Makes images of stylized dataframes so that when Jupyter notebooks are reopened, the stylized view is represented. Importantly, it includes tips that also enable it to be done without generating giant file sizes as well.
There is a more well-developed, related project that came after mine. See dataframe_image, which is described as "A python package for embedding pandas DataFrames as images into pdf and markdown documents." You may wish to check that out.
Click on a launch binder badge on this page to launch a session where everything needed for this process is set up and the steps in the process are demonstrated.
Pandas dataframe styling is a great feature for visualizing tabular data right in a notebook; however, at the time I wrote this I was finding it annoying that it gets removed from rendering when you reopen the notebook. (At least every place I have seen so far for apllied cell and text coloring while style.format settings such as style.format("{:.2%}" to show floats as percents with two-decimals, seem to survive. Always to be the case for cell and text coloring? UPDATE: I SHOULD ADD THAT I HAVEN'T TESTED HOW SURVIVAL OF THESE FEATURES IS NOW. IT COULD BE BETTER.) I wanted a way to have at least the information visible when I reload. So an image would at least be viewable, albeit not as interactive.
STEPS:
Pandas dataframe --> Pandas dataframe with conditional stylized cells ---> Pandas dataframe converted to HTML --> image file --> image viewable when notebook reopened.
This builds on the approach from Shovalt here, with some advanced styling following Andy Lane's post.
Addressing the default image size and quality of output was worked out with advice found here (thanks especially to landonwilkins) and here (thanks especially to jarrekk).
Related:
Keep in mind for some styles you can save it as HTML, e.g., here, and could probably place that code in as HTML cell or wrapped in markdown cell.
There is a more well-developed, related project that came after mine. See dataframe_image, which is described as "A python package for embedding pandas DataFrames as images into pdf and markdown documents." You may wish to check that out. (I learned of it when researching this topic again after seeing someone recommend my script, and then finding this thread, where I first encountered dataframe_image and also pandas.plotting.table which looks useful to know about for efforts with pandas tables displaying in conjunction with plots, as well as possibly elsewhere.
Click on a launch binder badge on this page to launch a session where everything needed for this process is set up and the steps are demonstrated.