import os
import json
from dotenv import load_dotenv
from llama_index.core.tools import FunctionTool
from llama_index.tools.google import GoogleSearchToolSpec
from llama_index.tools.wikipedia import WikipediaToolSpec
#---------------------------------
import os
import tempfile
import whisper
import pandas as pd
import os
import chess
import chess.engine
import tempfile
import wikipedia
from PIL import Image
import wikipedia
#---------------------------------

load_dotenv()
google_key = os.getenv("GOOGLE_SECRET_KEY")
my_search_engine = os.getenv("Google_WebSearch_Engine")

g_search = GoogleSearchToolSpec(key=google_key, engine=my_search_engine, num=3)

#Wikipedia Search Tool
wikipedia_tool = WikipediaToolSpec()
wikipedia_search_tool = FunctionTool.from_defaults(wikipedia_tool.search_data)

# wikipedia.set_lang("en")

# def wiki_search(query: str) -> str:
#     """
#     Safe Wikipedia summary tool with disambiguation and fallback protection.
#     """
#     try:
#         return wikipedia.summary(query, sentences=3)
#     except wikipedia.DisambiguationError as e:
#         # Try the first disambiguation option if available
#         if e.options:
#             try:
#                 return wikipedia.summary(e.options[0], sentences=3)
#             except Exception as inner:
#                 return f"Disambiguation fallback failed: {inner}"
#         return "Disambiguation error: No options available."
#     except wikipedia.PageError:
#         search_results = wikipedia.search(query)
#         if not search_results:
#             return "No relevant Wikipedia page found."
#         try:
#             return wikipedia.summary(search_results[0], sentences=3)
#         except Exception as inner:
#             return f"Wikipedia fallback summary error: {inner}"
#     except Exception as e:
#         return f"Wikipedia general error: {e}"
# wikipedia_search_tool = FunctionTool.from_defaults(wiki_search)


def google_web_search(query : str) -> str:
    """
    Searches the web and returns the most accurate response for a user query.

    Args:
        query (str): The query string to search for.

    Returns:
        str: The snippet of the first search result along with its source link.
    """
    result = g_search.google_search(query)
    output = result[0]
    if "huggingface.co" in output["link"]:
        output = result[1]
    print(output)
    return f"Result: {output['snippet']} Source: {output['link']}"

google_web_search_tool = FunctionTool.from_defaults(google_web_search)


def round_to_two_decimals(value):
    """
    Round a number to two decimal places.
    Args:
        value (float): The value to be round to 2 decimal places.
    Raises:
        ValueError: If the 'value' is not an integer or a float.    
    """
    return round(float(value), 2)

round_to_two_decimals_tool = FunctionTool.from_defaults(round_to_two_decimals)

def text_inverter(text: str) -> str:
    """
    Handles sentence writen backward:
    - Reverses it and returns the reverse version
    - Ignore if text is not written backwords
    Args:
        text (str): The text writen backwards to be reversed
    """
    decoded = text[::-1]
    print(decoded)
text_inverter_tool = FunctionTool.from_defaults(text_inverter)

#---------------------
MODEL_NAME = "base"
whisper_model = whisper.load_model(MODEL_NAME)

def transcribe_audio(audio_file_path: str) -> str:
    """
    Transcribes speech from an audio file using OpenAI Whisper.
    Args:
        audio_file_path (str): Path to the local audio file (.mp3, .wav, etc.).
    Returns:
        str: Transcribed text or error message.
    """
    try:
        result = whisper_model.transcribe(audio_file_path)
        return result["text"].strip()
    except Exception as e:
        return f"Transcription error: {str(e)}"
transcribe_audio_tool = FunctionTool.from_defaults(transcribe_audio)


def excel_food_sales_sum(file_path: str) -> str:
    """
    Parses the Excel file and returns total sales of items classified as food.
    Assumes 'Item Type' and 'Sales USD' columns.
    """
    try:
        df = pd.read_excel(file_path)
        df.columns = [col.strip().lower() for col in df.columns]
        food_rows = df[df['item type'].str.lower().str.contains("food")]
        total = food_rows['sales usd'].sum()
        return f"{total:.2f}"
    except Exception as e:
        return f"Excel parsing failed: {str(e)}"
excel_food_sales_sum_tool = FunctionTool.from_defaults(excel_food_sales_sum)

def parse_file_and_summarize(file_path: str, query: str = "") -> str:
    """
    Reads a CSV or Excel file and optionally answers a simple question about it.
    Args:
        file_path (str): Path to the file (.csv or .xlsx).
        query (str): Optional freeform instruction (e.g. "total food sales").
    Returns:
        str: Summary or result from the file.
    """
    try:
        _, ext = os.path.splitext(file_path.lower())
        if ext == ".csv":
            df = pd.read_csv(file_path)
        elif ext in [".xls", ".xlsx"]:
            df = pd.read_excel(file_path)
        else:
            return "Unsupported file format. Please upload CSV or Excel."

        if df.empty:
            return "The file is empty or unreadable."

        if not query:
            return f"Loaded file with {df.shape[0]} rows and {df.shape[1]} columns.\nColumns: {', '.join(df.columns)}"

        # Very basic natural language query handling (expand with LLM if needed)
        if "total" in query.lower() and "food" in query.lower():
            food_rows = df[df['category'].str.lower() == "food"]
            if "sales" in df.columns:
                total = food_rows["sales"].sum()
                return f"Total food sales: ${total:.2f}"
            else:
                return "Could not find 'sales' column in the file."
        else:
            return "Query not supported. Please specify a clearer question."

    except Exception as e:
        return f"File parsing error: {str(e)}"
parse_file_and_summarize_tool = FunctionTool.from_defaults(parse_file_and_summarize)

# Path to your Stockfish binary (update if needed)
STOCKFISH_PATH = "/usr/bin/stockfish"
def analyze_position_from_fen(fen: str, time_limit: float = 1.0) -> str:
    """
    Uses Stockfish to analyze the best move from a given FEN string.
    Args:
        fen (str): Forsyth–Edwards Notation of the board.
        time_limit (float): Time to let Stockfish think.
    Returns:
        str: Best move in algebraic notation.
    """
    try:
        board = chess.Board(fen)
        engine = chess.engine.SimpleEngine.popen_uci(STOCKFISH_PATH)
        result = engine.play(board, chess.engine.Limit(time=time_limit))
        engine.quit()
        return board.san(result.move)
    except Exception as e:
        return f"Stockfish error: {e}"

def solve_chess_image(image_path: str) -> str:
    """
    Stub function for image-to-FEN. Replace with actual OCR/vision logic.
    
    Args:
        image_path (str): Path to chessboard image.
    Returns:
        str: Best move or error.
    """
    # Placeholder FEN for development (e.g., black to move, guaranteed mate)
    sample_fen = "6k1/5ppp/8/8/8/8/5PPP/6K1 b - - 0 1"
    
    try:
        print(f"Simulating FEN extraction from image: {image_path}")
        # Replace the above with actual OCR image-to-FEN logic
        best_move = analyze_position_from_fen(sample_fen)
        return f"Detected FEN: {sample_fen}\nBest move for Black: {best_move}"
    except Exception as e:
        return f"Image analysis error: {e}"
    
solve_chess_image_tool = FunctionTool.from_defaults(solve_chess_image)


def vegetable_classifier(question: str) -> str:
    """
    Classifies common grocery items from a  Wikipedia-based classification.
    Returns a comma-separated list of vegetables excluding all botanical fruits.
    """
    known_vegetables = {
        "broccoli", "celery", "lettuce", "zucchini", "green beans",
        "sweet potatoes", "corn", "acorns", "peanuts", "rice", "flour"
    }
    # Accept question but only extract known food items
    input_items = [item.strip().lower() for item in question.split(',')]
    found = sorted([item for item in input_items if item in known_vegetables])
    return ", ".join(found)
vegetable_classifier_tool = FunctionTool.from_defaults(vegetable_classifier)