import math
import os
import re
import sys
import warnings

from sqlalchemy import create_engine
from sqlalchemy.sql import text
import colorama
import pymysql


# Format: mysql+pymysql://<user>:<password>@<ip>/<db>
ENGINE_DSN = os.environ.get('ENGINE_DSN')

QUERY_INDEX_TABLE_INFORMATION = (
    'SELECT table_name, column_name, column_type '
    'FROM INFORMATION_SCHEMA.COLUMNS '
    'WHERE extra like "%auto_increment%"')

QUERY_COUNT_INDEX_ENTRIES = 'SELECT count({field}) from `{table}`'

MAP_COLUMN_TYPES_SIZE = {
    'bigint': 64,
    'int': 32,
    'mediumint': 24,
    'smallint': 16,
    'tinyint': 8,
}

engine = create_engine(ENGINE_DSN, echo=False)
connection = engine.connect()

# Ignore warnings from pymysql.err.Warning. This is because some of our comment
# fields have invalid utf8 values.
warnings.filterwarnings('ignore', category=pymysql.err.Warning)


def analyse_database():
    """Analyse a database.

    The systems runs a query that lists all indexes for each table that have
    an autoincrement field, and analyse the index.

    Yields:
        dict: represents the analysis of an index (see analyse_table_index)
            for more details.
    """

    tables = connection.execute(text(QUERY_INDEX_TABLE_INFORMATION))
    for table in tables:
        table_name, column_name, column_format = table
        yield analyse_table_index(table_name, column_name, column_format)


def analyse_table_index(table_name, column_name, column_format):
    """Analyse a table index.

    Args:
        table_name (str): Name of the table.
        column_name (str): Name of the column.
        column_format (str): Format of the column (usually "int(12) unsigned").

    Returns:
        dict: the dictionary contains the different information (
            percentage used, current number of items, the maximum number of
            items the table acn store, the name of the column and its format.)
    """

    query = QUERY_COUNT_INDEX_ENTRIES.format(
        table=table_name, field=column_name)
    query = connection.execute(text(query))
    total, = query.first()

    column_information = column_format.split(' ')
    sign = -1
    if len(column_information) == 2 and column_information[1] == 'unsigned':
        sign = 0

    column_information = re.match(r"(.+)\((.+)\)", column_information[0])
    column_type = column_information.group(1)

    max_total = int(math.pow(2, MAP_COLUMN_TYPES_SIZE[column_type] - sign))
    percentage = math.ceil(100 * total / max_total)
    return {
        'percentage': percentage,
        'total': total,
        'max_total': max_total,
        'table_name': table_name,
        'column_name': column_name,
        'column_format': column_format
    }


def csv_printer(analysis):
    """CSV Printer.

    Args:
        analysis (dict): the analysis.
    """

    print(
        '{table_name},{column_name},{percentage},{total},{max_total}'.format(
            **analysis))


def pretty_printer(analysis):
    """Pretty print an analysis.

    Args:
        analysis (dict): the analysis.
    """

    table_color = colorama.Fore.BLUE
    if analysis.get('percentage') > 80:
        table_color = colorama.Fore.RED
    print(
        table_color + analysis.get('table_name') +
        colorama.Style.RESET_ALL + ': ' +
        colorama.Fore.YELLOW + analysis.get('column_name'),
        colorama.Style.RESET_ALL)

    print('    Percentage:', analysis.get('percentage'))
    print('    Total:', analysis.get('total'))
    print('    Max entries:', analysis.get('max_total'))
    print()


if __name__ == '__main__':
    colorama.init()

    is_csv = '--csv' in sys.argv
    if is_csv:
        print('table name,column name,percentage,total,max total')

    for analysis in analyse_database():
        if is_csv:
            csv_printer(analysis)
            continue

        pretty_printer(analysis)

    connection.close()