nb_tool.py

from langchain.tools import tool
import requests
from bs4 import BeautifulSoup
from typing import Optional
import re
import sys
from constants import TAVILY_KEY
import os
from tavily import TavilyClient

# Initialize Tavily client (set your API key as an environment variable)
tavily = TavilyClient(TAVILY_KEY)

@tool
def get_npb_player_info(player_name: str) -> str:
    """
    Finds a Nippon Professional Baseball (NPB) player's jersey number and team name using Tavily search.

    Parameters:
    - player_name (str): Full name of the player (e.g., "Taisho Tamai")

    Returns:
    - str: The player's jersey number and team name, or an error message if not found.
    """
    return get_npb_player_info_(player_name)

def get_npb_player_info_(player_name: str) -> str:
    try:
        # Search using Tavily for the NPB English profile page
        query = f"site:npb.jp/bis/eng/players {player_name}"
        search_results = tavily.search(query=query, include_answer=False)

        # Find first relevant link to the NPB player profile
        player_url = next(
            (result["url"] for result in search_results["results"]
            if re.match(r"https://npb\.jp/bis/eng/players/\d+\.html", result["url"])),
            None
        )

        if not player_url:
            return f"No player page found for '{player_name}'."

        # Fetch the player page and extract number and team
        headers = {"User-Agent": "Mozilla/5.0"}
        response = requests.get(player_url, headers=headers)
        soup = BeautifulSoup(response.text, "html.parser")

        number = soup.find("li", {"id": "pc_v_no"})
        team = soup.find("li", {"id": "pc_v_team"})

        if number and team:
            return f"{player_name} - #{number.text.strip()} - Team: {team.text.strip()}"
        else:
            return f"Player page found, but info not parsed properly: {player_url}"

    except Exception as e:
        return f"Error fetching info for '{player_name}': {e}"

#if __name__ == "__main__":
#    print(get_npb_player_info("Taisho Tamai"))

@tool
def get_team_roster(team_name: str, year: int) -> list[dict]:
    """
    Returns the team roster for a given NPB team and season.
    
    Each player is represented as a dict with:
    - name: "Last, First"
    - number: jersey number (str or None if not found)
    - role (pitcher, etc.)

    Parameters:
    - team_name (str): e.g., "Yomiuri Giants"
    - year (int): e.g., 2023

    Returns:
    - List of dicts representing players.
    """
    roster = []
    player_names = get_team_players_by_season(team_name, year)

    for player_name in player_names:
        # Get player info
        info = get_npb_player_info_(player_name)
        
        # Try to extract number from the response
        match = re.search(r"#(\d+)", info)
        number = match.group(1) if match else None

        roster.append({
            "name": player_name,
            "number": number,
            "role": "Pitcher"  # Static for now
        })

    return roster



def get_team_players_by_season(team_name: str, year: int) -> list[str]:
    """
    Given a team name and year, returns the list of player names for that NPB season.

    Parameters:
    - team_name (str): Team name (e.g., "Yomiuri Giants")
    - year (int): Season year (e.g., 2023)

    Returns:
    - List of player names (strings) in the format last name, first name
    """
    try:
        query = f"site:npb.jp/bis/eng/{year}/stats {team_name}"
        search_results = tavily.search(query=query, include_answer=False)

        # Find the correct stats page URL for that year
        stats_url = next(
            (result["url"] for result in search_results["results"]
             if re.match(rf"https://npb\.jp/bis/eng/{year}/stats/.+\.html", result["url"])),
            None
        )

        if not stats_url:
            print(f"No stats page found for {team_name} in {year}.")
            return []
        
        print(stats_url)

        headers = {"User-Agent": "Mozilla/5.0"}
        response = requests.get(stats_url, headers=headers)
        safe_text = response.text.encode(sys.stdout.encoding, errors='replace').decode(sys.stdout.encoding)
        

        response = requests.get(stats_url, headers=headers)

        soup = BeautifulSoup(response.text, "html.parser")

        # Find all <td class="stplayer"> elements
        
        player_tds = soup.find_all("td", class_="stplayer")

        player_names = []
        for td in player_tds:
            player_name = td.get_text(strip=True)
            player_names.append(player_name)

        # Remove duplicates preserving order
        player_names = list(dict.fromkeys(player_names))

        return player_names

    except Exception as e:
        print(f"Error fetching players for {team_name} {year}: {e}")
        return []

if __name__ == "__main__":
    team = "Hokkaido Nippon-Ham Fighters"
    season_year = 2023
    players = get_team_players_by_season(team, season_year)
    print(f"Players for {team} in {season_year} season:\n", players)