aboutsummaryrefslogtreecommitdiffstats
path: root/bsfs/query
diff options
context:
space:
mode:
Diffstat (limited to 'bsfs/query')
-rw-r--r--bsfs/query/__init__.py15
-rw-r--r--bsfs/query/ast/__init__.py23
-rw-r--r--bsfs/query/ast/fetch.py169
-rw-r--r--bsfs/query/ast/filter_.py516
-rw-r--r--bsfs/query/matcher.py361
-rw-r--r--bsfs/query/validator.py351
6 files changed, 1435 insertions, 0 deletions
diff --git a/bsfs/query/__init__.py b/bsfs/query/__init__.py
new file mode 100644
index 0000000..58ff03a
--- /dev/null
+++ b/bsfs/query/__init__.py
@@ -0,0 +1,15 @@
+
+# imports
+import typing
+
+# inner-module imports
+from . import ast
+from . import validator as validate
+
+# exports
+__all__: typing.Sequence[str] = (
+ 'ast',
+ 'validate',
+ )
+
+## EOF ##
diff --git a/bsfs/query/ast/__init__.py b/bsfs/query/ast/__init__.py
new file mode 100644
index 0000000..bceaac0
--- /dev/null
+++ b/bsfs/query/ast/__init__.py
@@ -0,0 +1,23 @@
+"""Query AST components.
+
+The query AST consists of a Filter and a Fetch syntax trees.
+
+Classes beginning with an underscore (_) represent internal type hierarchies
+and should not be used for parsing. Note that the AST structures do not
+(and cannot) check semantic validity or consistency with a given schema.
+
+"""
+# imports
+import typing
+
+# inner-module imports
+from . import fetch
+from . import filter_ as filter # pylint: disable=redefined-builtin
+
+# exports
+__all__: typing.Sequence[str] = (
+ 'fetch',
+ 'filter',
+ )
+
+## EOF ##
diff --git a/bsfs/query/ast/fetch.py b/bsfs/query/ast/fetch.py
new file mode 100644
index 0000000..66d94e1
--- /dev/null
+++ b/bsfs/query/ast/fetch.py
@@ -0,0 +1,169 @@
+
+# imports
+from collections import abc
+import typing
+
+# bsfs imports
+from bsfs.utils import URI, typename, normalize_args
+
+# exports
+__all__ : typing.Sequence[str] = (
+ 'All',
+ 'Fetch',
+ 'FetchExpression',
+ 'Node',
+ 'This',
+ 'Value',
+ )
+
+
+## code ##
+
+class FetchExpression(abc.Hashable):
+ """Generic Fetch expression."""
+
+ def __repr__(self) -> str:
+ """Return the expressions's string representation."""
+ return f'{typename(self)}()'
+
+ def __hash__(self) -> int:
+ """Return the expression's integer representation."""
+ return hash(type(self))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ """Return True if *self* and *other* are equivalent."""
+ return isinstance(other, type(self))
+
+
+class All(FetchExpression):
+ """Fetch all child expressions."""
+
+ # child expressions.
+ expr: typing.Set[FetchExpression]
+
+ def __init__(self, *expr):
+ # unpack child expressions
+ unfolded = set(normalize_args(*expr))
+ # check child expressions
+ if len(unfolded) == 0:
+ raise AttributeError('expected at least one expression, found none')
+ if not all(isinstance(itm, FetchExpression) for itm in unfolded):
+ raise TypeError(expr)
+ # initialize
+ super().__init__()
+ # assign members
+ self.expr = unfolded
+
+ def __iter__(self) -> typing.Iterator[FetchExpression]:
+ return iter(self.expr)
+
+ def __len__(self) -> int:
+ return len(self.expr)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.expr})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), tuple(sorted(self.expr, key=repr))))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) and self.expr == other.expr
+
+
+class _Branch(FetchExpression):
+ """Branch along a predicate."""
+
+ # FIXME: Use a Predicate (like in ast.filter) so that we can also reverse them!
+
+ # predicate to follow.
+ predicate: URI
+
+ def __init__(self, predicate: URI):
+ if not isinstance(predicate, URI):
+ raise TypeError(predicate)
+ self.predicate = predicate
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.predicate})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.predicate))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) and self.predicate == other.predicate
+
+
+class Fetch(_Branch):
+ """Follow a predicate before evaluating a child epxression."""
+
+ # child expression.
+ expr: FetchExpression
+
+ def __init__(self, predicate: URI, expr: FetchExpression):
+ # check child expressions
+ if not isinstance(expr, FetchExpression):
+ raise TypeError(expr)
+ # initialize
+ super().__init__(predicate)
+ # assign members
+ self.expr = expr
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.predicate}, {self.expr})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.expr))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) and self.expr == other.expr
+
+
+class _Named(_Branch):
+ """Fetch a (named) symbol at a predicate."""
+
+ # symbol name.
+ name: str
+
+ def __init__(self, predicate: URI, name: str):
+ super().__init__(predicate)
+ self.name = str(name)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.predicate}, {self.name})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.name))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) and self.name == other.name
+
+
+class Node(_Named): # pylint: disable=too-few-public-methods
+ """Fetch a Node at a predicate."""
+ # FIXME: Is this actually needed?
+
+
+class Value(_Named): # pylint: disable=too-few-public-methods
+ """Fetch a Literal at a predicate."""
+
+
+class This(FetchExpression):
+ """Fetch the current Node."""
+
+ # symbol name.
+ name: str
+
+ def __init__(self, name: str):
+ super().__init__()
+ self.name = str(name)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.name})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.name))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) and self.name == other.name
+
+## EOF ##
diff --git a/bsfs/query/ast/filter_.py b/bsfs/query/ast/filter_.py
new file mode 100644
index 0000000..610fdb4
--- /dev/null
+++ b/bsfs/query/ast/filter_.py
@@ -0,0 +1,516 @@
+"""Filter AST.
+
+Note that it is easily possible to construct an AST that is inconsistent with
+a given schema. Furthermore, it is possible to construct a semantically invalid
+AST which that cannot be parsed correctly or includes contradicting statements.
+The AST nodes do not (and cannot) check such issues.
+
+For example, consider the following AST:
+
+>>> Any(ns.bse.collection,
+... And(
+... Equals('hello'),
+... Is('hello world'),
+... Any(ns.bse.tag, Equals('world')),
+... Any(ns.bst.label, Equals('world')),
+... All(ns.bst.label, Not(Equals('world'))),
+... )
+... )
+
+This AST has multiple issues that are not verified upon its creation:
+* A condition on a non-literal.
+* A Filter on a literal.
+* Conditions exclude each other
+* The predicate along the branch have incompatible domains and ranges.
+
+"""
+# imports
+from collections import abc
+import typing
+
+# bsfs imports
+from bsfs.utils import URI, typename, normalize_args
+
+# exports
+__all__ : typing.Sequence[str] = (
+ # base classes
+ 'FilterExpression',
+ 'PredicateExpression',
+ # predicate expressions
+ 'OneOf',
+ 'Predicate',
+ # branching
+ 'All',
+ 'Any',
+ # aggregators
+ 'And',
+ 'Or',
+ # value matchers
+ 'Equals',
+ 'Substring',
+ 'EndsWith',
+ 'StartsWith',
+ # range matchers
+ 'GreaterThan',
+ 'LessThan',
+ # misc
+ 'Has',
+ 'Is',
+ 'Not',
+ )
+
+
+## code ##
+
+# pylint: disable=too-few-public-methods # Many expressions use mostly magic methods
+
+class _Expression(abc.Hashable):
+ def __repr__(self) -> str:
+ """Return the expressions's string representation."""
+ return f'{typename(self)}()'
+
+ def __hash__(self) -> int:
+ """Return the expression's integer representation."""
+ return hash(type(self))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ """Return True if *self* and *other* are equivalent."""
+ return isinstance(other, type(self))
+
+
+class FilterExpression(_Expression):
+ """Generic Filter expression."""
+
+
+class PredicateExpression(_Expression):
+ """Generic Predicate expression."""
+
+
+class _Branch(FilterExpression):
+ """Branch the filter along a predicate."""
+
+ # predicate to follow.
+ predicate: PredicateExpression
+
+ # child expression to evaluate.
+ expr: FilterExpression
+
+ def __init__(
+ self,
+ predicate: typing.Union[PredicateExpression, URI],
+ expr: FilterExpression,
+ ):
+ # process predicate argument
+ if isinstance(predicate, URI):
+ predicate = Predicate(predicate)
+ elif not isinstance(predicate, PredicateExpression):
+ raise TypeError(predicate)
+ # process expression argument
+ if not isinstance(expr, FilterExpression):
+ raise TypeError(expr)
+ # assign members
+ self.predicate = predicate
+ self.expr = expr
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.predicate}, {self.expr})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.predicate, self.expr))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) \
+ and self.predicate == other.predicate \
+ and self.expr == other.expr
+
+class Any(_Branch):
+ """Any (and at least one) triple matches."""
+
+
+class All(_Branch):
+ """All (and at least one) triples match."""
+
+
+class _Agg(FilterExpression, abc.Collection):
+ """Combine multiple expressions."""
+
+ # child expressions
+ expr: typing.Set[FilterExpression]
+
+ def __init__(
+ self,
+ *expr: typing.Union[FilterExpression,
+ typing.Iterable[FilterExpression],
+ typing.Iterator[FilterExpression]]
+ ):
+ # unfold arguments
+ unfolded = set(normalize_args(*expr))
+ # check type
+ if not all(isinstance(e, FilterExpression) for e in unfolded):
+ raise TypeError(expr)
+ # FIXME: Require at least one child expression?
+ # assign member
+ self.expr = unfolded
+
+ def __contains__(self, expr: typing.Any) -> bool:
+ """Return True if *expr* is among the child expressions."""
+ return expr in self.expr
+
+ def __iter__(self) -> typing.Iterator[FilterExpression]:
+ """Iterator over child expressions."""
+ return iter(self.expr)
+
+ def __len__(self) -> int:
+ """Number of child expressions."""
+ return len(self.expr)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.expr})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), tuple(sorted(self.expr, key=repr))))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) and self.expr == other.expr
+
+
+class And(_Agg):
+ """All conditions match."""
+
+
+class Or(_Agg):
+ """At least one condition matches."""
+
+
+class Not(FilterExpression):
+ """Invert a statement."""
+
+ # child expression
+ expr: FilterExpression
+
+ def __init__(self, expr: FilterExpression):
+ # check argument
+ if not isinstance(expr, FilterExpression):
+ raise TypeError(expr)
+ # assign member
+ self.expr = expr
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.expr})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.expr))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) and self.expr == other.expr
+
+
+class Has(FilterExpression):
+ """Has predicate N times"""
+
+ # predicate to follow.
+ predicate: PredicateExpression
+
+ # target count
+ count: FilterExpression
+
+ def __init__(
+ self,
+ predicate: typing.Union[PredicateExpression, URI],
+ count: typing.Optional[typing.Union[FilterExpression, int]] = None,
+ ):
+ # check predicate
+ if isinstance(predicate, URI):
+ predicate = Predicate(predicate)
+ elif not isinstance(predicate, PredicateExpression):
+ raise TypeError(predicate)
+ # check count
+ if count is None:
+ count = GreaterThan(1, strict=False)
+ elif isinstance(count, int):
+ count = Equals(count)
+ elif not isinstance(count, FilterExpression):
+ raise TypeError(count)
+ # assign members
+ self.predicate = predicate
+ self.count = count
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.predicate}, {self.count})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.predicate, self.count))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) \
+ and self.predicate == other.predicate \
+ and self.count == other.count
+
+
+class _Value(FilterExpression):
+ """Matches some value."""
+
+ # target value.
+ value: typing.Any
+
+ def __init__(self, value: typing.Any):
+ self.value = value
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.value})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.value))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) and self.value == other.value
+
+
+class Is(_Value):
+ """Match the URI of a node."""
+
+
+class Equals(_Value):
+ """Value matches exactly.
+ NOTE: Value must correspond to literal type.
+ """
+
+
+class Substring(_Value):
+ """Value matches a substring
+ NOTE: value must be a string.
+ """
+
+
+class StartsWith(_Value):
+ """Value begins with a given string."""
+
+
+class EndsWith(_Value):
+ """Value ends with a given string."""
+
+
+class Distance(FilterExpression):
+ """Distance to a reference is (strictly) below a threshold. Assumes a Feature literal."""
+
+ # FIXME:
+ # (a) pass a node/predicate as anchor instead of a value.
+ # Then we don't need to materialize the reference.
+ # (b) pass a FilterExpression (_Bounded) instead of a threshold.
+ # Then, we could also query values greater than a threshold.
+
+ # reference value.
+ reference: typing.Any
+
+ # distance threshold.
+ threshold: float
+
+ # closed (True) or open (False) bound.
+ strict: bool
+
+ def __init__(
+ self,
+ reference: typing.Any,
+ threshold: float,
+ strict: bool = False,
+ ):
+ self.reference = reference
+ self.threshold = float(threshold)
+ self.strict = bool(strict)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.reference}, {self.threshold}, {self.strict})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), tuple(self.reference), self.threshold, self.strict))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) \
+ and self.reference == other.reference \
+ and self.threshold == other.threshold \
+ and self.strict == other.strict
+
+
+class _Bounded(FilterExpression):
+ """Value is bounded by a threshold. Assumes a Number literal."""
+
+ # bound.
+ threshold: float
+
+ # closed (True) or open (False) bound.
+ strict: bool
+
+ def __init__(
+ self,
+ threshold: float,
+ strict: bool = True,
+ ):
+ self.threshold = float(threshold)
+ self.strict = bool(strict)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.threshold}, {self.strict})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.threshold, self.strict))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) \
+ and self.threshold == other.threshold \
+ and self.strict == other.strict
+
+
+
+class LessThan(_Bounded):
+ """Value is (strictly) smaller than threshold. Assumes a Number literal."""
+
+
+class GreaterThan(_Bounded):
+ """Value is (strictly) larger than threshold. Assumes a Number literal."""
+
+
+class Predicate(PredicateExpression):
+ """A single predicate."""
+
+ # predicate URI
+ predicate: URI
+
+ # reverse the predicate's direction
+ reverse: bool
+
+ def __init__(
+ self,
+ predicate: URI,
+ reverse: typing.Optional[bool] = False,
+ ):
+ # check arguments
+ if not isinstance(predicate, URI):
+ raise TypeError(predicate)
+ # assign members
+ self.predicate = predicate
+ self.reverse = bool(reverse)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.predicate}, {self.reverse})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.predicate, self.reverse))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) \
+ and self.predicate == other.predicate \
+ and self.reverse == other.reverse
+
+
+class OneOf(PredicateExpression, abc.Collection):
+ """A set of predicate alternatives.
+
+ The predicates' domains must be ascendants or descendants of each other.
+ The overall domain is the most specific one.
+
+ The predicate's domains must be ascendants or descendants of each other.
+ The overall range is the most generic one.
+ """
+
+ # predicate alternatives
+ expr: typing.Set[PredicateExpression]
+
+ def __init__(self, *expr: typing.Union[PredicateExpression, URI]):
+ # unfold arguments
+ unfolded = set(normalize_args(*expr)) # type: ignore [arg-type] # this is getting too complex...
+ # check arguments
+ if len(unfolded) == 0:
+ raise AttributeError('expected at least one expression, found none')
+ # ensure PredicateExpression
+ unfolded = {Predicate(e) if isinstance(e, URI) else e for e in unfolded}
+ # check type
+ if not all(isinstance(e, PredicateExpression) for e in unfolded):
+ raise TypeError(expr)
+ # assign member
+ self.expr = unfolded
+
+ def __contains__(self, expr: typing.Any) -> bool:
+ """Return True if *expr* is among the child expressions."""
+ return expr in self.expr
+
+ def __iter__(self) -> typing.Iterator[PredicateExpression]:
+ """Iterator over child expressions."""
+ return iter(self.expr)
+
+ def __len__(self) -> int:
+ """Number of child expressions."""
+ return len(self.expr)
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.expr})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), tuple(sorted(self.expr, key=repr))))
+
+ def __eq__(self, other) -> bool:
+ return super().__eq__(other) and self.expr == other.expr
+
+
+# Helpers
+# invalid-name is disabled since they explicitly mimic an expression
+
+def IsIn(*values) -> FilterExpression: # pylint: disable=invalid-name
+ """Match any of the given URIs."""
+ args = normalize_args(*values)
+ if len(args) == 0:
+ raise AttributeError('expected at least one value, found none')
+ if len(args) == 1:
+ return Is(args[0])
+ return Or(Is(value) for value in args)
+
+def IsNotIn(*values) -> FilterExpression: # pylint: disable=invalid-name
+ """Match none of the given URIs."""
+ return Not(IsIn(*values))
+
+
+def Between( # pylint: disable=invalid-name
+ lo: float = float('-inf'),
+ hi: float = float('inf'),
+ lo_strict: bool = True,
+ hi_strict: bool = True,
+ ) -> FilterExpression :
+ """Match numerical values between *lo* and *hi*. Include bounds if strict is False."""
+ if abs(lo) == hi == float('inf'):
+ raise ValueError('range cannot be INF on both sides')
+ if lo > hi:
+ raise ValueError(f'lower bound ({lo}) cannot be less than upper bound ({hi})')
+ if lo == hi and not lo_strict and not hi_strict:
+ return Equals(lo)
+ if lo == hi: # either bound is strict
+ raise ValueError('bounds cannot be equal when either is strict')
+ if lo != float('-inf') and hi != float('inf'):
+ return And(GreaterThan(lo, lo_strict), LessThan(hi, hi_strict))
+ if lo != float('-inf'):
+ return GreaterThan(lo, lo_strict)
+ # hi != float('inf'):
+ return LessThan(hi, hi_strict)
+
+
+def Includes(*values, approx: bool = False) -> FilterExpression: # pylint: disable=invalid-name
+ """Match any of the given *values*. Uses `Substring` if *approx* is set."""
+ args = normalize_args(*values)
+ cls = Substring if approx else Equals
+ if len(args) == 0:
+ raise AttributeError('expected at least one value, found none')
+ if len(args) == 1:
+ return cls(args[0])
+ return Or(cls(v) for v in args)
+
+
+def Excludes(*values, approx: bool = False) -> FilterExpression: # pylint: disable=invalid-name
+ """Match none of the given *values*. Uses `Substring` if *approx* is set."""
+ args = normalize_args(*values)
+ cls = Substring if approx else Equals
+ if len(args) == 0:
+ raise AttributeError('expected at least one value, found none')
+ if len(args) == 1:
+ return Not(cls(args[0]))
+ return Not(Or(cls(v) for v in args))
+
+
+## EOF ##
diff --git a/bsfs/query/matcher.py b/bsfs/query/matcher.py
new file mode 100644
index 0000000..17c9c8e
--- /dev/null
+++ b/bsfs/query/matcher.py
@@ -0,0 +1,361 @@
+
+# imports
+from collections import defaultdict
+from itertools import product
+from time import time
+import random
+import threading
+import typing
+
+# external imports
+from hopcroftkarp import HopcroftKarp
+
+# bsfs imports
+from bsfs.utils import errors, typename
+
+# inner-module imports
+from . import ast
+
+# exports
+__all__ : typing.Sequence[str] = (
+ 'Filter',
+ )
+
+
+## code ##
+
+class Any(ast.filter.FilterExpression, ast.filter.PredicateExpression):
+ """Match any ast class.
+
+ Note that Any instances are unique, i.e. they do not compare, and
+ can hence be repeated in a set:
+ >>> Any() == Any()
+ False
+ >>> len({Any(), Any(), Any(), Any()})
+ 4
+
+ """
+
+ # unique instance id
+ _uid: typing.Tuple[int, int, float, float]
+
+ def __init__(self):
+ self._uid = (
+ id(self),
+ id(threading.current_thread()),
+ time(),
+ random.random(),
+ )
+
+ def __eq__(self, other: typing.Any):
+ return super().__eq__(other) and self._uid == other._uid
+
+ def __hash__(self):
+ return hash((super().__hash__(), self._uid))
+
+
+class Rest(ast.filter.FilterExpression, ast.filter.PredicateExpression):
+ """Match the leftovers in a set of items to be compared.
+
+ Rest can be used in junction with aggregating expressions such as ast.filter.And,
+ ast.filter.Or, ast.filter.OneOf. It controls childs expressions that were not yet
+ consumed by other matching rules. Rest may match to only a specific expression.
+ The expresssion defaults to Any().
+
+ For example, the following to ast structures would match since Rest
+ allows an arbitrary repetition of ast.filter.Equals statements.
+
+ >>> And(Equals('hello'), Equals('world'), Equals('foobar'))
+ >>> And(Equals('world'), Rest(Partial(Equals)))
+
+ """
+
+ # child expression for the Rest.
+ expr: typing.Union[ast.filter.FilterExpression, ast.filter.PredicateExpression]
+
+ def __init__(
+ self,
+ expr: typing.Optional[typing.Union[ast.filter.FilterExpression, ast.filter.PredicateExpression]] = None,
+ ):
+ if expr is None:
+ expr = Any()
+ self.expr = expr
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.expr})'
+
+ def __hash__(self) -> int:
+ return hash((super().__hash__(), self.expr))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) and self.expr == other.expr
+
+
+class Partial(ast.filter.FilterExpression, ast.filter.PredicateExpression):
+ """Match a partially defined ast expression.
+
+ Literal values might be irrelevant or unknown when comparing two ast
+ structures. Partial allows to constrain the matcher to a certain
+ ast class, while leaving some of its members unspecified.
+
+ Pass the class (not instance) and its members as keyword arguments
+ to Partial. Note that the arguments are not validated.
+
+ For example, the following instance matches any ast.filter.Equals,
+ irrespective of its value:
+
+ >>> Partial(ast.filter.Equals)
+
+ Likewise, the following instance matches any ast.filter.LessThan
+ that has a strict bounds, but makes no claim about the threshold:
+
+ >>> Partial(ast.filter.LessThan, strict=False)
+
+ """
+
+ # target node type.
+ node: typing.Type
+
+ # node construction args.
+ kwargs: typing.Dict[str, typing.Any]
+
+ def __init__(
+ self,
+ node: typing.Type,
+ **kwargs,
+ ):
+ self.node = node
+ self.kwargs = kwargs
+
+ def __repr__(self) -> str:
+ return f'{typename(self)}({self.node.__name__}, {self.kwargs})'
+
+ def __hash__(self) -> int:
+ kwargs = tuple((key, self.kwargs[key]) for key in sorted(self.kwargs))
+ return hash((super().__hash__(), self.node, kwargs))
+
+ def __eq__(self, other: typing.Any) -> bool:
+ return super().__eq__(other) \
+ and self.node == other.node \
+ and self.kwargs == other.kwargs
+
+ def match(
+ self,
+ name: str,
+ value: typing.Any,
+ ) -> bool:
+ """Return True if *name* is unspecified or matches *value*."""
+ return name not in self.kwargs or self.kwargs[name] == value
+
+
+T_ITEM_TYPE = typing.TypeVar('T_ITEM_TYPE') # pylint: disable=invalid-name
+
+def _set_matcher(
+ query: typing.Collection[T_ITEM_TYPE],
+ reference: typing.Collection[T_ITEM_TYPE],
+ cmp: typing.Callable[[T_ITEM_TYPE, T_ITEM_TYPE], bool],
+ ) -> bool:
+ """Compare two sets of child expressions.
+
+ This check has a best-case complexity of O(|N|**2) and worst-case
+ complexity of O(|N|**3), with N the number of child expressions.
+ """
+ # get reference items
+ r_items = list(reference)
+ # deal with Rest
+ r_rest = {itm for itm in r_items if isinstance(itm, Rest)}
+ if len(r_rest) > 1:
+ raise errors.BackendError(f'there must be at most one Rest instance per set, found {len(r_rest)}')
+ if len(r_rest) == 1:
+ # replace Rest by filling the reference up with rest's expression
+ # NOTE: convert r_items to list so that items can be repeated
+ expr = next(iter(r_rest)).expr # type: ignore [attr-defined]
+ r_items = [itm for itm in r_items if not isinstance(itm, Rest)]
+ r_items += [expr for _ in range(len(query) - len(r_items))] # type: ignore [misc]
+ # sanity check: cannot match if the item sizes differ:
+ # either a reference item is unmatched (len(r_items) > len(query))
+ # or a query item is unmatched (len(r_items) < len(query))
+ if len(query) != len(r_items):
+ return False
+
+ # To have a positive match between the query and the reference,
+ # each query expr has to match any reference expr.
+ # However, each reference expr can only be "consumed" once even
+ # if it matches multiple query exprs (e.g., the Any expression matches
+ # every query expr).
+ # This is a bipartide matching problem (Hall's marriage problem)
+ # and the Hopcroft-Karp-Karzanov algorithm finds a maximum
+ # matching. While there might be multiple maximum matchings,
+ # we only need to know whether (at least) one complete matching
+ # exists. The hopcroftkarp module provides this functionality.
+ # The HKK algorithm has worst-case complexity of O(|N|**2 * sqrt(|N|))
+ # and we also need to compare expressions pairwise, hence O(|N|**2).
+ num_items = len(r_items)
+ graph = defaultdict(set)
+ # build the bipartide graph as {lhs: {rhs}, ...}
+ # lhs and rhs must be disjoint identifiers.
+ for (ridx, ref), (nidx, node) in product(enumerate(r_items), enumerate(query)):
+ # add edges for equal expressions
+ if cmp(node, ref):
+ graph[ridx].add(num_items + nidx)
+
+ # maximum_matching returns the matches for all nodes in the graph
+ # ({ref_itm: node_itm}), hence a complete matching's size is
+ # the number of reference's child expressions.
+ return len(HopcroftKarp(graph).maximum_matching(keys_only=True)) == num_items
+
+
+class Filter():
+ """Compare a bsfs.query.ast.filter` query's structure to a reference ast.
+
+ The reference ast may include `Rest`, `Partial`, or `Any` to account for irrelevant
+ or unknown ast pieces.
+
+ This is only a structural comparison, not a semantic one. For example, the
+ two following queries are semantically identical, but structurally different,
+ and would therefore not match:
+
+ >>> ast.filter.OneOf(ast.filter.Predicate(ns.bse.name))
+ >>> ast.filter.Predicate(ns.bse.name)
+
+ """
+
+ def __call__(self, query: ast.filter.FilterExpression, reference: ast.filter.FilterExpression) -> bool:
+ """Compare a *query* to a *reference* ast structure.
+ Return True if both are structurally equivalent.
+ """
+ if not isinstance(query, ast.filter.FilterExpression):
+ raise errors.BackendError(f'expected filter expression, found {query}')
+ if not isinstance(reference, ast.filter.FilterExpression):
+ raise errors.BackendError(f'expected filter expression, found {reference}')
+ return self._parse_filter_expression(query, reference)
+
+ def _parse_filter_expression(
+ self,
+ node: ast.filter.FilterExpression,
+ reference: ast.filter.FilterExpression,
+ ) -> bool:
+ """Route *node* to the handler of the respective FilterExpression subclass."""
+ # generic checks: reference type must be Any or match node type
+ if isinstance(reference, Any):
+ return True
+ # node-specific checks
+ if isinstance(node, ast.filter.Not):
+ return self._not(node, reference)
+ if isinstance(node, ast.filter.Has):
+ return self._has(node, reference)
+ if isinstance(node, ast.filter.Distance):
+ return self._distance(node, reference)
+ if isinstance(node, (ast.filter.Any, ast.filter.All)):
+ return self._branch(node, reference)
+ if isinstance(node, (ast.filter.And, ast.filter.Or)):
+ return self._agg(node, reference)
+ if isinstance(node, (ast.filter.Is, ast.filter.Equals, ast.filter.Substring,
+ ast.filter.StartsWith, ast.filter.EndsWith)):
+ return self._value(node, reference)
+ if isinstance(node, (ast.filter.LessThan, ast.filter.GreaterThan)):
+ return self._bounded(node, reference)
+ # invalid node
+ raise errors.BackendError(f'expected filter expression, found {node}')
+
+ def _parse_predicate_expression(
+ self,
+ node: ast.filter.PredicateExpression,
+ reference: ast.filter.PredicateExpression,
+ ) -> bool:
+ """Route *node* to the handler of the respective PredicateExpression subclass."""
+ if isinstance(reference, Any):
+ return True
+ if isinstance(node, ast.filter.Predicate):
+ return self._predicate(node, reference)
+ if isinstance(node, ast.filter.OneOf):
+ return self._one_of(node, reference)
+ # invalid node
+ raise errors.BackendError(f'expected predicate expression, found {node}')
+
+ def _one_of(self, node: ast.filter.OneOf, reference: ast.filter.PredicateExpression) -> bool:
+ if not isinstance(reference, type(node)):
+ return False
+ return _set_matcher(node, reference, self._parse_predicate_expression)
+
+ def _predicate(self, node: ast.filter.Predicate, reference: ast.filter.PredicateExpression) -> bool:
+ if not isinstance(reference, (Partial, type(node))):
+ return False
+ # partial check
+ if isinstance(reference, Partial):
+ if not isinstance(node, reference.node):
+ return False
+ return reference.match('predicate', node.predicate) \
+ and reference.match('reverse', node.reverse)
+ # full check
+ return node.predicate == reference.predicate \
+ and node.reverse == reference.reverse
+
+ def _branch(self,
+ node: typing.Union[ast.filter.Any, ast.filter.All],
+ reference: ast.filter.FilterExpression,
+ ) -> bool:
+ if not isinstance(reference, type(node)):
+ return False
+ if not self._parse_predicate_expression(node.predicate, reference.predicate): # type: ignore [attr-defined]
+ return False
+ if not self._parse_filter_expression(node.expr, reference.expr): # type: ignore [attr-defined]
+ return False
+ return True
+
+ def _agg(self, node: typing.Union[ast.filter.And, ast.filter.Or], reference: ast.filter.FilterExpression) -> bool:
+ if not isinstance(reference, type(node)):
+ return False
+ return _set_matcher(node, reference, self._parse_filter_expression) # type: ignore [arg-type]
+
+ def _not(self, node: ast.filter.Not, reference: ast.filter.FilterExpression) -> bool:
+ if not isinstance(reference, type(node)):
+ return False
+ return self._parse_filter_expression(node.expr, reference.expr)
+
+ def _has(self, node: ast.filter.Has, reference: ast.filter.FilterExpression) -> bool:
+ if not isinstance(reference, type(node)):
+ return False
+ return self._parse_predicate_expression(node.predicate, reference.predicate) \
+ and self._parse_filter_expression(node.count, reference.count)
+
+ def _distance(self, node: ast.filter.Distance, reference: ast.filter.FilterExpression) -> bool:
+ if not isinstance(reference, (Partial, type(node))):
+ return False
+ # partial check
+ if isinstance(reference, Partial):
+ if not isinstance(node, reference.node):
+ return False
+ return reference.match('reference', node.reference) \
+ and reference.match('threshold', node.threshold) \
+ and reference.match('strict', node.strict)
+ # full check
+ return node.reference == reference.reference \
+ and node.threshold == reference.threshold \
+ and node.strict == reference.strict
+
+ def _value(self, node: ast.filter._Value, reference: ast.filter.FilterExpression) -> bool:
+ if not isinstance(reference, (Partial, type(node))):
+ return False
+ # partial check
+ if isinstance(reference, Partial):
+ if not isinstance(node, reference.node):
+ return False
+ return reference.match('value', node.value)
+ # full ckeck
+ return node.value == reference.value
+
+ def _bounded(self, node: ast.filter._Bounded, reference: ast.filter.FilterExpression) -> bool:
+ if not isinstance(reference, (Partial, type(node))):
+ return False
+ # partial check
+ if isinstance(reference, Partial):
+ if not isinstance(node, reference.node):
+ return False
+ return reference.match('threshold', node.threshold) \
+ and reference.match('strict', node.strict)
+ # full check
+ return node.threshold == reference.threshold \
+ and node.strict == reference.strict
+
+## EOF ##
diff --git a/bsfs/query/validator.py b/bsfs/query/validator.py
new file mode 100644
index 0000000..10ca492
--- /dev/null
+++ b/bsfs/query/validator.py
@@ -0,0 +1,351 @@
+
+# imports
+import typing
+
+# bsfs imports
+from bsfs import schema as bsc
+from bsfs.namespace import ns
+from bsfs.utils import errors, typename
+
+# inner-module imports
+from . import ast
+
+# exports
+__all__ : typing.Sequence[str] = (
+ 'Filter',
+ )
+
+# FIXME: Split into a submodule and the two classes into their own respective files.
+
+## code ##
+
+class Filter():
+ """Validate a `bsfs.query.ast.filter` query's structure and schema compliance.
+
+ * Conditions (Bounded, Value) can only be applied on literals
+ * Branches, Id, and Has can only be applied on nodes
+ * Predicates' domain and range must match
+ * Predicate paths must follow the schema
+ * Referenced types are present in the schema
+
+ """
+
+ # schema to validate against.
+ schema: bsc.Schema
+
+ def __init__(self, schema: bsc.Schema):
+ self.schema = schema
+
+ def __call__(self, root_type: bsc.Node, query: ast.filter.FilterExpression) -> bool:
+ """Alias for `Filter.validate`."""
+ return self.validate(root_type, query)
+
+ def validate(self, root_type: bsc.Node, query: ast.filter.FilterExpression) -> bool:
+ """Validate a filter *query*, assuming the subject having *root_type*.
+
+ Raises a `bsfs.utils.errors.ConsistencyError` if the query violates the schema.
+ Raises a `bsfs.utils.errors.BackendError` if the query structure is invalid.
+
+ """
+ # root_type must be a schema.Node
+ if not isinstance(root_type, bsc.Node):
+ raise TypeError(f'expected a node, found {typename(root_type)}')
+ # root_type must exist in the schema
+ if root_type not in self.schema.nodes():
+ raise errors.ConsistencyError(f'{root_type} is not defined in the schema')
+ # check root expression
+ self._parse_filter_expression(root_type, query)
+ # all tests passed
+ return True
+
+
+ ## routing methods
+
+ def _parse_filter_expression(self, type_: bsc.Vertex, node: ast.filter.FilterExpression):
+ """Route *node* to the handler of the respective FilterExpression subclass."""
+ if isinstance(node, ast.filter.Is):
+ return self._is(type_, node)
+ if isinstance(node, ast.filter.Not):
+ return self._not(type_, node)
+ if isinstance(node, ast.filter.Has):
+ return self._has(type_, node)
+ if isinstance(node, ast.filter.Distance):
+ return self._distance(type_, node)
+ if isinstance(node, (ast.filter.Any, ast.filter.All)):
+ return self._branch(type_, node)
+ if isinstance(node, (ast.filter.And, ast.filter.Or)):
+ return self._agg(type_, node)
+ if isinstance(node, (ast.filter.Equals, ast.filter.Substring, ast.filter.StartsWith, ast.filter.EndsWith)):
+ return self._value(type_, node)
+ if isinstance(node, (ast.filter.LessThan, ast.filter.GreaterThan)):
+ return self._bounded(type_, node)
+ # invalid node
+ raise errors.BackendError(f'expected filter expression, found {node}')
+
+ def _parse_predicate_expression(self, node: ast.filter.PredicateExpression) -> typing.Tuple[bsc.Vertex, bsc.Vertex]:
+ """Route *node* to the handler of the respective PredicateExpression subclass."""
+ if isinstance(node, ast.filter.Predicate):
+ return self._predicate(node)
+ if isinstance(node, ast.filter.OneOf):
+ return self._one_of(node)
+ # invalid node
+ raise errors.BackendError(f'expected predicate expression, found {node}')
+
+
+ ## predicate expressions
+
+ def _predicate(self, node: ast.filter.Predicate) -> typing.Tuple[bsc.Vertex, bsc.Vertex]:
+ # predicate exists in the schema
+ if not self.schema.has_predicate(node.predicate):
+ raise errors.ConsistencyError(f'predicate {node.predicate} is not in the schema')
+ # determine domain and range
+ pred = self.schema.predicate(node.predicate)
+ if not isinstance(pred.range, (bsc.Node, bsc.Literal)):
+ raise errors.BackendError(f'the range of predicate {pred} is undefined')
+ dom, rng = pred.domain, pred.range
+ if node.reverse:
+ dom, rng = rng, dom # type: ignore [assignment] # variable re-use confuses mypy
+ # return domain and range
+ return dom, rng
+
+ def _one_of(self, node: ast.filter.OneOf) -> typing.Tuple[bsc.Vertex, bsc.Vertex]:
+ # determine domain and range types
+ # NOTE: select the most specific domain and the most generic range
+ dom, rng = None, None
+ for pred in node:
+ # parse child expression
+ subdom, subrng = self._parse_predicate_expression(pred)
+ # determine overall domain
+ if dom is None or subdom < dom: # pick most specific domain
+ dom = subdom
+ # domains must be related across all child expressions
+ if not subdom <= dom and not subdom >= dom:
+ raise errors.ConsistencyError(f'domains {subdom} and {dom} are not related')
+ # determine overall range
+ if rng is None or subrng > rng: # pick most generic range
+ rng = subrng
+ # ranges must be related across all child expressions
+ if not subrng <= rng and not subrng >= rng:
+ raise errors.ConsistencyError(f'ranges {subrng} and {rng} are not related')
+ # OneOf guarantees at least one expression, dom and rng are always bsc.Vertex.
+ # mypy does not realize this, hence we ignore the warning.
+ return dom, rng # type: ignore [return-value]
+
+
+ ## intermediates
+
+ def _branch(self, type_: bsc.Vertex, node: ast.filter._Branch):
+ # type is a Node
+ if not isinstance(type_, bsc.Node):
+ raise errors.ConsistencyError(f'expected a Node, found {type_}')
+ # type exists in the schema
+ # FIXME: Isn't it actually guaranteed that the type (except the root type) is part of the schema?
+ # all types can be traced back to (a) root_type, (b) predicate, or (c) manually set (e.g. in _is).
+ # For (a), we do (and have to) perform a check. For (c), the code base should be consistent throughout
+ # the module, so this is an assumption that has to be ensured in schema.Schema. For (b), we know (and
+ # check) that the predicate is in the schema, hence all node/literals derived from it are also in the
+ # schema by construction of the schema.Schema class. So, why do we check this every time?
+ if type_ not in self.schema.nodes():
+ raise errors.ConsistencyError(f'node {type_} is not in the schema')
+ # predicate is valid
+ dom, rng = self._parse_predicate_expression(node.predicate)
+ # type_ is a subtype of the predicate's domain
+ if not type_ <= dom:
+ raise errors.ConsistencyError(f'expected type {dom} or subtype thereof, found {type_}')
+ # child expression is valid
+ self._parse_filter_expression(rng, node.expr)
+
+ def _agg(self, type_: bsc.Vertex, node: ast.filter._Agg):
+ for expr in node:
+ # child expression is valid
+ self._parse_filter_expression(type_, expr)
+
+ def _not(self, type_: bsc.Vertex, node: ast.filter.Not):
+ # child expression is valid
+ self._parse_filter_expression(type_, node.expr)
+
+ def _has(self, type_: bsc.Vertex, node: ast.filter.Has):
+ # type is a Node
+ if not isinstance(type_, bsc.Node):
+ raise errors.ConsistencyError(f'expected a Node, found {type_}')
+ # type exists in the schema
+ if type_ not in self.schema.nodes():
+ raise errors.ConsistencyError(f'node {type_} is not in the schema')
+ # predicate is valid
+ dom, _= self._parse_predicate_expression(node.predicate)
+ # type_ is a subtype of the predicate's domain
+ if not type_ <= dom:
+ raise errors.ConsistencyError(f'expected type {dom}, found {type_}')
+ # node.count is a numerical expression
+ self._parse_filter_expression(self.schema.literal(ns.bsl.Number), node.count)
+
+ def _distance(self, type_: bsc.Vertex, node: ast.filter.Distance):
+ # type is a Literal
+ if not isinstance(type_, bsc.Feature):
+ raise errors.ConsistencyError(f'expected a Feature, found {type_}')
+ # type exists in the schema
+ if type_ not in self.schema.literals():
+ raise errors.ConsistencyError(f'literal {type_} is not in the schema')
+ # reference matches type_
+ if len(node.reference) != type_.dimension:
+ raise errors.ConsistencyError(f'reference has dimension {len(node.reference)}, expected {type_.dimension}')
+ # FIXME: test dtype
+
+
+ ## conditions
+
+ def _is(self, type_: bsc.Vertex, node: ast.filter.Is): # pylint: disable=unused-argument # (node)
+ if not isinstance(type_, bsc.Node):
+ raise errors.ConsistencyError(f'expected a Node, found {type_}')
+ if type_ not in self.schema.nodes():
+ raise errors.ConsistencyError(f'node {type_} is not in the schema')
+
+ def _value(self, type_: bsc.Vertex, node: ast.filter._Value): # pylint: disable=unused-argument # (node)
+ # type is a literal
+ if not isinstance(type_, bsc.Literal):
+ raise errors.ConsistencyError(f'expected a Literal, found {type_}')
+ # type exists in the schema
+ if type_ not in self.schema.literals():
+ raise errors.ConsistencyError(f'literal {type_} is not in the schema')
+ # FIXME: Check if node.value corresponds to type_
+ # FIXME: A specific literal might be requested (i.e., a numeric type when used in Has)
+
+ def _bounded(self, type_: bsc.Vertex, node: ast.filter._Bounded): # pylint: disable=unused-argument # (node)
+ # type is a literal
+ if not isinstance(type_, bsc.Literal):
+ raise errors.ConsistencyError(f'expected a Literal, found {type_}')
+ # type exists in the schema
+ if type_ not in self.schema.literals():
+ raise errors.ConsistencyError(f'literal {type_} is not in the schema')
+ # type must be a numerical
+ if not type_ <= self.schema.literal(ns.bsl.Number):
+ raise errors.ConsistencyError(f'expected a number type, found {type_}')
+ # FIXME: Check if node.value corresponds to type_
+
+
+class Fetch():
+ """Validate a `bsfs.query.ast.fetch` query's structure and schema compliance.
+
+ * Value can only be applied on literals
+ * Node can only be applied on nodes
+ * Names must be non-empty
+ * Branching nodes' predicates must match the type
+ * Symbols must be in the schema
+ * Predicates must follow the schema
+
+ """
+
+ # schema to validate against.
+ schema: bsc.Schema
+
+ def __init__(self, schema: bsc.Schema):
+ self.schema = schema
+
+ def __call__(self, root_type: bsc.Node, query: ast.fetch.FetchExpression) -> bool:
+ """Alias for `Fetch.validate`."""
+ return self.validate(root_type, query)
+
+ def validate(self, root_type: bsc.Node, query: ast.fetch.FetchExpression) -> bool:
+ """Validate a fetch *query*, assuming the subject having *root_type*.
+
+ Raises a `bsfs.utils.errors.ConsistencyError` if the query violates the schema.
+ Raises a `bsfs.utils.errors.BackendError` if the query structure is invalid.
+
+ """
+ # root_type must be a schema.Node
+ if not isinstance(root_type, bsc.Node):
+ raise TypeError(f'expected a node, found {typename(root_type)}')
+ # root_type must exist in the schema
+ if root_type not in self.schema.nodes():
+ raise errors.ConsistencyError(f'{root_type} is not defined in the schema')
+ # query must be a FetchExpression
+ if not isinstance(query, ast.fetch.FetchExpression):
+ raise TypeError(f'expected a fetch expression, found {typename(query)}')
+ # check root expression
+ self._parse_fetch_expression(root_type, query)
+ # all tests passed
+ return True
+
+ def _parse_fetch_expression(self, type_: bsc.Vertex, node: ast.fetch.FetchExpression):
+ """Route *node* to the handler of the respective FetchExpression subclass."""
+ if isinstance(node, (ast.fetch.Fetch, ast.fetch.Value, ast.fetch.Node)):
+ # NOTE: don't return so that checks below are executed
+ self._branch(type_, node)
+ if isinstance(node, (ast.fetch.Value, ast.fetch.Node)):
+ # NOTE: don't return so that checks below are executed
+ self._named(type_, node)
+ if isinstance(node, ast.fetch.All):
+ return self._all(type_, node)
+ if isinstance(node, ast.fetch.Fetch):
+ return self._fetch(type_, node)
+ if isinstance(node, ast.fetch.Value):
+ return self._value(type_, node)
+ if isinstance(node, ast.fetch.Node):
+ return self._node(type_, node)
+ if isinstance(node, ast.fetch.This):
+ return self._this(type_, node)
+ # invalid node
+ raise errors.BackendError(f'expected fetch expression, found {node}')
+
+ def _all(self, type_: bsc.Vertex, node: ast.fetch.All):
+ # check child expressions
+ for expr in node:
+ self._parse_fetch_expression(type_, expr)
+
+ def _branch(self, type_: bsc.Vertex, node: ast.fetch._Branch):
+ # type is a node
+ if not isinstance(type_, bsc.Node):
+ raise errors.ConsistencyError(f'expected a Node, found {type_}')
+ # node exists in the schema
+ if type_ not in self.schema.nodes():
+ raise errors.ConsistencyError(f'node {type_} is not in the schema')
+ # predicate exists in the schema
+ if not self.schema.has_predicate(node.predicate):
+ raise errors.ConsistencyError(f'predicate {node.predicate} is not in the schema')
+ pred = self.schema.predicate(node.predicate)
+ # type_ must be a subclass of domain
+ if not type_ <= pred.domain:
+ raise errors.ConsistencyError(
+ f'expected type {pred.domain} or subtype thereof, found {type_}')
+
+ def _fetch(self, type_: bsc.Vertex, node: ast.fetch.Fetch): # pylint: disable=unused-argument # type_ was considered in _branch
+ # range must be a node
+ rng = self.schema.predicate(node.predicate).range
+ if not isinstance(rng, bsc.Node):
+ raise errors.ConsistencyError(
+ f'expected the predicate\'s range to be a Node, found {rng}')
+ # child expression must be valid
+ self._parse_fetch_expression(rng, node.expr)
+
+ def _named(self, type_: bsc.Vertex, node: ast.fetch._Named): # pylint: disable=unused-argument # type_ was considered in _branch
+ # name must be set
+ if node.name.strip() == '':
+ raise errors.BackendError('node name cannot be empty')
+ # FIXME: check for double name use?
+
+ def _node(self, type_: bsc.Vertex, node: ast.fetch.Node): # pylint: disable=unused-argument # type_ was considered in _branch
+ # range must be a node
+ rng = self.schema.predicate(node.predicate).range
+ if not isinstance(rng, bsc.Node):
+ raise errors.ConsistencyError(
+ f'expected the predicate\'s range to be a Node, found {rng}')
+
+ def _value(self, type_: bsc.Vertex, node: ast.fetch.Value): # pylint: disable=unused-argument # type_ was considered in _branch
+ # range must be a literal
+ rng = self.schema.predicate(node.predicate).range
+ if not isinstance(rng, bsc.Literal):
+ raise errors.ConsistencyError(
+ f'expected the predicate\'s range to be a Literal, found {rng}')
+
+ def _this(self, type_: bsc.Vertex, node: ast.fetch.This):
+ # type is a node
+ if not isinstance(type_, bsc.Node):
+ raise errors.ConsistencyError(f'expected a Node, found {type_}')
+ # node exists in the schema
+ if type_ not in self.schema.nodes():
+ raise errors.ConsistencyError(f'node {type_} is not in the schema')
+ # name must be set
+ if node.name.strip() == '':
+ raise errors.BackendError('node name cannot be empty')
+
+## EOF ##