Update pages/2player_comparison.py

pages/2player_comparison.py

@@ -1,184 +1,133 @@
 import streamlit as st
 import pandas as pd
-import numpy as np
 import pickle
-from PIL import Image
-
-# ========== Custom CSS Styling ==========
-st.markdown("""
-<style>
-.stApp {
-    background-color: #f0f4f8;  /* Light clean background */
-}
-.title {
-    text-align: center;
-    color: #222222;
-    font-size: 40px;
-    font-family: 'Times New Roman';
-    font-weight: bold;
-}
-.header {
-    font-size: 32px;
-    font-family: 'Times New Roman';
-    color: #1a1a1a;
-    font-style: italic;
-    font-weight: bold;
-}
-.header1 {
-    font-size: 28px;
-    font-family: 'Times New Roman';
-    color: #333333;
-    font-style: italic;
-    font-weight: bold;
-}
-.header2 {
-    font-size: 20px;
-    font-family: 'Times New Roman';
-    color: #444444;
-    font-style: italic;
-    font-weight: bold;
-}
-.content {
-    font-size: 20px;
-    font-family: 'Times New Roman';
-    line-height: 1.6;
-    color: #222222;
-}
-.example {
-    font-size: 16px;
-    font-family: 'Times New Roman';
-    font-style: italic;
-    color: #666666;
-}
-</style>
-""", unsafe_allow_html=True)
-
-# ========== Load Model and Data ==========
+import matplotlib.pyplot as plt
+import seaborn as sns
+import numpy as np
+
+# Load model and encoder
 @st.cache_resource
-def load_model_and_data():
-    with open("best_rf_pipeline.pkl", "rb") as f:
+def load_model_and_encoder():
+    with open('best_rf_pipeline.pkl', 'rb') as f:
         model = pickle.load(f)
-    with open("label_encoder.pkl", "rb") as f:
+    with open('label_encoder.pkl', 'rb') as f:
         encoder = pickle.load(f)
-    stats_df = pd.read_csv("Reduced_final_teams.csv")
-    return model, encoder, stats_df
-
-pipeline, label_encoder, stats_df = load_model_and_data()
-
-# ========== Plot Functions ==========
-def plot_bar_comparison(x_labels, p1_values, p2_values, p1_name, p2_name, title, y_axis_label):
-    fig = px.bar(
-        x=x_labels * 2,
-        y=p1_values + p2_values,
-        color=[p1_name] * len(x_labels) + [p2_name] * len(x_labels),
-        barmode="group",
-        labels={"x": "Format", "y": y_axis_label, "color": "Player"},
-        title=title,
-        text=p1_values + p2_values
-    )
-    fig.update_traces(textposition='outside')
-    st.plotly_chart(fig, use_container_width=True)
-
-def plot_line_comparison(x_labels, p1_values, p2_values, p1_name, p2_name, title, y_axis_label):
-    fig = px.line(
-        x=x_labels * 2,
-        y=p1_values + p2_values,
-        color=[p1_name] * len(x_labels) + [p2_name] * len(x_labels),
-        markers=True,
-        labels={"x": "Format", "y": y_axis_label, "color": "Player"},
-        title=title
-    )
-    st.plotly_chart(fig, use_container_width=True)
-
-# ========== App Layout ==========
-st.markdown("<div class='title'> Player Comparison📊 </div><br>", unsafe_allow_html=True)
-
-col_img1, col_img2 = st.columns(2)
-
-with col_img1:
-    st.markdown("<div class='header'> Upload Player Image 1 </div><br>", unsafe_allow_html=True)
-    uploaded_file1 = st.file_uploader("", type=["jpg", "png", "jpeg"], key="img1", label_visibility="collapsed")
-    if uploaded_file1 is not None:
-        display_img1 = Image.open(uploaded_file1).convert('RGB')
-        st.image(display_img1, caption="Player 1", use_container_width=True)
-
-with col_img2:
-    st.markdown("<div class='header'> Upload Player Image 2 </div><br>", unsafe_allow_html=True)
-    uploaded_file2 = st.file_uploader("", type=["jpg", "png", "jpeg"], key="img2", label_visibility="collapsed")
-    if uploaded_file2 is not None:
-        display_img2 = Image.open(uploaded_file2).convert('RGB')
-        st.image(display_img2, caption="Player 2", use_container_width=True)
-
-# ========== Predict and Compare ==========
-if uploaded_file1 and uploaded_file2:
-    gray_img1 = Image.open(uploaded_file1).convert('L').resize((50, 50))
-    gray_img2 = Image.open(uploaded_file2).convert('L').resize((50, 50))
-
-    arr1 = np.array(gray_img1).flatten().reshape(1, -1)
-    arr2 = np.array(gray_img2).flatten().reshape(1, -1)
-
-    model, encoder, stats_df = load_model_and_data()
-
-    pred1 = model.predict(arr1)[0]
-    pred2 = model.predict(arr2)[0]
-
-    player_1 = encoder.inverse_transform([pred1])[0]
-    player_2 = encoder.inverse_transform([pred2])[0]
-
-    st.success(f"Player 1: {player_1}")
-    st.success(f"Player 2: {player_2}")
-
-    p1_data = stats_df[stats_df["Player"] == player_1].iloc[0]
-    p2_data = stats_df[stats_df["Player"] == player_2].iloc[0]
-
-    colA, colB = st.columns(2)
-    with colA:
-        show_batting = st.button("Show Batting Comparison 🏏")
-    with colB:
-        show_bowling = st.button("Show Bowling Comparison 🎯")
-
-    if show_batting:
-        st.markdown(f"<div class='header'>Batting Comparison: {player_1} vs {player_2} 🏏</div><br>", unsafe_allow_html=True)
-        formats = ["Test", "ODI", "T20", "IPL"]
-
-        batting_bar_metrics = {
-            "Runs": "batting_Runs_",
-            "50s": "batting_50s_",
-            "100s": "batting_100s_"
-        }
-        for metric_name, prefix in batting_bar_metrics.items():
-            p1_vals = [p1_data.get(f"{prefix}{fmt}", 0) for fmt in formats]
-            p2_vals = [p2_data.get(f"{prefix}{fmt}", 0) for fmt in formats]
-            plot_bar_comparison(formats, p1_vals, p2_vals, player_1, player_2, f"{metric_name} by Format", metric_name)
-
-        batting_line_metrics = {
-            "Batting Average": "batting_Average_",
-            "Strike Rate": "batting_SR_"
-        }
-        for metric_name, prefix in batting_line_metrics.items():
-            p1_vals = [p1_data.get(f"{prefix}{fmt}", 0) for fmt in formats]
-            p2_vals = [p2_data.get(f"{prefix}{fmt}", 0) for fmt in formats]
-            plot_line_comparison(formats, p1_vals, p2_vals, player_1, player_2, f"{metric_name} by Format", metric_name)
-
-    if show_bowling:
-        st.markdown(f"<div class='header'>Bowling Comparison: {player_1} vs {player_2} 🎯</div><br>", unsafe_allow_html=True)
-        formats = ["Test", "ODI", "T20", "IPL"]
-
-        bowling_bar_metrics = {
-            "Wickets": "_Wickets",
-            "Maidens": "_Maidens",
-            "Economy Rate": "_Eco"
-        }
-        for metric_name, suffix in bowling_bar_metrics.items():
-            p1_vals = [p1_data.get(f"bowling_{fmt}{suffix}", 0) for fmt in formats]
-            p2_vals = [p2_data.get(f"bowling_{fmt}{suffix}", 0) for fmt in formats]
-            plot_bar_comparison(formats, p1_vals, p2_vals, player_1, player_2, f"{metric_name} by Format", metric_name)
-
-        bowling_line_metrics = {
-            "Bowling Average": "_Avg",
-            "Strike Rate": "_SR"
-        }
-        for metric_name, suffix in bowling_line_metrics.items():
-            p1_vals = [p1_data.get(f"bowling_{fmt}{suffix}", 0) for fmt in formats]
-            p2_vals = [p2_data.get(f"bowling_{fmt}{suffix}", 0) for fmt in formats]
-            plot_line_comparison(formats, p1_vals, p2_vals, player_1, player_2, f"{metric_name} by Format", metric_name)
+    return model, encoder
+
+# Load player dataset
+@st.cache_data
+def load_data():
+    return pd.read_csv('Reduced_final_teams.csv')
+
+# Fuzzy match player name
+def get_matching_player(name_from_file, player_list):
+    name_lower = name_from_file.lower()
+    for player in player_list:
+        if player.lower() == name_lower:
+            return player
+    return None
+
+# Horizontal bar chart
+def plot_horizontal_bar(df, player1, player2):
+    st.subheader("📊 Stat Comparison - Horizontal Bar Chart")
+    num_cols = df.select_dtypes(include='number').columns
+    df_num = df[num_cols].T
+    df_num.columns = [player1, player2]
+    df_num = df_num.sort_values(by=player1, ascending=False).head(15)
+
+    fig, ax = plt.subplots(figsize=(10, 7))
+    df_num.plot(kind='barh', ax=ax)
+    ax.set_title(f"{player1} vs {player2} - Key Stats")
+    ax.set_xlabel("Value")
+    ax.set_ylabel("Metric")
+    ax.legend(loc="lower right")
+    st.pyplot(fig)
+
+# Pie chart comparison
+def plot_pie_charts(player1_data, player2_data, player1, player2):
+    st.subheader("🥧 Batting vs Bowling Contribution")
+    col1, col2 = st.columns(2)
+
+    for col, player_data, player_name in zip([col1, col2], [player1_data, player2_data], [player1, player2]):
+        batting_total = player_data.get('Runs_ODI', 0) + player_data.get('Runs_T20', 0) + player_data.get('Runs_Test', 0)
+        bowling_total = player_data.get('Wickets_ODI', 0) + player_data.get('Wickets_T20', 0) + player_data.get('Wickets_Test', 0)
+
+        labels = ['Batting', 'Bowling']
+        sizes = [batting_total, bowling_total]
+
+        fig, ax = plt.subplots()
+        ax.pie(sizes, labels=labels, autopct='%1.1f%%', startangle=90, colors=['#4CAF50', '#2196F3'])
+        ax.axis('equal')
+        col.pyplot(fig)
+        col.caption(f"{player_name}'s Batting vs Bowling")
+
+# Bowling metric bar chart
+def plot_bowling_comparison(df, player1, player2):
+    st.subheader("🎯 Bowling Metrics Comparison")
+
+    bowling_cols = [col for col in df.columns if 'Wickets' in col or 'Economy' in col or 'Bowling_Average' in col]
+    df_bowling = df[bowling_cols].T
+    df_bowling.columns = [player1, player2]
+    df_bowling = df_bowling.dropna().sort_values(by=player1, ascending=False)
+
+    fig, ax = plt.subplots(figsize=(10, 6))
+    df_bowling.plot(kind='bar', ax=ax)
+    ax.set_title("Bowling Stats")
+    ax.set_ylabel("Value")
+    ax.set_xticklabels(df_bowling.index, rotation=45, ha='right')
+    ax.legend(loc="upper right")
+    st.pyplot(fig)
+
+# Main app
+def main():
+    st.set_page_config(layout="wide")
+    st.title("Cricket Player Comparison Tool 🏏")
+
+    df = load_data()
+    model, encoder = load_model_and_encoder()
+    player_list = df['Player'].tolist()
+
+    # Upload images
+    col1, col2 = st.columns(2)
+    with col1:
+        img1 = st.file_uploader("Upload Image for Player 1", type=['png', 'jpg', 'jpeg'], key='img1')
+    with col2:
+        img2 = st.file_uploader("Upload Image for Player 2", type=['png', 'jpg', 'jpeg'], key='img2')
+
+    if img1 and img2:
+        name1_raw = img1.name.rsplit('.', 1)[0]
+        name2_raw = img2.name.rsplit('.', 1)[0]
+
+        player1_name = get_matching_player(name1_raw, player_list)
+        player2_name = get_matching_player(name2_raw, player_list)
+
+        if player1_name and player2_name and player1_name != player2_name:
+            player1_data = df[df['Player'].str.lower() == player1_name.lower()].squeeze()
+            player2_data = df[df['Player'].str.lower() == player2_name.lower()].squeeze()
+
+            st.success(f"Comparing **{player1_name}** vs **{player2_name}**")
+
+            col3, col4 = st.columns(2)
+            with col3:
+                st.image(img1, caption=player1_name, use_container_width=True)
+            with col4:
+                st.image(img2, caption=player2_name, use_container_width=True)
+
+            # Display comparison table
+            comparison_df = pd.DataFrame([player1_data, player2_data])
+            comparison_df.set_index('Player', inplace=True)
+            st.subheader("📋 Full Stats Table")
+            st.dataframe(comparison_df.T)
+
+            # Visualizations
+            plot_horizontal_bar(comparison_df, player1_name, player2_name)
+            plot_pie_charts(player1_data, player2_data, player1_name, player2_name)
+            plot_bowling_comparison(comparison_df, player1_name, player2_name)
+
+        else:
+            st.error("❌ Player names from image files don't match or are the same. Please check file names.")
+    else:
+        st.info("📸 Please upload two player images to continue.")
+
+if __name__ == "__main__":
+    main()