onetick-py 1.177.0__py3-none-any.whl

onetick/py/core/per_tick_script.py

@@ -0,0 +1,2256 @@
+import ast
+import inspect
+import textwrap
+import types
+import operator
+import tokenize
+import warnings
+from typing import Callable, Union, Any, Optional, Iterable, Type, Tuple, Dict, List, TypeVar
+from copy import deepcopy
+from functools import wraps
+
+from onetick.py.backports import astunparse, cached_property
+
+from collections import deque
+from contextlib import contextmanager
+
+from .. import types as ott
+from .column_operations.base import _Operation
+from .column import _Column
+from .lambda_object import _EmulateObject, _LambdaIfElse, _EmulateStateVars, _CompareTrackScope
+from ._internal._state_objects import (
+    _TickSequence, _TickSequenceTickBase, _TickListTick, _TickSetTick, _TickDequeTick, _DynamicTick
+)
+
+
+class Static:
+    """
+    Class for declaring static local variable in per-tick script.
+    Static variables are defined once and save their values between
+    arrival of the input ticks.
+    """
+    def __init__(self, value):
+        self.value = value
+
+    # these functions needed mostly for linters
+    def __getattr__(self, item):
+        return self.value.__getattr__(item)
+
+    def __getitem__(self, item):
+        return operator.getitem(self.value, item)
+
+    def __setitem__(self, key, value):
+        return operator.setitem(self.value, key, value)
+
+
+class LocalVariable(_Operation):
+    """
+    Class for inner representation of local variable in per-tick script.
+    Only simple values are supported, tick sequences are represented by another class.
+    """
+    def __init__(self, name, dtype=None):
+        super().__init__(op_str=f'LOCAL::{name}', dtype=dtype)
+        self.name = name
+
+
+class TickDescriptorFields(_TickSequence):
+    """
+    Class for declaring tick descriptor fields in per-tick script.
+    Can only be iterated, doesn't have methods and parameters.
+
+    See also
+    --------
+    :py:class:`TickDescriptorField <onetick.py.core.per_tick_script.TickDescriptorField>`
+
+    Examples
+    --------
+    >>> t = otp.Tick(A=1)
+    >>> def fun(tick):
+    ...     for field in otp.tick_descriptor_fields():
+    ...         tick['NAME'] = field.get_name()
+    >>> t = t.script(fun)
+    >>> otp.run(t)
+            Time  A NAME
+    0 2003-12-01  1    A
+    """
+    def __init__(self):
+        # we don't want to inherit parent class init
+        pass
+
+    def __str__(self):
+        return 'LOCAL::INPUT_TICK_DESCRIPTOR_FIELDS'
+
+    @property
+    def _tick_class(self):
+        return TickDescriptorField
+
+
+def tick_list_tick():
+    """
+    Can be used only in per-tick script function
+    to define a tick list tick local variable.
+
+    Tick list ticks can be used with some methods
+    of tick lists :py:class:`onetick.py.state.tick_list`.
+
+    See also
+    --------
+    :py:class:`onetick.py.state.tick_list`.
+
+    Returns
+    -------
+    :py:class:`onetick.py.static` value with tick object.
+
+    Examples
+    --------
+    >>> def fun(tick):
+    ...    t = otp.tick_list_tick()
+    ...    tick.state_vars['LIST'].push_back(t)
+    """
+    return Static(_TickListTick(None))
+
+
+def tick_set_tick():
+    """
+    Can be used only in per-tick script function
+    to define a tick set tick local variable.
+
+    Tick set ticks can be used with some methods
+    of tick sets :py:class:`onetick.py.state.tick_set`.
+
+    See also
+    --------
+    :py:class:`onetick.py.state.tick_set`.
+
+    Returns
+    -------
+    :py:class:`onetick.py.static` value with tick object.
+
+    Examples
+    --------
+    >>> def fun(tick):
+    ...    t = otp.tick_set_tick()
+    ...    if tick.state_vars['SET'].find(t, -1):
+    ...        tick['RES'] = '-1'
+    """
+    return Static(_TickSetTick(None))
+
+
+def tick_deque_tick():
+    """
+    Can be used only in per-tick script function
+    to define a tick deque tick local variable.
+
+    Tick deque ticks can be used with some methods
+    of tick deques :py:class:`onetick.py.state.tick_deque`.
+
+    See also
+    --------
+    :py:class:`onetick.py.state.tick_deque`.
+
+    Returns
+    -------
+    :py:class:`onetick.py.static` value with tick object.
+
+    Examples
+    --------
+    >>> def fun(tick):
+    ...    t = otp.tick_deque_tick()
+    ...    tick.state_vars['DEQUE'].get_tick(0, t)
+    """
+    return Static(_TickDequeTick(None))
+
+
+def dynamic_tick():
+    """
+    Can be used only in per-tick script function
+    to define a dynamic tick local variable.
+
+    Dynamic ticks can be used with some methods
+    of all tick sequences.
+
+    See also
+    --------
+    :py:class:`onetick.py.state.tick_list`
+    :py:class:`onetick.py.state.tick_set`
+    :py:class:`onetick.py.state.tick_deque`
+
+    Returns
+    -------
+    :py:class:`onetick.py.static` value with tick object.
+
+    Examples
+    --------
+    >>> def fun(tick):
+    ...    t = otp.dynamic_tick()
+    ...    t['X'] = tick['SUM']
+    """
+    return Static(_DynamicTick(None))
+
+
+class TickDescriptorField(_TickSequenceTickBase):
+    """
+    Tick descriptor field object.
+    Can be accessed only while iterating over
+    :py:class:`otp.tick_descriptor_fields <onetick.py.core.per_tick_script.TickDescriptorFields>`
+    in per-tick script.
+
+    Examples
+    --------
+    >>> t = otp.Tick(A=2, B='B', C=1.2345)
+    >>> def fun(tick):
+    ...     tick['NAMES'] = ''
+    ...     tick['TYPES'] = ''
+    ...     tick['SIZES'] = ''
+    ...     for field in otp.tick_descriptor_fields():
+    ...         tick['NAMES'] += field.get_name() + ','
+    ...         tick['TYPES'] += field.get_type() + ','
+    ...         tick['SIZES'] += field.get_size().apply(str) + ','
+    >>> t = t.script(fun)
+    >>> otp.run(t)
+            Time  A  B       C   NAMES                TYPES    SIZES
+    0 2003-12-01  2  B  1.2345  A,B,C,  long,string,double,  8,64,8,
+    """
+
+    _definition = 'TICK_DESCRIPTOR_FIELD'
+
+    def get_field_name(self):
+        """
+        Get the name of the field.
+
+        Returns
+        -------
+        onetick.py.Operation
+        """
+        return _Operation(op_str=f'{self}.GET_FIELD_NAME()', dtype=str)
+
+    def get_name(self):
+        """
+        Get the name of the field.
+
+        Returns
+        -------
+        onetick.py.Operation
+        """
+        return self.get_field_name()
+
+    def get_size(self):
+        """
+        Get the size of the type of the field.
+
+        Returns
+        -------
+        onetick.py.Operation
+        """
+        return _Operation(op_str=f'{self}.GET_SIZE()', dtype=int)
+
+    def get_type(self):
+        """
+        Get the name of the type of the field.
+
+        Returns
+        -------
+        onetick.py.Operation
+        """
+        return _Operation(op_str=f'{self}.GET_TYPE()', dtype=str)
+
+
+class Expression:
+    """
+    Class to save per-tick-script expressions along with their possible values.
+
+    Parameters
+    ----------
+    expr
+        string expression that will be saved to per tick script
+    values:
+        values that this expression can take.
+        For example, bool operation can take many values.
+    lhs:
+        True if expression is left hand expression.
+        In this case value of expression must be callable.
+        Calling it with right hand expression value as an argument
+        should be the same as execute the whole expression.
+    """
+    def __init__(self, *values: Any, expr: Optional[str] = None, lhs: bool = False):
+        self.values = values
+        self._expr = expr
+        self.lhs = lhs
+        if self.lhs:
+            assert callable(self.value)
+            assert self.expr
+
+    @property
+    def expr(self):
+        if self._expr:
+            return self._expr
+        if self.is_emulator:
+            self._expr = 'LOCAL::OUTPUT_TICK'
+        elif self.is_column:
+            self._expr = str(self.value)
+        elif self.values:
+            self._expr = self.value_to_onetick(self.value)
+        return self._expr
+
+    @property
+    def value(self):
+        length = len(self.values)
+        if length == 0:
+            raise ValueError(f"Expression '{self}' doesn't have values.")
+        if length > 1:
+            raise ValueError(f"Expression '{self}' have more than one value.")
+        return self.values[0]
+
+    @cached_property
+    def dtype(self):
+        return ott.get_type_by_objects(self.values)
+
+    @property
+    def is_emulator(self) -> bool:
+        try:
+            return isinstance(self.value, _EmulateObject)
+        except ValueError:
+            return False
+
+    @property
+    def is_state_vars(self) -> bool:
+        try:
+            return isinstance(self.value, _EmulateStateVars)
+        except ValueError:
+            return False
+
+    @property
+    def is_static(self) -> bool:
+        try:
+            return isinstance(self.value, Static)
+        except ValueError:
+            return False
+
+    @property
+    def is_dynamic_tick(self) -> bool:
+        try:
+            return isinstance(self.value, _DynamicTick)
+        except ValueError:
+            return False
+
+    @property
+    def is_tick(self) -> bool:
+        try:
+            return isinstance(self.value, _TickSequenceTickBase)
+        except ValueError:
+            return False
+
+    @property
+    def is_column(self) -> bool:
+        try:
+            return isinstance(self.value, _Column)
+        except ValueError:
+            return False
+
+    @property
+    def is_local_variable(self) -> bool:
+        try:
+            return isinstance(self.value, LocalVariable)
+        except ValueError:
+            return False
+
+    @property
+    def is_operation(self) -> bool:
+        try:
+            return isinstance(self.value, _Operation)
+        except ValueError:
+            return False
+
+    @property
+    def predefined(self) -> bool:
+        """Check if the value of expression is known before the execution of query"""
+        return not self.is_operation
+
+    @property
+    def expressible(self) -> bool:
+        return bool(self.expr)
+
+    def __str__(self):
+        if not self.expressible:
+            raise ValueError("This Expression can't be expressed in OneTick or is undefined yet")
+        return self.expr
+
+    def convert_to_operation(self):
+        """
+        Convert otp.Column to otp.Operation.
+        Needed to convert expressions like:
+            if tick['X']:
+        to
+            if (X != 0) {
+        """
+        if self.is_column:
+            self.values = [self.value._make_python_way_bool_expression()]
+            self._expr = str(self.value)
+
+    @staticmethod
+    def value_to_onetick(value: Union[str, int, float, bool, None, _Operation]) -> str:
+        """
+        Python value will be converted accordingly to OneTick syntax
+        (lowercase boolean values, string in quotes, etc.)
+        """
+        if value is None:
+            return str(ott.nan)
+        if isinstance(value, bool):
+            return str(value).lower()
+        return ott.value2str(value)
+
+
+class CaseOperatorParser:
+    """
+    Class with methods to convert ast operators to their string or python representations.
+    Only ast operators that can be used in OneTick's CASE function are accepted.
+    """
+
+    @staticmethod
+    def py_operator(op: Union[ast.operator, ast.cmpop, ast.unaryop, ast.boolop]) -> Callable:
+        """
+        Convert ast operator to python function for this operator.
+
+        Parameters
+        ----------
+        op
+            ast operator object
+        """
+        return {
+            # binary
+            ast.Add: operator.add,
+            ast.Sub: operator.sub,
+            ast.Mult: operator.mul,
+            ast.Div: operator.truediv,
+            ast.BitAnd: operator.and_,
+            ast.BitOr: operator.or_,
+            ast.Mod: operator.mod,
+            # unary
+            ast.UAdd: operator.pos,
+            ast.USub: operator.neg,
+            ast.Not: operator.not_,
+            ast.Invert: operator.invert,
+            # compare
+            ast.Lt: operator.lt,
+            ast.LtE: operator.le,
+            ast.Gt: operator.gt,
+            ast.GtE: operator.ge,
+            ast.Eq: operator.eq,
+            ast.NotEq: operator.ne,
+            # bool
+            ast.And: lambda x, y: x and y,
+            ast.Or: lambda x, y: x or y,
+        }[type(op)]  # type: ignore[return-value]
+
+    @staticmethod
+    def operator(op: Union[ast.operator, ast.cmpop, ast.unaryop, ast.boolop]) -> str:
+        """
+        Convert ast operator to OneTick's string representation.
+
+        Parameters
+        ----------
+        op
+            ast operator object
+        """
+        return {
+            # binary
+            ast.Add: '+',
+            ast.Sub: '-',
+            ast.Mult: '*',
+            ast.Div: '/',
+            ast.Mod: '%',
+            # unary
+            ast.UAdd: '+',
+            ast.USub: '-',
+            # compare
+            ast.Lt: '<',
+            ast.LtE: '<=',
+            ast.Gt: '>',
+            ast.GtE: '>=',
+            ast.Eq: '=',
+            ast.NotEq: '!=',
+            # bool
+            ast.And: 'AND',
+            ast.Or: 'OR',
+        }[type(op)]
+
+
+class OperatorParser(CaseOperatorParser):
+    """
+    Class with methods to convert ast operators to their string or python representations.
+    Only ast operators that can be used in OneTick's per tick script are accepted.
+    """
+
+    @staticmethod
+    def py_operator(op: Union[ast.operator, ast.cmpop, ast.unaryop, ast.boolop],
+                    aug: bool = False, **kwargs) -> Callable:
+        """
+        Convert ast operator to python function for this operator.
+
+        Parameters
+        ----------
+        op
+            ast operator object
+        aug
+            ast don't have separate inplace operators (+=, -=, etc.)
+            If this parameter is True then operator is inplace and otherwise if False.
+        """
+        if aug:
+            assert isinstance(op, ast.operator)
+            return {
+                ast.Add: operator.iadd,
+                ast.Sub: operator.isub,
+                ast.Mult: operator.imul,
+                ast.Div: operator.itruediv,
+            }[type(op)]
+        return CaseOperatorParser.py_operator(op, **kwargs)
+
+    @staticmethod
+    def operator(op: Union[ast.operator, ast.cmpop, ast.unaryop, ast.boolop],
+                 aug: bool = False) -> str:
+        """
+        Convert ast operator to its string representation.
+
+        Parameters
+        ----------
+        op
+            ast operator object
+        aug
+            ast don't have separate inplace binary operators (+=, -=, etc.)
+            If this parameter is True then parameter is inplace and otherwise if False.
+        """
+        if aug:
+            assert isinstance(op, ast.operator)
+            return {
+                ast.Add: '+=',
+                ast.Sub: '-=',
+                ast.Mult: '*=',
+                ast.Div: '/=',
+            }[type(op)]
+        try:
+            return {
+                ast.Eq: '==',
+                ast.And: '&&',
+                ast.Or: '||',
+            }[type(op)]
+        except KeyError:
+            return CaseOperatorParser.operator(op)
+
+
+class ExpressionParser:
+    """
+    Class with methods to convert ast expressions to OneTick's script or function syntax.
+    """
+    def __init__(self, fun: 'FunctionParser'):
+        self.fun = fun
+        self.operator_parser = OperatorParser()
+
+    @contextmanager
+    def _replace_context(self, closure_vars: inspect.ClosureVars):
+        """
+        Temporarily change closure variables in self.fun.
+        Variables will be replaced with those from closure_vars parameter.
+        """
+        nonlocals, globals_, *_ = closure_vars
+        assert isinstance(self.fun.closure_vars.globals, dict)
+        assert isinstance(self.fun.closure_vars.nonlocals, dict)
+        saved_globals = self.fun.closure_vars.globals.copy()
+        saved_nonlocals = self.fun.closure_vars.nonlocals.copy()
+        self.fun.closure_vars.globals.update(globals_)
+        self.fun.closure_vars.nonlocals.update(nonlocals)
+        yield
+        self.fun.closure_vars.globals.update(saved_globals)
+        self.fun.closure_vars.nonlocals.update(saved_nonlocals)
+
+    def constant(self, expr: ast.Constant) -> Expression:
+        """Some basic constant value: string, integer, float."""
+        return Expression(expr.value)
+
+    def string(self, expr: "ast.Str") -> Expression:
+        """String (for backward compatibility with Python 3.7)."""
+        return Expression(expr.s)
+
+    def number(self, expr: "ast.Str") -> Expression:
+        """Number (for backward compatibility with Python 3.7)."""
+        return Expression(expr.n)
+
+    def name(self, expr: ast.Name) -> Expression:
+        """
+        Name of the variable.
+        Every variable in per-tick script function, if defined correctly,
+        is considered to be local per-tick script variable.
+        If variable with this name is not found it will be captured from function context.
+        """
+        if self.fun.arg_name and expr.id == self.fun.arg_name:
+            value = self.fun.emulator if self.fun.emulator is not None else expr.id
+            return Expression(value)
+
+        if not isinstance(expr.ctx, ast.Load):
+            # local variable, left-hand side
+            return Expression(LocalVariable(expr.id))
+
+        for dict_name in ('LOCAL_VARS', 'STATIC_VARS'):
+            # local or static variable, right-hand side
+            variables = getattr(self.fun.emulator, dict_name, {})
+            if expr.id in variables:
+                dtype = ott.get_type_by_objects([variables[expr.id]])
+                if issubclass(dtype, _TickSequenceTickBase):
+                    # ticks have schema, owner and methods, so using saved value
+                    return Expression(variables[expr.id])
+                return Expression(LocalVariable(expr.id, dtype))
+
+        if not self.fun._from_args_annotations and expr.id in self.fun.args_annotations:
+            # parameter of the per-tick script function
+            return Expression(_Operation(op_str=expr.id, dtype=self.fun.args_annotations[expr.id]))
+
+        # pylint: disable-next=eval-used
+        value = eval(expr.id, self.fun.closure_vars.globals, self.fun.closure_vars.nonlocals)  # type: ignore[arg-type]
+        return Expression(value)
+
+    def index(self, expr: ast.Index) -> Expression:
+        """Proxy object in ast.Subscript in python < 3.9"""
+        return self.expression(expr.value)  # type: ignore[attr-defined]
+
+    def slice(self, expr: ast.Slice) -> Expression:
+        """
+        Slice of the list.
+        For example:
+            a = [1, 2, 3, 4]
+            a[2:4]
+        Here, 2:4 is the slice.
+        """
+        lower = self.expression(expr.lower).value if expr.lower else None
+        upper = self.expression(expr.upper).value if expr.upper else None
+        step = self.expression(expr.step).value if expr.step else None
+        return Expression(slice(lower, upper, step))
+
+    def subscript(self, expr: ast.Subscript) -> Expression:
+        """
+        Expression like: tick['X'].
+        Setting items of ticks and state variables is supported.
+        Getting items supported for any captured variable.
+        """
+        val = self.expression(expr.value)
+        item = self.expression(expr.slice)
+
+        if isinstance(expr.ctx, ast.Load):
+            v = val.value[item.value]
+            return Expression(v)
+
+        # index of per tick script function parameter or tick sequence tick is column name
+        if not (val.is_emulator or val.is_tick or val.is_state_vars):
+            raise ValueError(f"Setting items supported only for "
+                             f"'{self.fun.arg_name}' function argument, "
+                             f"tick sequences' ticks and state variables object")
+
+        return Expression(
+            lambda rhs: val.value.__setitem__(item.value, rhs),
+            expr=item.value,
+            lhs=True,
+        )
+
+    def attribute(self, expr: ast.Attribute) -> Expression:
+        """
+        Expression like: tick.X
+        For now we only support setting attributes of first function parameter.
+        Getting attributes supported for any captured variable.
+        """
+        val = self.expression(expr.value)
+        attr = expr.attr
+
+        if isinstance(expr.ctx, ast.Load):
+            v = getattr(val.value, attr)
+            return Expression(v)
+
+        # attribute of per tick script function parameter or tick sequence tick is column name
+        if not (val.is_emulator or val.is_tick):
+            raise ValueError(f"Setting attributes supported only for "
+                             f"'{self.fun.arg_name}' function argument and tick sequences' ticks")
+
+        return Expression(
+            lambda rhs: val.value.__setattr__(attr, rhs),
+            expr=attr,
+            lhs=True,
+        )
+
+    def bin_op(self, expr: ast.BinOp) -> Expression:
+        """
+        Binary operation expression: 2 + 2, tick['X'] * 2, etc.
+        """
+        left = self.expression(expr.left)
+        py_op = self.operator_parser.py_operator(expr.op)
+        right = self.expression(expr.right)
+        value = py_op(left.value, right.value)
+        return Expression(value)
+
+    def unary_op(self, expr: ast.UnaryOp) -> Expression:
+        """
+        Unary operation expression: -1, -tick['X'], not tick['X'], ~tick['X'], etc.
+        """
+        py_op = self.operator_parser.py_operator(expr.op)
+        operand = self.expression(expr.operand)
+        if operand.is_operation:
+            # special case for negative otp.Columns and otp.Operations
+            if isinstance(expr.op, (ast.Not, ast.Invert)):
+                operand.convert_to_operation()
+            if isinstance(expr.op, ast.Not):
+                py_op = self.operator_parser.py_operator(ast.Invert())
+        value = py_op(operand.value)
+        return Expression(value)
+
+    def bool_op(self, expr: ast.BoolOp) -> Expression:
+        """
+        Bool operation expression: True and tick['X'], etc.
+        Note that
+            * all python values will be checked inplace and will not be written to the script
+            * short-circuit logic will work for python values
+        For example:
+            True   and  0     and tick['X'] == 1  -------> false
+            'true' or   False or  tick['X'] == 1  -------> true
+            True   and  True  and tick['X'] == 1  -------> X == 1
+        """
+        value = None
+        for e in expr.values:
+            expression = self.expression(e)
+            expression.convert_to_operation()
+            v = expression.value
+
+            if not expression.is_operation:
+                # short-circuit logic, return as early as possible
+                if isinstance(expr.op, ast.And) and not v:
+                    # TODO: return v, not True or False
+                    # TODO: there can be many values if operations are present
+                    value = False
+                    break
+                if isinstance(expr.op, ast.Or) and v:
+                    value = True
+                    break
+                continue
+
+            if value is None:
+                value = v
+                continue
+
+            if isinstance(value, _Operation) or expression.is_operation:
+                # change operator for operations
+                py_op = self.operator_parser.py_operator({
+                    ast.And: ast.BitAnd(),
+                    ast.Or: ast.BitOr(),
+                }[type(expr.op)])
+            else:
+                py_op = self.operator_parser.py_operator(expr.op)
+
+            value = py_op(value, v)
+        return Expression(value)
+
+    def _convert_in_to_bool_op(self, expr: ast.Compare) -> Union[ast.Compare, ast.BoolOp]:
+        """
+        Convert expressions like:
+            tick['X']     in [1, 2]   ----->   tick['X'] == 1  or tick['X'] == 2
+            tick['X'] not in [1, 2]   ----->   tick['X'] != 1 and tick['X'] != 2
+        """
+        left, op, right = expr.left, expr.ops[0], expr.comparators[0]
+        if not isinstance(op, (ast.In, ast.NotIn)):
+            return expr
+
+        assert len(expr.ops) == 1
+        assert len(expr.comparators) == 1
+
+        right_value = self.expression(right).value
+        if isinstance(right_value, range) and right_value.step == 1:
+            # replace tick['X'] in range(5, 10) to X >=5 AND X < 10
+            return ast.BoolOp(
+                op=ast.And(),
+                values=[
+                    ast.Compare(left=left, ops=[ast.GtE()], comparators=[ast.Constant(right_value.start)]),
+                    ast.Compare(left=left, ops=[ast.Lt()], comparators=[ast.Constant(right_value.stop)]),
+                ]
+            )
+
+        right_values = [ast.Constant(r) for r in right_value]
+
+        bool_op: ast.boolop
+        compare_op: ast.cmpop
+        if isinstance(op, ast.In):
+            bool_op, compare_op = ast.Or(), ast.Eq()
+        else:
+            bool_op, compare_op = ast.And(), ast.NotEq()
+
+        values: List[ast.expr] = [
+            ast.Compare(left=left, ops=[compare_op], comparators=[r])
+            for r in right_values
+        ]
+        return ast.BoolOp(op=bool_op, values=values)
+
+    def _convert_many_comparators_to_bool_op(self, expr: ast.Compare) -> Union[ast.Compare, ast.BoolOp]:
+        """
+        OneTick don't support compare expressions with many comparators
+        so replacing them with several simple expressions.
+        For example:
+            1 < tick['X'] < 3,   ----->   tick['X'] > 1 AND tick['X'] < 3
+        """
+        if len(expr.comparators) == 1 and len(expr.ops) == 1:
+            return expr
+
+        comparators = []
+        comparators.append(expr.left)
+        ops = []
+        for op, right in zip(expr.ops, expr.comparators):
+            ops.append(op)
+            comparators.append(right)
+
+        bool_operands: List[ast.expr] = []
+        for i in range(len(comparators) - 1):
+            left, op, right = comparators[i], ops[i], comparators[i + 1]
+            bool_operands.append(
+                ast.Compare(left=left, ops=[op], comparators=[right])
+            )
+        return ast.BoolOp(op=ast.And(), values=bool_operands)
+
+    def compare(self, expr: ast.Compare) -> Expression:
+        """
+        Compare operation expression: tick['X'] > 1, 1 < 2 < 3, tick['X'] in [1, 2] etc.
+        """
+        if len(expr.ops) > 1:
+            return self.expression(
+                self._convert_many_comparators_to_bool_op(expr)
+            )
+
+        op = expr.ops[0]
+        if isinstance(op, (ast.In, ast.NotIn)):
+            return self.expression(
+                self._convert_in_to_bool_op(expr)
+            )
+
+        left = self.expression(expr.left)
+        right = self.expression(expr.comparators[0])
+
+        py_op = self.operator_parser.py_operator(op)
+        value = py_op(left.value, right.value)
+        return Expression(value)
+
+    def keyword(self, expr: ast.keyword) -> Tuple[str, Any]:
+        """
+        Keyword argument expression from function call: func(key=value).
+        Not converted to per tick script in any way, needed only in self.call() function.
+        """
+        arg = expr.arg
+        val = self.expression(expr.value)
+        assert arg is not None
+        return arg, val.value
+
+    def call(self, expr: ast.Call) -> Expression:
+        """
+        Any call expression, like otp.nsectime().
+        The returned value of the call will be inserted in script.
+        """
+        func = self.expression(expr.func)
+
+        # TODO: refactor, this code is copy-pasted from CaseExpressionParser.call()
+        need_to_parse = False
+        if not isinstance(func.value, types.BuiltinMethodType) and expr.args:
+            node = expr.args[0]
+            if isinstance(node, ast.Name) and node.id == self.fun.arg_name:
+                # we will parse inner function call to OneTick per-tick script function
+                # only if one of the function call arguments is 'tick' parameter of the original function
+                need_to_parse = True
+
+        if need_to_parse:
+            err = None
+            func_name = func.value.__name__
+            if str(func.value) not in self.fun.emulator.FUNCTIONS:
+                try:
+                    fp = FunctionParser(func.value, emulator=self.fun.emulator,
+                                        check_arg_name=self.fun.arg_name, inner_function=True)
+                except Exception as e:
+                    err = e
+                else:
+                    value = fp.per_tick_script()
+                    self.fun.emulator.FUNCTIONS[str(func.value)] = (value, fp.args_annotations, fp.return_annotation)
+            if err is None:
+                _, args_annotations, return_type = self.fun.emulator.FUNCTIONS[str(func.value)]
+                if expr.keywords:
+                    raise ValueError(
+                        f"Passing keyword parameters to per-tick script function '{func_name}' is not supported"
+                    )
+                args = []
+                for arg, (name, dtype) in zip(expr.args[1:], args_annotations.items()):
+                    if isinstance(arg, ast.Starred):
+                        raise ValueError(
+                            f"Passing starred parameter to per-tick script function '{func_name}' is not supported"
+                        )
+                    arg_expr = self.expression(arg)
+                    msg = (f"In function '{func_name}' parameter '{name}'"
+                           f" has type annotation '{dtype.__name__}',"
+                           f" but the type of passed argument is '{arg_expr.dtype.__name__}'")
+                    try:
+                        widest_type = ott.get_type_by_objects([arg_expr.dtype, dtype])
+                    except TypeError as e:
+                        raise TypeError(msg) from e
+                    if widest_type is not arg_expr.dtype:
+                        raise TypeError(msg)
+                    args.append(arg_expr)
+                str_args = ', '.join(map(str, args))
+                return Expression(_Operation(op_str=f'{func_name}({str_args})', dtype=return_type))
+
+        args = []
+        for arg in expr.args:
+            # TODO: support starred in CaseExpressionParser.call()
+            if isinstance(arg, ast.Starred):
+                args.extend(self.expression(arg.value).value)
+            else:
+                args.append(self.expression(arg).value)
+        keywords = dict(self.keyword(keyword) for keyword in expr.keywords)
+        value = func.value(*args, **keywords)
+        return Expression(value)
+
+    def formatted_value(self, expr: ast.FormattedValue) -> Expression:
+        """
+        Block from the f-string in curly brackets, e.g.
+            {tick['A']} and {123} in f"{tick['A']} {123}"
+        """
+        return self.expression(expr.value)
+
+    def joined_str(self, expr: ast.JoinedStr) -> Expression:
+        """
+        F-string expression, like: f"{tick['A']} {123}"
+        """
+        expressions = [self.expression(value) for value in expr.values]
+        value = None
+        for expression in expressions:
+            v = expression.value
+            if expression.is_operation:
+                v = v.apply(str)
+            else:
+                v = str(v)
+            if value is None:
+                value = v
+            else:
+                value = value + v
+        return Expression(value)
+
+    def list(self, expr: ast.List) -> Expression:
+        """
+        List expression, like: [1, 2, 3, 4, 5]
+        """
+        value = []
+        for e in expr.elts:
+            if isinstance(e, ast.Starred):
+                value.extend(self.expression(e.value).value)
+            else:
+                value.append(self.expression(e).value)
+        return Expression(value, expr=None)
+
+    def tuple(self, expr: ast.Tuple) -> Expression:
+        """
+        Tuple expression, like: (1, 2, 3, 4, 5)
+        """
+        expression = self.list(expr)  # type: ignore[arg-type]
+        expression.values = (tuple(expression.value),)
+        return expression
+
+    @property
+    def _expression(self) -> dict:
+        """Mapping from ast expression to parser functions"""
+        mapping = {
+            ast.Constant: self.constant,
+            ast.Name: self.name,
+            ast.Attribute: self.attribute,
+            ast.Index: self.index,
+            ast.Subscript: self.subscript,
+            ast.BinOp: self.bin_op,
+            ast.UnaryOp: self.unary_op,
+            ast.BoolOp: self.bool_op,
+            ast.Compare: self.compare,
+            ast.Call: self.call,
+            ast.FormattedValue: self.formatted_value,
+            ast.JoinedStr: self.joined_str,
+            ast.List: self.list,
+            ast.Tuple: self.tuple,
+            ast.Slice: self.slice,
+        }
+        deprecated = {
+            'NameConstant': self.constant,
+            'Str': self.string,
+            'Num': self.number,
+        }
+        with warnings.catch_warnings():
+            warnings.simplefilter('ignore', DeprecationWarning)
+            for name, callback in deprecated.items():
+                if hasattr(ast, name):
+                    mapping[getattr(ast, name)] = callback
+        return mapping
+
+    def expression(self, expr: ast.expr) -> Expression:
+        """Return parsed expression according to its type."""
+        return self._expression[type(expr)](expr)
+
+
+class CaseExpressionParser(ExpressionParser):
+    """
+    Class with methods to convert ast expressions to CASE function.
+    """
+    def __init__(self, fun: 'FunctionParser'):
+        super().__init__(fun)
+        self.operator_parser = CaseOperatorParser()  # type: ignore[assignment]
+
+    def _convert_bool_op_to_if_expr(self, expr: ast.expr) -> ast.expr:
+        """
+        Special case to convert bool operation to if expression.
+        For example:
+            lambda row: row['A'] or -1
+        will be converted to:
+            CASE(A != 0, 1, A, -1)
+        """
+        if not isinstance(expr, ast.BoolOp):
+            return expr
+
+        def get_if_expr(first, second):
+            if isinstance(expr.op, ast.Or):
+                return ast.IfExp(test=first, body=first, orelse=second)
+            else:
+                return ast.IfExp(test=first, body=second, orelse=first)
+
+        first = None
+        for i in range(len(expr.values) - 1):
+            if first is None:
+                first = expr.values[i]
+                first = self._convert_bool_op_to_if_expr(first)
+            second = expr.values[i + 1]
+            second = self._convert_bool_op_to_if_expr(second)
+            first = get_if_expr(first, second)
+        assert first is not None
+        return first
+
+    def if_expr(self, expr: ast.IfExp) -> Expression:
+        """
+        If expression: 'A' if tick['X'] > 0 else 'B'.
+        Do not confuse with if statement.
+        Will be converted to OneTick case function: CASE(X > 0, 1, 'A', 'B').
+        If condition value can be deduced before execution of script,
+        then if or else value will be returned without using CASE() function.
+        For example:
+            tick['A'] if False else 3  -----------> 3
+        """
+        test = self.expression(expr.test)
+        if test.predefined:
+            # we can remove unnecessary branch if condition value is already known
+            if test.value:
+                return self.expression(expr.body)
+            return self.expression(expr.orelse)
+        body = self.expression(expr.body)
+        orelse = self.expression(expr.orelse)
+        test.convert_to_operation()
+
+        str_expr = f'CASE({test}, 1, {body}, {orelse})'
+        value = _LambdaIfElse(str_expr, ott.get_type_by_objects([*body.values, *orelse.values]))
+        return Expression(value, expr=str_expr)
+
+    def call(self, expr: ast.Call) -> Expression:
+        """
+        For CASE() function we support using inner functions that return valid case expression.
+        """
+        func = self.expression(expr.func)
+
+        need_to_parse = False
+        if not isinstance(func.value, types.BuiltinMethodType):
+            for node in expr.args + [kw.value for kw in expr.keywords]:
+                if isinstance(node, ast.Name) and node.id == self.fun.arg_name:
+                    # we will parse inner function call to OneTick expression
+                    # only if one of the function call arguments is
+                    # 'tick' or 'row' parameter of the original function
+                    need_to_parse = True
+                    break
+        orig_err = None
+        if not need_to_parse:
+            with _CompareTrackScope(emulation_enabled=False):
+                try:
+                    return super().call(expr)
+                except Exception as err:
+                    orig_err = err
+                    uw = UserWarning(
+                        f"Function '{astunparse(expr)}' can't be called in python, "
+                        "will try to parse it to OneTick expression. "
+                        f"Use '{self.fun.arg_name}' in function's signature to indicate "
+                        "that this function can be parsed to OneTick expression."
+                    )
+                    uw.__cause__ = err
+                    warnings.warn(uw)
+
+        fp = FunctionParser(func.value, check_arg_name=False)
+
+        kwargs = {}
+        args = fp.ast_node.args.args
+        if fp.is_method:
+            args = args[1:]
+        for arg, default in zip(reversed(args), reversed(fp.ast_node.args.defaults)):
+            kwargs[arg.arg] = default
+        kwargs.update({keyword.arg: keyword.value for keyword in expr.keywords})  # type: ignore[misc]
+        for arg, arg_value in zip(args, expr.args):
+            kwargs[arg.arg] = arg_value
+
+        try:
+            value = fp.compress()
+        except Exception as err:
+            try:
+                return super().call(expr)
+            except Exception:
+                if orig_err is not None:
+                    raise err from orig_err
+                raise ValueError(
+                    f"Can't convert function '{astunparse(expr)}' to CASE() expression."
+                ) from err
+        try:
+            with self._replace_context(fp.closure_vars):
+                # replace function parameters with calculated values
+                value = fp.case_statement_parser._replace_nodes(value, replace_name=kwargs)
+                return self.expression(value)
+        except Exception as err:
+            if orig_err is not None:
+                raise err from orig_err
+            raise err
+
+    @property
+    def _expression(self) -> dict:
+        return dict(super()._expression.items() | {
+            ast.IfExp: self.if_expr,
+        }.items())
+
+
+class CaseStatementParser:
+    """
+    Class with methods to convert ast statements to CASE function.
+    """
+    def __init__(self, fun: 'FunctionParser'):
+        self.fun = fun
+        self.expression_parser = CaseExpressionParser(fun=fun)
+        self.operator_parser = CaseOperatorParser()
+
+    T = TypeVar('T', bound=ast.AST)
+
+    @staticmethod
+    def _replace_nodes(node: T,
+                       replace_name: Optional[Dict[str, ast.expr]] = None,
+                       replace_break: Union[ast.stmt, Exception, Type[Exception], None] = None,
+                       inplace: bool = False) -> T:
+        """
+        Function to replace expressions and statements inside ast.For node.
+
+        Parameters
+        ----------
+        node
+            ast node in which expressions and statements will be replaced
+        inplace
+            if True `node` object will be modified else it will be copied and the copy will be returned
+        replace_name
+            mapping from ast.Name ids to ast expressions.
+            ast.Name nodes with these ids will be replaced with corresponding expressions.
+        replace_break
+            replace break statement with another statement.
+            We can't execute for loop on real data here so we can't allow break statements at all.
+            So we will replace them with statements from code after the for loop.
+            If replace_break is Exception then exception will be raised when visiting ast.Break nodes.
+        """
+        class RewriteName(ast.NodeTransformer):
+            def visit_Name(self, n: ast.Name):
+                return (replace_name or {}).get(n.id) or n
+
+            def visit_Continue(self, n: ast.Continue):
+                # TODO: pass is not continue, we must allow only bodies with one statement in this case
+                return ast.Pass()
+
+            def visit_Break(self, n: ast.Break):
+                if replace_break is None:
+                    return n
+                if inspect.isclass(replace_break) and issubclass(replace_break, Exception):
+                    raise replace_break("Break is found in for loop and replacer is not provided")
+                if isinstance(replace_break, Exception):
+                    raise replace_break
+                assert isinstance(replace_break, ast.stmt)
+                return CaseStatementParser._replace_nodes(replace_break, replace_name=replace_name)
+
+        if not inplace:
+            node = deepcopy(node)
+        RewriteName().visit(node)
+        return node
+
+    def _flatten_for_stmt(self,
+                          stmt: ast.For,
+                          replace_break: Union[ast.stmt, Exception, Type[Exception], None] = None,
+                          stmt_after_for: Optional[ast.stmt] = None) -> List[ast.stmt]:
+        """
+        Convert for statement to list of copy-pasted statements from the body for each iteration.
+        """
+        stmts = []
+        target = stmt.target
+        assert isinstance(target, (ast.Name, ast.Tuple)), (
+            f"Unsupported expression '{astunparse(target)}' is used in for statement."
+            " Please, use variable or tuple of variables instead."
+        )
+        if isinstance(target, ast.Tuple):
+            for t in target.elts:
+                assert isinstance(t, ast.Name)
+            targets = target.elts
+        else:
+            targets = [target]
+        replace_name = {}
+        iter_ = self.expression_parser.expression(stmt.iter)
+        for iter_value in iter_.value:
+            if not isinstance(iter_value, Iterable) or isinstance(iter_value, str):
+                iter_value = [iter_value]
+            replace_name = {
+                target.id: ast.Constant(value)  # type: ignore[attr-defined]
+                for target, value in zip(targets, iter_value)
+            }
+            for s in stmt.body:
+                stmts.append(
+                    self._replace_nodes(s,
+                                        replace_name=replace_name,  # type: ignore[arg-type]
+                                        replace_break=replace_break)
+                )
+        if stmt_after_for and replace_name:
+            stmts.append(self._replace_nodes(stmt_after_for, replace_name=replace_name))  # type: ignore[arg-type]
+        return stmts
+
+    def _flatten_for_stmts(self, stmts: List[ast.stmt]) -> List[Union[ast.If, ast.Return, ast.Pass]]:
+        """
+        Find ast.For statements in list of statements and flatten them.
+        Return list of statements without ast.For.
+        Additionally raise exception if unsupported statement is found.
+        """
+        # TODO: support ast.For statements on deeper levels
+        res_stmts = []
+        for i, stmt in enumerate(stmts):
+            if not isinstance(stmt, (ast.If, ast.Return, ast.For, ast.Pass)):
+                raise ValueError(
+                    "this function can't be converted to CASE function, "
+                    "only for, if, return and pass statements are allowed"
+                )
+            if isinstance(stmt, ast.For):
+                try:
+                    res_stmts.extend(self._flatten_for_stmt(stmt, replace_break=ValueError))
+                except ValueError:
+                    stmts_left = len(stmts[i + 1:])
+                    assert stmts_left in (0, 1), "Can't be more than one statement after break"
+                    replace_break: Union[ast.Return, ast.Pass]
+                    if stmts_left == 0:
+                        stmt_after_for = None
+                        replace_break = ast.Pass()
+                    else:
+                        stmt_after_for = stmts[i + 1]
+                        assert isinstance(stmt_after_for, (ast.Return, ast.Pass)), (
+                            'Can only use pass and return statements after for loop with break'
+                        )
+                        replace_break = stmt_after_for
+                    res_stmts.extend(self._flatten_for_stmt(stmt,
+                                                            replace_break=replace_break,
+                                                            stmt_after_for=stmt_after_for))
+                    break
+            else:
+                res_stmts.append(stmt)
+        return res_stmts  # type: ignore[return-value]
+
+    def _compress_stmts_to_one_stmt(self, stmts: List[ast.stmt], filler=None) -> Union[ast.If, ast.Return]:
+        """
+        List of if statements will be converted to one if statement.
+        For example:
+            if tick['X'] <= 1:
+                if tick['X'] > 0:
+                    return 1
+                else:
+                    pass
+            else:
+                if tick['X'] < 3:
+                    return 2
+            if tick['X'] <= 3:
+                return 3
+            return 4
+        will be converted to:
+            if tick['X'] <= 1:
+                if tick['X'] > 0:
+                    return 1
+                else:
+                    if tick['X'] <= 3:
+                        return 3
+                    else:
+                        return 4
+            else:
+                if tick['X'] < 3:
+                    return 2
+                else:
+                    if tick['X'] <= 3:
+                        return 3
+                    else:
+                        return 4
+        """
+        filler = filler or ast.Pass()
+        if not stmts:
+            return filler
+        stmt, *others = stmts
+        if isinstance(stmt, ast.Return):
+            return stmt
+        if isinstance(stmt, ast.Pass):
+            return filler
+        filler = self._compress_stmts_to_one_stmt(others, filler=filler)
+        if isinstance(stmt, ast.If):
+            stmt.body = [self._compress_stmts_to_one_stmt(stmt.body, filler=filler)]
+            if stmt.orelse:
+                stmt.orelse = [self._compress_stmts_to_one_stmt(stmt.orelse, filler=filler)]
+            elif filler:
+                stmt.orelse = [filler]
+            assert stmt.orelse
+            return stmt
+        raise ValueError(
+            "this function can't be converted to CASE function, "
+            "only for, if, return and pass statements are allowed"
+        )
+
+    def _replace_local_variables(self, stmts: List[ast.stmt]) -> List[ast.stmt]:
+        """
+        We support local variables only by calculating their value and
+        replacing all it's occurrences in the code after variable definition.
+        For example:
+            a = 12345
+            if a:
+                return a
+            return 0
+        will be converted to:
+            if 12345:
+                return 12345
+            return 0
+        """
+        replace_name = {}
+        res_stmts = []
+        for stmt in stmts:
+            if isinstance(stmt, ast.Assign):
+                assert len(stmt.targets) == 1, 'Unpacking local variables is not supported yet'
+                var, val = stmt.targets[0], stmt.value
+                assert isinstance(var, ast.Name)
+                replace_name[var.id] = val
+                continue
+            res_stmts.append(
+                self._replace_nodes(stmt, replace_name=replace_name)
+            )
+        return res_stmts
+
+    def if_stmt(self, stmt: ast.If) -> ast.IfExp:
+        """
+        Classic if statement with limited set of allowed statements in the body:
+          * only one statement in the body
+          * statement can be return or another if statement with same rules as above
+
+        For example:
+            if tick['X'] > 0:
+                return 'POS'
+            elif tick['X'] == 0:
+                return 'ZERO'
+            else:
+                return 'NEG'
+        will be converted to OneTick's CASE function:
+            CASE(X > 0, 1, 'POS', CASE(X = 0, 1, 'ZERO', 'NEG'))
+        """
+        # TODO: support many statements in body
+        if len(stmt.body) != 1:
+            raise ValueError("this function can't be converted to CASE function, "
+                             "too many statements in if body")
+        body = self.statement(stmt.body[0])
+
+        if len(stmt.orelse) > 1:
+            raise ValueError("this function can't be converted to CASE function, "
+                             "too many statements in else body")
+        if stmt.orelse and not isinstance(stmt.orelse[0], ast.Pass):
+            orelse = self.statement(stmt.orelse[0])
+        else:
+            e = self.expression_parser.expression(body)
+            orelse = ast.Constant(ott.default_by_type(ott.get_type_by_objects(e.values)))
+        return ast.IfExp(test=stmt.test, body=body, orelse=orelse)
+
+    def return_stmt(self, stmt: ast.Return) -> ast.expr:
+        """
+        Return statement.
+        Will be converted to value according to OneTick's syntax.
+        """
+        if stmt.value is None:
+            raise ValueError('return statement must have value when converting to CASE function')
+        return stmt.value
+
+    def pass_stmt(self, _stmt: ast.Pass) -> ast.Constant:
+        """
+        Pass statement.
+        Will be converted to None according to OneTick's syntax.
+        """
+        return ast.Constant(None)
+
+    def compress(self, stmts: List[ast.stmt]) -> Union[ast.If, ast.Return]:
+        """
+        Compress list of statements to single statement.
+        This is possible only if simple if and return statements are used.
+        """
+        stmts = self._replace_local_variables(stmts)
+        flat_stmts = self._flatten_for_stmts(stmts)
+        stmt = self._compress_stmts_to_one_stmt(flat_stmts)  # type: ignore[arg-type]
+        return stmt
+
+    def statement(self, stmt: ast.stmt) -> ast.expr:
+        """Return statement converted to expression."""
+        return {
+            ast.If: self.if_stmt,
+            ast.Return: self.return_stmt,
+            ast.Pass: self.pass_stmt,
+        }[type(stmt)](stmt)  # type: ignore[operator]
+
+
+class StatementParser(CaseStatementParser):
+    """
+    Class with methods to convert ast statements to per tick script lines.
+    """
+
+    def __init__(self, fun: 'FunctionParser'):
+        super().__init__(fun)
+        self.expression_parser = ExpressionParser(fun=fun)  # type: ignore[assignment]
+        self.operator_parser = OperatorParser()
+        self._for_counter = 0
+
+    @staticmethod
+    def _transform_if_expr_to_if_stmt(stmt: Union[ast.Assign, ast.AugAssign]) -> ast.If:
+        """
+        Per tick script do not support if expressions, so converting it to if statement.
+        For example:
+            tick['X'] = 'A' if tick['S'] > 0 else 'B'
+        will be converted to:
+            if (S > 0) {
+                X = 'A';
+            }
+            else {
+                X = 'B';
+            }
+        """
+        if not isinstance(stmt.value, ast.IfExp):
+            raise ValueError()
+        if_expr: ast.IfExp = stmt.value
+        body, orelse = deepcopy(stmt), deepcopy(stmt)
+        body.value = if_expr.body
+        orelse.value = if_expr.orelse
+
+        return ast.If(
+            test=if_expr.test,
+            body=[body],
+            orelse=[orelse],
+        )
+
+    def assign(self, stmt: ast.Assign) -> str:
+        """
+        Assign statement: tick['X'] = 1
+        Will be converted to OneTick syntax: X = 1;
+        """
+        assert len(stmt.targets) == 1, 'Unpacking variables is not yet supported'
+
+        if isinstance(stmt.value, ast.IfExp):
+            if_stmt = self._transform_if_expr_to_if_stmt(stmt)
+            return self.statement(if_stmt)
+
+        var = self.expression_parser.expression(stmt.targets[0])
+        val = self.expression_parser.expression(stmt.value)
+
+        default_expr = f'{var} = {val};'
+
+        if var.lhs:
+            expr = var.value(val.value)
+            return expr or default_expr
+
+        if var.is_local_variable:
+            var_name = var.value.name
+            if val.is_static:
+                val = Expression(val.value.value)
+                if var_name in self.fun.emulator.STATIC_VARS:
+                    raise ValueError(f"Trying to define static variable '{var_name}' more than once")
+                if var_name in self.fun.emulator.LOCAL_VARS:
+                    raise ValueError(f"Can't redefine variable '{var_name}' as static")
+
+                if self.fun.emulator.NEW_VALUES:
+                    raise ValueError('Mixed definition of static variables and new columns is not supported')
+
+                if val.is_tick:
+                    # recreating tick object here, because it doesn't have name yet
+                    self.fun.emulator.STATIC_VARS[var_name] = val.dtype(var_name)
+                    return f'static {val.value._definition} {var};'
+                self.fun.emulator.STATIC_VARS[var_name] = val.value
+                return f'static {ott.type2str(val.dtype)} {var} = {val};'
+
+            variables = None
+            if var_name in self.fun.emulator.STATIC_VARS:
+                variables = self.fun.emulator.STATIC_VARS
+            elif var_name in self.fun.emulator.LOCAL_VARS:
+                variables = self.fun.emulator.LOCAL_VARS
+
+            if variables is None:
+                if val.is_tick:
+                    raise ValueError('Only primitive types are allowed for non static local variables.')
+                if self.fun.emulator.NEW_VALUES:
+                    raise ValueError('Mixed definition of local variables and new columns is not supported')
+                self.fun.emulator.LOCAL_VARS[var_name] = val.value
+                return f'{ott.type2str(val.dtype)} {var} = {val};'
+
+            dtype = ott.get_type_by_objects([variables[var_name]])
+            if val.dtype != dtype:
+                raise ValueError(f"Wrong type for variable '{var_name}': should be {dtype}, got {val.dtype}")
+
+        return default_expr
+
+    def aug_assign(self, stmt: ast.AugAssign) -> str:
+        """
+        Assign with inplace operation statement: tick['X'] += 1.
+        Will be converted to OneTick syntax: X = X + 1;
+        """
+        target = deepcopy(stmt.target)
+        target.ctx = ast.Load()
+        return self.assign(
+            ast.Assign(
+                targets=[stmt.target],
+                value=ast.BinOp(
+                    left=target,
+                    op=stmt.op,
+                    right=stmt.value,
+                )
+            )
+        )
+
+    def if_stmt(self, stmt: ast.If) -> str:  # type: ignore[override]
+        """
+        Classic if statement:
+            if tick['X'] > 0:
+                tick['Y'] = 1
+            elif tick['X'] == 0:
+                tick['Y'] = 0
+            else:
+                tick['Y'] = -1
+        Will be converted to:
+            if (X > 0) {
+                Y = 1;
+            }
+            else {
+                if (X == 0) {
+                    Y = 0;
+                }
+                else {
+                    Y = -1;
+                }
+            }
+        """
+        test = self.expression_parser.expression(stmt.test)
+        test.convert_to_operation()
+        body = [self.statement(s) for s in stmt.body]
+        orelse = [self.statement(s) for s in stmt.orelse]
+        if test.predefined:
+            if test.value:
+                return '\n'.join(body)
+            return '\n'.join(orelse)
+        lines = []
+        lines.append('if (%s) {' % test)
+        lines.extend(body)
+        lines.append('}')
+        if orelse:
+            lines.append('else {')
+            lines.extend(orelse)
+            lines.append('}')
+        return '\n'.join(lines)
+
+    def return_stmt(self, stmt: ast.Return) -> str:  # type: ignore[override]
+        """
+        Return statement.
+        For now we support returning only boolean values or nothing.
+        Will be converted to: return true;
+        """
+        # if return is empty then it is not filter
+        v = stmt.value if stmt.value is not None else ast.Constant(value=True)
+        value = self.expression_parser.expression(v)
+        dtype = ott.get_object_type(value.value)
+        if not self.fun.inner_function:
+            if dtype is not bool:
+                raise TypeError(f"Not supported return type {dtype}")
+            if stmt.value is not None:
+                self.fun.returns = True
+        else:
+            assert isinstance(self.fun.ast_node, ast.FunctionDef)
+            msg = (f"Function '{self.fun.ast_node.name}'"
+                   f" has return annotation '{self.fun.return_annotation.__name__}',"
+                   f" but the type of statement ({astunparse(stmt)}) is '{dtype.__name__}'")
+            try:
+                widest_type = ott.get_type_by_objects([dtype, self.fun.return_annotation])
+            except TypeError as e:
+                raise TypeError(msg) from e
+            if widest_type is not self.fun.return_annotation:
+                raise TypeError(msg)
+            self.fun.returns = True
+        return f'return {value};'
+
+    @staticmethod
+    def _check_break(*nodes) -> bool:
+        """
+        Check if break statement in the list of nodes (recursively).
+        """
+        class FoundBreakException(Exception):
+            pass
+
+        class FindBreak(ast.NodeVisitor):
+            def visit_Break(self, n: ast.Break):
+                raise FoundBreakException()
+
+        try:
+            for node in nodes:
+                FindBreak().visit(node)
+            return False
+        except FoundBreakException:
+            return True
+
+    def while_stmt(self, stmt: ast.While) -> str:
+        """
+        Classic while statement:
+            while tick['X'] > 0:
+                tick['Y'] = 1
+        Will be converted to:
+            while (X > 0) {
+                Y = 1;
+            }
+        """
+        test = self.expression_parser.expression(stmt.test)
+        test.convert_to_operation()
+        body = [self.statement(s) for s in stmt.body]
+        if test.predefined:
+            is_break_found = self._check_break(*stmt.body)
+            if not is_break_found:
+                raise ValueError(f'The condition of while statement always evaluates to {bool(test.value)}'
+                                 ' and there is no break statement in the loop body.'
+                                 ' That will result in infinite loop.'
+                                 ' Change condition or add break statements.')
+        lines = []
+        lines.append('while (%s) {' % test)
+        lines.extend(body)
+        lines.append('}')
+        return '\n'.join(lines)
+
+    def for_stmt(self, stmt: ast.For) -> str:
+        """
+        For now for statement in most cases will not be converted to per tick script's for statement.
+        Instead, the statements from the body of the for statement will be duplicated
+        for each iteration.
+
+        For example:
+            for i in [1, 2, 3]:
+                tick['X'] += i
+        will be converted to:
+            X += 1;
+            X += 2;
+            X += 3;
+
+        But simple case with range object will be translated more correctly:
+        For example:
+            for i in range(1, 4):
+                tick['X'] += i
+        will be converted to:
+            for (LOCAL::i = 1; LOCAL::i < 4; LOCAL::i += 1) {
+                X += LOCAL::i;
+            }
+        """
+        lines = []
+        iter_ = self.expression_parser.expression(stmt.iter)
+        if isinstance(iter_.value, _TickSequence):
+            target = stmt.target
+            assert isinstance(target, ast.Name), "Tuples can't be used while iterating on tick sequences"
+            state_tick = iter_.value._tick_obj(target.id)
+            # TODO: ugly
+            state_tick_name = f"_______state_tick_{self._for_counter}_______"
+            self._for_counter += 1
+            ast_tick = ast.Name(state_tick_name, ctx=ast.Load())
+            lines.append('for (%s %s : %s) {' % (state_tick._definition, state_tick, iter_.value))
+            with self.expression_parser._replace_context(
+                inspect.ClosureVars({}, {state_tick_name: state_tick}, {}, set())
+            ):
+                for s in stmt.body:
+                    s = self._replace_nodes(s, replace_name={target.id: ast_tick})
+                    lines.append(self.statement(s))
+            lines.append('}')
+        elif isinstance(iter_.value, range):
+            target = stmt.target
+            assert isinstance(target, ast.Name), "Tuples can't be used in for loop"
+            var_name = target.id
+            if var_name not in self.fun.emulator.LOCAL_VARS:
+                # initialize counter variable
+                self.fun.emulator.LOCAL_VARS[var_name] = int
+                self.fun.emulator.LOCAL_VARS_NEW_VALUES[var_name].append(0)
+            elif self.fun.emulator.LOCAL_VARS[var_name] is not int:
+                raise ValueError(f'Variable {var_name} was declared before with conflicting type.')
+            counter_var = LocalVariable(var_name, int)
+            counter_var_str = str(counter_var)
+            range_obj = iter_.value
+            start_expr = f'{counter_var_str} = {range_obj.start}'
+            if (range_obj.start < range_obj.stop and range_obj.step <= 0 or
+                    range_obj.start > range_obj.stop and range_obj.step >= 0):
+                raise ValueError(f'Range object {range_obj} will result in infinite loop')
+            if range_obj.start < range_obj.stop:
+                condition_expr = f'{counter_var_str} < {range_obj.stop}'
+            else:
+                condition_expr = f'{counter_var_str} > {range_obj.stop}'
+            increment_expr = f'{counter_var_str} += {range_obj.step}'
+            lines.append('for (%s; %s; %s) {' % (start_expr, condition_expr, increment_expr))
+            try:
+                for s in stmt.body:
+                    lines.append(self.statement(s))
+                lines.append('}')
+            except Exception:
+                if var_name in self.fun.emulator.LOCAL_VARS_NEW_VALUES:
+                    self.fun.emulator.LOCAL_VARS_NEW_VALUES.pop(var_name)
+                    self.fun.emulator.LOCAL_VARS.pop(var_name)
+                lines = [self.statement(s) for s in self._flatten_for_stmt(stmt)]
+        else:
+            lines = [self.statement(s) for s in self._flatten_for_stmt(stmt)]
+        return '\n'.join(lines)
+
+    def break_stmt(self, _stmt: ast.Break) -> str:
+        return 'break;'
+
+    def continue_stmt(self, _stmt: ast.Continue) -> str:
+        return 'continue;'
+
+    def pass_stmt(self, stmt: ast.Pass) -> str:  # type: ignore[override]
+        """Pass statement is not converted to anything"""
+        return ''
+
+    def yield_expr(self, expr: ast.Yield) -> Expression:
+        """
+        Yield expression, like: yield
+        Values for yield are not supported.
+        Will be translated to PROPAGATE_TICK() function.
+        Can be used only as a statement, so this function is here and not in ExpressionParser.
+        """
+        if expr.value is not None:
+            raise ValueError("Passing value with yield expression is not supported.")
+        return Expression('PROPAGATE_TICK();')
+
+    def expression(self, stmt: ast.Expr) -> str:
+        """
+        Here goes raw strings and yield expression.
+        For example:
+            if tick['A'] == 0:
+                'return 0;'
+        Here 'return 0;' is used as a statement and an expression.
+        Expression's returned value *must* be a string and
+        this string will be injected in per tick script directly.
+        """
+        if isinstance(stmt.value, ast.Yield):
+            expression = self.yield_expr(stmt.value)
+        else:
+            expression = self.expression_parser.expression(stmt.value)
+        assert isinstance(expression.value, (str, _Operation)), (
+            f"The statement '{astunparse(stmt)}' can't be used here"
+            " because the value of such statement can be string only"
+            " as it's value will be injected directly in per tick script."
+        )
+        value = str(expression.value)
+        if value and value[-1] != ';':
+            value += ';'
+        return value
+
+    def with_stmt(self, stmt: ast.With) -> str:
+        """
+        Used only with special context managers. Currently only `_ONCE` is supported.
+        """
+        if len(stmt.items) != 1:
+            raise ValueError('Currently it is possible to use only one context manager in single with statement')
+        with_item = stmt.items[0]
+        if with_item.optional_vars:
+            raise ValueError('It is not allowed to use "as" in with statements for per-tick script')
+        context_expr = with_item.context_expr
+        if isinstance(context_expr, ast.Call):
+            expr = self.expression_parser.expression(context_expr.func)
+        else:
+            raise ValueError(f'{context_expr} is not called')
+        if not issubclass(expr.value, once):
+            raise ValueError(f'{expr.value} is not supported in per-tick script with statements')
+        return expr.value().get_str('\n'.join([self.statement(s) for s in stmt.body]))
+
+    def statement(self, stmt: ast.stmt) -> str:  # type: ignore[override]
+        """Return parsed statement according to its type."""
+        return {
+            ast.Assign: self.assign,
+            ast.AugAssign: self.aug_assign,
+            ast.If: self.if_stmt,
+            ast.Return: self.return_stmt,
+            ast.While: self.while_stmt,
+            ast.For: self.for_stmt,
+            ast.Break: self.break_stmt,
+            ast.Continue: self.continue_stmt,
+            ast.Pass: self.pass_stmt,
+            ast.Expr: self.expression,
+            ast.With: self.with_stmt,
+        }[type(stmt)](stmt)  # type: ignore[operator]
+
+
+class EndOfBlock(Exception):
+    pass
+
+
+class LambdaBlockFinder:
+    """
+    This is simplified version of
+    inspect.BlockFinder
+    that supports multiline lambdas.
+    """
+
+    OPENING_BRACKETS = {
+        '[': ']',
+        '(': ')',
+        '{': '}',
+    }
+    CLOSING_BRACKETS = {c: o for o, c in OPENING_BRACKETS.items()}
+    BRACKETS_MATCHING = dict(OPENING_BRACKETS.items() | CLOSING_BRACKETS.items())
+
+    def __init__(self):
+        # current indentation level
+        self.indent = 0
+        # row and column index for the start of lambda expression
+        self.start = None
+        # row and column index for the end of lambda expression
+        self.end = None
+        # stack with brackets
+        self.brackets = deque()
+        self.prev = None
+        self.current = None
+
+    def tokeneater(self, type, token, srowcol, erowcol, line, start_row=0):
+        srowcol = (srowcol[0] + start_row, srowcol[1])
+        erowcol = (erowcol[0] + start_row, erowcol[1])
+        self.prev = self.current
+        self.current = tokenize.TokenInfo(type, token, srowcol, erowcol, line)
+        self.end = erowcol
+        if token == 'lambda':
+            self.start = srowcol
+        elif type == tokenize.INDENT:
+            self.indent += 1
+        elif type == tokenize.DEDENT:
+            self.indent -= 1
+            # the end of matching indent/dedent pairs ends a block
+            if self.indent <= 0:
+                raise EndOfBlock
+        elif not self.start:
+            self.indent = 0
+        elif type == tokenize.NEWLINE:
+            if self.indent == 0 or (
+                # if lambda is the argument of the function
+                self.prev and self.prev.type == tokenize.OP and self.prev.string == ','
+            ):
+                raise EndOfBlock
+        elif token in self.OPENING_BRACKETS:
+            self.brackets.append(token)
+        elif token in self.CLOSING_BRACKETS:
+            try:
+                assert self.brackets.pop() == self.CLOSING_BRACKETS[token]
+            except (IndexError, AssertionError):
+                self.end = self.prev.end
+                raise EndOfBlock  # noqa: W0707
+
+
+def get_lambda_source(lines):
+    """Extract the block of lambda code at the top of the given list of lines."""
+    blockfinder = LambdaBlockFinder()
+    start_row = 0
+    while True:
+        try:
+            tokens = tokenize.generate_tokens(iter(lines[start_row:]).__next__)
+            for _token in tokens:
+                blockfinder.tokeneater(*_token, start_row=start_row)
+            break
+        except IndentationError as e:
+            # indentation errors are possible because
+            # we started eating tokens from line with lambda
+            # not from the start of the statement
+            # trying to eat again from the current row in this case
+            start_row = e.args[1][1] - 1
+            continue
+        except EndOfBlock:
+            break
+    start_row, start_column = blockfinder.start
+    end_row, end_column = blockfinder.end
+    # crop block to get rid of tokens from the context around lambda
+    lines = lines[start_row - 1: end_row]
+    lines[-1] = lines[-1][:end_column]
+    lines[0] = lines[0][start_column:]
+    # add brackets around lambda in case it is multiline lambda
+    return ''.join(['(', *lines, ')'])
+
+
+def is_lambda(lambda_f) -> bool:
+    return isinstance(lambda_f, types.LambdaType) and lambda_f.__name__ == '<lambda>'
+
+
+def get_source(lambda_f) -> str:
+    """
+    Get source code of the function or lambda.
+    """
+    if is_lambda(lambda_f):
+        # that's a hack for multiline lambdas in brackets
+        # inspect.getsource parse them wrong
+        source_lines, lineno = inspect.findsource(lambda_f)
+        if 'lambda' not in source_lines[lineno]:
+            # inspect.findsource fails sometimes too
+            lineno = lambda_f.__code__.co_firstlineno + 1
+            while 'lambda' not in source_lines[lineno]:
+                lineno -= 1
+        source = get_lambda_source(source_lines[lineno:])
+    else:
+        source = inspect.getsource(lambda_f)
+    # doing dedent because self.ast_node do not like indented source code
+    return textwrap.dedent(source)
+
+
+class FunctionParser:
+    """
+    Class to parse callable objects (lambdas and functions) to
+    OneTick's per tick script or case functions.
+    Only simple functions corresponding to OneTick syntax supported
+    (without inner functions, importing modules, etc.)
+    You can call simple functions inside,
+    do operations with captured variables (without assigning to them),
+    but using non-pure functions is not recommended because
+    the code in function may not be executed in the order you expect.
+    """
+    SOURCE_CODE_ATTRIBUTE = '___SOURCE_CODE___'
+
+    def __init__(self, lambda_f, emulator=None, check_arg_name=True, inner_function=False):
+        """
+        Parameters
+        ----------
+        emulator
+            otp.Source emulator that will be tracking changes made to source
+        check_arg_name
+            if True, only callables with zero or one parameter will be allowed.
+        inner_function
+            if True, then function is treated like inner per-tick script function.
+            First argument will be checked and more arguments will be allowed too.
+        """
+
+        assert isinstance(lambda_f, (types.LambdaType, types.FunctionType, types.MethodType)), (
+            f"It is expected to get a function, method or lambda, but got '{type(lambda_f)}'"
+        )
+        self.lambda_f = lambda_f
+        self.emulator = emulator
+        self.check_arg_name = check_arg_name
+        self.inner_function = inner_function
+        if self.inner_function and not self.check_arg_name:
+            self.check_arg_name = self.check_arg_name or True
+        self.statement_parser = StatementParser(fun=self)
+        self.expression_parser = ExpressionParser(fun=self)
+        self.case_expression_parser = CaseExpressionParser(fun=self)
+        self.case_statement_parser = CaseStatementParser(fun=self)
+        # if the function returns some values or not
+        self.returns = False
+        # will be set to True when args_annotations will be calculated, need it to break recursion
+        self._from_args_annotations = False
+        # calling property here, so we can raise exception as early as possible
+        _ = self.arg_name
+
+    @cached_property
+    def is_method(self) -> bool:
+        return isinstance(self.lambda_f, types.MethodType)
+
+    @cached_property
+    def source_code(self) -> str:
+        """
+        Get source code of the function or lambda.
+        """
+        # first try to get code from special attribute else get code the usual way
+        return getattr(self.lambda_f, self.SOURCE_CODE_ATTRIBUTE, None) or get_source(self.lambda_f)
+
+    @cached_property
+    def closure_vars(self) -> inspect.ClosureVars:
+        """
+        Get closure variables of the function.
+        These are variables that were captured from the context before function definition.
+        For example:
+            A = 12345
+            def a():
+                print(A + 1)
+        In this function variable A is the captured variable.
+        We need closure variables, so we can use them when parsing ast tree.
+        """
+        return inspect.getclosurevars(self.lambda_f)
+
+    @cached_property
+    def ast_node(self) -> Union[ast.FunctionDef, ast.Lambda]:
+        """
+        Convert function or lambda to ast module statement.
+        """
+        source_code = self.source_code
+        tree = ast.parse(source_code)
+        for node in ast.walk(tree):
+            if isinstance(node, (ast.FunctionDef, ast.Lambda)):
+                if isinstance(node, ast.FunctionDef) and ast.get_docstring(node):
+                    # remove comment section from function body
+                    node.body.pop(0)
+                return node
+        raise ValueError("Can't find function or lambda in source code")
+
+    @cached_property
+    def arg_name(self) -> Optional[str]:
+        """Get name of the first function or lambda argument."""
+        node = self.ast_node
+        argv = list(node.args.args)
+        argc = len(argv)
+        if argc > 1 and argv[0].arg == 'self' and self.is_method:
+            argv.pop(0)
+            argc -= 1
+        if self.check_arg_name and argc > 1 and not self.inner_function:
+            raise ValueError(
+                "It is allowed to pass only functions or lambdas that take either one or"
+                f" zero parameters, but got {argc}"
+            )
+        arg_name = argv[0].arg if argv else None
+        if isinstance(self.check_arg_name, str) and arg_name != self.check_arg_name:
+            assert isinstance(node, ast.FunctionDef)
+            msg = f"Function '{node.name}' is expected to have first argument named '{self.check_arg_name}'"
+            if arg_name is None:
+                msg += ', but no argument is found'
+            else:
+                msg += f", but argument with name '{arg_name}' is found"
+            raise ValueError(msg)
+        return arg_name
+
+    @cached_property
+    def args_annotations(self) -> dict:
+        node = self.ast_node
+        if not isinstance(node, ast.FunctionDef):
+            return {}
+        argv = list(node.args.args)
+        if argv and argv[0].arg == 'self' and self.is_method:
+            argv.pop(0)
+        if argv and argv[0].arg == self.arg_name:
+            argv.pop(0)
+        if node.args.defaults:
+            raise ValueError("Default values for arguments are not supported"
+                             f" in per-tick script function '{node.name}'")
+        annotations = {}
+        for arg in argv:
+            name = arg.arg
+            annotation = getattr(arg, 'annotation', None)
+            if not annotation:
+                raise ValueError(f"Parameter '{name}' in function '{node.name}' doesn't have type annotation")
+            # TODO: remove hacking
+            self._from_args_annotations = True
+            dtype = self.expression_parser.expression(annotation).value
+            self._from_args_annotations = False
+            if not ott.is_type_basic(dtype):
+                raise ValueError(f"Parameter '{name}' in function '{node.name}' has unsupported type: {dtype}")
+            annotations[name] = dtype
+
+        return annotations
+
+    @cached_property
+    def return_annotation(self):
+        node = self.ast_node
+        annotation = getattr(node, 'returns')
+        if not annotation:
+            raise ValueError(f"Function '{node.name}' doesn't have return type annotation")
+        dtype = self.expression_parser.expression(annotation).value
+        if not ott.is_type_basic(dtype):
+            raise ValueError(f"Function '{node.name}' has unsupported return type: {dtype}")
+        return dtype
+
+    def per_tick_script(self) -> str:
+        """
+        Convert function to OneTick's per tick script.
+        """
+        node = self.ast_node
+
+        lines = []
+
+        assert isinstance(node, ast.FunctionDef), 'lambdas are not supported in per-tick-script yet'
+        function_def: ast.FunctionDef = node
+
+        for stmt in function_def.body:
+            line = self.statement_parser.statement(stmt)
+            if line:
+                lines.append(line)
+
+        if not self.inner_function and self.returns:
+            # if there were return statement anywhere in the code
+            # then we add default return at the end
+            # TODO: but the default behaviour in OneTick is to propagate all ticks?
+            #       changing that will break backward-compatibility
+            lines.append(self.statement_parser.statement(ast.Return(ast.Constant(False))))
+
+        if self.emulator is not None and not self.inner_function:
+            # per tick script syntax demand that we declare variables before using them
+            # so we get all new variables from emulator and declare them.
+
+            def var_definition(key, values):
+                dtype = ott.get_type_by_objects(values)
+                return f'{ott.type2str(dtype)} {str(key)} = {ott.value2str(ott.default_by_type(dtype))};'
+
+            new_columns = []
+            for key, values in self.emulator.NEW_VALUES.items():
+                new_columns.append(var_definition(key, values))
+            new_local_vars = []
+            for key, values in self.emulator.LOCAL_VARS_NEW_VALUES.items():
+                new_local_vars.append(var_definition(LocalVariable(key), values))
+            lines = new_columns + new_local_vars + lines
+
+        if not lines:
+            raise ValueError("The resulted body of PER TICK SCRIPT is empty")
+
+        if not self.inner_function:
+            lines = ['long main() {'] + lines + ['}']
+            for function, *_ in self.emulator.FUNCTIONS.values():
+                lines.append(function)
+        else:
+            if not self.returns:
+                raise ValueError(f"Function '{node.name}' must return values")
+            return_type = ott.type2str(self.return_annotation)
+            args = [
+                f'{ott.type2str(dtype)} {name}'
+                for name, dtype in self.args_annotations.items()
+            ]
+            args: str = ', '.join(args)  # type: ignore[no-redef]
+            lines = [f'{return_type} {node.name}({args})' + ' {'] + lines + ['}']
+
+        return '\n'.join(lines) + '\n'
+
+    def compress(self) -> ast.expr:
+        """
+        Convert lambda or function to AST expression.
+        """
+        node = self.ast_node
+        if isinstance(node, ast.Lambda):
+            return node.body
+        stmt = self.case_statement_parser.compress(node.body)
+        return self.case_statement_parser.statement(stmt)
+
+    def case(self) -> Tuple[str, List[Any]]:
+        """
+        Convert lambda or function to OneTick's CASE() function.
+        """
+        expr = self.compress()
+        expr = self.case_expression_parser._convert_bool_op_to_if_expr(expr)
+        expression = self.case_expression_parser.expression(expr)
+        # this will raise type error if type of the expression is not supported
+        ott.default_by_type(ott.get_type_by_objects(expression.values))
+        return str(expression), expression.values
+
+
+def remote(fun):
+    """
+    This decorator is needed in case function ``fun``
+    is used in :py:meth:`~onetick.py.Source.apply` method in a `Remote OTP with Ray` context.
+
+    We want to get source code of the function locally
+    because we will not be able to get source code on the remote server.
+
+    See also
+    --------
+    :ref:`Remote OTP with Ray <ray-remote>`.
+    """
+    # see PY-424
+    @wraps(fun)
+    def wrapper(*args, **kwargs):
+        return fun(*args, **kwargs)
+    setattr(wrapper, FunctionParser.SOURCE_CODE_ATTRIBUTE, get_source(fun))
+    return wrapper
+
+
+class once:
+    """
+    Used with a statement or a code block to make it run only once (the first time control reaches to the statement).
+    """
+    def __enter__(self):
+        # __enter__ and __exit__ methods are only used to express syntax for per-tick script, thus no implementation
+        pass
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        # __enter__ and __exit__ methods are only used to express syntax for per-tick script, thus no implementation
+        pass
+
+    def get_str(self, string: str) -> str:
+        return f"_ONCE\n{{\n{string}\n}}"
+
+
+class Once(once):
+    def __init__(self):
+        warnings.warn('Using `otp.Once` is deprecated, please, use `otp.once` instead', FutureWarning)
+        super().__init__()
+
+
+def logf(message, severity, *args) -> str:
+    """
+    Call built-in OneTick ``LOGF`` function from per-tick script.
+
+    Parameters
+    ----------
+    message: str
+        Log message/format string. The underlying formatting engine is the Boost Format Library:
+        https://www.boost.org/doc/libs/1_53_0/libs/format/doc/format.html
+
+    severity: str
+        Severity of message. Supported values: ``ERROR``, ``WARNING`` and ``INFO``.
+
+    args: list
+        Parameters for format string (optional).
+
+    Returns
+    -------
+    str
+
+    Examples
+    --------
+    >>> t = otp.Ticks({'X': [1, 2, 3]})
+
+    >>> def test_script(tick):
+    ...     otp.logf("Tick with value X=%1% processed", "INFO", tick["X"])
+
+    >>> t = t.script(test_script)
+
+    See also
+    --------
+    :ref:`Per-Tick Script Guide <python callable parser>`
+    """
+
+    if severity not in {"ERROR", "WARNING", "INFO"}:
+        raise ValueError(f"Param severity expected to be one of ERROR, WARNING or INFO. Got \"{severity}\"")
+
+    message = ott.value2str(message)
+    severity = ott.value2str(severity)
+
+    if args:
+        params = ", ".join([ott.value2str(arg) for arg in args])
+        return f"LOGF({message}, {severity}, {params});"
+    else:
+        return f"LOGF({message}, {severity});"
+
+
+def throw_exception(message: str) -> str:
+    """
+    Call built-in OneTick ``THROW_EXCEPTION`` function from per-tick script.
+
+    Parameters
+    ----------
+    message: str
+        Message string that defines the error message to be thrown.
+
+    Returns
+    -------
+    str
+
+    Examples
+    --------
+    >>> t = otp.Ticks({'X': [1, -2, 6]})
+
+    >>> def test_script(tick):
+    ...     if tick["X"] <= 0:
+    ...         otp.throw_exception("Tick column X should be greater than zero.")
+
+    >>> t = t.script(test_script)
+
+    See also
+    --------
+    :ref:`Per-Tick Script Guide <python callable parser>`
+    """
+    message = ott.value2str(message)
+    return f"THROW_EXCEPTION({message});"