[IPython-dev] nbconvert: eqnarray environment missing in html output

Jakob Gager gager at ilsb.tuwien.ac.at
Mon Sep 23 02:05:26 EDT 2013

Hi Zoltán,

pandoc strips raw latex (not placed inside $s) when converting to html, but keeps it when converting to 
latex. Vice versa raw html is stripped when converting to latex but retained when converting to html.
So you are actually facing a pandoc shortcoming (I wouldn't call it a bug since this behavior seems to be 
intended). Has been discussed e.g. https://github.com/ipython/ipython/issues/4234

To overcome the png issue you could use svgs instead, try
%config InlineBackend.figure_format = 'svg'


On 09/20/2013 05:38 PM, Zoltán Vörös wrote:
> Hi all,
> Sorry for the spamming, for, in some sense, this message still concerns
> the issue that I raised a couple of hours ago. I converted the notebook
> to html, but it seems to me that the eqnarray environment is complete
> stripped from the output. Why should this happen? Is this an issue in
> pandoc, or in nbconvert? In any case, why is the behaviour different,
> depending on the output format? Any ideas as to what could go wrong?
> Side note: quite a few people were complaining here on the mailing list
> that the print option has been removed from the notebook, and the
> official line is that one could generate the pdf file by calling
>   > ipython nbconvert --to latex somenotebook.ipynb
> and then compiling it with latex. Apart from the above-mentioned issue,
> my experience is that the quality of the pdf is quite low, simply
> because latex doesn't handle png files very well. An alternative is to call
>   > ipython nbconvert --to html somenotebook.ipynb
> and load the output into the browser, and use the browser's
> print-to-file facilities to create the desired pdf file. While it is
> certainly true that figures can be cut into two at page boundaries, yet,
> the quality is much higher than with the latex approach.
> Cheers,
> Zoltán
> _______________________________________________
> IPython-dev mailing list
> IPython-dev at scipy.org
> http://mail.scipy.org/mailman/listinfo/ipython-dev

More information about the IPython-dev mailing list