nbconvert: Ability to specify name of output file

[jdfreder]

Brought up in a conversation between @Carreau and I. Maybe add a traitlet to the FilesWriter and add a convenience flag to the nbconvertapp.py.

[jdfreder]

Open for conversation

[minrk]

Makes sense - should be easy, right?

[minrk]

Does it make sense for it to be a configurable on FilesWriter? Or not, because one writer can be used to write out many notebooks.

[dpsanders]

+1

[jdfreder]

Well, I think the FilesWriter is supposed to be stateless... But I was not thinking about supporting the renaming individual outputs, I was more thinking of the ability to prefix/suffix each file written (not the extracted files).

Something like:

output_name = Unicode(u'{name}.{extension}', config=True)

Which would be useful if someone wanted to export reveal and html which doesn't work now since the output names would cause the later to overwrite the earlier.

[jdfreder]

@Carreau ?

[minrk]

But the question is where to you specify that, such that it is configurable per-notebook in a multi-export run. Perhaps in NbConvertApp, and only allow it when len(notebooks) == 1?

[jdfreder]

I personally was not thinking about putting it in nbconvertapp.py, I was just thinking about replacing this line with a str.format(...)

[jdfreder]

And not even caring if the user doesn't include the {name} key (or whatever it will be called)...

[minrk]

ok