Source code for idp_engine.Annotate

# Copyright 2019 Ingmar Dasseville, Pierre Carbonnelle
#
# This file is part of Interactive_Consultant.
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <https://www.gnu.org/licenses/>.
"""

Methods to annotate the Abstract Syntax Tree (AST) of an IDP-Z3 program.

"""

from copy import copy
from itertools import chain
import string
from typing import Dict

from .Parse import (Vocabulary, Import, TypeDeclaration, Declaration, Type,
                    SymbolDeclaration, Symbol, VarDeclaration,
                    TheoryBlock, Definition, Rule,
                    Structure, SymbolInterpretation, Enumeration, FunctionEnum,
                    Tuple, ConstructedFrom, Display)
from .Expression import (Expression, Constructor, AIfExpr, AQuantification, Quantee,
                         ARImplication, AImplication, AEquivalence,
                         Operator, AComparison, AUnary, AAggregate,
                         AppliedSymbol, UnappliedSymbol, Variable, Brackets,
                         FALSE, SymbolExpr, Number, NOT, EQUALS, AND, OR, TRUE, FALSE,
                         IMPLIES, RIMPLIES, EQUIV, FORALL, EXISTS, Extension)

from .utils import (BOOL, INT, REAL, DATE, CONCEPT, RESERVED_SYMBOLS,
                    OrderedSet, IDPZ3Error, DEF_SEMANTICS, Semantics)


# Class Vocabulary  #######################################################

def annotate(self, idp):
    self.idp = idp

    # process Import and determine the constructors of CONCEPT
    temp = {}  # contains the new self.declarations
    for s in self.declarations:
        if isinstance(s, Import):
            other = self.idp.vocabularies[s.name]
            for s1 in other.declarations:
                if s1.name in temp:
                    s.check(str(temp[s1.name]) == str(s1),
                            f"Inconsistent declaration for {s1.name}")
                temp[s1.name] = s1
        else:
            s.block = self
            s.check(s.name not in temp or s.name in RESERVED_SYMBOLS,
                    f"Duplicate declaration of {s.name}")
            temp[s.name] = s
    temp[CONCEPT].constructors=([Constructor(name=f"`{s}")
                                 for s in [BOOL, INT, REAL, DATE, CONCEPT]]
                               +[Constructor(name=f"`{s.name}")
                                 for s in temp.values()
                                 if s.name not in RESERVED_SYMBOLS
                                 and type(s) in Declaration.__args__])
    self.declarations = list(temp.values())

    # annotate declarations
    for s in self.declarations:
        s.annotate(self)  # updates self.symbol_decls

    concepts = self.symbol_decls[CONCEPT]
    for constructor in concepts.constructors:
        constructor.symbol = (Symbol(name=constructor.name[1:])
                                .annotate(self, {}))

    # populate .map of CONCEPT
    for c in concepts.constructors:
        assert not c.sorts
        concepts.map[str(c)] = UnappliedSymbol.construct(c)
Vocabulary.annotate = annotate


# Class TypeDeclaration  #######################################################

def annotate(self, voc):
    self.check(self.name not in voc.symbol_decls,
                f"duplicate declaration in vocabulary: {self.name}")
    voc.symbol_decls[self.name] = self
    for s in self.sorts:
        s.annotate(voc, {})
    self.out.annotate(voc, {})
    for c in self.constructors:
        c.type = self.name
        self.check(c.name not in voc.symbol_decls or self.name == CONCEPT,
                    f"duplicate '{c.name}' constructor for '{self.name}' type")
        voc.symbol_decls[c.name] = c
    if self.interpretation:
        self.interpretation.annotate(voc)
TypeDeclaration.annotate = annotate


# Class SymbolDeclaration  #######################################################

def annotate(self, voc):
    self.voc = voc
    self.check(self.name is not None, "Internal error")
    self.check(self.name not in voc.symbol_decls,
                f"duplicate declaration in vocabulary: {self.name}")
    voc.symbol_decls[self.name] = self
    for s in self.sorts:
        s.annotate(voc, {})
    self.out.annotate(voc, {})
    self.type = self.out.name

    for s in chain(self.sorts, [self.out]):
        self.check(s.name != CONCEPT or s == s, # use equality to check nested concepts
                   f"`Concept` must be qualified with a type signature in {self}")
    self.base_type = (None if self.out.name != BOOL or self.arity != 1 else
                      self.sorts[0].decl.base_type)
    return self
SymbolDeclaration.annotate = annotate


# Class VarDeclaration  #######################################################

def annotate(self, voc):
    self.check(self.name not in voc.symbol_decls,
                f"duplicate declaration in vocabulary: {self.name}")
    self.check(self.name == self.name.rstrip(string.digits),
                f"Variable {self.name} cannot be declared with a digital suffix.")
    voc.symbol_decls[self.name] = self
    self.subtype.annotate(voc, {})
    return self
VarDeclaration.annotate = annotate


# Class Symbol  #######################################################

def annotate(self, voc, q_vars):
    if self.name in q_vars:
        return q_vars[self.name]
    self.decl = voc.symbol_decls[self.name]
    self.type = self.decl.type
    return self
Symbol.annotate = annotate


# Class Type  #######################################################

def annotate(self, voc, q_vars={}):
    Symbol.annotate(self, voc, q_vars)
    if self.out:
        self.ins = [s.annotate(voc, q_vars) for s in self.ins]
        self.out = self.out.annotate(voc, q_vars)
    return self
Type.annotate = annotate


# Class TheoryBlock  #######################################################

def annotate(self, idp):
    self.check(self.vocab_name in idp.vocabularies,
                f"Unknown vocabulary: {self.vocab_name}")
    self.voc = idp.vocabularies[self.vocab_name]

    for i in self.interpretations.values():
        i.annotate(self)
    self.voc.add_voc_to_block(self)

    self.definitions = [e.annotate(self.voc, {}) for e in self.definitions]

    self.constraints = OrderedSet([e.annotate(self.voc, {})
                                    for e in self.constraints])
TheoryBlock.annotate = annotate


# Class Definition  #######################################################

def annotate(self, voc, q_vars):
    self.rules = [r.annotate(voc, q_vars) for r in self.rules]
    self.set_level_symbols(voc)

    # create common variables, and rename vars in rule
    self.canonicals = {}
    for r in self.rules:
        decl = voc.symbol_decls[r.definiendum.decl.name]
        if decl.name not in self.def_vars:
            name = f"${decl.name}$"
            q_v = {f"${decl.name}!{str(i)}$":
                    Variable(name=f"${decl.name}!{str(i)}$", sort=sort)
                    for i, sort in enumerate(decl.sorts)}
            if decl.out.name != BOOL:
                q_v[name] = Variable(name=name, sort=decl.out)
            self.def_vars[decl.name] = q_v
        new_rule = r.rename_args(self.def_vars[decl.name])
        self.canonicals.setdefault(decl, []).append(new_rule)

    # join the bodies of rules
    for decl, rules in self.canonicals.items():
        new_rule = copy(rules[0])
        exprs = [rule.body for rule in rules]
        new_rule.body = OR(exprs)
        self.clarks[decl] = new_rule
    return self
Definition.annotate = annotate

[docs]def get_instantiables(self, interpretations: Dict[str, SymbolInterpretation], extensions: Dict[str, Extension], for_explain=False): """ compute Definition.instantiables, with level-mapping if definition is inductive Uses implications instead of equivalence if `for_explain` is True Example: `{ p() <- q(). p() <- r().}` Result when not for_explain: `p() <=> q() | r()` Result when for_explain : `p() <= q(). p() <= r(). p() => (q() | r()).` Args: for_explain (Bool): Use implications instead of equivalence, for rule-specific explanations """ result = {} for decl, rules in self.canonicals.items(): rule = rules[0] rule.is_whole_domain = all(s.extension(interpretations, extensions)[0] is not None for s in rule.definiendum.decl.sorts) if not rule.is_whole_domain: self.check(rule.definiendum.symbol.decl not in self.level_symbols, f"Cannot have inductive definitions on infinite domain") else: if rule.out: expr = AppliedSymbol.make(rule.definiendum.symbol, rule.definiendum.sub_exprs[:-1]) expr.in_head = True head = EQUALS([expr, rule.definiendum.sub_exprs[-1]]) else: head = AppliedSymbol.make(rule.definiendum.symbol, rule.definiendum.sub_exprs) head.in_head = True inductive = (not rule.out and DEF_SEMANTICS != Semantics.COMPLETION and rule.definiendum.symbol.decl in rule.parent.level_symbols) # determine reverse implications, if any bodies, out = [], [] for r in rules: if not inductive: bodies.append(r.body) if for_explain and 1 < len(rules): # not simplified -> no need to make copies out.append(RIMPLIES([head, r.body], r.annotations)) else: new = r.body.split_equivalences() bodies.append(new) if for_explain: new = new.copy().add_level_mapping(rule.parent.level_symbols, rule.definiendum, False, False) out.append(RIMPLIES([head, new], r.annotations)) all_bodies = OR(bodies) if not inductive: if out: # already contains reverse implications out.append(IMPLIES([head, all_bodies], self.annotations)) else: out = [EQUIV([head, all_bodies], self.annotations)] else: if not out: # no reverse implication yet new = all_bodies.copy().add_level_mapping(rule.parent.level_symbols, rule.definiendum, False, False) out = [RIMPLIES([head.copy(), new], self.annotations)] all_bodies = all_bodies.copy().add_level_mapping(rule.parent.level_symbols, rule.definiendum, True, True) out.append(IMPLIES([head, all_bodies], self.annotations)) result[decl] = out return result
Definition.get_instantiables = get_instantiables # Class Rule ####################################################### def annotate(self, voc, q_vars): self.check(not self.definiendum.symbol.is_intentional(), f"No support for intentional objects in the head of a rule: " f"{self}") # create head variables q_v = {**q_vars} # copy for q in self.quantees: q.annotate(voc, q_vars) for vars in q.vars: for var in vars: var.sort = q.sub_exprs[0] if q.sub_exprs else None q_v[var.name] = var self.definiendum = self.definiendum.annotate(voc, q_v) self.body = self.body.annotate(voc, q_v) if self.out: self.out = self.out.annotate(voc, q_v) return self Rule.annotate = annotate
[docs]def rename_args(self, new_vars): """ for Clark's completion input : '!v: f(args) <- body(args)' output: '!nv: f(nv) <- nv=args & body(args)' """ self.check(len(self.definiendum.sub_exprs) == len(new_vars), "Internal error") vars = {var.name : var for q in self.quantees for vars in q.vars for var in vars} for i in range(len(self.definiendum.sub_exprs)): arg, nv = self.definiendum.sub_exprs[i], list(new_vars.values())[i] if type(arg) == Variable \ and arg.name in vars and arg.name not in new_vars: del vars[arg.name] self.body = self.body.instantiate([arg], [nv]) self.out = (self.out.instantiate([arg], [nv]) if self.out else self.out) for j in range(i, len(self.definiendum.sub_exprs)): self.definiendum.sub_exprs[j] = \ self.definiendum.sub_exprs[j].instantiate([arg], [nv]) else: eq = EQUALS([nv, arg]) self.body = AND([eq, self.body]) for v in vars.values(): self.body = EXISTS([Quantee.make(v, v.sort)], self.body) self.definiendum.sub_exprs = list(new_vars.values()) self.quantees = [Quantee.make(v, v.sort) for v in new_vars.values()] return self
Rule.rename_args = rename_args # Class Structure ####################################################### def annotate(self, idp): """ Annotates the structure with the enumerations found in it. Every enumeration is converted into an assignment, which is added to `self.assignments`. :arg idp: a `Parse.IDP` object. :returns None: """ if self.vocab_name not in idp.vocabularies: raise IDPZ3Error(f"Unknown vocabulary: {self.vocab_name}") self.voc = idp.vocabularies[self.vocab_name] for i in self.interpretations.values(): i.annotate(self) self.voc.add_voc_to_block(self) Structure.annotate = annotate # Class SymbolInterpretation ####################################################### def annotate(self, block): """ Annotate the symbol. :arg block: a Structure object :returns None: """ voc = block.voc self.block = block self.symbol = Symbol(name=self.name).annotate(voc, {}) enumeration = self.enumeration # shorthand # create constructors if it is a type enumeration self.is_type_enumeration = (type(self.symbol.decl) != SymbolDeclaration) if self.is_type_enumeration and enumeration.constructors: # create Constructors before annotating the tuples for c in enumeration.constructors: c.type = self.name self.check(c.name not in voc.symbol_decls, f"duplicate '{c.name}' constructor for '{self.name}' symbol") voc.symbol_decls[c.name] = c #TODO risk of side-effects => use local decls ? issue #81 enumeration.annotate(voc) self.check(self.is_type_enumeration or all(s.name not in [INT, REAL, DATE] # finite domain #TODO for s in self.symbol.decl.sorts) or self.default is None, f"Can't use default value for '{self.name}' on infinite domain nor for type enumeration.") self.check(not(self.symbol.decl.out.decl.base_type.name == BOOL and type(enumeration) == FunctionEnum), f"Can't use function enumeration for predicates '{self.name}' (yet)") # predicate enumeration have FALSE default if type(enumeration) != FunctionEnum and self.default is None: self.default = FALSE if self.default is not None: self.default = self.default.annotate(voc, {}) self.check(self.default.value is not None, f"Default value for '{self.name}' must be ground: {self.default}") SymbolInterpretation.annotate = annotate # Class Enumeration ####################################################### def annotate(self, voc): if self.tuples: for t in self.tuples: t.annotate(voc) Enumeration.annotate = annotate # Class Tuple ####################################################### def annotate(self, voc): self.args = [arg.annotate(voc, {}) for arg in self.args] self.check(all(a.value is not None for a in self.args), f"Tuple must be ground : ({self})") Tuple.annotate = annotate # Class ConstructedFrom ####################################################### def annotate(self, voc): for c in self.constructors: for i, ts in enumerate(c.sorts): if ts.accessor is None: ts.accessor = Symbol(name=f"{c.name}_{i}") if ts.accessor.name in self.accessors: self.check(self.accessors[ts.accessor.name] == i, "Accessors used at incompatible indices") else: self.accessors[ts.accessor.name] = i c.annotate(voc) ConstructedFrom.annotate = annotate # Class Constructor ####################################################### def annotate(self, voc): for a in self.sorts: self.check(a.type in voc.symbol_decls, f"Unknown type: {a.type}" ) a.decl = SymbolDeclaration(annotations='', name=a.accessor, sorts=[Type(name=self.type)], out=Type(name=a.type)) a.decl.annotate(voc) self.tester = SymbolDeclaration(annotations='', name=Symbol(name=f"is_{self.name}"), sorts=[Type(name=self.type)], out=Type(name=BOOL)) self.tester.annotate(voc) Constructor.annotate = annotate # Class Display ####################################################### def annotate(self, idp): self.voc = idp.vocabulary # add display predicates viewType = TypeDeclaration(name='_ViewType', constructors=[Constructor(name='normal'), Constructor(name='expanded')]) viewType.annotate(self.voc) # Check the AST for any constructors that belong to open types. # For now, the only open types are `unit` and `heading`. open_constructors = {'unit': [], 'heading': []} for constraint in self.constraints: constraint.generate_constructors(open_constructors) # Next, we convert the list of constructors to actual types. open_types = {} for name, constructors in open_constructors.items(): # If no constructors were found, then the type is not used. if not constructors: open_types[name] = None continue type_name = name.capitalize() # e.g. type Unit (not unit) open_type = TypeDeclaration(name=type_name, constructors=constructors) open_type.annotate(self.voc) open_types[name] = Symbol(name=type_name) for name, out in [ ('expand', Symbol(name=BOOL)), ('hide', Symbol(name=BOOL)), ('view', Symbol(name='_ViewType')), ('moveSymbols', Symbol(name=BOOL)), ('optionalPropagation', Symbol(name=BOOL)), ('manualPropagation', Symbol(name=BOOL)), ('optionalRelevance', Symbol(name=BOOL)), ('manualRelevance', Symbol(name=BOOL)), ('unit', open_types['unit']), ('heading', open_types['heading']), ('noOptimization', Symbol(name=BOOL)) ]: symbol_decl = SymbolDeclaration(annotations='', name=Symbol(name=name), sorts=[out] if out else [], out=out) symbol_decl.annotate(self.voc) # annotate constraints and interpretations for constraint in self.constraints: constraint.annotate(self.voc, {}) for i in self.interpretations.values(): i.annotate(self) Display.annotate = annotate # Class Expression ####################################################### def annotate(self, voc, q_vars): """annotate tree after parsing Resolve names and determine type as well as variables in the expression Args: voc (Vocabulary): the vocabulary q_vars (Dict[str, Variable]): the quantifier variables that may appear in the expression Returns: Expression: an equivalent AST node, with updated type, .variables """ self.sub_exprs = [e.annotate(voc, q_vars) for e in self.sub_exprs] return self.annotate1() Expression.annotate = annotate def annotate1(self): " annotations that are common to __init__ and make() " self.variables = set() if self.value is not None: pass if self.simpler is not None: self.variables = self.simpler.variables else: for e in self.sub_exprs: self.variables.update(e.variables) return self Expression.annotate1 = annotate1 # Class AIfExpr ####################################################### def annotate1(self): self.type = self.sub_exprs[AIfExpr.THEN].type return Expression.annotate1(self) AIfExpr.annotate1 = annotate1 # Class AQuantification ####################################################### def annotate(self, voc, q_vars): # also called by AAgregate.annotate q_v = {**q_vars} # copy for q in self.quantees: q.annotate(voc, q_vars) for vars in q.vars: for var in vars: self.check(var.name not in voc.symbol_decls or type(voc.symbol_decls[var.name]) == VarDeclaration, f"the quantified variable '{var.name}' cannot have" f" the same name as another symbol") var.sort = q.sub_exprs[0] if q.sub_exprs else None var_decl = voc.symbol_decls.get(var.name.rstrip(string.digits), None) if var_decl: subtype = var_decl.subtype self.check(var.sort is None or var.sort.name == subtype.name, f"Can't use declared {var.name} as a " f"{var.sort.name if var.sort else ''}") if var.sort is None: q.sub_exprs = [subtype.annotate(voc, {})] q_v[var.name] = var self.sub_exprs = [e.annotate(voc, q_v) for e in self.sub_exprs] return self.annotate1() AQuantification.annotate = annotate def annotate1(self): Expression.annotate1(self) for q in self.quantees: # remove declared variables for vs in q.vars: for v in vs: self.variables.discard(v.name) for q in self.quantees: # add variables in sort expression for sort in q.sub_exprs: self.variables.update(sort.variables) return self AQuantification.annotate1 = annotate1 # Class Operator ####################################################### def annotate1(self): if self.type is None: self.type = REAL if any(e.type == REAL for e in self.sub_exprs) \ else INT if any(e.type == INT for e in self.sub_exprs) \ else self.sub_exprs[0].type # constructed type, without arithmetic return Expression.annotate1(self) Operator.annotate1 = annotate1 # Class AImplication ####################################################### def annotate1(self): self.check(len(self.sub_exprs) == 2, "Implication is not associative. Please use parenthesis.") self.type = BOOL return Expression.annotate1(self) AImplication.annotate1 = annotate1 # Class AEquivalence ####################################################### def annotate1(self): self.check(len(self.sub_exprs) == 2, "Equivalence is not associative. Please use parenthesis.") self.type = BOOL return Expression.annotate1(self) AEquivalence.annotate1 = annotate1 # Class ARImplication ####################################################### def annotate(self, voc, q_vars): # reverse the implication self.sub_exprs.reverse() out = AImplication(sub_exprs=self.sub_exprs, operator=['⇒']*len(self.operator)) if hasattr(self, "block"): out.block = self.block return out.annotate(voc, q_vars) ARImplication.annotate = annotate # Class AComparison ####################################################### def annotate(self, voc, q_vars): out = Operator.annotate(self, voc, q_vars) out.type = BOOL # a≠b --> Not(a=b) if len(self.sub_exprs) == 2 and self.operator == ['≠']: out = NOT(EQUALS(self.sub_exprs)) return out AComparison.annotate = annotate # Class AUnary ####################################################### def annotate1(self): if len(self.operators) % 2 == 0: # negation of negation return self.sub_exprs[0] self.type = self.sub_exprs[0].type return Expression.annotate1(self) AUnary.annotate1 = annotate1 # Class AAggregate ####################################################### def annotate(self, voc, q_vars): self = AQuantification.annotate(self, voc, q_vars) if not self.annotated: assert len(self.sub_exprs) == 1, "Internal error" if self.aggtype == "#": self.sub_exprs = [AIfExpr.make(self.sub_exprs[0], Number(number='1'), Number(number='0'))] self.type = INT else: self.type = self.sub_exprs[0].type if self.aggtype in ["min", "max"]: # the `min` aggregate in `!y in T: min(lamda x in type: term(x,y))=0` # is replaced by `_*(y)` with the following co-constraint: # !y in T: ( ?x in type: term(x) = _*(y) # !x in type: term(x) =< _*(y). name = "_" + self.str if name in voc.symbol_decls: symbol_decl = voc.symbol_decls[name] to_create = False else: symbol_decl = SymbolDeclaration.make( "_"+self.str, # name `_ *` len(q_vars), # arity [Type(name=v.sort.code) for v in q_vars.values()], Type(name=self.type)).annotate(voc) # output_domain to_create = True symbol = Symbol(name=symbol_decl.name) applied = AppliedSymbol.make(symbol, q_vars.values()) applied = applied.annotate(voc, q_vars) if to_create: coc1 = EXISTS(self.quantees, EQUALS([applied.copy(), self.sub_exprs[0]])) op = '≤' if self.aggtype == "min" else '≥' coc2 = FORALL(self.quantees.copy(), AComparison.make(op, [applied.copy(), self.sub_exprs[0].copy()])) coc = AND([coc1, coc2]) quantees = [Quantee.make(v, v.sort) for v in q_vars.values()] applied.co_constraint = FORALL(quantees, coc).annotate(voc, q_vars) applied.co_constraint.annotations['reading'] = f"Calculation of {self.code}" return applied self.annotated = True return self AAggregate.annotate = annotate AAggregate.annotate1 = AQuantification.annotate1 # Class AppliedSymbol ####################################################### def annotate(self, voc, q_vars): self.symbol = self.symbol.annotate(voc, q_vars) self.check((not self.symbol.decl or type(self.symbol.decl) != Constructor or 0 < self.symbol.decl.arity), f"Constructor `{self.symbol}` cannot be applied to argument(s)") self.sub_exprs = [e.annotate(voc, q_vars) for e in self.sub_exprs] if self.in_enumeration: self.in_enumeration.annotate(voc) out = self.annotate1() # move the negation out if 'not' in self.is_enumerated: out = AppliedSymbol.make(out.symbol, out.sub_exprs, is_enumerated='is enumerated') out = NOT(out) elif 'not' in self.is_enumeration: out = AppliedSymbol.make(out.symbol, out.sub_exprs, is_enumeration='in', in_enumeration=out.in_enumeration) out = NOT(out) return out AppliedSymbol.annotate = annotate def annotate1(self): out = Expression.annotate1(self) out.symbol = out.symbol.annotate1() out.variables.update(out.symbol.variables) return out.simplify1() AppliedSymbol.annotate1 = annotate1 # Class SymbolExpr ####################################################### def annotate(self, voc, q_vars): out = Expression.annotate(self, voc, q_vars) return out.simplify1() SymbolExpr.annotate = annotate # Class Variable ####################################################### def annotate(self, voc, q_vars): self.type = self.sort.decl.name if self.sort and self.sort.decl else '' return self Variable.annotate = annotate # Class Number ####################################################### def annotate(self, voc, q_vars): self.decl = voc.symbol_decls[self.type] return self Number.annotate = annotate # Class UnappliedSymbol ####################################################### def annotate(self, voc, q_vars): if self.name in q_vars: # ignore VarDeclaration return q_vars[self.name] if self.name in voc.symbol_decls: self.decl = voc.symbol_decls[self.name] self.variables = {} self.check(type(self.decl) == Constructor, f"{self} should be applied to arguments (or prefixed with a back-tick)") return self # elif self.name in voc.symbol_decls: # in symbol_decls # out = AppliedSymbol.make(self.s, self.sub_exprs) # return out.annotate(voc, q_vars) # If this code is reached, an undefined symbol was present. if self.name.rstrip(string.digits) in q_vars: # after considering it as a declared symbol return self self.check(False, f"Symbol not in vocabulary: {self}") UnappliedSymbol.annotate = annotate # Class Brackets ####################################################### def annotate1(self): if not self.annotations: return self.sub_exprs[0] # remove the bracket self.type = self.sub_exprs[0].type if self.annotations['reading']: self.sub_exprs[0].annotations = self.annotations self.variables = self.sub_exprs[0].variables return self Brackets.annotate1 = annotate1 Done = True