Python visuals doesn't work on online reading view but on desktop does

Hi everyone, 

Since a couple of day maybe week, i didn't noticed that one of my own python visual didn't work on the online version. On my dekstop app, nothing, it work well but only when i share the link and open it via my workspace the visual doesn't work anymore. 

there erreur that i have is this one : 

Script Runtime Error

[S-12c5e384-a869-4a16-9c4a-12dd6770c94c][S-12c5e384-a869-4a16-9c4a-12dd6770c94c]AttributeError: 'NoneType' object has no attribute 'points_to_pixels'
Please try again later or contact support. If you contact support, please provide these details.

Activity ID: fb2167c7-a9c2-4229-ba95-dd2df542479d
Request ID: d0f51534-072d-4d2e-a767-651859b64a82
Correlation ID: 101ac569-2fa3-434e-5997-79102e36e5a0
Time: Mon Dec 02 2024 18:25:33 GMT+0100 (Central European Standard Time)
Service version: 13.0.24766.27
Client version: 2411.3.21845-train
Cluster URI: https://wabi-west-europe-b-primary-redirect.analysis.windows.net/

My code is but i am not sure that it come from him because nothing change since a couple of months... 

import matplotlib.pyplot as plt
import pandas as pd
import matplotlib.dates as mdates
from matplotlib.dates import DateFormatter, MonthLocator

# Supposons que dataset soit un DataFrame pandas déjà défini
dataset['Date update'] = pd.to_datetime(dataset['Date update'])
dataset['Forecast'] = pd.to_datetime(dataset['Forecast'])

# Pivot table
pivot = pd.pivot_table(dataset, aggfunc='first', index='Date update', values='Forecast', columns='Milestones name')
fig, ax = plt.subplots(figsize=(12, 8))

# Plot each column
for column in pivot.columns:
    clean = pivot[[column]].dropna()
    ax.plot_date(clean.index, clean[column], xdate=True, ydate=True, ls='-', lw=1, label=column)

# Ajuster les limites des axes pour découpler Date update (axe x) et Forecast (axe y)
min_date_update = dataset['Date update'].min()
max_date_update = dataset['Date update'].max()

min_forecast = dataset['Forecast'].min()
max_forecast = dataset['Forecast'].max()

# Dates limites pour les axes
max_date_forecast = max(dataset['Forecast'])
next_year_forecast = max_date_forecast.year + 1
q1_next_year_forecast = pd.Timestamp(year=next_year_forecast, month=1, day=1)

min_date_forecast = min(dataset['Forecast'])
min_year_forecast = min_date_forecast.year
q1_actual_year_forecast = pd.Timestamp(year=min_year_forecast, month=1, day=1)

min_date_update = min(dataset['Date update'])
min_year_dateupdate = min_date_update.year
q1_actual_year_dateupdate = pd.Timestamp(year=min_year_dateupdate, month=1, day=1)

# Ajouter la ligne 45 degrés (y=x)
# Créer des listes de dates pour la ligne 45 degrés
line_dates = pd.date_range(start=q1_actual_year_dateupdate, end=q1_next_year_forecast, freq='D')
ax.plot(line_dates, line_dates, color='black', linestyle='--', label='')

# Coloration après les courbes
adjusted_min_date_update = q1_actual_year_dateupdate
adjusted_max_forecast = q1_next_year_forecast
if len(line_dates) > 0:
    ax.fill_between(line_dates, adjusted_min_date_update, line_dates, color='white', zorder=2)

# -----------------------------------------------------------------------------------------------

# Définir les localisateurs pour les mois (trimestres)

month_locator = MonthLocator(bymonth=[1, 4, 7, 10])

plt.gca().xaxis.set_major_locator(month_locator)
plt.gca().yaxis.set_major_locator(month_locator)

# Création du formateur personnalisé pour afficher les trimestres
class QuarterFormatter(DateFormatter):
    def __init__(self, fmt='%Y-%m', **kwargs):
        super().__init__(fmt, **kwargs)

    def __call__(self, x, pos=0):
        date = mdates.num2date(x)
        # Si c'est janvier, afficher l'année en dessous de Q1
        if date.month == 1:
            return f'{date.year} - Q1'
        elif date.month == 4:
            return 'Q2'
        elif date.month == 7:
            return 'Q3'
        elif date.month == 10:
            return 'Q4'
        return ''

# Appliquer le formateur de trimestre personnalisé
quarter_formatter = QuarterFormatter()
plt.gca().xaxis.set_major_formatter(quarter_formatter)
plt.gca().yaxis.set_major_formatter(quarter_formatter)

# -----------------------------------------------------------------------------------------------

# Rotation des étiquettes pour une meilleure lisibilité
plt.xticks(rotation=90, ha='center')

# Ajouter les lignes de grille pour les années uniquement
start_date_update = pd.Timestamp(year=q1_actual_year_dateupdate.year, month=1, day=1)
end_date_forecast = pd.Timestamp(year=q1_next_year_forecast.year, month=1, day=1)
for year in range(start_date_update.year, end_date_forecast.year):
    january_start = pd.Timestamp(year=year, month=1, day=1)
    ax.axvline(x=january_start, color='gray', linestyle='--', linewidth=0.5)
    ax.axhline(y=january_start, color='gray', linestyle='--', linewidth=0.5)

# -----------------------------------------------------------------------------------------------

# Ajuster les limites des axes
plt.xlim(q1_actual_year_dateupdate, q1_next_year_forecast)  # Limites pour l'axe des x (Date update)
plt.ylim(q1_actual_year_dateupdate, q1_next_year_forecast)   # Limites pour l'axe des y (Forecast)

# -----------------------------------------------------------------------------------------------
legend = ax.legend(loc='upper left', bbox_to_anchor=(1.01, 1), frameon=False)
# Ajuster la légende
def adjust_legend_fontsize(ax):
    legend = ax.get_legend()  # Récupère la légende existante
    if legend is None:  # Si aucune légende n'est définie, arrêter
        return 10

    fig = ax.get_window_extent()
    fig_height = fig.height 
    fontsize = 10
    adjusted = False
    while not adjusted:
        legend = ax.legend(loc='upper left', bbox_to_anchor=(1, 1), fontsize=fontsize)
        legend_bbox = legend.get_window_extent()
        legend_height = legend_bbox.height
        if legend_height <= fig_height:
            adjusted = True
        else:
            fontsize -= 1
            if fontsize < 8:
                break
    return fontsize
ax = plt.gca()
fontsize = adjust_legend_fontsize(ax)
myLegend = ax.legend(loc='upper left', bbox_to_anchor=(1.01, 1), frameon=False, fontsize=fontsize)

# Désactiver les ajustements automatiques des marges
plt.subplots_adjust(left=0.08, right=0.75, top=0.95, bottom=0.1)

# Afficher le graphique
plt.show()

If you have an idea, that coudl help me !! 
thanks you very much in advance, and have a nice day !

Flan