MindsDB 1.2.9__py3-none-any.whl → 24.12.4.0__py3-none-any.whl

mindsdb/api/executor/command_executor.py

@@ -0,0 +1,2047 @@
+import datetime
+from pathlib import Path
+from textwrap import dedent
+from typing import Optional
+from functools import reduce
+
+import pandas as pd
+from mindsdb_evaluator.accuracy.general import evaluate_accuracy
+from mindsdb_sql_parser import parse_sql
+from mindsdb_sql_parser.ast import (
+    Alter,
+    ASTNode,
+    BinaryOperation,
+    CommitTransaction,
+    Constant,
+    CreateTable,
+    Delete,
+    Describe,
+    DropDatabase,
+    DropTables,
+    DropView,
+    Explain,
+    Function,
+    Identifier,
+    Insert,
+    NativeQuery,
+    Operation,
+    RollbackTransaction,
+    Select,
+    Set,
+    Show,
+    Star,
+    StartTransaction,
+    Union,
+    Update,
+    Use,
+    Variable,
+    Tuple,
+)
+
+# typed models
+from mindsdb_sql_parser.ast.mindsdb import (
+    CreateAgent,
+    CreateAnomalyDetectionModel,
+    CreateChatBot,
+    CreateDatabase,
+    CreateJob,
+    CreateKnowledgeBase,
+    CreateMLEngine,
+    CreatePredictor,
+    CreateSkill,
+    CreateTrigger,
+    CreateView,
+    DropAgent,
+    DropChatBot,
+    DropDatasource,
+    DropJob,
+    DropKnowledgeBase,
+    DropMLEngine,
+    DropPredictor,
+    DropSkill,
+    DropTrigger,
+    Evaluate,
+    FinetunePredictor,
+    RetrainPredictor,
+    UpdateAgent,
+    UpdateChatBot,
+    UpdateSkill
+)
+
+import mindsdb.utilities.profiler as profiler
+
+from mindsdb.integrations.utilities.query_traversal import query_traversal
+from mindsdb.api.executor import Column, SQLQuery, ResultSet
+from mindsdb.api.executor.data_types.answer import ExecuteAnswer
+from mindsdb.api.mysql.mysql_proxy.libs.constants.mysql import (
+    CHARSET_NUMBERS,
+    SERVER_VARIABLES,
+    TYPES,
+)
+
+from .exceptions import (
+    ExecutorException,
+    BadDbError,
+    NotSupportedYet,
+    WrongArgumentError,
+    TableNotExistError,
+)
+from mindsdb.api.executor.utilities.functions import download_file
+from mindsdb.api.executor.utilities.sql import query_df
+from mindsdb.integrations.libs.const import (
+    HANDLER_CONNECTION_ARG_TYPE,
+    PREDICTOR_STATUS,
+)
+from mindsdb.integrations.libs.response import HandlerStatusResponse
+from mindsdb.interfaces.chatbot.chatbot_controller import ChatBotController
+from mindsdb.interfaces.database.projects import ProjectController
+from mindsdb.interfaces.jobs.jobs_controller import JobsController
+from mindsdb.interfaces.model.functions import (
+    get_model_record,
+    get_model_records,
+    get_predictor_integration,
+)
+from mindsdb.interfaces.query_context.context_controller import query_context_controller
+from mindsdb.interfaces.triggers.triggers_controller import TriggersController
+from mindsdb.utilities.context import context as ctx
+from mindsdb.utilities.functions import mark_process, resolve_model_identifier, get_handler_install_message
+from mindsdb.utilities.exception import EntityExistsError, EntityNotExistsError
+from mindsdb.utilities import log
+from mindsdb.api.mysql.mysql_proxy.utilities import ErParseError
+
+logger = log.getLogger(__name__)
+
+
+def _get_show_where(
+    statement: ASTNode,
+    from_name: Optional[str] = None,
+    like_name: Optional[str] = None,
+    initial: Optional[ASTNode] = None,
+) -> ASTNode:
+    """combine all possible show filters to single 'where' condition
+    SHOW category [FROM name] [LIKE filter] [WHERE filter]
+
+    Args:
+        statement (ASTNode): 'show' query statement
+        from_name (str): name of column for 'from' filter
+        like_name (str): name of column for 'like' filter,
+        initial (ASTNode): initial 'where' filter
+    Returns:
+        ASTNode: 'where' statemnt
+    """
+    where = []
+    if initial is not None:
+        where.append(initial)
+    if statement.from_table is not None and from_name is not None:
+        where.append(
+            BinaryOperation(
+                "=",
+                args=[Identifier(from_name), Constant(statement.from_table.parts[-1])],
+            )
+        )
+    if statement.like is not None and like_name is not None:
+        where.append(
+            BinaryOperation(
+                "like", args=[Identifier(like_name), Constant(statement.like)]
+            )
+        )
+    if statement.where is not None:
+        where.append(statement.where)
+
+    if len(where) > 0:
+        return reduce(
+            lambda prev, next: BinaryOperation("and", args=[prev, next]), where
+        )
+    return None
+
+
+class ExecuteCommands:
+    def __init__(self, session, context=None):
+        if context is None:
+            context = {}
+
+        self.context = context
+        self.session = session
+
+        self.charset_text_type = CHARSET_NUMBERS["utf8_general_ci"]
+        self.datahub = session.datahub
+
+    @profiler.profile()
+    def execute_command(self, statement, database_name: str = None) -> ExecuteAnswer:
+        sql = None
+        if isinstance(statement, ASTNode):
+            sql = statement.to_string()
+        sql_lower = sql.lower()
+
+        if database_name is None:
+            database_name = self.session.database
+
+        if type(statement) is CreateDatabase:
+            return self.answer_create_database(statement)
+        elif type(statement) is CreateMLEngine:
+            name = statement.name.parts[-1]
+
+            return self.answer_create_ml_engine(
+                name,
+                handler=statement.handler,
+                params=statement.params,
+                if_not_exists=getattr(statement, "if_not_exists", False)
+            )
+        elif type(statement) is DropMLEngine:
+            return self.answer_drop_ml_engine(statement)
+        elif type(statement) is DropPredictor:
+            return self.answer_drop_model(statement, database_name)
+
+        elif type(statement) is DropTables:
+            return self.answer_drop_tables(statement, database_name)
+        elif type(statement) is DropDatasource or type(statement) is DropDatabase:
+            return self.answer_drop_database(statement)
+        elif type(statement) is Describe:
+            # NOTE in sql 'describe table' is same as 'show columns'
+            obj_type = statement.type
+
+            if obj_type is None or obj_type.upper() in ('MODEL', 'PREDICTOR'):
+                return self.answer_describe_predictor(statement.value, database_name)
+            else:
+                return self.answer_describe_object(obj_type.upper(), statement.value, database_name)
+
+        elif type(statement) is RetrainPredictor:
+            return self.answer_retrain_predictor(statement, database_name)
+        elif type(statement) is FinetunePredictor:
+            return self.answer_finetune_predictor(statement, database_name)
+        elif type(statement) is Show:
+            sql_category = statement.category.lower()
+            if hasattr(statement, "modes"):
+                if isinstance(statement.modes, list) is False:
+                    statement.modes = []
+                statement.modes = [x.upper() for x in statement.modes]
+            if sql_category == "ml_engines":
+                new_statement = Select(
+                    targets=[Star()],
+                    from_table=Identifier(parts=["information_schema", "ml_engines"]),
+                    where=_get_show_where(statement, like_name="name"),
+                )
+
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category == "handlers":
+                new_statement = Select(
+                    targets=[Star()],
+                    from_table=Identifier(parts=["information_schema", "handlers"]),
+                    where=_get_show_where(statement, like_name="name"),
+                )
+
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category == "plugins":
+                if statement.where is not None or statement.like:
+                    raise ExecutorException(
+                        "'SHOW PLUGINS' query should be used without filters"
+                    )
+                new_statement = Select(
+                    targets=[Star()],
+                    from_table=Identifier(parts=["information_schema", "PLUGINS"]),
+                )
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category in ("databases", "schemas"):
+                new_statement = Select(
+                    targets=[Identifier(parts=["NAME"], alias=Identifier("Database"))],
+                    from_table=Identifier(parts=["information_schema", "DATABASES"]),
+                    where=_get_show_where(statement, like_name="Database"),
+                )
+
+                if "FULL" in statement.modes:
+                    new_statement.targets.extend(
+                        [
+                            Identifier(parts=["TYPE"], alias=Identifier("TYPE")),
+                            Identifier(parts=["ENGINE"], alias=Identifier("ENGINE")),
+                        ]
+                    )
+
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category in ("tables", "full tables"):
+                schema = database_name or "mindsdb"
+                if (
+                    statement.from_table is not None
+                    and statement.in_table is not None
+                ):
+                    raise ExecutorException(
+                        "You have an error in your SQL syntax: 'from' and 'in' cannot be used together"
+                    )
+
+                if statement.from_table is not None:
+                    schema = statement.from_table.parts[-1]
+                    statement.from_table = None
+                if statement.in_table is not None:
+                    schema = statement.in_table.parts[-1]
+                    statement.in_table = None
+
+                table_types = [Constant(t) for t in ['MODEL', 'BASE TABLE', 'SYSTEM VIEW', 'VIEW']]
+                where = BinaryOperation(
+                    "and",
+                    args=[
+                        BinaryOperation("=", args=[Identifier("table_schema"), Constant(schema)]),
+                        BinaryOperation("in", args=[Identifier("table_type"), Tuple(table_types)])
+                    ]
+                )
+
+                new_statement = Select(
+                    targets=[
+                        Identifier(
+                            parts=["table_name"],
+                            alias=Identifier(f"Tables_in_{schema}"),
+                        )
+                    ],
+                    from_table=Identifier(parts=["information_schema", "TABLES"]),
+                    where=_get_show_where(
+                        statement, like_name=f"Tables_in_{schema}", initial=where
+                    ),
+                )
+
+                if "FULL" in statement.modes:
+                    new_statement.targets.append(
+                        Identifier(parts=["TABLE_TYPE"], alias=Identifier("Table_type"))
+                    )
+
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category in (
+                "variables",
+                "session variables",
+                "session status",
+                "global variables",
+            ):
+                new_statement = Select(
+                    targets=[
+                        Identifier(parts=["Variable_name"]),
+                        Identifier(parts=["Value"]),
+                    ],
+                    from_table=Identifier(parts=["dataframe"]),
+                    where=_get_show_where(statement, like_name="Variable_name"),
+                )
+
+                data = {}
+                is_session = "session" in sql_category
+                for var_name, var_data in SERVER_VARIABLES.items():
+                    var_name = var_name.replace("@@", "")
+                    if is_session and var_name.startswith("session.") is False:
+                        continue
+                    if var_name.startswith("session.") or var_name.startswith(
+                        "GLOBAL."
+                    ):
+                        name = var_name.replace("session.", "").replace("GLOBAL.", "")
+                        data[name] = var_data[0]
+                    elif var_name not in data:
+                        data[var_name] = var_data[0]
+
+                df = pd.DataFrame(data.items(), columns=["Variable_name", "Value"])
+                df2 = query_df(df, new_statement)
+
+                return ExecuteAnswer(
+                    data=ResultSet().from_df(df2, table_name="session_variables")
+                )
+            elif sql_category == "search_path":
+                return ExecuteAnswer(
+                    data=ResultSet(
+                        columns=[
+                            Column(name="search_path", table_name="search_path", type="str")
+                        ],
+                        values=[['"$user", public']]
+                    )
+                )
+            elif "show status like 'ssl_version'" in sql_lower:
+                return ExecuteAnswer(
+                    data=ResultSet(
+                        columns=[
+                            Column(
+                                name="Value", table_name="session_variables", type="str"
+                            ),
+                            Column(
+                                name="Value", table_name="session_variables", type="str"
+                            ),
+                        ],
+                        values=[["Ssl_version", "TLSv1.1"]],
+                    )
+                )
+            elif sql_category in ("function status", "procedure status"):
+                # SHOW FUNCTION STATUS WHERE Db = 'MINDSDB';
+                # SHOW PROCEDURE STATUS WHERE Db = 'MINDSDB'
+                # SHOW FUNCTION STATUS WHERE Db = 'MINDSDB' AND Name LIKE '%';
+                return self.answer_function_status()
+            elif sql_category in ("index", "keys", "indexes"):
+                # INDEX | INDEXES | KEYS are synonyms
+                # https://dev.mysql.com/doc/refman/8.0/en/show-index.html
+                new_statement = Select(
+                    targets=[
+                        Identifier("TABLE_NAME", alias=Identifier("Table")),
+                        Identifier("NON_UNIQUE", alias=Identifier("Non_unique")),
+                        Identifier("INDEX_NAME", alias=Identifier("Key_name")),
+                        Identifier("SEQ_IN_INDEX", alias=Identifier("Seq_in_index")),
+                        Identifier("COLUMN_NAME", alias=Identifier("Column_name")),
+                        Identifier("COLLATION", alias=Identifier("Collation")),
+                        Identifier("CARDINALITY", alias=Identifier("Cardinality")),
+                        Identifier("SUB_PART", alias=Identifier("Sub_part")),
+                        Identifier("PACKED", alias=Identifier("Packed")),
+                        Identifier("NULLABLE", alias=Identifier("Null")),
+                        Identifier("INDEX_TYPE", alias=Identifier("Index_type")),
+                        Identifier("COMMENT", alias=Identifier("Comment")),
+                        Identifier("INDEX_COMMENT", alias=Identifier("Index_comment")),
+                        Identifier("IS_VISIBLE", alias=Identifier("Visible")),
+                        Identifier("EXPRESSION", alias=Identifier("Expression")),
+                    ],
+                    from_table=Identifier(parts=["information_schema", "STATISTICS"]),
+                    where=statement.where,
+                )
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            # FIXME if have answer on that request, then DataGrip show warning '[S0022] Column 'Non_unique' not found.'
+            elif "show create table" in sql_lower:
+                # SHOW CREATE TABLE `MINDSDB`.`predictors`
+                table = sql[sql.rfind(".") + 1:].strip(" .;\n\t").replace("`", "")
+                return self.answer_show_create_table(table)
+            elif sql_category in ("character set", "charset"):
+                new_statement = Select(
+                    targets=[
+                        Identifier("CHARACTER_SET_NAME", alias=Identifier("Charset")),
+                        Identifier(
+                            "DEFAULT_COLLATE_NAME", alias=Identifier("Description")
+                        ),
+                        Identifier(
+                            "DESCRIPTION", alias=Identifier("Default collation")
+                        ),
+                        Identifier("MAXLEN", alias=Identifier("Maxlen")),
+                    ],
+                    from_table=Identifier(
+                        parts=["INFORMATION_SCHEMA", "CHARACTER_SETS"]
+                    ),
+                    where=_get_show_where(statement, like_name="CHARACTER_SET_NAME"),
+                )
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category == "warnings":
+                return self.answer_show_warnings()
+            elif sql_category == "engines":
+                new_statement = Select(
+                    targets=[Star()],
+                    from_table=Identifier(parts=["information_schema", "ENGINES"]),
+                )
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category == "collation":
+                new_statement = Select(
+                    targets=[
+                        Identifier("COLLATION_NAME", alias=Identifier("Collation")),
+                        Identifier("CHARACTER_SET_NAME", alias=Identifier("Charset")),
+                        Identifier("ID", alias=Identifier("Id")),
+                        Identifier("IS_DEFAULT", alias=Identifier("Default")),
+                        Identifier("IS_COMPILED", alias=Identifier("Compiled")),
+                        Identifier("SORTLEN", alias=Identifier("Sortlen")),
+                        Identifier("PAD_ATTRIBUTE", alias=Identifier("Pad_attribute")),
+                    ],
+                    from_table=Identifier(parts=["INFORMATION_SCHEMA", "COLLATIONS"]),
+                    where=_get_show_where(statement, like_name="Collation"),
+                )
+                query = SQLQuery(new_statement, session=self.session, database=database_name)
+                return self.answer_select(query)
+            elif sql_category == "table status":
+                # TODO improve it
+                # SHOW TABLE STATUS LIKE 'table'
+                table_name = None
+                if statement.like is not None:
+                    table_name = statement.like
+                # elif condition == 'from' and type(expression) == Identifier:
+                #     table_name = expression.parts[-1]
+                if table_name is None:
+                    err_str = f"Can't determine table name in query: {sql}"
+                    logger.warning(err_str)
+                    raise TableNotExistError(err_str)
+                return self.answer_show_table_status(table_name)
+            elif sql_category == "columns":
+                is_full = statement.modes is not None and "full" in statement.modes
+                return self.answer_show_columns(
+                    statement.from_table,
+                    statement.where,
+                    statement.like,
+                    is_full=is_full,
+                    database_name=database_name,
+                )
+
+            elif sql_category in ("agents", "jobs", "skills", "chatbots", "triggers", "views",
+                                  "knowledge_bases", "knowledge bases", "predictors", "models"):
+
+                if sql_category == "knowledge bases":
+                    sql_category = "knowledge_bases"
+
+                if sql_category == "predictors":
+                    sql_category = "models"
+
+                db_name = database_name
+                if statement.from_table is not None:
+                    db_name = statement.from_table.parts[-1]
+
+                where = BinaryOperation(op='=', args=[Identifier('project'), Constant(db_name)])
+
+                select_statement = Select(
+                    targets=[Star()],
+                    from_table=Identifier(
+                        parts=["information_schema", sql_category]
+                    ),
+                    where=_get_show_where(statement, like_name="name", initial=where),
+                )
+                query = SQLQuery(select_statement, session=self.session)
+                return self.answer_select(query)
+
+            elif sql_category == "projects":
+                where = BinaryOperation(op='=', args=[Identifier('type'), Constant('project')])
+                select_statement = Select(
+                    targets=[Identifier(parts=["NAME"], alias=Identifier('project'))],
+                    from_table=Identifier(
+                        parts=["information_schema", "DATABASES"]
+                    ),
+                    where=_get_show_where(statement, like_name="project", from_name="project", initial=where),
+                )
+
+                query = SQLQuery(select_statement, session=self.session)
+                return self.answer_select(query)
+            else:
+                raise NotSupportedYet(f"Statement not implemented: {sql}")
+        elif type(statement) in (
+            StartTransaction,
+            CommitTransaction,
+            RollbackTransaction,
+        ):
+            return ExecuteAnswer()
+        elif type(statement) is Set:
+            category = (statement.category or "").lower()
+            if category == "" and isinstance(statement.name, Identifier):
+                param = statement.name.parts[0].lower()
+
+                value = None
+                if isinstance(statement.value, Constant):
+                    value = statement.value.value
+
+                if param == "profiling":
+                    self.session.profiling = value in (1, True)
+                    if self.session.profiling is True:
+                        profiler.enable()
+                    else:
+                        profiler.disable()
+                elif param == "predictor_cache":
+                    self.session.predictor_cache = value in (1, True)
+                elif param == "context":
+                    if value in (0, False, None):
+                        # drop context
+                        query_context_controller.drop_query_context(None)
+                elif param == "show_secrets":
+                    self.session.show_secrets = value in (1, True)
+
+                return ExecuteAnswer()
+            elif category == "autocommit":
+                return ExecuteAnswer()
+            elif category == "names":
+                # set names utf8;
+                charsets = {
+                    "utf8": CHARSET_NUMBERS["utf8_general_ci"],
+                    "utf8mb4": CHARSET_NUMBERS["utf8mb4_general_ci"],
+                }
+                self.charset = statement.value.value
+                self.charset_text_type = charsets.get(self.charset)
+                if self.charset_text_type is None:
+                    logger.warning(
+                        f"Unknown charset: {self.charset}. Setting up 'utf8_general_ci' as charset text type."
+                    )
+                    self.charset_text_type = CHARSET_NUMBERS["utf8_general_ci"]
+                return ExecuteAnswer(
+                    state_track=[
+                        ["character_set_client", self.charset],
+                        ["character_set_connection", self.charset],
+                        ["character_set_results", self.charset],
+                    ],
+                )
+            elif category == "active":
+                return self.answer_update_model_version(statement.value, database_name)
+
+            else:
+                logger.warning(
+                    f"SQL statement is not processable, return OK package: {sql}"
+                )
+                return ExecuteAnswer()
+        elif type(statement) is Use:
+            db_name = statement.value.parts[-1]
+            self.change_default_db(db_name)
+            return ExecuteAnswer()
+        elif type(statement) in (
+            CreatePredictor,
+            CreateAnomalyDetectionModel,  # we may want to specialize these in the future
+        ):
+            return self.answer_create_predictor(statement, database_name)
+        elif type(statement) is CreateView:
+            return self.answer_create_view(statement, database_name)
+        elif type(statement) is DropView:
+            return self.answer_drop_view(statement, database_name)
+        elif type(statement) is Delete:
+            SQLQuery(statement, session=self.session, execute=True, database=database_name)
+            return ExecuteAnswer()
+
+        elif type(statement) is Insert:
+            SQLQuery(statement, session=self.session, execute=True, database=database_name)
+            return ExecuteAnswer()
+        elif type(statement) is Update:
+            SQLQuery(statement, session=self.session, execute=True, database=database_name)
+            return ExecuteAnswer()
+        elif (
+            type(statement) is Alter
+            and ("disable keys" in sql_lower)
+            or ("enable keys" in sql_lower)
+        ):
+            return ExecuteAnswer()
+        elif type(statement) is Select:
+            if statement.from_table is None:
+                return self.answer_single_row_select(statement, database_name)
+            query = SQLQuery(statement, session=self.session, database=database_name)
+            return self.answer_select(query)
+        elif type(statement) is Union:
+            query = SQLQuery(statement, session=self.session, database=database_name)
+            return self.answer_select(query)
+        elif type(statement) is Explain:
+            return self.answer_show_columns(statement.target, database_name=database_name)
+        elif type(statement) is CreateTable:
+            return self.answer_create_table(statement, database_name)
+        # -- jobs --
+        elif type(statement) is CreateJob:
+            return self.answer_create_job(statement, database_name)
+        elif type(statement) is DropJob:
+            return self.answer_drop_job(statement, database_name)
+        # -- triggers --
+        elif type(statement) is CreateTrigger:
+            return self.answer_create_trigger(statement, database_name)
+        elif type(statement) is DropTrigger:
+            return self.answer_drop_trigger(statement, database_name)
+        # -- chatbots
+        elif type(statement) is CreateChatBot:
+            return self.answer_create_chatbot(statement, database_name)
+        elif type(statement) is UpdateChatBot:
+            return self.answer_update_chatbot(statement, database_name)
+        elif type(statement) is DropChatBot:
+            return self.answer_drop_chatbot(statement, database_name)
+        elif type(statement) is CreateKnowledgeBase:
+            return self.answer_create_kb(statement, database_name)
+        elif type(statement) is DropKnowledgeBase:
+            return self.answer_drop_kb(statement, database_name)
+        elif type(statement) is CreateSkill:
+            return self.answer_create_skill(statement, database_name)
+        elif type(statement) is DropSkill:
+            return self.answer_drop_skill(statement, database_name)
+        elif type(statement) is UpdateSkill:
+            return self.answer_update_skill(statement, database_name)
+        elif type(statement) is CreateAgent:
+            return self.answer_create_agent(statement, database_name)
+        elif type(statement) is DropAgent:
+            return self.answer_drop_agent(statement, database_name)
+        elif type(statement) is UpdateAgent:
+            return self.answer_update_agent(statement, database_name)
+        elif type(statement) is Evaluate:
+            statement.data = parse_sql(statement.query_str)
+            return self.answer_evaluate_metric(statement, database_name)
+        else:
+            logger.warning(f"Unknown SQL statement: {sql}")
+            raise NotSupportedYet(f"Unknown SQL statement: {sql}")
+
+    def answer_create_trigger(self, statement, database_name):
+        triggers_controller = TriggersController()
+
+        name = statement.name
+        trigger_name = statement.name.parts[-1]
+        project_name = name.parts[-2] if len(name.parts) > 1 else database_name
+
+        triggers_controller.add(
+            trigger_name,
+            project_name,
+            statement.table,
+            statement.query_str,
+            statement.columns,
+        )
+        return ExecuteAnswer()
+
+    def answer_drop_trigger(self, statement, database_name):
+        triggers_controller = TriggersController()
+
+        name = statement.name
+        trigger_name = statement.name.parts[-1]
+        project_name = name.parts[-2] if len(name.parts) > 1 else database_name
+
+        triggers_controller.delete(trigger_name, project_name)
+
+        return ExecuteAnswer()
+
+    def answer_create_job(self, statement: CreateJob, database_name):
+        jobs_controller = JobsController()
+
+        name = statement.name
+        job_name = name.parts[-1]
+        project_name = name.parts[-2] if len(name.parts) > 1 else database_name
+
+        try:
+            jobs_controller.create(job_name, project_name, statement)
+        except EntityExistsError:
+            if getattr(statement, "if_not_exists", False) is False:
+                raise
+
+        return ExecuteAnswer()
+
+    def answer_drop_job(self, statement, database_name):
+        jobs_controller = JobsController()
+
+        name = statement.name
+        job_name = name.parts[-1]
+        project_name = name.parts[-2] if len(name.parts) > 1 else database_name
+        try:
+            jobs_controller.delete(job_name, project_name)
+        except EntityNotExistsError:
+            if statement.if_exists is False:
+                raise
+        except Exception as e:
+            raise e
+
+        return ExecuteAnswer()
+
+    def answer_create_chatbot(self, statement, database_name):
+        chatbot_controller = ChatBotController()
+
+        name = statement.name
+        project_name = name.parts[-2] if len(name.parts) > 1 else database_name
+        is_running = statement.params.pop("is_running", True)
+
+        database = self.session.integration_controller.get(statement.database.parts[-1])
+        if database is None:
+            raise ExecutorException(f"Database not found: {statement.database}")
+
+        # Database ID cannot be null
+        database_id = database["id"] if database is not None else -1
+
+        model_name = None
+        if statement.model is not None:
+            model_name = statement.model.parts[-1]
+
+        agent_name = None
+        if statement.agent is not None:
+            agent_name = statement.agent.parts[-1]
+        chatbot_controller.add_chatbot(
+            name.parts[-1],
+            project_name=project_name,
+            model_name=model_name,
+            agent_name=agent_name,
+            database_id=database_id,
+            is_running=is_running,
+            params=statement.params,
+        )
+        return ExecuteAnswer()
+
+    def answer_update_chatbot(self, statement, database_name):
+        chatbot_controller = ChatBotController()
+
+        name = statement.name
+        name_no_project = name.parts[-1]
+        project_name = name.parts[-2] if len(name.parts) > 1 else database_name
+
+        # From SET keyword parameters
+        updated_name = statement.params.pop("name", None)
+        model_name = statement.params.pop("model", None)
+        agent_name = statement.params.pop("agent", None)
+        database_name = statement.params.pop("database", None)
+        is_running = statement.params.pop("is_running", None)
+
+        database_id = None
+        if database_name is not None:
+            database = self.session.integration_controller.get(database_name)
+            if database is None:
+                raise ExecutorException(f"Database with name {database_name} not found")
+            database_id = database["id"]
+
+        updated_chatbot = chatbot_controller.update_chatbot(
+            name_no_project,
+            project_name=project_name,
+            name=updated_name,
+            model_name=model_name,
+            agent_name=agent_name,
+            database_id=database_id,
+            is_running=is_running,
+            params=statement.params,
+        )
+        if updated_chatbot is None:
+            raise ExecutorException(f"Chatbot with name {name_no_project} not found")
+        return ExecuteAnswer()
+
+    def answer_drop_chatbot(self, statement, database_name):
+        chatbot_controller = ChatBotController()
+
+        name = statement.name
+        project_name = name.parts[-2] if len(name.parts) > 1 else database_name
+
+        chatbot_controller.delete_chatbot(name.parts[-1], project_name=project_name)
+        return ExecuteAnswer()
+
+    def answer_evaluate_metric(self, statement, database_name):
+        try:
+            sqlquery = SQLQuery(statement.data, session=self.session, database=database_name)
+        except Exception as e:
+            raise Exception(
+                f'Nested query failed to execute with error: "{e}", please check and try again.'
+            )
+        result = sqlquery.fetch('dataframe')
+        df = result["result"]
+        df.columns = [
+            str(t.alias) if hasattr(t, "alias") else str(t.parts[-1])
+            for t in statement.data.targets
+        ]
+
+        for col in ["actual", "prediction"]:
+            assert (
+                col in df.columns
+            ), f"`{col}` column was not provided, please try again."
+            assert (
+                df[col].isna().sum() == 0
+            ), f"There are missing values in the `{col}` column, please try again."
+
+        metric_name = statement.name.parts[-1]
+        target_series = df.pop("prediction")
+        using_clause = statement.using if statement.using is not None else {}
+        metric_value = evaluate_accuracy(
+            df,
+            target_series,
+            metric_name,
+            target="actual",
+            ts_analysis=using_clause.get("ts_analysis", {}),  # will be deprecated soon
+            n_decimals=using_clause.get("n_decimals", 3),
+        )  # 3 decimals by default
+        return ExecuteAnswer(
+            data=ResultSet(
+                columns=[Column(name=metric_name, table_name="", type="str")],
+                values=[[metric_value]],
+            )
+        )
+
+    def answer_describe_object(self, obj_type: str, obj_name: Identifier, database_name: str):
+
+        project_objects = ("AGENTS", "JOBS", "SKILLS", "CHATBOTS", "TRIGGERS", "VIEWS",
+                           "KNOWLEDGE_BASES", "PREDICTORS", "MODELS")
+
+        global_objects = ("DATABASES", "PROJECTS", "HANDLERS", "ML_ENGINES")
+
+        all_objects = project_objects + global_objects
+
+        # is not plural?
+        if obj_type not in all_objects:
+            if obj_type + 'S' in all_objects:
+                obj_type = obj_type + 'S'
+            elif obj_type + 'ES' in all_objects:
+                obj_type = obj_type + 'ES'
+            else:
+                raise WrongArgumentError(f'Unknown describe type: {obj_type}')
+
+        name = obj_name.parts[-1]
+        where = BinaryOperation(op='=', args=[
+            Identifier('name'),
+            Constant(name)
+        ])
+
+        if obj_type in project_objects:
+            where = BinaryOperation(op='and', args=[
+                where,
+                BinaryOperation(op='=', args=[Identifier('project'), Constant(database_name)])
+            ])
+
+        select_statement = Select(
+            targets=[Star()],
+            from_table=Identifier(
+                parts=["information_schema", obj_type]
+            ),
+
+            where=where,
+        )
+        query = SQLQuery(select_statement, session=self.session)
+        return self.answer_select(query)
+
+    def answer_describe_predictor(self, obj_name, database_name):
+        value = obj_name.parts.copy()
+        # project.model.version.?attrs
+        parts = value[:3]
+        attrs = value[3:]
+        model_info = self._get_model_info(Identifier(parts=parts), except_absent=False, database_name=database_name)
+        if model_info is None:
+            # project.model.?attrs
+            parts = value[:2]
+            attrs = value[2:]
+            model_info = self._get_model_info(Identifier(parts=parts), except_absent=False, database_name=database_name)
+            if model_info is None:
+                # model.?attrs
+                parts = value[:1]
+                attrs = value[1:]
+                model_info = self._get_model_info(Identifier(parts=parts), except_absent=False, database_name=database_name)
+
+        if model_info is None:
+            raise ExecutorException(f"Model not found: {obj_name}")
+
+        if len(attrs) == 1:
+            attrs = attrs[0]
+        elif len(attrs) == 0:
+            attrs = None
+
+        df = self.session.model_controller.describe_model(
+            self.session,
+            model_info["project_name"],
+            model_info["model_record"].name,
+            attribute=attrs,
+            version=model_info['model_record'].version
+        )
+
+        return ExecuteAnswer(
+            data=ResultSet().from_df(df, table_name="")
+        )
+
+    def _get_model_info(self, identifier, except_absent=True, database_name=None):
+        if len(identifier.parts) == 1:
+            identifier.parts = [database_name, identifier.parts[0]]
+
+        database_name, model_name, model_version = resolve_model_identifier(identifier)
+        if database_name is None:
+            database_name = database_name
+
+        if model_name is None:
+            if except_absent:
+                raise Exception(f"Model not found: {identifier.to_string()}")
+            else:
+                return
+
+        model_record = get_model_record(
+            name=model_name,
+            project_name=database_name,
+            except_absent=except_absent,
+            version=model_version,
+            active=True if model_version is None else None,
+        )
+        if not model_record:
+            return None
+        return {"model_record": model_record, "project_name": database_name}
+
+    def _sync_predictor_check(self, phase_name):
+        """Checks if there is already a predictor retraining or fine-tuning
+        Do not allow to run retrain if there is another model in training process in less that 1h
+        """
+        is_cloud = self.session.config.get("cloud", False)
+        if is_cloud and ctx.user_class == 0:
+            models = get_model_records(active=None)
+            shortest_training = None
+            for model in models:
+                if (
+                    model.status
+                    in (PREDICTOR_STATUS.GENERATING, PREDICTOR_STATUS.TRAINING)
+                    and model.training_start_at is not None
+                    and model.training_stop_at is None
+                ):
+                    training_time = datetime.datetime.now() - model.training_start_at
+                    if shortest_training is None or training_time < shortest_training:
+                        shortest_training = training_time
+
+            if (
+                shortest_training is not None
+                and shortest_training < datetime.timedelta(hours=1)
+            ):
+                raise ExecutorException(
+                    f"Can't start {phase_name} process while any other predictor is in status 'training' or 'generating'"
+                )
+
+    def answer_retrain_predictor(self, statement, database_name):
+        model_record = self._get_model_info(statement.name, database_name=database_name)["model_record"]
+
+        if statement.query_str is None:
+            if model_record.data_integration_ref is not None:
+                if model_record.data_integration_ref["type"] == "integration":
+                    integration = self.session.integration_controller.get_by_id(
+                        model_record.data_integration_ref["id"]
+                    )
+                    if integration is None:
+                        raise EntityNotExistsError(
+                            "The database from which the model was trained no longer exists"
+                        )
+        elif statement.integration_name is None:
+            # set to current project
+            statement.integration_name = Identifier(database_name)
+
+        ml_handler = None
+        if statement.using is not None:
+            # repack using with lower names
+            statement.using = {k.lower(): v for k, v in statement.using.items()}
+
+            if "engine" in statement.using:
+                ml_integration_name = statement.using.pop("engine")
+                ml_handler = self.session.integration_controller.get_ml_handler(
+                    ml_integration_name
+                )
+
+        # use current ml handler
+        if ml_handler is None:
+            integration_record = get_predictor_integration(model_record)
+            if integration_record is None:
+                raise EntityNotExistsError("ML engine model was trained with does not esxists")
+            ml_handler = self.session.integration_controller.get_ml_handler(
+                integration_record.name
+            )
+
+        self._sync_predictor_check(phase_name="retrain")
+        df = self.session.model_controller.retrain_model(statement, ml_handler)
+
+        return ExecuteAnswer(
+            data=ResultSet().from_df(df)
+        )
+
+    @profiler.profile()
+    @mark_process("learn")
+    def answer_finetune_predictor(self, statement, database_name):
+        model_record = self._get_model_info(statement.name, database_name=database_name)["model_record"]
+
+        if statement.using is not None:
+            # repack using with lower names
+            statement.using = {k.lower(): v for k, v in statement.using.items()}
+
+        if statement.query_str is not None and statement.integration_name is None:
+            # set to current project
+            statement.integration_name = Identifier(database_name)
+
+        # use current ml handler
+        integration_record = get_predictor_integration(model_record)
+        if integration_record is None:
+            raise Exception(
+                "The ML engine that the model was trained with does not exist."
+            )
+        ml_handler = self.session.integration_controller.get_ml_handler(
+            integration_record.name
+        )
+
+        self._sync_predictor_check(phase_name="finetune")
+        df = self.session.model_controller.finetune_model(statement, ml_handler)
+
+        return ExecuteAnswer(
+            data=ResultSet().from_df(df)
+        )
+
+    def _create_integration(self, name: str, engine: str, connection_args: dict):
+        # we have connection checkers not for any db. So do nothing if fail
+        # TODO return rich error message
+
+        if connection_args is None:
+            connection_args = {}
+        status = HandlerStatusResponse(success=False)
+
+        storage = None
+        try:
+            handler_meta = self.session.integration_controller.get_handler_meta(engine)
+            if handler_meta is None:
+                raise ExecutorException(f"There is no engine '{engine}'")
+
+            if handler_meta.get("import", {}).get("success") is not True:
+                raise ExecutorException(f"The '{engine}' handler isn't installed.\n" + get_handler_install_message(engine))
+
+            accept_connection_args = handler_meta.get("connection_args")
+            if accept_connection_args is not None and connection_args is not None:
+                for arg_name, arg_value in connection_args.items():
+                    if arg_name not in accept_connection_args:
+                        continue
+                    arg_meta = accept_connection_args[arg_name]
+                    arg_type = arg_meta.get("type")
+                    if arg_type == HANDLER_CONNECTION_ARG_TYPE.PATH:
+                        # arg may be one of:
+                        # str: '/home/file.pem'
+                        # dict: {'path': '/home/file.pem'}
+                        # dict: {'url': 'https://host.com/file'}
+                        arg_value = connection_args[arg_name]
+                        if isinstance(arg_value, (str, dict)) is False:
+                            raise ExecutorException(f"Unknown type of arg: '{arg_value}'")
+                        if isinstance(arg_value, str) or "path" in arg_value:
+                            path = (
+                                arg_value
+                                if isinstance(arg_value, str)
+                                else arg_value["path"]
+                            )
+                            if Path(path).is_file() is False:
+                                raise ExecutorException(f"File not found at: '{path}'")
+                        elif "url" in arg_value:
+                            path = download_file(arg_value["url"])
+                        else:
+                            raise ExecutorException(
+                                f"Argument '{arg_name}' must be path or url to the file"
+                            )
+                        connection_args[arg_name] = path
+
+            handler = self.session.integration_controller.create_tmp_handler(
+                name=name,
+                engine=engine,
+                connection_args=connection_args
+            )
+            status = handler.check_connection()
+            if status.copy_storage:
+                storage = handler.handler_storage.export_files()
+        except Exception as e:
+            status.error_message = str(e)
+
+        if status.success is False:
+            raise ExecutorException(f"Can't connect to db: {status.error_message}")
+
+        integration = self.session.integration_controller.get(name)
+        if integration is not None:
+            raise EntityExistsError('Database already exists', name)
+        try:
+            integration = ProjectController().get(name=name)
+        except EntityNotExistsError:
+            pass
+        if integration is not None:
+            raise EntityExistsError('Project exists with this name', name)
+
+        self.session.integration_controller.add(name, engine, connection_args)
+        if storage:
+            handler = self.session.integration_controller.get_data_handler(name, connect=False)
+            handler.handler_storage.import_files(storage)
+
+    def answer_create_ml_engine(self, name: str, handler: str, params: dict = None, if_not_exists=False):
+
+        integrations = self.session.integration_controller.get_all()
+        if name in integrations:
+            if not if_not_exists:
+                raise EntityExistsError('Integration already exists', name)
+            else:
+                return ExecuteAnswer()
+
+        handler_module_meta = self.session.integration_controller.get_handler_meta(handler)
+
+        if handler_module_meta is None:
+            raise ExecutorException(f"There is no engine '{handler}'")
+
+        params_out = {}
+        if params:
+            for key, value in params.items():
+                # convert ast types to string
+                if isinstance(value, (Constant, Identifier)):
+                    value = value.to_string()
+                params_out[key] = value
+
+        try:
+            self.session.integration_controller.add(
+                name=name, engine=handler, connection_args=params_out
+            )
+        except Exception as e:
+            msg = str(e)
+            if type(e) in (ImportError, ModuleNotFoundError):
+                msg = dedent(
+                    f"""\
+                    The '{handler_module_meta['name']}' handler cannot be used. Reason is:
+                        {handler_module_meta['import']['error_message']}
+                """
+                )
+                is_cloud = self.session.config.get("cloud", False)
+                if is_cloud is False and "No module named" in handler_module_meta['import']['error_message']:
+                    logger.info(get_handler_install_message(handler_module_meta['name']))
+            ast_drop = DropMLEngine(name=Identifier(name))
+            self.answer_drop_ml_engine(ast_drop)
+            logger.info(msg)
+            raise ExecutorException(msg)
+
+        return ExecuteAnswer()
+
+    def answer_drop_ml_engine(self, statement: ASTNode):
+        name = statement.name.parts[-1]
+        integrations = self.session.integration_controller.get_all()
+        if name not in integrations:
+            if not statement.if_exists:
+                raise EntityNotExistsError('Integration does not exists', name)
+            else:
+                return ExecuteAnswer()
+        self.session.integration_controller.delete(name)
+        return ExecuteAnswer()
+
+    def answer_create_database(self, statement: ASTNode):
+        """create new handler (datasource/integration in old terms)
+        Args:
+            statement (ASTNode): data for creating database/project
+        """
+
+        if len(statement.name.parts) != 1:
+            raise Exception("Database name should contain only 1 part.")
+
+        database_name = statement.name.parts[0]
+        engine = statement.engine
+        if engine is None:
+            engine = "mindsdb"
+        engine = engine.lower()
+        connection_args = statement.parameters
+
+        if engine == "mindsdb":
+            try:
+                ProjectController().add(database_name)
+            except EntityExistsError:
+                if statement.if_not_exists is False:
+                    raise
+        else:
+            try:
+                self._create_integration(database_name, engine, connection_args)
+            except EntityExistsError:
+                if getattr(statement, "if_not_exists", False) is False:
+                    raise
+
+        return ExecuteAnswer()
+
+    def answer_drop_database(self, statement):
+        if len(statement.name.parts) != 1:
+            raise Exception("Database name should contain only 1 part.")
+        db_name = statement.name.parts[0]
+        try:
+            self.session.database_controller.delete(db_name)
+        except EntityNotExistsError:
+            if statement.if_exists is not True:
+                raise
+        return ExecuteAnswer()
+
+    def answer_drop_tables(self, statement, database_name):
+        """answer on 'drop table [if exists] {name}'
+        Args:
+            statement: ast
+        """
+
+        for table in statement.tables:
+            if len(table.parts) > 1:
+                db_name = table.parts[0]
+                table = Identifier(parts=table.parts[1:])
+            else:
+                db_name = database_name
+
+            dn = self.session.datahub[db_name]
+            if db_name is not None:
+                dn.drop_table(table, if_exists=statement.if_exists)
+
+            elif db_name in self.session.database_controller.get_dict(filter_type="project"):
+                # TODO do we need feature: delete object from project via drop table?
+
+                project = self.session.database_controller.get_project(db_name)
+                project_tables = {
+                    key: val
+                    for key, val in project.get_tables().items()
+                    if val.get("deletable") is True
+                }
+                table_name = table.to_string()
+
+                if table_name in project_tables:
+                    self.session.model_controller.delete_model(
+                        table_name, project_name=db_name
+                    )
+                elif statement.if_exists is False:
+                    raise ExecutorException(
+                        f"Cannot delete a table from database '{db_name}': table does not exists"
+                    )
+            else:
+                raise ExecutorException(
+                    f"Cannot delete a table from database '{db_name}'"
+                )
+
+        return ExecuteAnswer()
+
+    def answer_create_view(self, statement, database_name):
+        project_name = database_name
+        # TEMP
+        if isinstance(statement.name, Identifier):
+            parts = statement.name.parts
+        else:
+            parts = statement.name.split(".")
+
+        view_name = parts[-1]
+        if len(parts) == 2:
+            project_name = parts[0]
+
+        query_str = statement.query_str
+        query = parse_sql(query_str)
+
+        if isinstance(statement.from_table, Identifier):
+            query = Select(
+                targets=[Star()],
+                from_table=NativeQuery(
+                    integration=statement.from_table, query=statement.query_str
+                ),
+            )
+            query_str = str(query)
+
+        if isinstance(query, Select):
+            # check create view sql
+            query.limit = Constant(1)
+
+            query_context_controller.set_context(
+                query_context_controller.IGNORE_CONTEXT
+            )
+            try:
+                sqlquery = SQLQuery(query, session=self.session, database=database_name)
+                if sqlquery.fetch()["success"] is not True:
+                    raise ExecutorException("Wrong view query")
+            finally:
+                query_context_controller.release_context(
+                    query_context_controller.IGNORE_CONTEXT
+                )
+
+        project = self.session.database_controller.get_project(project_name)
+        try:
+            project.create_view(view_name, query=query_str)
+        except EntityExistsError:
+            if getattr(statement, "if_not_exists", False) is False:
+                raise
+        return ExecuteAnswer()
+
+    def answer_drop_view(self, statement, database_name):
+        names = statement.names
+
+        for name in names:
+            view_name = name.parts[-1]
+            if len(name.parts) > 1:
+                db_name = name.parts[0]
+            else:
+                db_name = database_name
+            project = self.session.database_controller.get_project(db_name)
+
+            try:
+                project.drop_view(view_name)
+            except EntityNotExistsError:
+                if statement.if_exists is not True:
+                    raise
+
+        return ExecuteAnswer()
+
+    def answer_create_kb(self, statement: CreateKnowledgeBase, database_name: str):
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        if statement.storage is not None:
+            if len(statement.storage.parts) != 2:
+                raise ExecutorException(
+                    f"Invalid vectordatabase table name: {statement.storage}"
+                    "Need the form 'database_name.table_name'"
+                )
+
+        if statement.from_query is not None:
+            # TODO: implement this
+            raise ExecutorException(
+                "Create a knowledge base from a select is not supported yet"
+            )
+
+        kb_name = statement.name.parts[-1]
+
+        # create the knowledge base
+        _ = self.session.kb_controller.add(
+            name=kb_name,
+            project_name=project_name,
+            embedding_model=statement.model,
+            storage=statement.storage,
+            params=statement.params,
+            if_not_exists=statement.if_not_exists,
+        )
+
+        return ExecuteAnswer()
+
+    def answer_drop_kb(self, statement: DropKnowledgeBase, database_name: str):
+        name = statement.name.parts[-1]
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        # delete the knowledge base
+        self.session.kb_controller.delete(
+            name=name,
+            project_name=project_name,
+            if_exists=statement.if_exists,
+        )
+
+        return ExecuteAnswer()
+
+    def answer_create_skill(self, statement, database_name):
+        name = statement.name.parts[-1]
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        try:
+            _ = self.session.skills_controller.add_skill(
+                name,
+                project_name,
+                statement.type,
+                statement.params
+            )
+        except ValueError as e:
+            # Project does not exist or skill already exists.
+            raise ExecutorException(str(e))
+
+        return ExecuteAnswer()
+
+    def answer_drop_skill(self, statement, database_name):
+        name = statement.name.parts[-1]
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        try:
+            self.session.skills_controller.delete_skill(name, project_name)
+        except ValueError as e:
+            # Project does not exist or skill does not exist.
+            raise ExecutorException(str(e))
+
+        return ExecuteAnswer()
+
+    def answer_update_skill(self, statement, database_name):
+        name = statement.name.parts[-1]
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        type = statement.params.pop('type', None)
+        try:
+            _ = self.session.skills_controller.update_skill(
+                name,
+                project_name=project_name,
+                type=type,
+                params=statement.params
+            )
+        except ValueError as e:
+            # Project does not exist or skill does not exist.
+            raise ExecutorException(str(e))
+
+        return ExecuteAnswer()
+
+    def answer_create_agent(self, statement, database_name):
+        name = statement.name.parts[-1]
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        skills = statement.params.pop('skills', [])
+        provider = statement.params.pop('provider', None)
+        try:
+            _ = self.session.agents_controller.add_agent(
+                name=name,
+                project_name=project_name,
+                model_name=statement.model,
+                skills=skills,
+                provider=provider,
+                params=statement.params
+            )
+        except ValueError as e:
+            # Project does not exist or agent already exists.
+            raise ExecutorException(str(e))
+
+        return ExecuteAnswer()
+
+    def answer_drop_agent(self, statement, database_name):
+        name = statement.name.parts[-1]
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        try:
+            self.session.agents_controller.delete_agent(name, project_name)
+        except ValueError as e:
+            # Project does not exist or agent does not exist.
+            raise ExecutorException(str(e))
+
+        return ExecuteAnswer()
+
+    def answer_update_agent(self, statement, database_name):
+        name = statement.name.parts[-1]
+        project_name = (
+            statement.name.parts[0]
+            if len(statement.name.parts) > 1
+            else database_name
+        )
+
+        model = statement.params.pop('model', None)
+        skills_to_add = statement.params.pop('skills_to_add', [])
+        skills_to_remove = statement.params.pop('skills_to_remove', [])
+        try:
+            _ = self.session.agents_controller.update_agent(
+                name,
+                project_name=project_name,
+                model_name=model,
+                skills_to_add=skills_to_add,
+                skills_to_remove=skills_to_remove,
+                params=statement.params
+            )
+        except (EntityExistsError, EntityNotExistsError, ValueError) as e:
+            # Project does not exist or agent does not exist.
+            raise ExecutorException(str(e))
+
+        return ExecuteAnswer()
+
+    @mark_process("learn")
+    def answer_create_predictor(self, statement: CreatePredictor, database_name):
+        integration_name = database_name
+
+        # allow creation in non-active projects, e.g. 'create mode proj.model' works whether `proj` is active or not
+        if len(statement.name.parts) > 1:
+            integration_name = statement.name.parts[0]
+        model_name = statement.name.parts[-1]
+        statement.name.parts = [integration_name.lower(), model_name]
+
+        ml_integration_name = "lightwood"  # default
+        if statement.using is not None:
+            # repack using with lower names
+            statement.using = {k.lower(): v for k, v in statement.using.items()}
+
+            ml_integration_name = statement.using.pop("engine", ml_integration_name)
+
+        if statement.query_str is not None and statement.integration_name is None:
+            # set to current project
+            statement.integration_name = Identifier(database_name)
+
+        try:
+            ml_handler = self.session.integration_controller.get_ml_handler(
+                ml_integration_name
+            )
+        except EntityNotExistsError:
+            # not exist, try to create it with same name as handler
+            self.answer_create_ml_engine(ml_integration_name, handler=ml_integration_name)
+
+            ml_handler = self.session.integration_controller.get_ml_handler(
+                ml_integration_name
+            )
+
+        if getattr(statement, "is_replace", False) is True:
+            # try to delete
+            try:
+                self.session.model_controller.delete_model(
+                    model_name,
+                    project_name=integration_name
+                )
+            except EntityNotExistsError:
+                pass
+
+        try:
+            df = self.session.model_controller.create_model(statement, ml_handler)
+
+            return ExecuteAnswer(data=ResultSet().from_df(df))
+        except EntityExistsError:
+            if getattr(statement, "if_not_exists", False) is True:
+                return ExecuteAnswer()
+            raise
+
+    def answer_show_columns(
+        self,
+        target: Identifier,
+        where: Optional[Operation] = None,
+        like: Optional[str] = None,
+        is_full=False,
+        database_name=None,
+    ):
+        if len(target.parts) > 1:
+            db = target.parts[0]
+        elif isinstance(database_name, str) and len(database_name) > 0:
+            db = database_name
+        else:
+            db = "mindsdb"
+        table_name = target.parts[-1]
+
+        new_where = BinaryOperation(
+            "and",
+            args=[
+                BinaryOperation("=", args=[Identifier("TABLE_SCHEMA"), Constant(db)]),
+                BinaryOperation(
+                    "=", args=[Identifier("TABLE_NAME"), Constant(table_name)]
+                ),
+            ],
+        )
+        if where is not None:
+            new_where = BinaryOperation("and", args=[new_where, where])
+        if like is not None:
+            like = BinaryOperation("like", args=[Identifier("View"), Constant(like)])
+            new_where = BinaryOperation("and", args=[new_where, like])
+
+        targets = [
+            Identifier("COLUMN_NAME", alias=Identifier("Field")),
+            Identifier("COLUMN_TYPE", alias=Identifier("Type")),
+            Identifier("IS_NULLABLE", alias=Identifier("Null")),
+            Identifier("COLUMN_KEY", alias=Identifier("Key")),
+            Identifier("COLUMN_DEFAULT", alias=Identifier("Default")),
+            Identifier("EXTRA", alias=Identifier("Extra")),
+        ]
+        if is_full:
+            targets.extend(
+                [
+                    Constant("COLLATION", alias=Identifier("Collation")),
+                    Constant("PRIVILEGES", alias=Identifier("Privileges")),
+                    Constant("COMMENT", alias=Identifier("Comment")),
+                ]
+            )
+        new_statement = Select(
+            targets=targets,
+            from_table=Identifier(parts=["information_schema", "COLUMNS"]),
+            where=new_where,
+        )
+
+        query = SQLQuery(new_statement, session=self.session, database=database_name)
+        return self.answer_select(query)
+
+    def answer_single_row_select(self, statement, database_name):
+
+        def adapt_query(node, is_table, **kwargs):
+            if is_table:
+                return
+
+            if isinstance(node, Identifier):
+                if node.parts[-1].lower() == "session_user":
+                    return Constant(self.session.username, alias=node)
+                if node.parts[-1].lower() == '$$':
+                    # NOTE: sinve version 9.0 mysql client sends query 'select $$'.
+                    # Connection can be continued only if answer is parse error.
+                    raise ErParseError(
+                        "You have an error in your SQL syntax; check the manual that corresponds to your server "
+                        "version for the right syntax to use near '$$' at line 1"
+                    )
+
+            if isinstance(node, Function):
+                function_name = node.op.lower()
+
+                functions_results = {
+                    "database": database_name,
+                    "current_user": self.session.username,
+                    "user": self.session.username,
+                    "version": "8.0.17",
+                    "current_schema": "public",
+                    "connection_id": self.context.get('connection_id')
+                }
+                if function_name in functions_results:
+                    return Constant(functions_results[function_name], alias=Identifier(parts=[function_name]))
+
+            if isinstance(node, Variable):
+                var_name = node.value
+                column_name = f"@@{var_name}"
+                result = SERVER_VARIABLES.get(column_name)
+                if result is None:
+                    logger.error(f"Unknown variable: {column_name}")
+                    raise Exception(f"Unknown variable '{var_name}'")
+                else:
+                    return Constant(result[0], alias=Identifier(parts=[column_name]))
+
+        query_traversal(statement, adapt_query)
+
+        statement.from_table = Identifier('t')
+        df = query_df(pd.DataFrame([[0]]), statement, session=self.session)
+
+        return ExecuteAnswer(
+            data=ResultSet().from_df(df, table_name="")
+        )
+
+    def answer_show_create_table(self, table):
+        columns = [
+            Column(table_name="", name="Table", type=TYPES.MYSQL_TYPE_VAR_STRING),
+            Column(
+                table_name="", name="Create Table", type=TYPES.MYSQL_TYPE_VAR_STRING
+            ),
+        ]
+        return ExecuteAnswer(
+            data=ResultSet(
+                columns=columns,
+                values=[[table, f"create table {table} ()"]],
+            )
+        )
+
+    def answer_function_status(self):
+        columns = [
+            Column(
+                name="Db",
+                alias="Db",
+                table_name="schemata",
+                table_alias="ROUTINES",
+                type="str",
+                database="mysql",
+                charset=self.charset_text_type,
+            ),
+            Column(
+                name="Db",
+                alias="Db",
+                table_name="routines",
+                table_alias="ROUTINES",
+                type="str",
+                database="mysql",
+                charset=self.charset_text_type,
+            ),
+            Column(
+                name="Type",
+                alias="Type",
+                table_name="routines",
+                table_alias="ROUTINES",
+                type="str",
+                database="mysql",
+                charset=CHARSET_NUMBERS["utf8_bin"],
+            ),
+            Column(
+                name="Definer",
+                alias="Definer",
+                table_name="routines",
+                table_alias="ROUTINES",
+                type="str",
+                database="mysql",
+                charset=CHARSET_NUMBERS["utf8_bin"],
+            ),
+            Column(
+                name="Modified",
+                alias="Modified",
+                table_name="routines",
+                table_alias="ROUTINES",
+                type=TYPES.MYSQL_TYPE_TIMESTAMP,
+                database="mysql",
+                charset=CHARSET_NUMBERS["binary"],
+            ),
+            Column(
+                name="Created",
+                alias="Created",
+                table_name="routines",
+                table_alias="ROUTINES",
+                type=TYPES.MYSQL_TYPE_TIMESTAMP,
+                database="mysql",
+                charset=CHARSET_NUMBERS["binary"],
+            ),
+            Column(
+                name="Security_type",
+                alias="Security_type",
+                table_name="routines",
+                table_alias="ROUTINES",
+                type=TYPES.MYSQL_TYPE_STRING,
+                database="mysql",
+                charset=CHARSET_NUMBERS["utf8_bin"],
+            ),
+            Column(
+                name="Comment",
+                alias="Comment",
+                table_name="routines",
+                table_alias="ROUTINES",
+                type=TYPES.MYSQL_TYPE_BLOB,
+                database="mysql",
+                charset=CHARSET_NUMBERS["utf8_bin"],
+            ),
+            Column(
+                name="character_set_client",
+                alias="character_set_client",
+                table_name="character_sets",
+                table_alias="ROUTINES",
+                type=TYPES.MYSQL_TYPE_VAR_STRING,
+                database="mysql",
+                charset=self.charset_text_type,
+            ),
+            Column(
+                name="collation_connection",
+                alias="collation_connection",
+                table_name="collations",
+                table_alias="ROUTINES",
+                type=TYPES.MYSQL_TYPE_VAR_STRING,
+                database="mysql",
+                charset=self.charset_text_type,
+            ),
+            Column(
+                name="Database Collation",
+                alias="Database Collation",
+                table_name="collations",
+                table_alias="ROUTINES",
+                type=TYPES.MYSQL_TYPE_VAR_STRING,
+                database="mysql",
+                charset=self.charset_text_type,
+            ),
+        ]
+
+        return ExecuteAnswer(data=ResultSet(columns=columns))
+
+    def answer_show_table_status(self, table_name):
+        # NOTE at this moment parsed statement only like `SHOW TABLE STATUS LIKE 'table'`.
+        # NOTE some columns has {'database': 'mysql'}, other not. That correct. This is how real DB sends messages.
+        columns = [
+            {
+                "database": "mysql",
+                "table_name": "tables",
+                "name": "Name",
+                "alias": "Name",
+                "type": TYPES.MYSQL_TYPE_VAR_STRING,
+                "charset": self.charset_text_type,
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Engine",
+                "alias": "Engine",
+                "type": TYPES.MYSQL_TYPE_VAR_STRING,
+                "charset": self.charset_text_type,
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Version",
+                "alias": "Version",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "mysql",
+                "table_name": "tables",
+                "name": "Row_format",
+                "alias": "Row_format",
+                "type": TYPES.MYSQL_TYPE_VAR_STRING,
+                "charset": self.charset_text_type,
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Rows",
+                "alias": "Rows",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Avg_row_length",
+                "alias": "Avg_row_length",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Data_length",
+                "alias": "Data_length",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Max_data_length",
+                "alias": "Max_data_length",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Index_length",
+                "alias": "Index_length",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Data_free",
+                "alias": "Data_free",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Auto_increment",
+                "alias": "Auto_increment",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Create_time",
+                "alias": "Create_time",
+                "type": TYPES.MYSQL_TYPE_TIMESTAMP,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Update_time",
+                "alias": "Update_time",
+                "type": TYPES.MYSQL_TYPE_TIMESTAMP,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Check_time",
+                "alias": "Check_time",
+                "type": TYPES.MYSQL_TYPE_TIMESTAMP,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "mysql",
+                "table_name": "tables",
+                "name": "Collation",
+                "alias": "Collation",
+                "type": TYPES.MYSQL_TYPE_VAR_STRING,
+                "charset": self.charset_text_type,
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Checksum",
+                "alias": "Checksum",
+                "type": TYPES.MYSQL_TYPE_LONGLONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Create_options",
+                "alias": "Create_options",
+                "type": TYPES.MYSQL_TYPE_VAR_STRING,
+                "charset": self.charset_text_type,
+            },
+            {
+                "database": "",
+                "table_name": "tables",
+                "name": "Comment",
+                "alias": "Comment",
+                "type": TYPES.MYSQL_TYPE_BLOB,
+                "charset": self.charset_text_type,
+            },
+        ]
+        columns = [Column(**d) for d in columns]
+        data = [
+            [
+                table_name,  # Name
+                "InnoDB",  # Engine
+                10,  # Version
+                "Dynamic",  # Row_format
+                1,  # Rows
+                16384,  # Avg_row_length
+                16384,  # Data_length
+                0,  # Max_data_length
+                0,  # Index_length
+                0,  # Data_free
+                None,  # Auto_increment
+                datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S"),  # Create_time
+                datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S"),  # Update_time
+                None,  # Check_time
+                "utf8mb4_0900_ai_ci",  # Collation
+                None,  # Checksum
+                "",  # Create_options
+                "",  # Comment
+            ]
+        ]
+        return ExecuteAnswer(data=ResultSet(columns=columns, values=data))
+
+    def answer_show_warnings(self):
+        columns = [
+            {
+                "database": "",
+                "table_name": "",
+                "name": "Level",
+                "alias": "Level",
+                "type": TYPES.MYSQL_TYPE_VAR_STRING,
+                "charset": self.charset_text_type,
+            },
+            {
+                "database": "",
+                "table_name": "",
+                "name": "Code",
+                "alias": "Code",
+                "type": TYPES.MYSQL_TYPE_LONG,
+                "charset": CHARSET_NUMBERS["binary"],
+            },
+            {
+                "database": "",
+                "table_name": "",
+                "name": "Message",
+                "alias": "Message",
+                "type": TYPES.MYSQL_TYPE_VAR_STRING,
+                "charset": self.charset_text_type,
+            },
+        ]
+        columns = [Column(**d) for d in columns]
+        return ExecuteAnswer(data=ResultSet(columns=columns))
+
+    def answer_create_table(self, statement, database_name):
+        SQLQuery(statement, session=self.session, execute=True, database=database_name)
+        return ExecuteAnswer()
+
+    def answer_select(self, query):
+        data = query.fetch()
+
+        return ExecuteAnswer(data=data["result"])
+
+    def answer_update_model_version(self, model_version, database_name):
+        if not isinstance(model_version, Identifier):
+            raise ExecutorException(f'Please define version: {model_version}')
+
+        model_parts = model_version.parts
+        version = model_parts[-1]
+        if version.isdigit():
+            version = int(version)
+        else:
+            raise ExecutorException(f'Unknown version: {version}')
+
+        if len(model_parts) == 3:
+            project_name, model_name = model_parts[:2]
+        elif len(model_parts) == 2:
+            model_name = model_parts[0]
+            project_name = database_name
+        else:
+            raise ExecutorException(f'Unknown model: {model_version}')
+
+        self.session.model_controller.set_model_active_version(project_name, model_name, version)
+        return ExecuteAnswer()
+
+    def answer_drop_model(self, statement, database_name):
+
+        model_parts = statement.name.parts
+        version = None
+
+        # with version?
+        if model_parts[-1].isdigit():
+            version = int(model_parts[-1])
+            model_parts = model_parts[:-1]
+
+        if len(model_parts) == 2:
+            project_name, model_name = model_parts
+        elif len(model_parts) == 1:
+            model_name = model_parts[0]
+            project_name = database_name
+        else:
+            raise ExecutorException(f'Unknown model: {statement.name}')
+
+        if version is not None:
+            # delete version
+            try:
+                self.session.model_controller.delete_model_version(project_name, model_name, version)
+            except EntityNotExistsError as e:
+                if not statement.if_exists:
+                    raise e
+        else:
+            # drop model
+            try:
+                project = self.session.database_controller.get_project(project_name)
+                project.drop_model(model_name)
+            except Exception as e:
+                if not statement.if_exists:
+                    raise e
+
+        return ExecuteAnswer()
+
+    def change_default_db(self, db_name):
+        # That fix for bug in mssql: it keeps connection for a long time, but after some time mssql can
+        # send packet with COM_INIT_DB=null. In this case keep old database name as default.
+        if db_name != "null":
+            if self.session.database_controller.exists(db_name):
+                self.session.database = db_name
+            else:
+                raise BadDbError(f"Database {db_name} does not exists")