases (file with self-contained metadata vs jupytext.toml), but I did not have the bandwidth to develop complex solutions. I am open to discuss possible solutions. Your jupytext expertise and experience is likely very valuable (there might be things I did not consider simply because I am not aware of everything that jupytext is capable of).

Side note, there is also an open issue with very long discussion (probably not worth reading in detail) #5 that is related to respecting the jupytext.toml.

[mwouts]

I see! I will add a Python function in Jupytext that will give you the paired formats for a given file. Also, based on our other chat at #12 (comment) I will make sure that Jupytext leaves unpaired files unchanged - this way you won't need to run two commands, just one, and this will avoid having to read big notebooks twice.