import pandas as pd

class Comparison:

    @staticmethod
    def normalize_value(val):
        """Normalize non-numeric values (date, string)."""
        if pd.isna(val):
            return ""

        # Normalize dates
        if isinstance(val, pd.Timestamp):
            return val.date().isoformat()

        # Normalize strings
        return str(val).strip().upper().replace(" ", "")

    @staticmethod
    def round_amount(val, decimals=0):
        """Round numeric values safely for comparison."""
        if pd.isna(val):
            return ""
        try:
            return round(float(val), decimals)
        except Exception:
            return val

    @staticmethod
    def find_unmatched_rows(sheet1_df, sheet2_df):
        # Clean column names
        sheet1_df.columns = sheet1_df.columns.str.strip()
        sheet2_df.columns = sheet2_df.columns.str.strip()

        comparison_columns = [
            'Date',
            'GSTIN/UIN',
            'Voucher Ref. No.',
            'Total Tax',
            'Total Amount'
        ]

        amount_columns = ['Total Tax', 'Total Amount']

        # Validate required columns
        for col in comparison_columns:
            if col not in sheet1_df.columns or col not in sheet2_df.columns:
                raise ValueError(f"Missing column '{col}' in one of the sheets.")

        # Normalize values
        for col in comparison_columns:
            if col in amount_columns:
                sheet1_df[col] = sheet1_df[col].apply(
                    lambda x: Comparison.round_amount(x, 0)
                )
                sheet2_df[col] = sheet2_df[col].apply(
                    lambda x: Comparison.round_amount(x, 0)
                )
            else:
                sheet1_df[col] = sheet1_df[col].apply(Comparison.normalize_value)
                sheet2_df[col] = sheet2_df[col].apply(Comparison.normalize_value)

        # Create comparison keys
        sheet1_keys = sheet1_df[comparison_columns].apply(tuple, axis=1)
        sheet2_keys = sheet2_df[comparison_columns].apply(tuple, axis=1)

        # Find unmatched rows
        unmatched_sheet1 = sheet1_df[~sheet1_keys.isin(sheet2_keys)].copy()
        unmatched_sheet2 = sheet2_df[~sheet2_keys.isin(sheet1_keys)].copy()

        # Add source column
        unmatched_sheet1["Source"] = "Portal"
        unmatched_sheet2["Source"] = "Tally"

        return unmatched_sheet1, unmatched_sheet2