<Plug>JupyterExecute is executing the next cell instead of the current one

[IndianBoy42]

vim.api.nvim_buf_set_keymap(0, "n", "<localleader>j", "<Plug>JupyterExecute", {})

# %%
import numpy as np
import scipy as sp

# %%
print("Hello World")

Putting my cursor on import and hitting <localleader>j will cause the second cell to execute. What could be the problem?

[fecet]

same issue

[fecet]

A rough workaround is evoke function of vim-ipython-cell:

function! PrevExecNextCell()
    call IPythonCellPrevCell()
    call jupyter_ascending#execute()
    call IPythonCellNextCell()
endfunction

nmap <leader><CR> :call PrevExecNextCell()<CR>

hope someone have a more elegant solution

[jvivian]

hope someone have a more elegant solution

@fecet @IndianBoy42

After taking a look at the code, it appears to be an issue with the assumptions for how the file starts: https://github.com/untitled-ai/jupyter_ascending/blob/main/jupyter_ascending/requests/execute.py#L20

I don't know enough about jupytext as this is my first time trying to use it, but I assume you need the header at the top for it to work correctly, so you just need to remove the very first # %% in order to trigger the second specified case.

This worked for me, hopefully it does for you as well.

Cheers

[ViRu-ThE-ViRuS]

@jvivian this solutions works! Thanks for the this workaround :)

[IndianBoy42]

So if i'm understanding right, not having the first cell marker on the first line should make it work?

But my notebook files are generally created by jupytext which does add a leading cell marker, so could it be fixed in jupyter_ascending?

[jvivian]

So if i'm understanding right, not having the first cell marker on the first line should make it work?

I used the jupyter_ascending.scripts.make_pairs function to generate the .py/ipynb pairs, which puts a jupytext header at the top. If you remove the first # %% below the header (---), it should work. Example below

(base) jvivian@DESKTOP-8U6S9I6:~/Dropbox/Learning/Statistics/Bayesian-modeling-and-computation$ head Notebooks/03_Linear_models.sync.py -n 20
# ---
# jupyter:
#   jupytext:
#     text_representation:   
#       extension: .py       
#       format_name: percent 
#       format_version: '1.3'
#       jupytext_version: 1.3.4
#   kernelspec:
#     display_name: Python 3
#     language: python
#     name: python3
# ---

import pandas as pd

# %%
print('hello world')

[zplizzi]

This was maybe hopefully fixed in the most recent update of the jupyter_ascending plugin - there was a numbering bug. If anyone happens to be able to confirm or deny, lmk!