app.py

import pandas as pd
import streamlit as st
from datasets import load_dataset
from datetime import datetime


# Load data
@st.cache_data  # Cache data for performance improvement
def load_data(language="en"):
    dataset = load_dataset("santhosh/day_in_history", language, split="train")
    df = pd.DataFrame(dataset)
    return df


# Function to process user input (date) and return description and reference
def process_date(day, month, year=None):
    # Filter data based on selected date
    if year is None or year == "":
        filtered_data = df[(df["month"] == month) & (df["day"] == int(day))]
    else:
        filtered_data = df[
            (df["year"] == int(year)) & (df["month"] == month) & (df["day"] == int(day))
        ]

    if not filtered_data.empty:
        # Prepare empty lists to store descriptions and references
        descriptions = []
        references = []

        # Loop through filtered data and append descriptions and references
        for index, row in filtered_data.iterrows():
            descriptions.append(row["event_description"])
            references.append(row["reference"])

        # Return lists of descriptions and references
        return descriptions
    else:
        return [f"No data found for selected date {year} {month} {day}"]

supported_languages = {
    "en": "English",
    "ml": "Malayalam",
}
MONTHS = [
    "January",
    "February",
    "March",
    "April",
    "May",
    "June",
    "July",
    "August",
    "September",
    "October",
    "November",
    "December",
]


def main():
    # Page title and header
    st.title("Day in History")
    with st.form("my_form"):
        st.header("Select a date")
        selected_language = st.selectbox(
            "Select language",
            supported_languages.keys(),
            format_func=lambda x: supported_languages[x],
            index=0,
        )
        col1, col2, col3 = st.columns(3)
        # Datepicker
        with col1:
            selected_day = st.selectbox(
                "Select Day", range(1, 32), index=datetime.now().day - 1
            )
        with col2:
            selected_month = st.selectbox(
                "Select Month",
                range(0, 12),
                format_func=lambda x: MONTHS[x],
                index=datetime.now().month - 1,
            )
        with col3:
            selected_year = st.number_input(
                "Enter Year (optional)", min_value=0, max_value=9999, value=None
            )
        submitted = st.form_submit_button("Submit")
    if submitted:
        df = load_data(selected_language)
        # Process data based on selected date
        if selected_year is None or selected_year == "":
            filtered_data = df[
                (df["month"] == int(selected_month + 1))
                & (df["day"] == int(selected_day))
            ]
        else:
            filtered_data = df[
                (df["year"] == int(selected_year))
                & (df["month"] == selected_month + 1)
                & (df["day"] == int(selected_day))
            ]

        # Display results
        if not filtered_data.empty:
            st.subheader("Search Results")

            for index, row in filtered_data.iterrows():
                container = st.container(border=True)
                container.subheader(
                    f"{row['year']} {MONTHS[row['month']-1]} {row['day']}"
                )
                container.markdown(
                    f"{row['event_description']}", unsafe_allow_html=True
                )
                if row["reference"] is not None:
                    container.markdown(f"{row['reference']}", unsafe_allow_html=True)
        else:
            st.warning(
                f"No data found for selected date {selected_day} {MONTHS[selected_month]} {selected_year}"
            )


if __name__ == "__main__":
    main()