Nbconvert: jupyter nbconvert error: pattern matched no files, if filename contains `[]`

[ceefour]

If nbconvert-ing a file containing [ ] characters, it throws error. For example:

(base) E:\project_amanah\WorkshopML\Day 3>jupyter nbconvert "Day 3-1, Hands-on Deep Learning [2].ipynb" --to slides --post serve

[NbConvertApp] WARNING | pattern 'Day 3-1, Hands-on Deep Learning [2].ipynb' matched no files

This application is used to convert notebook files (*.ipynb) to various other
formats.
...

Version: jupyter 4.4.0, python 3.6, Windows 10, Anaconda

Workaround: Rename the filename to e.g. Day 3-1, Hands-on Deep Learning2.ipynb (remove the [ ] characters)

[montebhoover]

Confirming I also ran into this issue on on jupyter version 4.4.0, python 3.5.

[LoicAG]

Same issue confirmed on Linux, nbconvert version 5.4.0.dev0, jupyter 4.4.0, python 3.6.6

[takluyver]

It's trying to use the names you give it as glob patterns to match multiple files, and square brackets have meaning for glob patterns.

[LoicAG]

@takluyver we ended up solving this by replacing [ with [[] and ] with []]

[t-makaro]

In general, if you have a string that has a globing issue, you can do

import glob
print(glob.escape(...))

on your problem string to find the workaround.

Here is the relevant code:

https://github.com/jupyter/nbconvert/blob/a6f89c3ea214ec9abe126f85b5a810216d4c9ab4/nbconvert/nbconvertapp.py#L314

We could add a

globbed_files.append(glob.glob(glob.escape(pattern)))
globbed_files.append(glob.glob(glob.escape(pattern+'ipynb')))

to fix this completely.