django-tidb/django_tidb/tidb/base.py

"""
MySQL database backend for Django.

Requires mysqlclient: https://pypi.python.org/pypi/mysqlclient/
MySQLdb is supported for Python 2 only: http://sourceforge.net/projects/mysql-python
"""
from __future__ import unicode_literals

import datetime
import re
import sys
import warnings

from django.conf import settings
from django.db import utils
from django.db.backends import utils as backend_utils
from django.db.backends.base.base import BaseDatabaseWrapper
from django.utils import six, timezone
from django.utils.encoding import force_str
from django.utils.functional import cached_property
from django.utils.safestring import SafeBytes, SafeText

try:
    import MySQLdb as Database
except ImportError as e:
    from django.core.exceptions import ImproperlyConfigured
    raise ImproperlyConfigured("Error loading MySQLdb module: %s" % e)

from MySQLdb.constants import CLIENT, FIELD_TYPE                # isort:skip
from MySQLdb.converters import Thing2Literal, conversions       # isort:skip

# Some of these import MySQLdb, so import them after checking if it's installed.
from .client import DatabaseClient                          # isort:skip
from .creation import DatabaseCreation                      # isort:skip
from .features import DatabaseFeatures                      # isort:skip
from .introspection import DatabaseIntrospection            # isort:skip
from .operations import DatabaseOperations                  # isort:skip
from .schema import DatabaseSchemaEditor                    # isort:skip
from .validation import DatabaseValidation                  # isort:skip

# We want version (1, 2, 1, 'final', 2) or later. We can't just use
# lexicographic ordering in this check because then (1, 2, 1, 'gamma')
# inadvertently passes the version test.
version = Database.version_info
if (version < (1, 2, 1) or (version[:3] == (1, 2, 1) and
        (len(version) < 5 or version[3] != 'final' or version[4] < 2))):
    from django.core.exceptions import ImproperlyConfigured
    raise ImproperlyConfigured("MySQLdb-1.2.1p2 or newer is required; you have %s" % Database.__version__)


DatabaseError = Database.DatabaseError
IntegrityError = Database.IntegrityError

# It's impossible to import datetime_or_None directly from MySQLdb.times
parse_datetime = conversions[FIELD_TYPE.DATETIME]


def parse_datetime_with_timezone_support(value):
    dt = parse_datetime(value)
    # Confirm that dt is naive before overwriting its tzinfo.
    if dt is not None and settings.USE_TZ and timezone.is_naive(dt):
        dt = dt.replace(tzinfo=timezone.utc)
    return dt


def adapt_datetime_with_timezone_support(value, conv):
    # Equivalent to DateTimeField.get_db_prep_value. Used only by raw SQL.
    if settings.USE_TZ:
        if timezone.is_naive(value):
            warnings.warn("MySQL received a naive datetime (%s)"
                          " while time zone support is active." % value,
                          RuntimeWarning)
            default_timezone = timezone.get_default_timezone()
            value = timezone.make_aware(value, default_timezone)
        value = value.astimezone(timezone.utc).replace(tzinfo=None)
    return Thing2Literal(value.strftime("%Y-%m-%d %H:%M:%S.%f"), conv)

# MySQLdb-1.2.1 returns TIME columns as timedelta -- they are more like
# timedelta in terms of actual behavior as they are signed and include days --
# and Django expects time, so we still need to override that. We also need to
# add special handling for SafeText and SafeBytes as MySQLdb's type
# checking is too tight to catch those (see Django ticket #6052).
# Finally, MySQLdb always returns naive datetime objects. However, when
# timezone support is active, Django expects timezone-aware datetime objects.
django_conversions = conversions.copy()
django_conversions.update({
    FIELD_TYPE.TIME: backend_utils.typecast_time,
    FIELD_TYPE.DECIMAL: backend_utils.typecast_decimal,
    FIELD_TYPE.NEWDECIMAL: backend_utils.typecast_decimal,
    FIELD_TYPE.DATETIME: parse_datetime_with_timezone_support,
    datetime.datetime: adapt_datetime_with_timezone_support,
})

# This should match the numerical portion of the version numbers (we can treat
# versions like 5.0.24 and 5.0.24a as the same). Based on the list of version
# at http://dev.mysql.com/doc/refman/4.1/en/news.html and
# http://dev.mysql.com/doc/refman/5.0/en/news.html .
server_version_re = re.compile(r'(\d{1,2})\.(\d{1,2})\.(\d{1,2})')


# MySQLdb-1.2.1 and newer automatically makes use of SHOW WARNINGS on
# MySQL-4.1 and newer, so the MysqlDebugWrapper is unnecessary. Since the
# point is to raise Warnings as exceptions, this can be done with the Python
# warning module, and this is setup when the connection is created, and the
# standard backend_utils.CursorDebugWrapper can be used. Also, using sql_mode
# TRADITIONAL will automatically cause most warnings to be treated as errors.

class CursorWrapper(object):
    """
    A thin wrapper around MySQLdb's normal cursor class so that we can catch
    particular exception instances and reraise them with the right types.

    Implemented as a wrapper, rather than a subclass, so that we aren't stuck
    to the particular underlying representation returned by Connection.cursor().
    """
    codes_for_integrityerror = (1048,)

    def __init__(self, cursor):
        self.cursor = cursor

    def execute(self, query, args=None):
        try:
            # args is None means no string interpolation
            return self.cursor.execute(query, args)
        except Database.OperationalError as e:
            # Map some error codes to IntegrityError, since they seem to be
            # misclassified and Django would prefer the more logical place.
            if e.args[0] in self.codes_for_integrityerror:
                six.reraise(utils.IntegrityError, utils.IntegrityError(*tuple(e.args)), sys.exc_info()[2])
            raise

    def executemany(self, query, args):
        try:
            return self.cursor.executemany(query, args)
        except Database.OperationalError as e:
            # Map some error codes to IntegrityError, since they seem to be
            # misclassified and Django would prefer the more logical place.
            if e.args[0] in self.codes_for_integrityerror:
                six.reraise(utils.IntegrityError, utils.IntegrityError(*tuple(e.args)), sys.exc_info()[2])
            raise

    def __getattr__(self, attr):
        if attr in self.__dict__:
            return self.__dict__[attr]
        else:
            return getattr(self.cursor, attr)

    def __iter__(self):
        return iter(self.cursor)

    def __enter__(self):
        return self

    def __exit__(self, type, value, traceback):
        # Ticket #17671 - Close instead of passing thru to avoid backend
        # specific behavior.
        self.close()


class DatabaseWrapper(BaseDatabaseWrapper):
    vendor = 'mysql'
    # This dictionary maps Field objects to their associated MySQL column
    # types, as strings. Column-type strings can contain format strings; they'll
    # be interpolated against the values of Field.__dict__ before being output.
    # If a column type is set to None, it won't be included in the output.
    _data_types = {
        'AutoField': 'integer AUTO_INCREMENT',
        'BinaryField': 'longblob',
        'BooleanField': 'bool',
        'CharField': 'varchar(%(max_length)s)',
        'CommaSeparatedIntegerField': 'varchar(%(max_length)s)',
        'DateField': 'date',
        'DateTimeField': 'datetime',
        'DecimalField': 'numeric(%(max_digits)s, %(decimal_places)s)',
        'DurationField': 'bigint',
        'FileField': 'varchar(%(max_length)s)',
        'FilePathField': 'varchar(%(max_length)s)',
        'FloatField': 'double precision',
        'IntegerField': 'integer',
        'BigIntegerField': 'bigint',
        'IPAddressField': 'char(15)',
        'GenericIPAddressField': 'char(39)',
        'NullBooleanField': 'bool',
        'OneToOneField': 'integer',
        'PositiveIntegerField': 'integer UNSIGNED',
        'PositiveSmallIntegerField': 'smallint UNSIGNED',
        'SlugField': 'varchar(%(max_length)s)',
        'SmallIntegerField': 'smallint',
        'TextField': 'longtext',
        'TimeField': 'time',
        'UUIDField': 'char(32)',
    }

    @cached_property
    def data_types(self):
        if self.features.supports_microsecond_precision:
            return dict(self._data_types, DateTimeField='datetime(6)', TimeField='time(6)')
        else:
            return self._data_types

    operators = {
        'exact': '= %s',
        'iexact': 'LIKE %s',
        'contains': 'LIKE BINARY %s',
        'icontains': 'LIKE %s',
        'regex': 'REGEXP BINARY %s',
        'iregex': 'REGEXP %s',
        'gt': '> %s',
        'gte': '>= %s',
        'lt': '< %s',
        'lte': '<= %s',
        'startswith': 'LIKE BINARY %s',
        'endswith': 'LIKE BINARY %s',
        'istartswith': 'LIKE %s',
        'iendswith': 'LIKE %s',
    }

    # The patterns below are used to generate SQL pattern lookup clauses when
    # the right-hand side of the lookup isn't a raw string (it might be an expression
    # or the result of a bilateral transformation).
    # In those cases, special characters for LIKE operators (e.g. \, *, _) should be
    # escaped on database side.
    #
    # Note: we use str.format() here for readability as '%' is used as a wildcard for
    # the LIKE operator.
    pattern_esc = r"REPLACE(REPLACE(REPLACE({}, '\\', '\\\\'), '%%', '\%%'), '_', '\_')"
    pattern_ops = {
        'contains': "LIKE BINARY CONCAT('%%', {}, '%%')",
        'icontains': "LIKE CONCAT('%%', {}, '%%')",
        'startswith': "LIKE BINARY CONCAT({}, '%%')",
        'istartswith': "LIKE CONCAT({}, '%%')",
        'endswith': "LIKE BINARY CONCAT('%%', {})",
        'iendswith': "LIKE CONCAT('%%', {})",
    }

    Database = Database
    SchemaEditorClass = DatabaseSchemaEditor

    def __init__(self, *args, **kwargs):
        super(DatabaseWrapper, self).__init__(*args, **kwargs)

        self.features = DatabaseFeatures(self)
        self.ops = DatabaseOperations(self)
        self.client = DatabaseClient(self)
        self.creation = DatabaseCreation(self)
        self.introspection = DatabaseIntrospection(self)
        self.validation = DatabaseValidation(self)

    def get_connection_params(self):
        kwargs = {
            'conv': django_conversions,
            'charset': 'utf8',
        }
        if six.PY2:
            kwargs['use_unicode'] = True
        settings_dict = self.settings_dict
        if settings_dict['USER']:
            kwargs['user'] = settings_dict['USER']
        if settings_dict['NAME']:
            kwargs['db'] = settings_dict['NAME']
        if settings_dict['PASSWORD']:
            kwargs['passwd'] = force_str(settings_dict['PASSWORD'])
        if settings_dict['HOST'].startswith('/'):
            kwargs['unix_socket'] = settings_dict['HOST']
        elif settings_dict['HOST']:
            kwargs['host'] = settings_dict['HOST']
        if settings_dict['PORT']:
            kwargs['port'] = int(settings_dict['PORT'])
        # We need the number of potentially affected rows after an
        # "UPDATE", not the number of changed rows.
        kwargs['client_flag'] = CLIENT.FOUND_ROWS
        kwargs.update(settings_dict['OPTIONS'])
        return kwargs

    def get_new_connection(self, conn_params):
        conn = Database.connect(**conn_params)
        conn.encoders[SafeText] = conn.encoders[six.text_type]
        conn.encoders[SafeBytes] = conn.encoders[bytes]
        return conn

    def init_connection_state(self):
        with self.cursor() as cursor:
            # SQL_AUTO_IS_NULL in MySQL controls whether an AUTO_INCREMENT column
            # on a recently-inserted row will return when the field is tested for
            # NULL.  Disabling this value brings this aspect of MySQL in line with
            # SQL standards.
            cursor.execute('SET SQL_AUTO_IS_NULL = 0')

    def create_cursor(self):
        cursor = self.connection.cursor()
        return CursorWrapper(cursor)

    def _rollback(self):
        try:
            BaseDatabaseWrapper._rollback(self)
        except Database.NotSupportedError:
            pass

    def _set_autocommit(self, autocommit):
        with self.wrap_database_errors:
            self.connection.autocommit(autocommit)

    def disable_constraint_checking(self):
        """
        Disables foreign key checks, primarily for use in adding rows with forward references. Always returns True,
        to indicate constraint checks need to be re-enabled.
        """
        self.cursor().execute('SET foreign_key_checks=0')
        return True

    def enable_constraint_checking(self):
        """
        Re-enable foreign key checks after they have been disabled.
        """
        # Override needs_rollback in case constraint_checks_disabled is
        # nested inside transaction.atomic.
        self.needs_rollback, needs_rollback = False, self.needs_rollback
        try:
            self.cursor().execute('SET foreign_key_checks=1')
        finally:
            self.needs_rollback = needs_rollback

    def check_constraints(self, table_names=None):
        """
        Checks each table name in `table_names` for rows with invalid foreign
        key references. This method is intended to be used in conjunction with
        `disable_constraint_checking()` and `enable_constraint_checking()`, to
        determine if rows with invalid references were entered while constraint
        checks were off.

        Raises an IntegrityError on the first invalid foreign key reference
        encountered (if any) and provides detailed information about the
        invalid reference in the error message.

        Backends can override this method if they can more directly apply
        constraint checking (e.g. via "SET CONSTRAINTS ALL IMMEDIATE")
        """
        cursor = self.cursor()
        if table_names is None:
            table_names = self.introspection.table_names(cursor)
        for table_name in table_names:
            primary_key_column_name = self.introspection.get_primary_key_column(cursor, table_name)
            if not primary_key_column_name:
                continue
            key_columns = self.introspection.get_key_columns(cursor, table_name)
            for column_name, referenced_table_name, referenced_column_name in key_columns:
                cursor.execute("""
                    SELECT REFERRING.`%s`, REFERRING.`%s` FROM `%s` as REFERRING
                    LEFT JOIN `%s` as REFERRED
                    ON (REFERRING.`%s` = REFERRED.`%s`)
                    WHERE REFERRING.`%s` IS NOT NULL AND REFERRED.`%s` IS NULL"""
                    % (primary_key_column_name, column_name, table_name, referenced_table_name,
                    column_name, referenced_column_name, column_name, referenced_column_name))
                for bad_row in cursor.fetchall():
                    raise utils.IntegrityError("The row in table '%s' with primary key '%s' has an invalid "
                        "foreign key: %s.%s contains a value '%s' that does not have a corresponding value in %s.%s."
                        % (table_name, bad_row[0],
                        table_name, column_name, bad_row[1],
                        referenced_table_name, referenced_column_name))

    def is_usable(self):
        try:
            self.connection.ping()
        except Database.Error:
            return False
        else:
            return True

    @cached_property
    def mysql_version(self):
        with self.temporary_connection():
            server_info = self.connection.get_server_info()
        match = server_version_re.match(server_info)
        if not match:
            raise Exception('Unable to determine MySQL version from version string %r' % server_info)
        return tuple(int(x) for x in match.groups())