Passing arguments to a nested function in Python

I have two functions, called update_y_with_x and update_x

which are as follows:

update_y_with_x

def update_y_with_x(chosen_tissue):
    for i, trace in enumerate(f.data):
        if trace.type == "scatter":
            # reset all line width
            f.data[i].line.width = 1

            if chosen_tissue in trace.text:
                f.data[i].line.width = 1 + 9

update_x

def update_x(trace, points, selector):
    if len(points.point_inds) == 0:
        return

    # Remove existing bar shades and linked lines' highlights
    f.layout.shapes = ()

    for i, trace in enumerate(f.data):
        if trace.type == "scatter" and f.data[i].line.width != 1:
            f.data[i].line.width = 1

    # Get tissue identity
    chosen_tissue = f.data[points.trace_index].y[points.point_inds[0]]

    # Get the index of the chosen tissue
    possible_trace_indices = [
        i for i, trace in enumerate(f.data) if trace.type == "bar"
    ]
    chosen_tissue_ranks = [
        list(f.data[index].y).index(chosen_tissue) for index in possible_trace_indices
    ]

    for i, _ in enumerate(possible_trace_indices):
        # Specify the y-range around the clicked bar for shading
        y0 = chosen_tissue_ranks[i] + 0.5
        y1 = y0 - 1

        f.add_hrect(
            y0=y0,
            y1=y1,
            line_width=0,
            fillcolor="grey",
            opacity=0.3,
            row=1,
            col=i * 2 + 1,
        )

    # Update lines
    update_y_with_x(chosen_tissue)

as you can see, update_x calls the update_y_with_x function, at the moment, both use a hard-coded plotly go figurewidget, where

f = go.FigureWidget(fig)

and are called like so:

for i, trace in enumerate(f.data):
    if trace.type == 'bar':  # Check if it's a barplot
        f.data[i].on_click(update_x)

I would like to do two things:

1. Move these two functions to an independent function file and import them into the main.ipynb file I’m working with

2. Pass f as a parameter to each function, rather than having it hard – coded as it currently exists in the function, and make sure that it is passed to update_y_with_x prior to it being used in update_x

If I’ve missed something key please let me know and would be incredibly grateful for any help 🙂

Edit:
update_x click events on a bar plot. When a bar is clicked, it removes existing shapes, resets line widths for linked scatter plots, adds shaded rectangles around the clicked bars, and updates lines which link these bars together.

What I tried:
I initially exported these functions to a new file called "click_events.py", and added in the argument for each figure_widget, like so:

• def update_x(trace, points, selector, figure_widget): (replaced f in file)

• def update_y_with_x(chosen_tissue, figure_widget): (replaced f in file)

• modify the code in update_x so that update_y_with_x(chosen_tissue, figure_widget = figure_widget)

However, I believe that when I imported and try to run these functions in the way mentioned above the figure_widget isn’t being passed to ‘update_y_with_x’.

>Solution :

For the two goals you mentioned, moving the functions to an independent file and passing f as a parameter

Steps:

1. Create an Independent Function File:

Create a new Python file (let’s call it plot_utils.py for example) and move the functions update_y_with_x and update_bar into this file. The content of plot_utils.py would look like this:

# plot_utils.py

import plotly.graph_objects as go

def update_y_with_x(fig, chosen_tissue):
    for i, trace in enumerate(fig.data):
        if trace.type == "scatter":
            # reset all line width
            fig.data[i].line.width = 1

            if chosen_tissue in trace.text:
                fig.data[i].line.width = 1 + 9

def update_bar(fig, trace, points, selector):
    if len(points.point_inds) == 0:
        return

    # Remove existing bar shades and linked lines' highlights
    fig.layout.shapes = ()

    for i, trace in enumerate(fig.data):
        if trace.type == "scatter" and fig.data[i].line.width != 1:
            fig.data[i].line.width = 1

    # Get tissue identity
    chosen_tissue = fig.data[points.trace_index].y[points.point_inds[0]]

    # Get the index of the chosen tissue
    possible_trace_indices = [
        i for i, trace in enumerate(fig.data) if trace.type == "bar"
    ]
    chosen_tissue_ranks = [
        list(fig.data[index].y).index(chosen_tissue) for index in possible_trace_indices
    ]

    for i, _ in enumerate(possible_trace_indices):
        # Specify the y-range around the clicked bar for shading
        y0 = chosen_tissue_ranks[i] + 0.5
        y1 = y0 - 1

        fig.add_hrect(
            y0=y0,
            y1=y1,
            line_width=0,
            fillcolor="grey",
            opacity=0.3,
            row=1,
            col=i * 2 + 1,
        )

    # Update lines
    update_y_with_x(fig, chosen_tissue)

2. In your Main Jupyter Notebook (main.ipynb):

In your main Jupyter Notebook, you can import these functions and use them as follows:

# main.ipynb

import plot_utils
import plotly.graph_objects as go

# Create the initial figure
f = go.FigureWidget(fig)

# Register the on_click event
for i, trace in enumerate(f.data):
    if trace.type == 'bar':  # Check if it's a barplot
        f.data[i].on_click(lambda trace, points, selector: plot_utils.update_bar(f, trace, points, selector))

Now, both update_y_with_x and update_bar are in a separate file, and you can import and use them in your main notebook. The f figure is passed as a parameter to both functions, making them more flexible and avoiding hard coding.