src/streamlit_app.py

import streamlit as st
import json
import pandas as pd
import altair as alt

# ---------------------------
# CONFIG
# ---------------------------
FILE_PATH = "src/data_final_version.json"

st.set_page_config(layout="wide")
st.title("Dota Toxic Messages Dashboard")

# ---------------------------
# LOAD DATA
# ---------------------------
@st.cache_data
def load_data():
    with open(FILE_PATH, "r", encoding="utf-8") as f:
        raw = json.load(f)
    return pd.DataFrame(raw)

df = load_data()

# ---------------------------
# CLEAN DATA
# ---------------------------
df["lobby_type"] = df["lobby_type"].fillna("Unknown")
df["region"] = df["region"].fillna("Unknown")
df["message_classified"] = df["message_classified"].fillna("unknown")

# ---------------------------
# SIDEBAR FILTERS
# ---------------------------
st.sidebar.header("Filters")

regions = st.sidebar.multiselect(
    "Region",
    options=df["region"].unique(),
    default=df["region"].unique()
)

lobbies = st.sidebar.multiselect(
    "Lobby Type",
    options=df["lobby_type"].unique(),
    default=df["lobby_type"].unique()
)

df_filtered = df[
    (df["region"].isin(regions)) &
    (df["lobby_type"].isin(lobbies))
]

# ---------------------------
# KPI
# ---------------------------
col1, col2, col3 = st.columns(3)

col1.metric("Total Messages", len(df_filtered))
col2.metric("Unique Matches", df_filtered["match_id"].nunique())
col3.metric("Avg Duration", round(df_filtered["duration_game"].mean(), 1))

# ---------------------------
# CHART 1 - Toxic categories
# ---------------------------
st.subheader("Toxic Message Distribution")

chart1 = alt.Chart(df_filtered).mark_bar().encode(
    x=alt.X("count()", title="Count"),
    y=alt.Y("message_classified", sort="-x", title="Category"),
    color="message_classified"
)

st.altair_chart(chart1, use_container_width=True)

# ---------------------------
# CHART 2 - By Lobby Type
# ---------------------------
st.subheader("Toxicity by Lobby Type")

chart2 = alt.Chart(df_filtered).mark_bar().encode(
    x="lobby_type",
    y="count()",
    color="lobby_type"
)

st.altair_chart(chart2, use_container_width=True)

# ---------------------------
# CHART 3 - By Region
# ---------------------------
st.subheader("Toxicity by Region")

chart3 = alt.Chart(df_filtered).mark_bar().encode(
    x=alt.X("region", sort="-y"),
    y="count()",
    color="region"
)

st.altair_chart(chart3, use_container_width=True)

# ---------------------------
# RAW DATA (optional)
# ---------------------------
with st.expander("Show raw data"):
    st.dataframe(df_filtered)