import pandas as pd import streamlit as st # Set page title and favicon st.set_page_config(page_icon=":soccer:",layout="wide") st.markdown( """ """, unsafe_allow_html=True ) # Set title and create a new tab for league history st.title("⚽ SoccerTwos Challenge Analytics Extra!⚽ ") tab_team, tab_owners = st.tabs(["Form Table", "Games by Author",]) # Match Results MATCH_RESULTS_URL = "https://huggingface.co/datasets/huggingface-projects/bot-fight-data/raw/main/soccer_history.csv" @st.cache(ttl=1800) def fetch_match_history(): """ Fetch the match results from the last 24 hours. Cache the result for 30min to avoid unnecessary requests. Return a DataFrame. """ df = pd.read_csv(MATCH_RESULTS_URL) df["timestamp"] = pd.to_datetime(df.timestamp, unit="s") df = df[df["timestamp"] >= pd.Timestamp.now() - pd.Timedelta(hours=24)] df.columns = ["home", "away", "timestamp", "result"] return df match_df = fetch_match_history() # Define a function to calculate the total number of matches played def num_matches_played(): return match_df.shape[0] # Get a list of all teams that have played in the last 24 hours teams = sorted( list(pd.concat([match_df["home"], match_df["away"]]).unique()), key=str.casefold ) # Create the form table, which shows the win percentage for each team # st.header("Form Table") team_results = {} for i, row in match_df.iterrows(): home_team = row["home"] away_team = row["away"] result = row["result"] if home_team not in team_results: team_results[home_team] = [0, 0, 0] if away_team not in team_results: team_results[away_team] = [0, 0, 0] if result == 0: team_results[home_team][2] += 1 team_results[away_team][0] += 1 elif result == 1: team_results[home_team][0] += 1 team_results[away_team][2] += 1 else: team_results[home_team][1] += 1 team_results[away_team][1] += 1 # Create a DataFrame from the results dictionary and calculate the win percentage df = pd.DataFrame.from_dict( team_results, orient="index", columns=["wins", "draws", "losses"] ).sort_index() df[["owner", "team"]] = df.index.to_series().str.split("/", expand=True) df = df[["owner", "team", "wins", "draws", "losses"]] df["win_pct"] = (df["wins"] / (df["wins"] + df["draws"] + df["losses"])) * 100 # Get a list of all teams that have played in the last 24 hours @st.cache(ttl=1800) def fetch_owners(): """ Fetch a list of all owners who have played in the matches, along with the number of teams they own and the number of unique teams they played with. """ # Extract the owner name and team name from each home and away team name in the DataFrame team_owners = match_df["home"].apply(lambda x: x.split('/')[0]).tolist() + match_df['away'].apply(lambda x: x.split('/')[0]).tolist() teams = match_df["home"].apply(lambda x: x.split('/')[1]).tolist() + match_df['away'].apply(lambda x: x.split('/')[1]).tolist() # Count the number of games played by each owner and the number of unique teams they played with owner_team_counts = {} owner_team_set = {} for i, team_owner in enumerate(team_owners): owner = team_owner.split(' ')[0] if owner not in owner_team_counts: owner_team_counts[owner] = 1 owner_team_set[owner] = {teams[i]} else: owner_team_counts[owner] += 1 owner_team_set[owner].add(teams[i]) # Create a DataFrame from the dictionary owners_df = pd.DataFrame.from_dict(owner_team_counts, orient="index", columns=["Games played by owner"]) owners_df["Unique teams by owner"] = owners_df.index.map(lambda x: len(owner_team_set[x])) # Return the DataFrame return owners_df # Display the DataFrame as a table, sorted by win percentage with tab_team: st.write("Form Table for previous 24 hours, ranked by win percentage") stats = df.sort_values(by="win_pct", ascending=False) styled_stats = stats.style.set_table_attributes("style='font-size: 20px'").set_table_styles([dict(selector='th', props=[('max-width', '200px')])]) styled_stats = styled_stats.set_table_attributes("style='max-height: 1200px; overflow: auto'") st.dataframe(styled_stats) # Create a DataFrame from the list of owners and their number of teams owners_df = fetch_owners() # Display the DataFrame as a table with tab_owners: st.dataframe(owners_df)