main.py

# Import libraries
import streamlit as st
import pandas as pd
from streamlit_searchbox import st_searchbox
from typing import List
import plotly.express as px
import flagpy as fp
from PIL import Image
from utils.styles import styles


st.set_page_config(
    page_title="World Happiness Report,2015 up to 2022",
)

#  css part
st.write(styles, unsafe_allow_html=True)


@st.cache_data
def get_data():
    dataframe = pd.read_csv(
        './world_happiness_2015_2022/world-happiness-report-2015-2022.csv')
    flags_df = fp.get_flag_df()
    countires = set(dataframe["Country"])
    regions = set(dataframe["Region"])
    return dataframe, countires, flags_df, regions


@st.cache_data
def get_desc_by_country(dataframe):
    return dataframe.groupby('Country')['Happiness Rank'].describe()


def check_flag(country_name):
    flag_name_converter = {'Dominican Republic': 'The Dominican Republic', 'United Arab Emirates': 'The United Arab Emirates',
                           'United Kingdom': 'The United Kingdom', 'United States': 'The United States'}
    country_name = flag_name_converter.get(country_name, country_name)
    if country_name in fp.get_country_list():
        return True, country_name
    return False, country_name


def search_by_country(searchterm: str) -> List[any]:
    global countries
    return [c for c in countries if c.lower().startswith(searchterm.lower())] if searchterm else []


overall_df, countries, flags_df, regions = get_data()
check_boxes = {}
with st.sidebar:
    selected_country = st_searchbox(
        search_function=search_by_country,
        key="country_searchbox",
        placeholder="Search for a Country",

    )

    if not selected_country:
        st.markdown('<b>Filter By Region</b><hr/>', unsafe_allow_html=True)
        c9, c10 = st.columns([0.5, 0.5])
        for index, cb in enumerate(regions):
            if index % 2 == 0:
                with c9:
                    check_box = st.checkbox(cb)
            else:
                with c10:
                    check_box = st.checkbox(cb)
            check_boxes[cb] = check_box


overall_df = overall_df.loc[:, ~overall_df.columns.str.contains('^Unnamed')]

if all(not value for value in check_boxes.values()):
    check_boxes['All'] = True
else:
    check_boxes['All'] = False

if selected_country:
    col1, col2 = st.columns([0.5, 0.5])
    col3, col4 = st.columns([0.5, 0.5])
    with col1:
        st.write(selected_country)
    with col2:
        signal, country_name = check_flag(selected_country)
        if signal:
            img = Image.fromarray(flags_df["flag"].loc[country_name], 'RGB')
            img.save('flag.png')
            img = Image.fromarray(flags_df["flag"].loc[country_name], 'RGB')
            with Image.open('./flag.png') as image:
                st.image(image)

    df_by_selected_country = overall_df[overall_df.Country == selected_country]
    df_by_selected_country = df_by_selected_country.sort_values(
        by=["Year"], ascending=True)
    st.dataframe(df_by_selected_country.set_index(overall_df.columns[10]))
    with col3:
        st.write('Best Rank:', df_by_selected_country.loc[df_by_selected_country['Happiness Rank'].idxmin(
        )]['Happiness Rank'])
    with col4:
        st.write('Worst Rank:', df_by_selected_country.loc[df_by_selected_country['Happiness Rank'].idxmax(
        )]['Happiness Rank'])

    difference = df_by_selected_country.loc[df_by_selected_country['Year'].idxmin(
    )]['Happiness Rank'] - df_by_selected_country.loc[df_by_selected_country['Year'].idxmax()]['Happiness Rank']
    st.write(f"""Difference from <b>{df_by_selected_country.loc[df_by_selected_country['Year'].idxmin()]['Year']}</b> until
                <b>{df_by_selected_country.loc[df_by_selected_country['Year'].idxmax()]['Year']}</b> :
                <b style="color: {'green' if difference > 0 else 'gray' if difference == 0 else 'red'}">{"%+d" % (difference)}</b>""", unsafe_allow_html=True)

    fig = px.line(df_by_selected_country, x="Year",
                  y="Happiness Rank", title=f"Year / Happiness Rank")
    st.plotly_chart(fig)


else:

    selected_year = st.selectbox(
        'Choose The Year',
        (str(year) for year in range(2015, 2023)))
    st.markdown(
        f"You Selected Year: <b>{selected_year}</b>", unsafe_allow_html=True)

    df = overall_df[overall_df.Year == int(selected_year)]

    try:
        if not check_boxes['All']:
            filtered_region = list(
                filter(lambda k: check_boxes[k], check_boxes))
            st.write(
                f'<b>Filters:</b> <br/><span class="c-pill c-pill--warning">{",".join(filtered_region)}</span>', unsafe_allow_html=True)
            df = df[df['Region'].isin(filtered_region)]
    except KeyError:
        pass

    # Happiness Rank Column as Index
    df = df.sort_values(by=["Happiness Rank"], ascending=True)
    if df.shape[0] > 0:

        st.dataframe(df.set_index(df.columns[0]))

        c5, c6 = st.columns([0.5, 0.5])
        with c5:
            x_list = list(df.columns) + ['Happiness Score']
            x_list.remove('Generosity')
            x_axis = st.selectbox(
                'Choose X axis criterion',
                (f'x: {cri}' for cri in x_list[::-1]))
        with c6:
            y_list = list(df.columns) + ['Generosity']
            y_list.remove('Generosity')
            y_axis = st.selectbox(
                'Choose Y axis criterion',
                (f'y: {cri}' for cri in y_list[::-1]))
        try:
            try:
                corr = round(df[x_axis.replace('x: ', '')].apply(lambda x: float(x.split()[0].replace(',', ''))).corr(
                    df[y_axis.replace('y: ', '')].apply(lambda x: float(x.split()[0].replace(',', '')))), 2)
            except AttributeError:
                corr = round(df[x_axis.replace('x: ', '')].astype(float).corr(
                    df[y_axis.replace('y: ', '')].astype(float)), 2)
            if str(corr) != "nan":
                st.markdown(f"Correlation between <b>{x_axis.replace('x: ','')}</b> and <b>{y_axis.replace('y: ','')}</b> is : <b>{corr}</b>",
                            unsafe_allow_html=True)
        except ValueError:
            print("Error catched")
            pass
        fig = px.scatter(df, x=x_axis.replace('x: ', ''),
                         y=y_axis.replace('y: ', ''), title=f"{x_axis.replace('x: ','')} / {y_axis.replace('y: ','')}")

        st.plotly_chart(fig)
        st.header('Happiness Rank over 2015 until 2022')
        try:
            if not check_boxes['All']:
                st.dataframe(overall_df[overall_df['Region'].isin(filtered_region)].groupby(
                    'Country')['Happiness Rank'].describe())
            else:
                st.dataframe(get_desc_by_country(overall_df))
        except KeyError:
            pass

        st.write("""<small>
                    <b>Count:</b> It refers to the count number of observations (for example: Afghanistan 8 times and Angola 4 times from 2015 until 2022)
                    <br/>
                    <b>Mean:</b> The average rank
                    <br/>
                    <b>Std:</b> The standard deviation
                    <br/>
                    <b>Min:</b> Lowest Rank
                    <br/>
                    <b>25%:</b> The 25% percentile
                    <br/>
                    <b>50%:</b> The 50% percentile
                    <br/>
                    <b>75%:</b> The 75% percentile
                    <br/>
                    <b>Max:</b> Highest Rank
                    </small>
                    """, unsafe_allow_html=True)

    else:
        st.markdown("<b>No Data, Please Change Filters</b>",
                    unsafe_allow_html=True)