How to build this caluculated columns in Pandas

Question

I need to get a dataframe that shows me, in correspondence with the warehouse loads, when the stock would run out (excluding subsequent loads)

The starting dataframe is this: starting dataframe

Here you have the code to build it:

import pandas as pd

data = {'warehouse': ['M1', 'M1', 'M1', 'M1', 'M1', 'M1', 'M1', 'M1', 'M1', 'M2'],
        'item': ['Partxy', 'Partxy', 'Partxy', 'Partxy', 'Partxy', 'Partxy', 'Partxy', 'Partxy', 'Partxy', 'Partz'],
        'date': ['01/01/2023', '05/01/2023', '07/01/2023', '08/01/2023', '09/01/2023', '10/01/2023', '15/01/2023', '18/01/2023', '19/01/2023', '22/01/2023'],
        'tot_load_replenishment': [0, 0, 20, 0, 0, 50, 0, 50, 0, 0],
        'tot_unload': [0, -30, -15, -50, -10, -5, -30, -10, -5, -10],
        'stock': [100, 70, 75, 25, 15, 60, 30, 70, 65, 300],
        'stock_before_load': [None, None, 55, None, None, 10, None, 20, None, None]}


load_unload = pd.DataFrame(data)

load_unload['tot_load_replenishment'] = load_unload['tot_load_replenishment'].astype(int)
load_unload['tot_unload'] = load_unload['tot_unload'].astype(int)
load_unload['date'] = pd.to_datetime(load_unload['date'], format='%d/%m/%Y').dt.strftime('%Y-%m-%d')

What I want to obtain is this:

final result

I tried this code:

#pass one item in one warehouse
item = 'Partxy'
warehouse = 'M1'

#filter the dataframe
item_war = (load_unload['item'] == item) & (load_unload['warehouse'] == warehouse) 
load_unload_item_war = load_unload[item_war]

#consider only the row where I have some load for replenishment
mov_replenishment = load_unload_item_war[load_unload_item_war['tot_load_replenishment'] > 0]

#define a function that return the exhaustion_date, given the stock_before_load
def exhaustion_date(row, load_unload_item_war):

    #consider only lines from today forward
    f_load_unload_item_war = load_unload_item_war[(load_unload_item_war['date'] > row['date'])]
    
    #store the stock before load in a specific variable
    residual = row['stock_before_load']
    
    #calculate the cumulative unloaded
    f_load_unload_item_war['cum_unload'] = f_load_unload_item_war['tot_unload'].cumsum()*(-1)
    
    #filter the dataframe to find the first date in which stock would become negative
    exhaustion_date = f_load_unload_item_war.loc[f_load_unload_item_war['tot_unload'] < residual, 'date'].min()
    
    return exhaustion_date

#apply the function to obtain the exaustion date only when I have a replenishment
mov_replenishment['exhaustion_date'] = mov_replenishment.apply(lambda row: exhaustion_date(row, load_unload_item_war), axis=1)

What I obtain is this (not correct):

final result not correct

Thanks in advance to anyone who can help me!