search

sassoftware / sas_kernel

A Jupyter kernel for SAS. This opens up all the data manipulation and analytics capabilities of your SAS system within a notebook interface. Use the Jupyter Notebook interface to execute SAS code and view results inline.
Apache License 2.0
199 stars 82 forks source link

Update log rendering in kernel #68

Closed jld23 closed 4 years ago

jld23 commented 4 years ago

This PR completely changes how SAS logs are rendered. Currently, the SAS log and listing details are rendered as an HTML file which, through pygments, allows the log to be colored consistently with other SAS clients.

This caused problems in conversion to PDF, HTML, and other destinations and if you run from console instead of a notebook.

This PR renders the SAS log as text using ANSI escape sequences to color the log. The goal is that users will not notice any difference in routine usage and see a huge improvement in use cases beyond just entering code.

I've also taken advantage of metakernal functions to flag cells with ERROR messages as failing while producing the same output as before. This is a helpful visual cue to the user that there is an error and will aid in auto-grading from nbgrader through this PR jupyter/nbgrader/pull/1356

Thanks to @dsblank for committing my PR Calysto/metakernel/pull/205 of adding a new method Error_display that will show the error in the log with context and the listing output.

This PR also removes the dependency on pygments since everything can be done with IPython and text.

This update also addresses issues #13 and #42

Additional logging was also added to improve the ability to debug issues.

jld23 commented 4 years ago

Just updated the regex to fix some issues @tomweber-sas found. See below for examples: image

image

image

tomweber-sas commented 4 years ago

Hey Jared, this does fix the 3 issues I saw. As for the dull red, I still see that in jupyter, but the exported html is bold, and much brighter. I still am using Notebook on this system, so maybe it comes out brighter in Lab? What you pasted above is bright, like I see in the html, but that screenshots actually jupyter right? So mine is probably because it's an old notebook version? All in all though, this looks good! Tom

tomweber-sas commented 4 years ago

I installed in Lab on another machine and the REDs are bright! yea :)