vtlengine 1.1rc2__py3-none-any.whl → 1.2.0__py3-none-any.whl

vtlengine/AST/ASTConstructorModules/__init__.py

@@ -0,0 +1,50 @@
+from typing import Dict, Union
+
+from antlr4.ParserRuleContext import ParserRuleContext
+from antlr4.Token import CommonToken
+
+from vtlengine.AST.Grammar.lexer import Lexer
+
+
+def extract_token_info(token: Union[CommonToken, ParserRuleContext]) -> Dict[str, int]:
+    """
+    Extracts the token information from a token or ParserRuleContext.
+
+    The Token information includes:
+    - column_start: The starting column of the token.
+    - column_stop: The stopping column of the token.
+    - line_start: The starting line number of the token.
+    - line_stop: The stopping line number of the token.
+
+    The overall idea is to provide the information from which line and column,
+    and to which line and column, the text is referenced by the AST object, including children.
+
+    Important Note: the keys of the dict are the same as the class attributes of the AST Object.
+
+    Args:
+        token (Union[CommonToken, ParserRuleContext]): The token or ParserRuleContext to extract
+        information from.
+
+    Returns:
+        Dict[str, int]: A dictionary containing the token information.
+    """
+
+    if isinstance(token, ParserRuleContext):
+        return {
+            "column_start": token.start.column,
+            "column_stop": token.stop.column + len(token.stop.text),
+            "line_start": token.start.line,
+            "line_stop": token.stop.line,
+        }
+    line_start = token.line
+    line_stop = token.line
+    # For block comments, we need to add the lines inside the block, marked by \n, to the stop line.
+    # The ML_COMMENT does not take into account the final \n in its grammar.
+    if token.type == Lexer.ML_COMMENT:
+        line_stop = token.line + token.text.count("\n")
+    return {
+        "column_start": token.column,
+        "column_stop": token.column + len(token.text),
+        "line_start": line_start,
+        "line_stop": line_stop,
+    }

vtlengine/AST/ASTEncoders.py

@@ -1,12 +1,16 @@
 import json
 
 from vtlengine import AST
+from vtlengine.Model import Dataset
 
 
 class ComplexEncoder(json.JSONEncoder):
     def default(self, obj):
         if hasattr(obj, "toJSON"):
             return obj.toJSON()
+        # Makes a circular reference error if we do not check for this
+        elif isinstance(obj, Dataset):
+            return "dataset"
         else:
             return json.__dict__
 

vtlengine/AST/ASTString.py

@@ -0,0 +1,622 @@
+import copy
+from dataclasses import dataclass
+from typing import Any, Optional, Tuple, Union
+
+from vtlengine import AST
+from vtlengine.AST import Comment, DPRuleset, HRuleset, Operator, TimeAggregation
+from vtlengine.AST.ASTTemplate import ASTTemplate
+from vtlengine.AST.Grammar.lexer import Lexer
+from vtlengine.AST.Grammar.tokens import (
+    AGGREGATE,
+    ATTRIBUTE,
+    CAST,
+    CHARSET_MATCH,
+    CHECK_DATAPOINT,
+    CHECK_HIERARCHY,
+    DATE_ADD,
+    DATEDIFF,
+    DROP,
+    FILL_TIME_SERIES,
+    FILTER,
+    HAVING,
+    HIERARCHY,
+    IDENTIFIER,
+    INSTR,
+    INTERSECT,
+    LOG,
+    MAX,
+    MEASURE,
+    MEMBERSHIP,
+    MIN,
+    MINUS,
+    MOD,
+    NOT,
+    NVL,
+    PLUS,
+    POWER,
+    RANDOM,
+    REPLACE,
+    ROUND,
+    SETDIFF,
+    SUBSTR,
+    SYMDIFF,
+    TIMESHIFT,
+    TRUNC,
+    UNION,
+    VIRAL_ATTRIBUTE,
+)
+from vtlengine.DataTypes import SCALAR_TYPES_CLASS_REVERSE
+from vtlengine.Model import Component, Dataset
+
+nl = "\n"
+tab = "\t"
+
+
+def _handle_literal(value: Union[str, int, float, bool]):
+    if isinstance(value, str):
+        if '"' in value:
+            return value
+        return f'"{value}"'
+    elif isinstance(value, bool):
+        return "true" if value else "false"
+    elif isinstance(value, float):
+        decimal = str(value).split(".")[1]
+        if len(decimal) > 4:
+            return f"{value:f}".rstrip("0")
+        else:
+            return f"{value:g}"
+    return str(value)
+
+
+def _format_dataset_eval(dataset: Dataset) -> str:
+    def __format_component(component: Component) -> str:
+        return (
+            f"\n\t\t\t{component.role.value.lower()}"
+            f"<{SCALAR_TYPES_CLASS_REVERSE[component.data_type].lower()}> "
+            f"{component.name}"
+        )
+
+    return f"{{ {', '.join([__format_component(x) for x in dataset.components.values()])} \n\t\t}}"
+
+
+def _format_reserved_word(value: str):
+    reserved_words = {x.replace("'", ""): x for x in Lexer.literalNames}
+    if value in reserved_words:
+        return reserved_words[value]
+    elif value[0] == "_":
+        return f"'{value}'"
+    return value
+
+
+@dataclass
+class ASTString(ASTTemplate):
+    vtl_script: str = ""
+    pretty: bool = False
+    is_first_assignment: bool = False
+    is_from_agg: bool = False  # Handler to write grouping at aggr level
+
+    def render(self, ast: AST.AST) -> str:
+        self.vtl_script = ""
+        result = self.visit(ast)
+        if result:
+            self.vtl_script += result
+        return self.vtl_script
+
+    def visit_Start(self, node: AST.Start) -> Any:
+        transformations = [
+            x for x in node.children if not isinstance(x, (HRuleset, DPRuleset, Operator, Comment))
+        ]
+        for child in node.children:
+            if child in transformations:
+                self.is_first_assignment = True
+            self.visit(child)
+            self.vtl_script += "\n"
+
+    # ---------------------- Rulesets ----------------------
+    def visit_HRuleset(self, node: AST.HRuleset) -> None:
+        signature = f"{node.signature_type} rule {node.element.value}"
+        if self.pretty:
+            self.vtl_script += f"define hierarchical ruleset {node.name}({signature}) is{nl}"
+            for i, rule in enumerate(node.rules):
+                self.vtl_script += f"{tab}{self.visit(rule)}{nl}"
+                if rule.erCode:
+                    self.vtl_script += f"{tab}errorcode {_handle_literal(rule.erCode)}{nl}"
+                if rule.erLevel:
+                    self.vtl_script += f"{tab}errorlevel {rule.erLevel}"
+                    if i != len(node.rules) - 1:
+                        self.vtl_script += f";{nl}"
+                    self.vtl_script += nl
+            self.vtl_script += f"end hierarchical ruleset;{nl}"
+        else:
+            rules_strs = []
+            for rule in node.rules:
+                rule_str = self.visit(rule)
+                if rule.erCode:
+                    rule_str += f" errorcode {_handle_literal(rule.erCode)}"
+                if rule.erLevel:
+                    rule_str += f" errorlevel {rule.erLevel}"
+                rules_strs.append(rule_str)
+            rules_sep = "; " if len(rules_strs) > 1 else ""
+            rules = rules_sep.join(rules_strs)
+            self.vtl_script += (
+                f"define hierarchical ruleset {node.name} ({signature}) is {rules} "
+                f"end hierarchical ruleset;"
+            )
+
+    def visit_HRule(self, node: AST.HRule) -> str:
+        vtl_script = ""
+        if node.name is not None:
+            vtl_script += f"{node.name}: "
+        vtl_script += f"{self.visit(node.rule)}"
+        return vtl_script
+
+    def visit_HRBinOp(self, node: AST.HRBinOp) -> str:
+        if node.op == "when":
+            if self.pretty:
+                return (
+                    f"{tab * 3}when{nl}"
+                    f"{tab * 4}{self.visit(node.left)}{nl}"
+                    f"{tab * 3}then{nl}"
+                    f"{tab * 4}{self.visit(node.right)}"
+                )
+            else:
+                return f"{node.op} {self.visit(node.left)} then {self.visit(node.right)}"
+        return f"{self.visit(node.left)} {node.op} {self.visit(node.right)}"
+
+    def visit_HRUnOp(self, node: AST.HRUnOp) -> str:
+        return f"{node.op} {self.visit(node.operand)}"
+
+    def visit_DefIdentifier(self, node: AST.DefIdentifier) -> str:
+        return _format_reserved_word(node.value)
+
+    def visit_DPRule(self, node: AST.DPRule) -> str:
+        if self.pretty:
+            lines = []
+            if node.name is not None:
+                lines.append(f"{tab}{node.name}: ")
+            lines.append(self.visit(node.rule))
+            if node.erCode is not None:
+                lines.append(f"{tab * 3}errorcode  {_handle_literal(node.erCode)}")
+            if node.erLevel is not None:
+                lines.append(f"{tab * 3}errorlevel {node.erLevel}")
+            return nl.join(lines)
+        else:
+            vtl_script = ""
+            if node.name is not None:
+                vtl_script += f"{node.name}: "
+            vtl_script += f"{self.visit(node.rule)}"
+            if node.erCode is not None:
+                vtl_script += f" errorcode {_handle_literal(node.erCode)}"
+            if node.erLevel is not None:
+                vtl_script += f" errorlevel {node.erLevel}"
+            return vtl_script
+
+    def visit_DPRIdentifier(self, node: AST.DPRIdentifier) -> str:
+        vtl_script = f"{node.value}"
+        if node.alias is not None:
+            vtl_script += f" as {node.alias}"
+        return vtl_script
+
+    def visit_DPRuleset(self, node: AST.DPRuleset) -> None:
+        rules_sep = "; " if len(node.rules) > 1 else ""
+        signature_sep = ", " if len(node.params) > 1 else ""
+        signature = (
+            f"{node.signature_type} {signature_sep.join([self.visit(x) for x in node.params])}"
+        )
+
+        if self.pretty:
+            self.vtl_script += f"define datapoint ruleset {node.name}({signature}) is {nl}"
+            rules = ""
+            for i, rule in enumerate(node.rules):
+                rules += f"\t{self.visit(rule)}"
+                if i != len(node.rules) - 1:
+                    rules += f";{nl * 2}"
+                else:
+                    rules += f"{nl}"
+            self.vtl_script += rules
+            self.vtl_script += f"end datapoint ruleset;{nl}"
+        else:
+            rules = rules_sep.join([self.visit(x) for x in node.rules])
+            self.vtl_script += (
+                f"define datapoint ruleset {node.name} "
+                f"({signature}) is {rules} end datapoint ruleset;"
+            )
+
+    # ---------------------- User Defined Operators ----------------------
+
+    def visit_Argument(self, node: AST.Argument) -> str:
+        default = f" default {self.visit(node.default)}" if node.default is not None else ""
+
+        if isinstance(node.type_, Dataset):
+            argument_type = "dataset"
+        elif isinstance(node.type_, Component):
+            argument_type = "component"
+        else:
+            argument_type = node.type_.__name__.lower()
+
+        name = _format_reserved_word(node.name)
+
+        return f"{name} {argument_type}{default}"
+
+    def visit_Operator(self, node: AST.Operator) -> None:
+        signature_sep = ", " if len(node.parameters) > 1 else ""
+        signature = signature_sep.join([self.visit(x) for x in node.parameters])
+        if self.pretty:
+            self.vtl_script += f"define operator {node.op}({signature}){nl}"
+            self.vtl_script += f"\treturns {node.output_type.lower()} is{nl}"
+            expression = self.visit(node.expression)
+            if "(" in expression:
+                expression = expression.replace("(", f"({nl}{tab * 2}")
+                expression = expression.replace(")", f"{nl}{tab * 2})")
+
+            self.vtl_script += f"{tab * 2}{expression}{nl}"
+            self.vtl_script += f"end operator;{nl}"
+        else:
+            body = f"returns {node.output_type.lower()} is {self.visit(node.expression)}"
+            self.vtl_script += f"define operator {node.op}({signature}) {body} end operator;"
+
+    # ---------------------- Basic Operators ----------------------
+    def visit_Assignment(self, node: AST.Assignment) -> Optional[str]:
+        return_element = not copy.deepcopy(self.is_first_assignment)
+        is_first = self.is_first_assignment
+        if is_first:
+            self.is_first_assignment = False
+        if self.pretty:
+            if is_first:
+                expression = f"{self.visit(node.left)} {node.op}{nl}{tab}{self.visit(node.right)}"
+            else:
+                expression = f"{self.visit(node.left)} {node.op} {self.visit(node.right)}"
+        else:
+            expression = f"{self.visit(node.left)} {node.op} {self.visit(node.right)}"
+        if return_element:
+            return expression
+        self.vtl_script += f"{expression};"
+
+    def visit_PersistentAssignment(self, node: AST.PersistentAssignment) -> Optional[str]:
+        return self.visit_Assignment(node)
+
+    def visit_BinOp(self, node: AST.BinOp) -> str:
+        if node.op in [NVL, LOG, MOD, POWER, RANDOM, TIMESHIFT, DATEDIFF, CHARSET_MATCH]:
+            return f"{node.op}({self.visit(node.left)}, {self.visit(node.right)})"
+        elif node.op == MEMBERSHIP:
+            return f"{self.visit(node.left)}{node.op}{self.visit(node.right)}"
+        if self.pretty:
+            return f"{self.visit(node.left)} {node.op} {self.visit(node.right)}"
+
+        return f"{self.visit(node.left)} {node.op} {self.visit(node.right)}"
+
+    def visit_UnaryOp(self, node: AST.UnaryOp) -> str:
+        if node.op in [PLUS, MINUS]:
+            return f"{node.op}{self.visit(node.operand)}"
+        elif node.op in [IDENTIFIER, ATTRIBUTE, VIRAL_ATTRIBUTE, NOT]:
+            return f"{node.op} {self.visit(node.operand)}"
+        elif node.op == MEASURE:
+            return self.visit(node.operand)
+
+        return f"{node.op}({self.visit(node.operand)})"
+
+    def visit_MulOp(self, node: AST.MulOp) -> str:
+        sep = ", " if len(node.children) > 1 else ""
+        body = sep.join([self.visit(x) for x in node.children])
+        if self.pretty:
+            return f"{node.op}({body})"
+        return f"{node.op}({body})"
+
+    def visit_ParamOp(self, node: AST.ParamOp) -> str:
+        if node.op == HAVING:
+            return f"{node.op} {self.visit(node.params)}"
+        elif node.op in [SUBSTR, INSTR, REPLACE, ROUND, TRUNC, UNION, SETDIFF, SYMDIFF, INTERSECT]:
+            params_sep = ", " if len(node.params) > 1 else ""
+            return (
+                f"{node.op}({self.visit(node.children[0])}, "
+                f"{params_sep.join([self.visit(x) for x in node.params])})"
+            )
+        elif node.op in (CHECK_HIERARCHY, HIERARCHY):
+            operand = self.visit(node.children[0])
+            component_name = self.visit(node.children[1])
+            rule_name = self.visit(node.children[2])
+            param_mode_value = node.params[0].value
+            param_input_value = node.params[1].value
+            param_output_value = node.params[2].value
+
+            default_value_input = "dataset" if node.op == CHECK_HIERARCHY else "rule"
+            default_value_output = "invalid" if node.op == CHECK_HIERARCHY else "computed"
+
+            param_mode = f" {param_mode_value}" if param_mode_value != "non_null" else ""
+            param_input = (
+                f" {param_input_value}" if param_input_value != default_value_input else ""
+            )
+            param_output = (
+                f" {param_output_value}" if param_output_value != default_value_output else ""
+            )
+            if self.pretty:
+                return (
+                    f"{node.op}({nl}{tab * 2}{operand},{nl}{tab * 2}{rule_name}{nl}{tab * 2}rule "
+                    f"{component_name}"
+                    f"{param_mode}{param_input}{param_output})"
+                )
+            else:
+                return (
+                    f"{node.op}({operand}, {rule_name} rule {component_name}"
+                    f"{param_mode}{param_input}{param_output})"
+                )
+
+        elif node.op == CHECK_DATAPOINT:
+            operand = self.visit(node.children[0])
+            rule_name = node.children[1]
+            output = ""
+            if len(node.params) == 1 and node.params[0] != "invalid":
+                output = f"{node.params[0]}"
+            if self.pretty:
+                return f"{node.op}({nl}{tab}{operand},{nl}{tab}{rule_name}{nl}{tab}{output}{nl})"
+            else:
+                return f"{node.op}({operand}, {rule_name}{output})"
+        elif node.op == CAST:
+            operand = self.visit(node.children[0])
+            data_type = SCALAR_TYPES_CLASS_REVERSE[node.children[1]].lower()
+            mask = ""
+            if len(node.params) == 1:
+                mask = f", {_handle_literal(self.visit(node.params[0]))}"
+            if self.pretty:
+                return f"{node.op}({operand}, {data_type}{mask})"
+            else:
+                return f"{node.op}({operand}, {data_type}{mask})"
+        elif node.op == FILL_TIME_SERIES:
+            operand = self.visit(node.children[0])
+            param = node.params[0].value if node.params else "all"
+            if self.pretty:
+                return f"{node.op}({operand},{param})"
+            else:
+                return f"{node.op}({operand}, {param})"
+        elif node.op == DATE_ADD:
+            operand = self.visit(node.children[0])
+            shift_number = self.visit(node.params[0])
+            period_indicator = self.visit(node.params[1])
+            if self.pretty:
+                return (
+                    f"{node.op}({nl}{tab * 2}{operand},{nl}{tab * 2}{shift_number},"
+                    f"{nl}{tab * 2}"
+                    f"{period_indicator})"
+                )
+            else:
+                return f"{node.op}({operand}, {shift_number}, {period_indicator})"
+        return ""
+
+    # ---------------------- Individual operators ----------------------
+
+    def _handle_grouping_having(self, node: AST) -> Tuple[str, str]:
+        if self.is_from_agg:
+            return "", ""
+        grouping = ""
+        if node.grouping is not None:
+            grouping_sep = ", " if len(node.grouping) > 1 else ""
+            grouping_values = []
+            for grouping_value in node.grouping:
+                if isinstance(grouping_value, TimeAggregation):
+                    grouping_values.append(self.visit(grouping_value))
+                else:
+                    grouping_values.append(_format_reserved_word(grouping_value.value))
+            grouping = f" {node.grouping_op} {grouping_sep.join(grouping_values)}"
+        having = f" {self.visit(node.having_clause)}" if node.having_clause is not None else ""
+        return grouping, having
+
+    def visit_Aggregation(self, node: AST.Aggregation) -> str:
+        grouping, having = self._handle_grouping_having(node)
+        if self.pretty and node.op not in (MAX, MIN):
+            operand = self.visit(node.operand)
+            return f"{node.op}({nl}{tab * 2}{operand}{grouping}{having}{nl}{tab * 2})"
+        return f"{node.op}({self.visit(node.operand)}{grouping}{having})"
+
+    def visit_Analytic(self, node: AST.Analytic) -> str:
+        operand = "" if node.operand is None else self.visit(node.operand)
+        partition = ""
+        if node.partition_by:
+            partition_sep = ", " if len(node.partition_by) > 1 else ""
+            partition_values = [_format_reserved_word(x) for x in node.partition_by]
+            partition = f"partition by {partition_sep.join(partition_values)}"
+        order = ""
+        if node.order_by:
+            order_sep = ", " if len(node.order_by) > 1 else ""
+            order = f" order by {order_sep.join([self.visit(x) for x in node.order_by])}"
+        window = f" {self.visit(node.window)}" if node.window is not None else ""
+        params = ""
+        if node.params:
+            params = "" if len(node.params) == 0 else f", {int(node.params[0])}"
+        if self.pretty:
+            result = (
+                f"{node.op}({nl}{tab * 3}{operand}{params} over({partition}{order} {window})"
+                f"{nl}{tab * 2})"
+            )
+        else:
+            result = f"{node.op}({operand}{params} over ({partition}{order}{window}))"
+
+        return result
+
+    def visit_Case(self, node: AST.Case) -> str:
+        if self.pretty:
+            else_str = f"{nl}{tab * 2}else{nl}{tab * 3}{self.visit(node.elseOp)}"
+            body_sep = " " if len(node.cases) > 1 else ""
+            body = body_sep.join([self.visit(x) for x in node.cases])
+            return f"case {body} {else_str}"
+        else:
+            else_str = f"else {self.visit(node.elseOp)}"
+            body_sep = " " if len(node.cases) > 1 else ""
+            body = body_sep.join([self.visit(x) for x in node.cases])
+            return f"case {body} {else_str}"
+
+    def visit_CaseObj(self, node: AST.CaseObj) -> str:
+        if self.pretty:
+            return (
+                f"{nl}{tab * 2}when{nl}{tab * 3}{self.visit(node.condition)}{nl}{tab * 2}then"
+                f"{nl}{tab * 3}{self.visit(node.thenOp)}"
+            )
+        else:
+            return f"when {self.visit(node.condition)} then {self.visit(node.thenOp)}"
+
+    def visit_EvalOp(self, node: AST.EvalOp) -> str:
+        operand_sep = ", " if len(node.operands) > 1 else ""
+        if self.pretty:
+            operands = operand_sep.join([self.visit(x) for x in node.operands])
+            ext_routine = f"{nl}{tab * 2}{node.name}({operands})"
+            language = f"{nl}{tab * 2}language {_handle_literal(node.language)}{nl}"
+            output = f"{tab * 2}returns dataset {_format_dataset_eval(node.output)}"
+            return f"eval({ext_routine} {language} {output})"
+        else:
+            operands = operand_sep.join([self.visit(x) for x in node.operands])
+            ext_routine = f"{node.name}({operands})"
+            language = f"language {_handle_literal(node.language)}"
+            output = f"returns dataset {_format_dataset_eval(node.output)}"
+            return f"eval({ext_routine} {language} {output})"
+
+    def visit_If(self, node: AST.If) -> str:
+        if self.pretty:
+            else_str = (
+                f"else{nl}{tab * 5}{self.visit(node.elseOp)}" if node.elseOp is not None else ""
+            )
+            return (
+                f"{nl}{tab * 4}if {nl}{tab * 5}{self.visit(node.condition)} "
+                f"{nl}{tab * 4}then {nl}{tab * 5}{self.visit(node.thenOp)}{nl}{tab * 4}{else_str}"
+            )
+        else:
+            else_str = f"else {self.visit(node.elseOp)}" if node.elseOp is not None else ""
+            return f"if {self.visit(node.condition)} then {self.visit(node.thenOp)} {else_str}"
+
+    def visit_JoinOp(self, node: AST.JoinOp) -> str:
+        if self.pretty:
+            sep = f",{nl}{tab * 2}" if len(node.clauses) > 1 else ""
+            clauses = sep.join([self.visit(x) for x in node.clauses])
+            using = ""
+            if node.using is not None:
+                using_sep = ", " if len(node.using) > 1 else ""
+                using_values = [_format_reserved_word(x) for x in node.using]
+                using = f"using {using_sep.join(using_values)}"
+            return f"{node.op}({nl}{tab * 2}{clauses}{nl}{tab * 2}{using})"
+        else:
+            sep = ", " if len(node.clauses) > 1 else ""
+            clauses = sep.join([self.visit(x) for x in node.clauses])
+            using = ""
+            if node.using is not None:
+                using_sep = ", " if len(node.using) > 1 else ""
+                using_values = [_format_reserved_word(x) for x in node.using]
+                using = f" using {using_sep.join(using_values)}"
+            return f"{node.op}({clauses}{using})"
+
+    def visit_ParFunction(self, node: AST.ParFunction) -> str:
+        return f"({self.visit(node.operand)})"
+
+    def visit_RegularAggregation(self, node: AST.RegularAggregation) -> str:
+        child_sep = ", " if len(node.children) > 1 else ""
+        if node.op == AGGREGATE:
+            self.is_from_agg = True
+            body = child_sep.join([self.visit(x) for x in node.children])
+            self.is_from_agg = False
+            grouping, having = self._handle_grouping_having(node.children[0].right)
+            if self.pretty:
+                body = f"{nl}{tab * 3}{body}{nl}{tab * 3}{grouping}{having}{nl}{tab * 2}"
+            else:
+                body = f"{body}{grouping}{having}"
+        elif node.op == DROP and self.pretty:
+            drop_sep = f",{nl}{tab * 3}" if len(node.children) > 1 else ""
+            body = f"{drop_sep.join([self.visit(x) for x in node.children])}{nl}{tab * 2}"
+        elif node.op == FILTER and self.pretty:
+            condition = self.visit(node.children[0])
+            if " and " in condition or " or " in condition:
+                for op in (" and ", " or "):
+                    condition = condition.replace(op, f"{op}{nl}{tab * 5}")
+            body = f"{nl}{tab * 4}{condition}{nl}{tab * 2}"
+        else:
+            body = child_sep.join([self.visit(x) for x in node.children])
+        if isinstance(node.dataset, AST.JoinOp):
+            dataset = self.visit(node.dataset)
+            if self.pretty:
+                return f"{dataset[:-1]} {(node.op)} {body}{nl}{tab})"
+            else:
+                return f"{dataset[:-1]} {node.op} {body})"
+        else:
+            dataset = self.visit(node.dataset)
+            if self.pretty:
+                return f"{dataset}{nl}{tab * 2}[{node.op} {body}]"
+            else:
+                return f"{dataset}[{node.op} {body}]"
+
+    def visit_RenameNode(self, node: AST.RenameNode) -> str:
+        return f"{node.old_name} to {node.new_name}"
+
+    def visit_TimeAggregation(self, node: AST.TimeAggregation) -> str:
+        operand = self.visit(node.operand)
+        period_from = "_" if node.period_from is None else _handle_literal(node.period_from)
+        period_to = _handle_literal(node.period_to)
+        if self.pretty:
+            return f"{node.op}({period_to}, {period_from}, {operand})"
+        else:
+            return f"{node.op}({period_to}, {period_from}, {operand})"
+
+    def visit_UDOCall(self, node: AST.UDOCall) -> str:
+        params_sep = ", " if len(node.params) > 1 else ""
+        params = params_sep.join([self.visit(x) for x in node.params])
+        return f"{node.op}({params})"
+
+    def visit_Validation(self, node: AST.Validation) -> str:
+        operand = self.visit(node.validation)
+        imbalance = f" imbalance {self.visit(node.imbalance)}" if node.imbalance is not None else ""
+        error_code = (
+            f" errorcode {_handle_literal(node.error_code)}" if node.error_code is not None else ""
+        )
+        error_level = f" errorlevel {node.error_level}" if node.error_level is not None else ""
+        invalid = " invalid" if node.invalid else " all"
+        return f"{node.op}({operand}{error_code}{error_level}{imbalance}{invalid})"
+
+    # ---------------------- Constants and IDs ----------------------
+
+    def visit_VarID(self, node: AST.VarID) -> str:
+        return _format_reserved_word(node.value)
+
+    def visit_Identifier(self, node: AST.Identifier) -> Any:
+        return _format_reserved_word(node.value)
+
+    def visit_Constant(self, node: AST.Constant) -> str:
+        if node.value is None:
+            return "null"
+        return _handle_literal(node.value)
+
+    def visit_ParamConstant(self, node: AST.ParamConstant) -> Any:
+        if node.value is None:
+            return "null"
+        return node.value
+
+    def visit_Collection(self, node: AST.Collection) -> str:
+        if node.kind == "ValueDomain":
+            return node.name
+        sep = ", " if len(node.children) > 1 else ""
+        return f"{{{sep.join([self.visit(x) for x in node.children])}}}"
+
+    def visit_Windowing(self, node: AST.Windowing) -> str:
+        if (
+            node.type_ == "data"
+            and node.start == -1
+            and node.start_mode == "preceding"
+            and node.stop == 0
+            and node.stop_mode == "current"
+        ):
+            return ""
+        if node.start == -1:
+            start = f"unbounded {node.start_mode}"
+        elif node.start_mode == "current":
+            start = "current data point"
+        else:
+            start = f"{node.start} {node.start_mode}"
+        stop = f"{node.stop} {node.stop_mode}"
+        if node.stop_mode == "current":
+            stop = "current data point"
+        mode = "data points" if node.type_ == "data" else "range"
+        return f"{mode} between {start} and {stop}"
+
+    def visit_OrderBy(self, node: AST.OrderBy) -> str:
+        if node.order == "asc":
+            return f"{_format_reserved_word(node.component)}"
+        return f"{_format_reserved_word(node.component)} {node.order}"
+
+    def visit_Comment(self, node: AST.Comment) -> None:
+        value = copy.copy(node.value)
+        value = value[:-1] if value[-1] == "\n" else value
+        self.vtl_script += value