File size: 5,299 Bytes
95bad29
 
 
 
 
b398e78
 
 
 
 
 
 
 
 
 
 
 
 
95bad29
 
 
 
 
 
b398e78
 
 
 
 
95bad29
 
 
b398e78
 
 
 
 
95bad29
b398e78
 
 
 
 
 
 
 
 
 
 
 
 
 
95bad29
b398e78
95bad29
 
 
 
 
 
 
 
 
 
 
 
 
b398e78
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5858a75
 
 
 
 
 
 
 
 
 
 
 
 
 
b398e78
 
 
95bad29
5858a75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b398e78
 
 
 
 
 
95bad29
b398e78
 
 
 
 
95bad29
 
b398e78
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
import streamlit as st

from functions import Functions, StreamlitFunctions
import pandas as pd


def calculate_salary_parameters():
    st.session_state.type_to_calculate = "salary_parameters"


def calculate_initial_salary_parameter():
    st.session_state.type_to_calculate = "desired_salary"


def calculate_tax_on_current_salary():
    st.session_state.type_to_calculate = "tax_on_current_salary"


StreamlitFunctions.initialize_session_values()
StreamlitFunctions.print_tax_brackets()
StreamlitFunctions.reset_tax_brackets()

StreamlitFunctions.print_salary_parameters()
StreamlitFunctions.reset_salary_parameters()
st.button(
    "Calculate Based on Salary Parameters",
    use_container_width=True,
    on_click=calculate_salary_parameters,
)

StreamlitFunctions.initial_salary_parameter()
StreamlitFunctions.reset_initial_salary_parameter()
st.button(
    "Calculate Based on Desired Net Salary",
    use_container_width=True,
    on_click=calculate_initial_salary_parameter,
)

StreamlitFunctions.print_tax_on_current_salary()
StreamlitFunctions.reset_tax_on_current_salary()
st.button(
    "Calculate Tax on Current Salary",
    use_container_width=True,
    on_click=calculate_tax_on_current_salary,
)

if st.session_state.type_to_calculate is not None:
    if st.session_state.type_to_calculate == "tax_on_current_salary":
        initial_desired_net = Functions.calculated_current_salary_after_tax(
            st.session_state.tax_on_current_salary, st.session_state.tax_brackets
        )
    elif st.session_state.type_to_calculate == "desired_salary":
        initial_desired_net = st.session_state.user_initial_desired_net
    elif st.session_state.type_to_calculate == "salary_parameters":
        initial_desired_net = Functions.calculated_initial_desired_net(
            st.session_state.current_salary,
            st.session_state.desired_increment_percentage,
            st.session_state.daily_cost_of_travel,
            st.session_state.physical_days_per_week,
        )

    result = Functions.calculate_additional_amount(
        initial_desired_net, st.session_state.tax_brackets
    )

    # Display how initial_desired_net was determined
    st.markdown("---")
    if st.session_state.type_to_calculate == "tax_on_current_salary":
        st.success(
            "βœ… Calculation was done based on the selected value of 'Tax on Current Salary'"
        )
        summary_df = pd.DataFrame(
            {
                "Parameter": [
                    "Current Salary",
                    "Tax",
                    "Gross Salary",
                ],
                "Value": [
                    f"PKR {result['final_net_salary']:,.2f}",
                    f"PKR {result['tax']:,.2f}",
                    f"PKR {result['gross_salary_needed']:,.2f}",
                ],
            }
        )
    elif st.session_state.type_to_calculate == "desired_salary":
        st.success(
            "βœ… Calculation was done based on the selected value of 'Final Desired Net Salary'"
        )
        summary_df = pd.DataFrame(
            {
                "Parameter": [
                    "Final Net Salary",
                    "Tax",
                    "Gross Salary",
                ],
                "Value": [
                    f"PKR {result['final_net_salary']:,.2f}",
                    f"PKR {result['tax']:,.2f}",
                    f"PKR {result['gross_salary_needed']:,.2f}",
                ],
            }
        )
    elif st.session_state.type_to_calculate == "salary_parameters":
        st.success(
            "βœ… Calculation was done based on the selected values of 'Salary Parameters'"
        )
        summary_df = pd.DataFrame(
            {
                "Parameter": [
                    "Current Salary",
                    "Desired Increment",
                    "Daily Travel Cost",
                    "On-Site Days/Week",
                    "Gross Salary",
                    "Tax",
                    "Final Net Salary",
                ],
                "Value": [
                    f"PKR {st.session_state.current_salary:,.2f}",
                    f"{st.session_state.desired_increment_percentage:.2%}",
                    f"PKR {st.session_state.daily_cost_of_travel:,.2f}",
                    f"{st.session_state.physical_days_per_week}",
                    f"PKR {result['gross_salary_needed']:,.2f}",
                    f"PKR {result['tax']:,.2f}",
                    f"PKR {result['final_net_salary']:,.2f}",
                ],
            }
        )
    st.header("Salary Calculation Results")
    col1, col2 = st.columns(2)
    with col1:
        # custom_metric("Initial Desired Net Salary", result['initial_desired_net'])
        StreamlitFunctions.custom_metric("Final Net Salary", result["final_net_salary"])
        StreamlitFunctions.custom_metric("Tax", result["tax"])

    with col2:
        # custom_metric("Additional Amount Needed", result['additional_amount'])
        StreamlitFunctions.custom_metric(
            "Gross Salary Needed", result["gross_salary_needed"]
        )
    # Display a summary of the calculation
    st.subheader("Calculation Summary")
    st.data_editor(summary_df, use_container_width=True, hide_index=True)
    st.session_state.type_to_calculate = None