import streamlit as st import pandas as pd import numpy as np import requests import os from datetime import datetime from bs4 import BeautifulSoup, Comment # Re-added for custom BS scraping import re # Re-added for custom BS scraping (regex for comments) import plotly.express as px import plotly.graph_objects as go # —————————————————————————————————————————————— # Import BRScraper try: from BRScraper import nba BRSCRAPER_AVAILABLE = True except ImportError: BRSCRAPER_AVAILABLE = False st.error("BRScraper not found. Please install with: `pip install BRScraper`") # Page config st.set_page_config( page_title="NBA Analytics Hub", page_icon="🏀", layout="wide", initial_sidebar_state="expanded" ) # —————————————————————————————————————————————— # CSS (black & white theme) st.markdown(""" """, unsafe_allow_html=True) # Initialize chat history if 'chat_history' not in st.session_state: st.session_state.chat_history = [] # —————————————————————————————————————————————— # Custom BeautifulSoup Data Fetching Utilities (for teams) # —————————————————————————————————————————————— @st.cache_data(ttl=3600) def fetch_html(url): """Fetch raw HTML for a URL (with error handling).""" try: resp = requests.get(url, timeout=20) resp.raise_for_status() return resp.text except requests.exceptions.RequestException as e: st.error(f"Failed to fetch {url}: {e}") return "" except Exception as e: st.error(f"An unexpected error occurred while fetching {url}: {e}") return "" def parse_table(html, table_id=None): """ Given raw HTML and optional table_id, locate that , handling cases where it's commented out, then parse it with pandas.read_html. """ if not html: return pd.DataFrame() soup = BeautifulSoup(html, "lxml") # Using lxml for potentially faster parsing tbl_html = "" if table_id: # First, try to find the table directly tbl = soup.find("table", {"id": table_id}) if tbl: tbl_html = str(tbl) else: # If not found directly, search for it within HTML comments comment_pattern = re.compile(r'' % table_id, re.DOTALL) comment_match = comment_pattern.search(html) if comment_match: # Extract the content of the comment comment_content = comment_match.group(0) # Remove the comment tags comment_content = comment_content.replace('', '') # Parse the comment content as new HTML comment_soup = BeautifulSoup(comment_content, 'lxml') tbl = comment_soup.find('table', {'id': table_id}) if tbl: tbl_html = str(tbl) else: # fallback: first table on page (only if no table_id specified) first = soup.find("table") if first: tbl_html = str(first) if not tbl_html: return pd.DataFrame() try: # pd.read_html returns a list of DataFrames, we want the first one return pd.read_html(tbl_html)[0] except ValueError: # No tables found in the provided HTML string return pd.DataFrame() except Exception as e: st.error(f"Error parsing table with pandas: {e}") return pd.DataFrame() def clean_team_name(team_name): """ Clean and standardize team names from Basketball Reference. """ if pd.isna(team_name): return team_name # Remove any asterisks or other symbols team_name = str(team_name).strip().replace('*', '') # Handle special cases for team name variations (if needed, keep for consistency) team_mapping = { 'NOP': 'NO', # New Orleans Pelicans sometimes shown as NOP 'PHX': 'PHO', # Phoenix Suns sometimes shown as PHX 'BRK': 'BKN', # Brooklyn Nets sometimes shown as BRK 'CHA': 'CHO', # Charlotte sometimes inconsistent 'UTA': 'UTH' # Utah Jazz sometimes shown as UTA } return team_mapping.get(team_name, team_name) @st.cache_data(ttl=300) def get_team_stats_bs(year): """ Scrapes the league’s per‑game team stats table from: https://www.basketball-reference.com/leagues/NBA_{year}_per_game.html Returns cleaned DataFrame. """ url = f"https://www.basketball-reference.com/leagues/NBA_{year}_per_game.html" html = fetch_html(url) if not html: return pd.DataFrame() # Try multiple possible table IDs for team stats possible_table_ids = ["per_game-team", "per_game_team", "team-stats-per_game", "teams_per_game"] df = pd.DataFrame() for table_id in possible_table_ids: df = parse_table(html, table_id=table_id) if not df.empty: break # If no specific table found, try to find any table with team data if df.empty: soup = BeautifulSoup(html, "lxml") # Use lxml for consistency tables = soup.find_all("table") for table in tables: if table.find("th", string=lambda text: text and "team" in text.lower()): df = parse_table(str(table)) if not df.empty: break if df.empty: st.warning(f"Could not find team stats table for {year}") return pd.DataFrame() # Handle potential MultiIndex columns if isinstance(df.columns, pd.MultiIndex): df.columns = ['_'.join(str(col).strip() for col in cols if str(col).strip() and str(col).strip() != 'Unnamed: 0_level_0') for cols in df.columns.values] # Clean column names df.columns = [str(col).strip() for col in df.columns] # Find team column team_col = None for col in df.columns: if 'team' in col.lower() or col in ['Team', 'Tm']: team_col = col break if team_col is None: st.warning(f"Could not find team column in team stats. Available columns: {df.columns.tolist()}") return pd.DataFrame() # Rename team column to standard name if team_col != 'Team': df = df.rename(columns={team_col: 'Team'}) # Remove header rows df = df[df["Team"].astype(str) != "Team"].copy() df = df[df["Team"].notna()].copy() # Standardize column names column_mapping = { 'G': 'GP', 'MP': 'MIN', 'FG%': 'FG_PCT', '3P%': 'FG3_PCT', 'FT%': 'FT_PCT', 'TRB': 'REB', 'AST': 'AST', 'STL': 'STL', 'BLK': 'BLK', 'TOV': 'TO', 'PF': 'PF', 'PTS': 'PTS', 'Rk': 'RANK', 'W': 'WINS', 'L': 'LOSSES', 'W/L%': 'WIN_LOSS_PCT', 'FG': 'FGM', 'FGA': 'FGA', '3P': 'FG3M', '3PA': 'FG3A', '2P': 'FGM2', '2PA': 'FGA2', '2P%': 'FG2_PCT', 'eFG%': 'EFG_PCT', 'FT': 'FTM', 'FTA': 'FTA', 'ORB': 'OREB', 'DRB': 'DREB' } # Apply column mapping only for columns that exist for old_col, new_col in column_mapping.items(): if old_col in df.columns: df = df.rename(columns={old_col: new_col}) # Clean team names if 'Team' in df.columns: # Ensure 'Team' column exists before applying df['Team'] = df['Team'].apply(clean_team_name) # Convert numeric columns non_numeric_cols = {"Team", "RANK"} for col in df.columns: if col not in non_numeric_cols: df[col] = pd.to_numeric(df[col], errors="coerce") return df # —————————————————————————————————————————————— # BRScraper Data Fetching Utilities (for players) # —————————————————————————————————————————————— # Helper for dynamic season generation def get_available_seasons(num_seasons=6): """Generates a list of recent NBA seasons in 'YYYY–YY' format.""" current_year = datetime.now().year current_month = datetime.now().month # Determine the latest season end year. latest_season_end_year = current_year if current_month >= 7: # NBA season typically ends in June latest_season_end_year += 1 seasons_list = [] for i in range(num_seasons): end_year = latest_season_end_year - i start_year = end_year - 1 seasons_list.append(f"{start_year}–{end_year}") return sorted(seasons_list, reverse=True) # Sort to show most recent first @st.cache_data(ttl=3600) def get_player_index_brscraper(): """ Uses BRScraper to get a list of players from a recent season's stats. This serves as our player index for the multiselect. """ if not BRSCRAPER_AVAILABLE: return pd.DataFrame(columns=['name']) try: # Get the end year of the most recent season for BRScraper # Example: '2024–25' -> 2025 latest_season_end_year = int(get_available_seasons(1)[0].split('–')[1]) # Use nba.get_stats to get a list of players for the latest season # BRScraper's get_stats returns a 'Player' column df = nba.get_stats(latest_season_end_year, info='per_game', rename=False) if df.empty or 'Player' not in df.columns: st.warning(f"BRScraper could not fetch player list for {latest_season_end_year}. Falling back to common players.") # Fallback to a hardcoded list of common players for demo common_players = [ 'LeBron James', 'Stephen Curry', 'Kevin Durant', 'Giannis Antetokounmpo', 'Nikola Jokic', 'Joel Embiid', 'Jayson Tatum', 'Luka Doncic', 'Damian Lillard', 'Jimmy Butler', 'Kawhi Leonard', 'Paul George', 'Anthony Davis', 'Rudy Gobert', 'Donovan Mitchell', 'Trae Young', 'Devin Booker', 'Karl-Anthony Towns', 'Zion Williamson', 'Ja Morant' ] return pd.DataFrame({'name': common_players}) player_names = df['Player'].unique().tolist() return pd.DataFrame({'name': player_names}) except Exception as e: st.error(f"Error fetching player index with BRScraper: {e}. Falling back to common players.") # Fallback to hardcoded list fallback_players = [ 'LeBron James', 'Stephen Curry', 'Kevin Durant', 'Giannis Antetokounmpo', 'Nikola Jokic', 'Joel Embiid', 'Jayson Tatum', 'Luka Doncic' ] return pd.DataFrame({'name': fallback_players}) @st.cache_data(ttl=300) def get_player_career_stats_brscraper(player_name, seasons_to_check=10): """ Build a player's career stats by fetching per-game data season-by-season via nba.get_stats, instead of nba.get_player_stats (which was returning empty). """ if not BRSCRAPER_AVAILABLE: return pd.DataFrame() all_rows = [] seasons = get_available_seasons(seasons_to_check) for season_str in seasons: end_year = int(season_str.split('–')[1]) try: df_season = nba.get_stats(end_year, info='per_game', playoffs=False, rename=False) if 'Player' in df_season.columns: row = df_season[df_season['Player'] == player_name] if not row.empty: row = row.copy() # normalize the season label row['Season'] = season_str all_rows.append(row) except Exception as e: # skip any failures st.warning(f"Could not fetch {season_str} for {player_name}: {e}") if not all_rows: return pd.DataFrame() df = pd.concat(all_rows, ignore_index=True) # Standardize columns mapping = { 'G':'GP','GS':'GS','MP':'MIN', 'FG%':'FG_PCT','3P%':'FG3_PCT','FT%':'FT_PCT', 'TRB':'REB','AST':'AST','STL':'STL','BLK':'BLK','TOV':'TO', 'PF':'PF','PTS':'PTS','ORB':'OREB','DRB':'DREB', 'FG':'FGM','FGA':'FGA','3P':'FG3M','3PA':'FG3A', '2P':'FGM2','2PA':'FGA2','2P%':'FG2_PCT','eFG%':'EFG_PCT', 'FT':'FTM','FTA':'FTA' } df = df.rename(columns={o:n for o,n in mapping.items() if o in df.columns}) # Convert numeric columns non_num = {'Season','Player','Tm','Lg','Pos'} for col in df.columns: if col not in non_num: df[col] = pd.to_numeric(df[col], errors='coerce') df['Player'] = player_name return df # —————————————————————————————————————————————— # Perplexity API PERP_KEY = os.getenv("PERPLEXITY_API_KEY") PERP_URL = "https://api.perplexity.ai/chat/completions" def ask_perp(prompt, system="You are a helpful NBA analyst AI.", max_tokens=500, temp=0.2): if not PERP_KEY: st.error("Set PERPLEXITY_API_KEY env var.") return "" hdr = {'Authorization':f'Bearer {PERP_KEY}','Content-Type':'application/json'} payload = { "model":"sonar-medium-online", # Using a commonly available online model "messages":[{"role":"system","content":system},{"role":"user","content":prompt}], "max_tokens":max_tokens, "temperature":temp } try: r = requests.post(PERP_URL, json=payload, headers=hdr, timeout=45) r.raise_for_status() return r.json().get("choices", [])[0].get("message", {}).get("content", "") except Exception as e: st.error(f"Perplexity API error: {e}") return "" # —————————————————————————————————————————————— # Plotting functions def create_comparison_chart(data, players_names, metric): """Create comparison chart for players""" fig = go.Figure() for i, player in enumerate(players_names): if player in data['Player'].values: player_data = data[data['Player'] == player] fig.add_trace(go.Scatter( x=player_data['Season'], y=player_data[metric], mode='lines+markers', name=player, line=dict(width=3) )) fig.update_layout( title=f"{metric} Comparison", xaxis_title="Season", yaxis_title=metric, hovermode='x unified', height=500 ) return fig def create_radar_chart(player_stats, categories): """Create radar chart for player comparison""" fig = go.Figure() for player_name, stats in player_stats.items(): r_values = [stats.get(cat,0) for cat in categories] fig.add_trace(go.Scatterpolar( r=r_values, theta=categories, fill='toself', name=player_name, opacity=0.7 )) fig.update_layout( polar=dict(radialaxis=dict(visible=True, range=[0,100])), showlegend=True, title="Player Comparison Radar Chart" ) return fig # —————————————————————————————————————————————— # Main structure def main(): if not BRSCRAPER_AVAILABLE: st.warning("⚠️ BRScraper is not installed. Install with `pip install BRScraper`") st.markdown('

🏀 NBA Analytics Hub (BBR Edition)

', unsafe_allow_html=True) st.sidebar.title("Navigation") page = st.sidebar.radio("", [ "Player vs Player Comparison", "Team vs Team Analysis", "NBA Awards Predictor", "AI Chat & Insights", "Young Player Projections", "Similar Players Finder", "Roster Builder", "Trade Scenario Analyzer" ]) if page == "Player vs Player Comparison": player_vs_player() elif page == "Team vs Team Analysis": team_vs_team() elif page == "NBA Awards Predictor": awards_predictor() elif page == "AI Chat & Insights": ai_chat() elif page == "Young Player Projections": young_projections() elif page == "Similar Players Finder": similar_players() elif page == "Roster Builder": roster_builder() else: trade_analyzer() # —————————————————————————————————————————————— # Page implementations def player_vs_player(): st.markdown('

Player vs Player Comparison

', unsafe_allow_html=True) if not BRSCRAPER_AVAILABLE: st.error("BRScraper is required for this feature. Please install BRScraper.") return idx = get_player_index_brscraper() selected_players = st.multiselect("Select Players (up to 4)", idx['name'], max_selections=4) seasons = get_available_seasons() selected_seasons = st.multiselect("Select Seasons", seasons, default=[seasons[0]] if seasons else []) # ─── RAW DEBUG PREVIEW ────────────────────────────────────────────────────── # This section is for debugging and can be commented out or removed in production # if selected_players: # st.markdown("### 🧪 Raw Seasonal Data Preview") # for player in selected_players: # st.subheader(player) # # Fetching for a few seasons to show raw structure # raw = get_player_career_stats_brscraper(player) # if raw.empty: # st.write("❌ No rows returned—player may be named differently or data unavailable.") # else: # st.write("Columns:", raw.columns.tolist()) # st.dataframe(raw, use_container_width=True) # st.markdown("---") # ──────────────────────────────────────────────────────────────────────────── if st.button("Run Comparison"): if not selected_players: st.warning("Please select at least one player.") return all_data, no_data = [], [] with st.spinner("Fetching filtered data..."): for p in selected_players: # get_player_career_stats_brscraper fetches all career stats, then we filter df = get_player_career_stats_brscraper(p) if not df.empty: filt = df[df['Season'].isin(selected_seasons)] if not filt.empty: all_data.append(filt) else: no_data.append(p) else: no_data.append(p) if no_data: st.info(f"No data found for the selected seasons ({', '.join(selected_seasons)}) for: {', '.join(no_data)}.") if not all_data: st.error("No data available for any of the selected players and seasons to display. Please adjust your selections.") return comp_df = pd.concat(all_data, ignore_index=True) tabs = st.tabs(["Basic Stats", "Advanced Stats", "Visualizations"]) with tabs[0]: st.subheader("Basic Statistics") if len(selected_seasons) > 1: basic_display_df = comp_df.groupby('Player').mean(numeric_only=True).reset_index() basic_cols = ['Player','GP','MIN','PTS','REB','AST','STL','BLK','FG_PCT','FT_PCT','FG3_PCT'] else: basic_display_df = comp_df.copy() basic_cols = ['Player','Season','GP','MIN','PTS','REB','AST','STL','BLK','FG_PCT','FT_PCT','FG3_PCT'] st.dataframe(basic_display_df[[c for c in basic_cols if c in basic_display_df.columns]].round(2), use_container_width=True) with tabs[1]: st.subheader("Advanced Statistics") if not comp_df.empty: advanced_df = comp_df.copy() advanced_df['FGA'] = pd.to_numeric(advanced_df.get('FGA', 0), errors='coerce').fillna(0) advanced_df['FTA'] = pd.to_numeric(advanced_df.get('FTA', 0), errors='coerce').fillna(0) advanced_df['PTS'] = pd.to_numeric(advanced_df.get('PTS', 0), errors='coerce').fillna(0) advanced_df['TS_PCT'] = advanced_df.apply( lambda r: r['PTS'] / (2 * (r['FGA'] + 0.44 * r['FTA'])) if (r['FGA'] + 0.44 * r['FTA']) else 0, axis=1 ) if len(selected_seasons) > 1: advanced_display_df = advanced_df.groupby('Player').mean(numeric_only=True).reset_index() advanced_cols = ['Player','PTS','REB','AST','FG_PCT','TS_PCT'] else: advanced_display_df = advanced_df.copy() advanced_cols = ['Player','Season','PTS','REB','AST','FG_PCT','TS_PCT'] st.dataframe(advanced_display_df[[c for c in advanced_cols if c in advanced_display_df.columns]].round(3), use_container_width=True) else: st.info("No data available for advanced statistics.") with tabs[2]: st.subheader("Visualizations") if not comp_df.empty: metrics = [m for m in ['PTS','REB','AST','FG_PCT','FG3_PCT','FT_PCT','STL','BLK'] if m in comp_df.columns] if metrics: selected_metric = st.selectbox("Select Metric to Visualize", metrics) if selected_metric: if len(selected_players) == 1 and len(selected_seasons) > 1: player_trend_df = comp_df[comp_df['Player'] == selected_players[0]].sort_values('Season') fig = px.line(player_trend_df, x='Season', y=selected_metric, markers=True, title=f"{selected_players[0]} – {selected_metric} Trend") else: avg_comparison_df = comp_df.groupby('Player')[metrics].mean(numeric_only=True).reset_index() fig = px.bar(avg_comparison_df, x='Player', y=selected_metric, title=f"Average {selected_metric} Comparison (Selected Seasons)") st.plotly_chart(fig, use_container_width=True) radar_metrics_for_chart = [c for c in ['PTS','REB','AST','STL','BLK'] if c in comp_df.columns] if len(radar_metrics_for_chart) >= 3: radar_source_df = (comp_df.groupby('Player')[radar_metrics_for_chart].mean(numeric_only=True).reset_index() if len(selected_seasons) > 1 else comp_df.copy()) scaled_radar_df = radar_source_df.copy() for c in radar_metrics_for_chart: mn, mx = scaled_radar_df[c].min(), scaled_radar_df[c].max() scaled_radar_df[c] = ((scaled_radar_df[c] - mn) / (mx - mn) * 100) if mx > mn else 0 radar_data = {r['Player']: {c: r[c] for c in radar_metrics_for_chart} for _, r in scaled_radar_df.iterrows()} st.plotly_chart(create_radar_chart(radar_data, radar_metrics_for_chart), use_container_width=True) else: st.info("Need ≥3 metrics for radar chart.") else: st.info("No metrics available.") else: st.info("No data available for visualizations.") def team_vs_team(): st.markdown('

Team vs Team Analysis

', unsafe_allow_html=True) # This page uses the custom BeautifulSoup scraper, so no BRSCRAPER_AVAILABLE check here. seasons = get_available_seasons() selected_season_str = st.selectbox("Select Season", seasons, index=0) year_for_team_stats = int(selected_season_str.split('–')[1]) # Use the custom BeautifulSoup scraper for team stats tm_df = get_team_stats_bs(year_for_team_stats) if tm_df.empty: st.info(f"No team data available for the {selected_season_str} season. This might be because the season hasn't started or data is not yet available, or the scraper encountered an issue.") return teams = tm_df['Team'].unique().tolist() selected_teams = st.multiselect("Select Teams (up to 4)", teams, max_selections=4) if st.button("Run Comparison"): if not selected_teams: st.warning("Please select at least one team.") return stats = [] teams_with_no_data = [] with st.spinner("Fetching team data..."): for t in selected_teams: df = tm_df[tm_df.Team == t].copy() if not df.empty: df_dict = df.iloc[0].to_dict() df_dict['Season'] = selected_season_str # Add season string for display stats.append(df_dict) else: teams_with_no_data.append(t) if teams_with_no_data: st.info(f"No data found for the selected season ({selected_season_str}) for: {', '.join(teams_with_no_data)}. This might be because the season hasn't started or data is not yet available.") if not stats: st.error("No data available for the selected teams to display. Please adjust your selections.") return comp = pd.DataFrame(stats) for col in ['PTS', 'REB', 'AST', 'STL', 'BLK', 'FG_PCT', '3P_PCT', 'FT_PCT']: if col in comp.columns: comp[col] = pd.to_numeric(comp[col], errors='coerce') st.subheader("Team Statistics Comparison") cols = ['Team', 'Season', 'PTS', 'REB', 'AST', 'STL', 'BLK', 'FG_PCT', '3P_PCT', 'FT_PCT'] display_cols = [col for col in cols if col in comp.columns] st.dataframe(comp[display_cols].round(2), use_container_width=True) st.subheader("Team Performance Visualization") metric_options = ['PTS', 'REB', 'AST', 'FG_PCT', '3P_PCT', 'FT_PCT'] available_metrics = [m for m in metric_options if m in comp.columns] if available_metrics: selected_metric = st.selectbox("Select Metric", available_metrics) fig = px.bar( comp, x='Team', y=selected_metric, color='Team', # Color by team for clarity title=f"Team {selected_metric} Comparison ({selected_season_str} Season)", barmode='group' ) st.plotly_chart(fig, use_container_width=True) else: st.info("No common metrics available for visualization.") def awards_predictor(): st.markdown('

NBA Awards Predictor

', unsafe_allow_html=True) award = st.selectbox("Select Award", ["MVP","Defensive Player of the Year","Rookie of the Year","6th Man of the Year","All-NBA First Team"]) st.subheader(f"{award} Criteria") if award=="MVP": crit = { "PPG":st.slider("Min PPG",15.0,35.0,25.0), "Wins":st.slider("Min Team Wins",35,70,50), "PER":st.slider("Min PER",15.0,35.0,25.0), "WS":st.slider("Min Win Shares",5.0,20.0,10.0) } elif award=="Defensive Player of the Year": crit = { "BPG":st.slider("Min BPG",0.0,4.0,1.5), "SPG":st.slider("Min SPG",0.0,3.0,1.0), "DefRtgMax":st.slider("Max Def Rating",90.0,120.0,105.0), "DefRankMax":st.slider("Max Team Def Rank",1,30,10) } else: crit = { "PPG":st.slider("Min PPG",10.0,30.0,15.0), "Games":st.slider("Min Games",50,82,65), "FG%":st.slider("Min FG%",0.35,0.65,0.45) } if st.button("Generate Predictions"): p = f"Predict top 5 {award} candidates based on {crit}. Focus on 2024-25 season." resp = ask_perp(p, system="You are an NBA awards expert AI.", max_tokens=800) st.markdown("### Predictions") st.write(resp) def ai_chat(): st.markdown('

AI Chat & Insights

', unsafe_allow_html=True) for msg in st.session_state.chat_history: with st.chat_message(msg["role"]): st.write(msg["content"]) if prompt := st.chat_input("Ask me anything about NBA…"): st.session_state.chat_history.append({"role":"user","content":prompt}) with st.chat_message("user"): # Display user message immediately st.write(prompt) with st.chat_message("assistant"): ans = ask_perp(prompt) st.write(ans) st.session_state.chat_history.append({"role":"assistant","content":ans}) st.subheader("Quick Actions") c1, c2, c3 = st.columns(3) if c1.button("🏆 Contenders"): resp = ask_perp("Analyze the current NBA championship contenders for 2025.") st.write(resp) if c2.button("⭐ Rising Stars"): resp = ask_perp("Who are the most promising young NBA players to watch in 2025?") st.write(resp) if c3.button("📊 Trades"): resp = ask_perp("What are some potential NBA trades this season?") st.write(resp) def young_projections(): st.markdown('

Young Player Projections

', unsafe_allow_html=True) names = get_player_index_brscraper()['name'].tolist() sel = st.selectbox("Select or enter player", [""]+names) player = sel or st.text_input("Enter player name manually") if player: age = st.number_input("Current Age",18,25,21) yrs = st.number_input("Years in NBA",0,7,2) ppg = st.number_input("PPG",0.0,40.0,15.0) rpg = st.number_input("RPG",0.0,20.0,5.0) apg = st.number_input("APG",0.0,15.0,3.0) if st.button("Generate AI Projection"): prompt = ( f"Project {player}: Age={age}, Years={yrs}, PPG={ppg}, RPG={rpg}, APG={apg}." ) out = ask_perp(prompt, system="You are an NBA projection expert AI.", max_tokens=800) st.markdown("### Projection Analysis") st.write(out) years = [f"Year {i+1}" for i in range(5)] vals = [ppg*(1+0.1*i) for i in range(5)] fig = go.Figure() fig.add_trace(go.Scatter(x=years, y=vals, mode='lines+markers')) fig.update_layout(title=f"{player} – PPG Projection", xaxis_title="Year", yaxis_title="PPG") st.plotly_chart(fig, use_container_width=True) def similar_players(): st.markdown('

Similar Players Finder

', unsafe_allow_html=True) names = get_player_index_brscraper()['name'].tolist() tp = st.selectbox("Target Player", names) crit = st.multiselect("Criteria", ["Position","Height/Weight","Playing Style","Statistical Profile","Age/Experience"], default=["Playing Style","Statistical Profile"]) if tp and crit and st.button("Find Similar"): prompt = f"Find top 5 current and top 3 historical similar to {tp} based on {', '.join(crit)}." st.write(ask_perp(prompt, system="You are a similarity expert AI.")) st.subheader("Manual Compare") p1 = st.selectbox("Player 1", names, key="p1") p2 = st.selectbox("Player 2", names, key="p2") if p1 and p2 and p1!=p2 and st.button("Compare Players"): prompt = f"Compare {p1} vs {p2} in detail on stats, style, impact, etc." st.write(ask_perp(prompt, system="You are a comparison expert AI.")) def roster_builder(): st.markdown('

NBA Roster Builder

', unsafe_allow_html=True) cap = st.number_input("Salary Cap (Millions)",100,200,136) strat = st.selectbox("Strategy",["Championship Contender","Young Core Development","Balanced Veteran Mix","Small Ball","Defense First"]) pos = st.multiselect("Priority Positions",["Point Guard","Shooting Guard","Small Forward","Power Forward","Center"],default=["Point Guard","Center"]) st.subheader("Budget Allocation") cols = st.columns(5) alloc = {} total = 0 for i,p in enumerate(["PG","SG","SF","PF","C"]): val = cols[i].number_input(f"{p} Budget ($M)",0,50,20, key=f"b{p}") alloc[p] = val total += val st.write(f"Total Allocated: ${total}M / ${cap}M") if total > cap: st.error("Budget exceeds cap!") if st.button("Generate Roster Suggestions") and total <= cap: prompt = ( f"Build roster: Cap ${cap}M, Strategy {strat}, Positions {pos}, Budgets {alloc}." ) st.markdown("### AI Roster Recommendations") st.write(ask_perp(prompt, system="You are an NBA roster building expert AI.")) def trade_analyzer(): st.markdown('

Trade Scenario Analyzer

', unsafe_allow_html=True) t1 = st.text_input("Team 1 trades") t2 = st.text_input("Team 2 trades") if t1 and t2 and st.button("Analyze Trade"): prompt = f"Analyze trade: Team1 {t1}, Team2 {t2}." st.write(ask_perp(prompt, system="You are a trade analysis AI.")) if __name__ == "__main__": main()