Hide input cells

[ellisonbg]

I plan on implementing the hiding of input cells. Might not happen for 4.1, but I will be optimistic for now.

[dsblank]

Woot!

BTW, I like our variation of making a Tabbed Input/Output cell [1] such that you can see the code in the Input tab, but defaults to showing the Output tab initially. Also can display the input next to output.

[1] - https://www.youtube.com/watch?t=19&v=WwoTzvOkEJQ

[ellisonbg]

I can imagine that this view would be useful sometimes. But initially this
will be as simple as possible - possibly without UI (just in the command
palette).

On Tue, Oct 6, 2015 at 4:02 PM, Doug Blank notifications@github.com wrote:

Woot!

BTW, I like our variation of making a Tabbed Input/Output cell [1] such
that you can see the code in the Input tab, but defaults to showing the
Output tab initially. Also can display the input next to output.

[1] - https://www.youtube.com/watch?t=19&v=WwoTzvOkEJQ

—
Reply to this email directly or view it on GitHub
#534 (comment).

Brian E. Granger
Associate Professor of Physics and Data Science
Cal Poly State University, San Luis Obispo
@ellisonbg on Twitter and GitHub
bgranger@calpoly.edu and ellisonbg@gmail.com

[akhmerov]

Would it be possible to expose the JS so that it survives nbconvert? Otherwise any derived materials would need to re-implement show/hide machinery.

I'm also 👍 on the possibility of switching input on one cell only, imagine publishing lecture materials in a non-programming course, where you only might want to check how a particular thing is produced.

[Saynah]

Here's a snipped that works for me if I run as HTML in the 1st cell.

<script>
    var code_show=true; //true -> hide code at first

    function code_toggle() {
        $('div.prompt').hide(); // always hide prompt

        if (code_show){
            $('div.input').hide();
        } else {
            $('div.input').show();
        }
        code_show = !code_show
    }
    $( document ).ready(code_toggle);
</script>

Here's the button.

<a href="javascript:code_toggle()">[Toggle Code]</a>

[ecollins]

I just signed in to thank @Saynah for that great hack. I don't use Jupyter specifically because it doesn't have cell folding, but this is the closest I've found so far.

[kirbs-]

@Saynah, @ecollins, I created a code hiding plugin for Jupyter a couple of months ago. If you're interested in check out hide_code.

Edit: fixed a typo.

[ruffsl]

Thanks @kirbs- , had some issue the first time around the install. It'd definitely be nice to see this feature upstream.

[iondiode]

It seems like this would be nice done a cell magic, then it's easy to selectively hide cells, not sure though. still learning the system. Plus I seem to be running an ancient version

[takluyver]

I don't think a cell magic is the right thing for this:

• Hiding cells should work the same way for any kernel, and magics are not general.
• Cell magics usually affect what the kernel does with code, whereas here we don't want that to change

If you're on an old version, have a look at the upgrade instructions here:
http://jupyter.readthedocs.org/en/latest/install.html#upgrading

[cenveoanalyst]

Hiding input cells for HTML output would be greatly appreciated!

[caos21]

I second that!

[cenveoanalyst]

I have cell hiding on HTML export figured out with a helpful piece of code to trigger upon export; but now that I'm moving to deploying Jupyterhub in my organization I really would appreciate the ability to hide live input cells so users don't see them when they run their notebooks.

[amueller]

The hide_code notebook extension is nice, but it doesn't hide the code when exporting. Can an nbextension register preprocessors? As hide_code adds a variable to the meta-data, that should be trivial to implement.

[juhasch]

You can hide the code in nbconvert using a custom template.
Like this template hide_input_output.tpl for conversion to HTML. You have to change the metadata variable name in the template, because it was made for another notebook extension.

For conversion to TeX or PDF you can take a look at this template printviewlatex.tplx and modify it accordingly.

[amueller]

I think it would be cleaner to do this with a preprocessor (which I just did locally). I fell the hiding of cells should be independent of the template.

[BreitA]

It would be neat if in addition to the feature we could allow to export to pdf/html removing/hiding input cell code and prompt so we could have a clean pdf/html output without any code but with the beautiful graphs, markdown and results.
A man can dream.

[amueller]

@BreitA that's super easy. The hard part is getting rid of the prompt for the output. Lemme send a PR.

[amueller]

Hm actually I realized that I was doing the wrong thing and also removed the output of the cell. That was simple.

[Carreau]

Oops, ping @JamiesHQ

[ellisonbg]

I don't now the current state of how lab and the notebook are handling this metadata, we would have to do some tests.

[jrbrodie77]

I've been trying to get PDF export to hide input cells and have done this successfully with tplx templates etc.

My problem now is that hide_input seems to hide the cells even if I disable all extensions? Is hide_input actually being used by vanilla notebook? or do I have an extension mysteriously running somewhere?

[takluyver]

I don't think the notebook web interface uses any metadata for hiding input yet, though maybe it's gone in without me noticing.

Stuff to check: browser Javascript console, logging from the server, run jupyter nbextension list

[SimonEnsemble]

+10, would be perfect if we can edit the meta-data in Jupyter Lab to:

• hide input of cell and only show output
• hide cell numbers, looks tacky.

See Ref. here.

[ellisonbg]

It is my understanding that what @takluyver says is correct (not using the metadata in notebook yet). We do have work in progress to add that handling to lab though.

[MSeal]

I realize this has been open for a while and there were some difficulties here, but I have been getting a lot of questions regarding why classic jupyter doesn't respect the jupyter metadata source_hidden. JupyterLab, nbconvert, nbformat, and nteract all respect the field (if I recall correctly).

I see there's a lot of discussion on the DOM structure and how JupyterLab reimplemented that, but it seems like source_hidden and outputs_hidden wouldn't need drastic changes to enforce (even if classic couldn't edit the attribute easily). My front-end dev skills are a bit rusty but would adding a metadata check to

notebook/notebook/static/notebook/js/codecell.js

Line 181 in d61cd6e

var input_area = $('<div/>').addClass('input_area');

to set display: None be 80% of the work for source_hidden? Seems like it'd only need to capture initial render and metadata edit events to adjust the display attribute.

Thoughts/comments?

[Zsailer]

Thanks @ellisonbg! This is a great idea!

This is currently possible in JupyterLab and will be possible in the coming Notebook v7 (which inherits from JupyterLab's notebook package). Look out for this feature in our next major release!

We'll leave this issue open until v7 is published.

[andrii-i]

Notebook 7 is close to be published. JupyterLab adds ability to collapse an input cell so that only 1st line of the cell is visible. But this does not hide the input cell completely which is the desired feature here. Moving this issue to 7.0.x for further discussion and follow up, please let me know if I'm missing something (CC @Zsailer).

[rgbkrk]

Love it. That's a great start.