import logging
import duckdb
import pandas as pd
import polars as pl
import pyarrow as pa
import re
from coded_flows.types import Union, List, DataFrame, ArrowTable, Str

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)


def _coalesce(*values):
    return next((v for v in values if v is not None), None)


def _sanitize_identifier(identifier):
    """
    Sanitize SQL identifier by escaping special characters.
    Handles double quotes and other problematic characters.
    """
    return identifier.replace('"', '""')


def _is_boolean_type(duckdb_type):
    """
    Check if a DuckDB type is boolean.
    """
    type_lower = duckdb_type.lower()
    return "bool" in type_lower


def negate_boolean(
    data: Union[DataFrame, ArrowTable],
    columns: List = None,
    regex_pattern: Str = None,
    options=None,
) -> Union[DataFrame, ArrowTable]:
    brick_display_name = "Negate Boolean"
    options = options or {}
    verbose = options.get("verbose", True)
    columns = _coalesce(columns, options.get("columns", []))
    regex_pattern = _coalesce(regex_pattern, options.get("regex_pattern", ""))
    fill_all_columns = len(columns) == 0
    output_format = options.get("output_format", "pandas")
    safe_mode = options.get("safe_mode", False)
    result = None
    conn = None
    if not isinstance(columns, list) or not all((isinstance(c, str) for c in columns)):
        verbose and logger.error(
            f"[{brick_display_name}] Invalid columns format! Expected a list."
        )
        raise ValueError("Columns must be provided as a list!")
    try:
        negate_mode = None
        if regex_pattern:
            negate_mode = "regex_pattern"
        elif fill_all_columns:
            negate_mode = "all_columns"
        else:
            negate_mode = "column_list"
        verbose and logger.info(
            f"[{brick_display_name}] Detected mode: '{negate_mode}'. Starting boolean negation operation."
        )
        data_type = None
        if isinstance(data, pd.DataFrame):
            data_type = "pandas"
        elif isinstance(data, pl.DataFrame):
            data_type = "polars"
        elif isinstance(data, (pa.Table, pa.lib.Table)):
            data_type = "arrow"
        if data_type is None:
            verbose and logger.error(
                f"[{brick_display_name}] Input data must be a pandas DataFrame, Polars DataFrame, or Arrow Table"
            )
            raise ValueError(
                "Input data must be a pandas DataFrame, Polars DataFrame, or Arrow Table"
            )
        verbose and logger.info(
            f"[{brick_display_name}] Detected input format: {data_type}."
        )
        conn = duckdb.connect(":memory:")
        conn.register("input_table", data)
        column_info = conn.execute("DESCRIBE input_table").fetchall()
        all_columns = {col[0]: col[1] for col in column_info}
        verbose and logger.info(
            f"[{brick_display_name}] Total columns in data: {len(all_columns)}."
        )
        columns_to_negate = []
        if negate_mode == "column_list":
            if not safe_mode:
                missing_columns = [col for col in columns if col not in all_columns]
                if missing_columns:
                    verbose and logger.error(
                        f"[{brick_display_name}] Columns not found in data: {missing_columns}"
                    )
                    raise ValueError(f"Columns not found in data: {missing_columns}")
            columns_to_negate = [col for col in columns if col in all_columns]
            skipped = len(columns) - len(columns_to_negate)
            if safe_mode and skipped > 0:
                skipped_cols = [col for col in columns if col not in all_columns]
                verbose and logger.warning(
                    f"[{brick_display_name}] Safe mode: Skipped {skipped} non-existent columns: {skipped_cols}"
                )
            verbose and logger.info(
                f"[{brick_display_name}] Negating {len(columns_to_negate)} column(s): {columns_to_negate}."
            )
        elif negate_mode == "regex_pattern":
            try:
                pattern = re.compile(regex_pattern)
                columns_to_negate = [
                    col for col in all_columns.keys() if pattern.search(col)
                ]
                if not columns_to_negate:
                    verbose and logger.warning(
                        f"[{brick_display_name}] No columns matched regex pattern '{regex_pattern}'. Returning data unchanged."
                    )
                verbose and logger.info(
                    f"[{brick_display_name}] Regex pattern '{regex_pattern}' matched {len(columns_to_negate)} columns: {columns_to_negate}."
                )
            except re.error as e:
                verbose and logger.error(
                    f"[{brick_display_name}] Invalid regex pattern."
                )
                raise ValueError(f"Invalid regex pattern!")
        elif negate_mode == "all_columns":
            columns_to_negate = [
                col for col in all_columns.keys() if _is_boolean_type(all_columns[col])
            ]
            verbose and logger.info(
                f"[{brick_display_name}] Negating all boolean columns: {len(columns_to_negate)} columns."
            )
        select_parts = []
        negated_count = 0
        for col in all_columns.keys():
            sanitized_col = _sanitize_identifier(col)
            if col in columns_to_negate:
                col_type = all_columns[col]
                if not _is_boolean_type(col_type):
                    if safe_mode:
                        verbose and logger.warning(
                            f"[{brick_display_name}] Safe mode: Skipping non-boolean column '{col}' (type: {col_type})."
                        )
                        select_parts.append(f'"{sanitized_col}"')
                        continue
                    else:
                        verbose and logger.error(
                            f"[{brick_display_name}] Column '{col}' is not boolean (type: {col_type})."
                        )
                        raise ValueError(
                            f"Column '{col}' is not boolean (type: {col_type}). Cannot apply negation."
                        )
                try:
                    negation_expr = f'NOT "{sanitized_col}"'
                    select_parts.append(f'{negation_expr} AS "{sanitized_col}"')
                    verbose and logger.info(
                        f"[{brick_display_name}] Column '{col}' (type: {col_type}): applying negation (NOT)."
                    )
                    negated_count += 1
                except Exception as e:
                    if safe_mode:
                        verbose and logger.warning(
                            f"[{brick_display_name}] Safe mode: Skipping column '{col}' due to error: {str(e)}"
                        )
                        select_parts.append(f'"{sanitized_col}"')
                    else:
                        verbose and logger.error(
                            f"[{brick_display_name}] Error processing column '{col}'."
                        )
                        raise
            else:
                select_parts.append(f'"{sanitized_col}"')
        if negated_count == 0:
            verbose and logger.warning(
                f"[{brick_display_name}] No columns were negated. Returning data unchanged."
            )
            result = data
        else:
            select_clause = ", ".join(select_parts)
            query = f"SELECT {select_clause} FROM input_table"
            verbose and logger.info(
                f"[{brick_display_name}] Executing query to negate boolean values."
            )
            if output_format == "pandas":
                result = conn.execute(query).df()
                verbose and logger.info(
                    f"[{brick_display_name}] Converted result to pandas DataFrame."
                )
            elif output_format == "polars":
                result = conn.execute(query).pl()
                verbose and logger.info(
                    f"[{brick_display_name}] Converted result to Polars DataFrame."
                )
            elif output_format == "arrow":
                result = conn.execute(query).fetch_arrow_table()
                verbose and logger.info(
                    f"[{brick_display_name}] Converted result to Arrow Table."
                )
            else:
                verbose and logger.error(
                    f"[{brick_display_name}] Unsupported output format: {output_format}"
                )
                raise ValueError(f"Unsupported output format: {output_format}")
            verbose and logger.info(
                f"[{brick_display_name}] Boolean negation operation completed successfully. Negated {negated_count} column(s)."
            )
    except Exception as e:
        verbose and logger.error(
            f"[{brick_display_name}] Error during boolean negation operation: {str(e)}"
        )
        raise
    finally:
        if conn is not None:
            conn.close()
    return result