Require cell regex to be at beginning of line

[kylebarron]

Currently, the regular expression used to delimit cells matches the comment symbol plus %% or other symbols anywhere on the line. That means that using export-notebook on the following splits the string:

# %%

import pandas as pd
import numpy as np

print('# %%')
print('this should be one cell')

Before:

After:

Presumably, run-cell would also fail on this.

[kylebarron]

Do I need to use \\s to represent a space?

[n-riesco]

On a regexp, no, but that's a string!

[kylebarron]

So I do need to use \\s then?

[n-riesco]

yes, we need to escape the backslash.

[kylebarron]

Thanks. That does make sense.

[n-riesco]

I can't look at this right now, but I don't think changing the regexp is the right approach.

I think we should update this part of the code to skip regexp matches for which isComment is false.

[kylebarron]

That makes sense. I might try to find some time to look at that this weekend.

I'm also trying to figure out a good way to refactor some of this code that searches for code cell boundaries in order to carry some metadata about the cell, namely if it's a Markdown cell. Essentially all that's needed to fix #1296 is knowledge about which cells are Markdown and which are code.

[BenRussert]

Good catch. I think the approach is fine, my only hesitation is @n-riesco's comment.

[BenRussert]

Either way, it would be good to add a test to cover this.

[kylebarron]

See #1520