Passing arguments using commands.execute

JupyterLab
,
1

I am trying to address a specific use case:

Ask:
show a custom HTML in cell output for keyword.

Implementation:
created a custom React widget and bound it to python class similar to ipywidgets sample.
Lets call it example widget, so now I am able to call it by adding below steps in plugin’s execution method in frontend:
array = [‘sample1’, ‘sample2’]
INotebookTracker.activeCell.model.sharedModel.setSource(
from myExtension import exampleWidget exampleWidget(value=[${array}])
)
commands.execute(‘notebook:run-cell’)

The value of array is dynamically generated from a service call.
Initial Error:
exampleWidget(value=[sample1,sample2])
^
SyntaxError: invalid syntax. Perhaps you forgot a comma?

question: For my use case, Is my approach appropriate? (any feedback on this solution is welcome)

2

That one lacks quotes? Is that what you actually run or do you run:

exampleWidget(value=['sample1','sample2'])

when you run it ‘manually’? They aren’t equivalent.

3

When I manually run the command and pass array of string as input it works.

Edit:
It might be a issue with my syntax itself, I need to test it. I’m still curious about question 2.

I’ll update the post once I test my theory and possibly solve questions 1

Edit2:
Thanks for responding @formightez
turns out, it was syntax error. Now only question I have is to get insights/feedback from Jupyter community on my solution to the use.

4

I meant by ‘They aren’t equivalent.’ that exampleWidget(value=[sample1,sample2]) you get when you try with the code and exampleWidget(value=['sample1','sample2']) that I assume you use when manual, aren’t equivalent.

Anyway, I guess you sorted the syntax error.

At this point, what is the solution? You fixed something and didn’t post it.

5

Good point, I’ll add it back to the POST, It might benefit any other lost soul.

for my fix:

It is syntax error, in ts when using template literals along with interpolation, we need to wrap objects or arrays in string

here array should be JSON.stringify([‘sample1’, ‘sample2’]),

INotebookTracker.activeCell.model.sharedModel.setSource(
from myExtension import exampleWidget exampleWidget(value=[${JSON.stringify(['sample1', 'sample2'])}])
)
This fixes the syntax error.

That being said,
I wanted to know if the approach I took is the right one. What I wanted to implement is a hook to cell input and directly manipulate the cell output region.

For example:

  1. we have cell A of type RAW, a custom Jupyterlab extension will check all the inputs user types in this cell
  2. user enters a text: ‘start example’
  3. Our extension will have a logic to render exampleWidget as this cell’s output for this text.

I couldn’t find a way to achieve point 3 directly through INotebookTracker or other classes, so I created a workaround which would

  1. create a new cell under this cell
  2. add the text from myExtension import exampleWidget exampleWidget(value=[${array}]) to the new cell’s input
  3. executes this cell to render ExampleWidget
  4. hide the new cell’s input area.
1 Like