Enabling inline for SAS outputs in notebook

[SundarKrishnan]

I downloaded the notebook as PDF and I do not see the SAS outputs in the file. It shows the following in the PDF output "<IPython.core.display.HTML object>" instead of the actual output and graphs. Is there a way to enable inline magics options as the same way we do in Python notebook, so that the plots can also be include when we download the notebook as PDF.

SAS_Hello_Word.pdf

Foreg: %matplotlib inline
After I execute the above command using Python Shell, I can see graph when I download the notebook as PDF.

[jld23]

@SundarKrishnan, I've been looking into this but can't replicate your results.
I assume you're arriving at this pdf by doing file --> Download As --> PDF via latex
Can you share the jupyter version (jupyter --version) and the OS (uname -a).
If you do a print preview and then print the preview to pdf do you get the same results?
Does the download as HTML work as expected?

[SundarKrishnan]

Yes you are right, I use the PDF via Latex option to generate the PDF file. I tried to do the preview and then download as PDF. It works fine. But it adds extra work to the user. And HTML will also work. But we are looking for the inbuilt option in Jupyter to work so that it will be handy to users. Again, there is some floating table of contents extensions that we have installed, so when downloading it as a PDF is there a way to have the table of contents extensions as well pop-up in the output file? (Attached the image with this for floating table of contents)

OS - Linux ebiapps-wc-rf.sys.comcast.net 3.10.0-123.20.1.el7.x86_64 #1 SMP Thu Jan 29 18:05:33 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

Jupyter Version - 4.1.0

Thanks.

[jld23]

@SundarKrishnan, I'm still looking into this issue but wanted to provide an update. The problem is not with the sas_kernel (at least not exclusively). It is the rendering of HTML results by IPython.display. See the attached pdf:
Untitled8 (7).pdf

I also love the table of contents nbextension, but it doesn't appear to be included with the PDF or HTML output from Jupyter. The subsystem Jupyter uses for this type of conversion it nbconvert.

[jacobmatthewmurray]

I am working with the SAS kernel in jupyter, but am having the issue described here when trying to download to pdf. Have there been any updates?

[mgdondon]

Hello,

The problem is not with the sas_kernel (at least not exclusively). It is the rendering of HTML results by IPython.display.

It's related to this issue : jupyter/nbconvert#474.

An alternative could be to allow listing output as in SASmarkdown package :

SASmarkdown_notebook.pdf

Thanks,

[mgdondon]

The notebook-as-pdf extension allows to export html to pdf [notebook_html_export_pdf_via_html.pdf] and therefore to generate beautiful pdf [notebook_Jupyter_SAS.pdf].

Regards,

[fsmunoz]

Hello,

I've found this issue again, and I'm not sure if it's better to reopen a new one or to add to this (or even if this is something that is related to sas kernel or SASpy) - please advice, I'll follow whichever path is more adequate.

I've been experimenting with markdown-based systems for a while, and specifically trying to use them with SAS. I've got some examples of using it with Rmd and a discussion about how it would be great to support non-HTML target at the SASpy layer. More recently, I've been experimenting with Quarto, with some success.

Here is where the <IPython.core.display.HTML object> issue raises its head again. I haven't checked how Quarto does it (I'm assuming that there is some existing plumbing at the Python and R integration level that changes the output format based on the detected destination), but when exporting to PDF (and probably all others), the HTML output is changed to whatever can be included (i.e., an image is present in the PDF, as it is in HTML).

With the SAS jupyter kernel, the table looks great (so there is some markdown/LaTeX conversion going on at some level), but the plot is wrapped inside HTML and, alas, isn't displayed:

In the SASpy example before, I've used some Python code (BeautifulSoup based) to naively extract the image and make it work, but perhaps considering ODS rich output capabilities there's a different way? In other words, is the ability to change the type of output depending on a detected output format something possible, and if so, would it be at the level of sas_kernel, SASpy, or both? I'm willing to dig on it a bit if I have some pointers.

PS: the notebook-as-pdf is a workaround but not a solution to the underlying issue, since it doesn't allow SAS to be used in document production systems that allow far greater control on the formatting, output, etc.

[tomweber-sas]

Hey @fsmunoz , yes, would you mind changing this to a new issue that refers to this older one? That's helpful for me. I'll take a look into this to research it a bit to get up to speed on this.
Thanks!
Tom