diff options
author | Matthias Baumgartner <dev@igsor.net> | 2023-01-16 21:43:38 +0100 |
---|---|---|
committer | Matthias Baumgartner <dev@igsor.net> | 2023-01-16 21:43:38 +0100 |
commit | e12cd52ad267563c8046a593ad551b1dd089a702 (patch) | |
tree | d94cdf7ac540eb82630f78cbf564682b66007f51 | |
parent | 7f5a2920ef311b2077300714d7700313077a0bf6 (diff) | |
parent | 3504609e1ba1f7f653fa79910474bebd3ec24d8a (diff) | |
download | bsfs-e12cd52ad267563c8046a593ad551b1dd089a702.tar.gz bsfs-e12cd52ad267563c8046a593ad551b1dd089a702.tar.bz2 bsfs-e12cd52ad267563c8046a593ad551b1dd089a702.zip |
Merge branch 'mb/features' into develop
33 files changed, 2329 insertions, 632 deletions
diff --git a/bsfs/apps/migrate.py b/bsfs/apps/migrate.py index 91c1661..b9d019f 100644 --- a/bsfs/apps/migrate.py +++ b/bsfs/apps/migrate.py @@ -42,15 +42,15 @@ def main(argv): graph = bsfs.Open(config) # initialize schema - schema = bsfs.schema.Schema.Empty() + schema = bsfs.schema.Schema() if len(args.schema) == 0: # assemble schema from standard input - schema = schema + bsfs.schema.Schema.from_string(sys.stdin.read()) + schema = schema + bsfs.schema.from_string(sys.stdin.read()) else: # assemble schema from input files for pth in args.schema: with open(pth, mode='rt', encoding='UTF-8') as ifile: - schema = schema + bsfs.schema.Schema.from_string(ifile.read()) + schema = schema + bsfs.schema.from_string(ifile.read()) # migrate schema graph.migrate(schema, not args.remove) diff --git a/bsfs/graph/graph.py b/bsfs/graph/graph.py index f030fed..2210755 100644 --- a/bsfs/graph/graph.py +++ b/bsfs/graph/graph.py @@ -10,7 +10,7 @@ import typing # bsfs imports from bsfs.query import ast, validate -from bsfs.schema import Schema +from bsfs import schema as bsc from bsfs.triple_store import TripleStoreBase from bsfs.utils import URI, typename @@ -67,11 +67,11 @@ class Graph(): return f'{typename(self)}({str(self._backend)}, {self._user})' @property - def schema(self) -> Schema: + def schema(self) -> bsc.Schema: """Return the store's local schema.""" return self._backend.schema - def migrate(self, schema: Schema, append: bool = True) -> 'Graph': + def migrate(self, schema: bsc.Schema, append: bool = True) -> 'Graph': """Migrate the current schema to a new *schema*. Appends to the current schema by default; control this via *append*. @@ -79,14 +79,14 @@ class Graph(): """ # check args - if not isinstance(schema, Schema): + if not isinstance(schema, bsc.Schema): raise TypeError(schema) # append to current schema if append: schema = schema + self._backend.schema # add Graph schema requirements with open(os.path.join(os.path.dirname(__file__), 'schema.nt'), mode='rt', encoding='UTF-8') as ifile: - schema = schema + Schema.from_string(ifile.read()) + schema = schema + bsc.from_string(ifile.read()) # migrate schema in backend # FIXME: consult access controls! self._backend.schema = schema diff --git a/bsfs/graph/resolve.py b/bsfs/graph/resolve.py index feb0855..00b778b 100644 --- a/bsfs/graph/resolve.py +++ b/bsfs/graph/resolve.py @@ -37,8 +37,6 @@ class Filter(): """ - T_VERTEX = typing.Union[bsc.Node, bsc.Literal] - def __init__(self, schema): self.schema = schema @@ -47,7 +45,7 @@ class Filter(): def _parse_filter_expression( self, - type_: T_VERTEX, + type_: bsc.Vertex, node: ast.filter.FilterExpression, ) -> ast.filter.FilterExpression: """Route *node* to the handler of the respective FilterExpression subclass.""" @@ -65,6 +63,8 @@ class Filter(): return self._and(type_, node) if isinstance(node, ast.filter.Or): return self._or(type_, node) + if isinstance(node, ast.filter.Distance): + return self._distance(type_, node) if isinstance(node, (ast.filter.Equals, ast.filter.Substring, \ ast.filter.StartsWith, ast.filter.EndsWith)): return self._value(type_, node) @@ -73,7 +73,7 @@ class Filter(): # invalid node raise errors.BackendError(f'expected filter expression, found {node}') - def _parse_predicate_expression(self, node: ast.filter.PredicateExpression) -> T_VERTEX: + def _parse_predicate_expression(self, node: ast.filter.PredicateExpression) -> bsc.Vertex: """Route *node* to the handler of the respective PredicateExpression subclass.""" if isinstance(node, ast.filter.Predicate): return self._predicate(node) @@ -82,7 +82,7 @@ class Filter(): # invalid node raise errors.BackendError(f'expected predicate expression, found {node}') - def _predicate(self, node: ast.filter.Predicate) -> T_VERTEX: + def _predicate(self, node: ast.filter.Predicate) -> bsc.Vertex: 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) @@ -91,49 +91,52 @@ class Filter(): dom, rng = rng, dom return rng - def _one_of(self, node: ast.filter.OneOf) -> T_VERTEX: + def _one_of(self, node: ast.filter.OneOf) -> bsc.Vertex: # determine domain and range types rng = None for pred in node: # parse child expression subrng = self._parse_predicate_expression(pred) # determine the next type - try: - if rng is None or subrng > rng: # pick most generic range - rng = subrng - except TypeError as err: - raise errors.ConsistencyError(f'ranges {subrng} and {rng} are not related') from err - if rng is None: - raise errors.UnreachableError() + if rng is None or subrng > rng: # pick most generic range + rng = subrng + # check range consistency + if not subrng <= rng and not subrng >= rng: + raise errors.ConsistencyError(f'ranges {subrng} and {rng} are not related') + if not isinstance(rng, (bsc.Node, bsc.Literal)): + raise errors.BackendError(f'the range of node {node} is undefined') return rng - def _any(self, type_: T_VERTEX, node: ast.filter.Any) -> ast.filter.Any: # pylint: disable=unused-argument + def _any(self, type_: bsc.Vertex, node: ast.filter.Any) -> ast.filter.Any: # pylint: disable=unused-argument next_type = self._parse_predicate_expression(node.predicate) return ast.filter.Any(node.predicate, self._parse_filter_expression(next_type, node.expr)) - def _all(self, type_: T_VERTEX, node: ast.filter.All) -> ast.filter.All: # pylint: disable=unused-argument + def _all(self, type_: bsc.Vertex, node: ast.filter.All) -> ast.filter.All: # pylint: disable=unused-argument next_type = self._parse_predicate_expression(node.predicate) return ast.filter.All(node.predicate, self._parse_filter_expression(next_type, node.expr)) - def _and(self, type_: T_VERTEX, node: ast.filter.And) -> ast.filter.And: + def _and(self, type_: bsc.Vertex, node: ast.filter.And) -> ast.filter.And: return ast.filter.And({self._parse_filter_expression(type_, expr) for expr in node}) - def _or(self, type_: T_VERTEX, node: ast.filter.Or) -> ast.filter.Or: + def _or(self, type_: bsc.Vertex, node: ast.filter.Or) -> ast.filter.Or: return ast.filter.Or({self._parse_filter_expression(type_, expr) for expr in node}) - def _not(self, type_: T_VERTEX, node: ast.filter.Not) -> ast.filter.Not: + def _not(self, type_: bsc.Vertex, node: ast.filter.Not) -> ast.filter.Not: return ast.filter.Not(self._parse_filter_expression(type_, node.expr)) - def _has(self, type_: T_VERTEX, node: ast.filter.Has) -> ast.filter.Has: # pylint: disable=unused-argument + def _has(self, type_: bsc.Vertex, node: ast.filter.Has) -> ast.filter.Has: # pylint: disable=unused-argument + return node + + def _distance(self, type_: bsc.Vertex, node: ast.filter.Distance): # pylint: disable=unused-argument return node - def _value(self, type_: T_VERTEX, node: ast.filter._Value) -> ast.filter._Value: # pylint: disable=unused-argument + def _value(self, type_: bsc.Vertex, node: ast.filter._Value) -> ast.filter._Value: # pylint: disable=unused-argument return node - def _bounded(self, type_: T_VERTEX, node: ast.filter._Bounded) -> ast.filter._Bounded: # pylint: disable=unused-argument + def _bounded(self, type_: bsc.Vertex, node: ast.filter._Bounded) -> ast.filter._Bounded: # pylint: disable=unused-argument return node - def _is(self, type_: T_VERTEX, node: ast.filter.Is) -> typing.Union[ast.filter.Or, ast.filter.Is]: + def _is(self, type_: bsc.Vertex, node: ast.filter.Is) -> typing.Union[ast.filter.Or, ast.filter.Is]: # check if action is needed if not isinstance(node.value, nodes.Nodes): return node diff --git a/bsfs/graph/schema.nt b/bsfs/graph/schema.nt index 8612681..f619746 100644 --- a/bsfs/graph/schema.nt +++ b/bsfs/graph/schema.nt @@ -8,7 +8,8 @@ prefix bsfs: <http://bsfs.ai/schema/> prefix bsm: <http://bsfs.ai/schema/Meta#> # literals -xsd:integer rdfs:subClassOf bsfs:Literal . +bsfs:Number rdfs:subClassOf bsfs:Literal . +xsd:integer rdfs:subClassOf bsfs:Number . # predicates bsm:t_created rdfs:subClassOf bsfs:Predicate ; diff --git a/bsfs/namespace/namespace.py b/bsfs/namespace/namespace.py index f652dcd..1d443c1 100644 --- a/bsfs/namespace/namespace.py +++ b/bsfs/namespace/namespace.py @@ -59,7 +59,7 @@ class Namespace(): return hash((type(self), self.prefix, self.fsep, self.psep)) def __str__(self) -> str: - return f'{typename(self)}({self.prefix})' + return str(self.prefix) def __repr__(self) -> str: return f'{typename(self)}({self.prefix}, {self.fsep}, {self.psep})' diff --git a/bsfs/query/ast/filter_.py b/bsfs/query/ast/filter_.py index b129ded..2f0270c 100644 --- a/bsfs/query/ast/filter_.py +++ b/bsfs/query/ast/filter_.py @@ -252,8 +252,7 @@ class Has(FilterExpression): class _Value(FilterExpression): - """ - """ + """Matches some value.""" # target value. value: typing.Any @@ -277,13 +276,13 @@ class Is(_Value): class Equals(_Value): """Value matches exactly. - NOTE: Value format must correspond to literal type; can be a string, a number, or a Node + NOTE: Value must correspond to literal type. """ class Substring(_Value): """Value matches a substring - NOTE: value format must be a string + NOTE: value must be a string. """ @@ -295,9 +294,49 @@ 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 @@ -327,15 +366,11 @@ class _Bounded(FilterExpression): class LessThan(_Bounded): - """Value is (strictly) smaller than threshold. - NOTE: only on numerical literals - """ + """Value is (strictly) smaller than threshold. Assumes a Number literal.""" class GreaterThan(_Bounded): - """Value is (strictly) larger than threshold - NOTE: only on numerical literals - """ + """Value is (strictly) larger than threshold. Assumes a Number literal.""" class Predicate(PredicateExpression): diff --git a/bsfs/query/validator.py b/bsfs/query/validator.py index 352203a..904ac14 100644 --- a/bsfs/query/validator.py +++ b/bsfs/query/validator.py @@ -34,9 +34,6 @@ class Filter(): """ - # vertex types - T_VERTEX = typing.Union[bsc.Node, bsc.Literal] # FIXME: Shouldn't this be in the schema? - # schema to validate against. schema: bsc.Schema @@ -64,7 +61,7 @@ class Filter(): ## routing methods - def _parse_filter_expression(self, type_: T_VERTEX, node: ast.filter.FilterExpression): + 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) @@ -72,6 +69,8 @@ class Filter(): 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)): @@ -83,7 +82,7 @@ class Filter(): # invalid node raise errors.BackendError(f'expected filter expression, found {node}') - def _parse_predicate_expression(self, node: ast.filter.PredicateExpression) -> typing.Tuple[T_VERTEX, T_VERTEX]: + 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) @@ -95,58 +94,47 @@ class Filter(): ## predicate expressions - def _predicate(self, node: ast.filter.Predicate) -> typing.Tuple[T_VERTEX, T_VERTEX]: + 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 rng is None: - # FIXME: It is a design error that Predicates can have a None range... - raise errors.BackendError(f'predicate {pred} has no 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[T_VERTEX, T_VERTEX]: + 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) - try: - # 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') - except TypeError as err: # compared literal vs. node - raise errors.ConsistencyError(f'domains {subdom} and {dom} are not of the same type') from err - - try: - # 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') - except TypeError as err: # compared literal vs. node - raise errors.ConsistencyError(f'ranges {subrng} and {rng} are not of the same type') from err - # check domain and range - if dom is None or rng is None: - # OneOf guarantees at least one expression, these two cases cannot happen - raise errors.UnreachableError() - # return domain and range - return dom, rng + # 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_: T_VERTEX, node: ast.filter._Branch): + 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_}') @@ -167,16 +155,16 @@ class Filter(): # child expression is valid self._parse_filter_expression(rng, node.expr) - def _agg(self, type_: T_VERTEX, node: ast.filter._Agg): + 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_: T_VERTEX, node: ast.filter.Not): + 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_: T_VERTEX, node: ast.filter.Has): + 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_}') @@ -189,19 +177,30 @@ class Filter(): if not type_ <= dom: raise errors.ConsistencyError(f'expected type {dom}, found {type_}') # node.count is a numerical expression - # FIXME: We have to ensure that ns.xsd.integer is always known in the schema! - self._parse_filter_expression(self.schema.literal(ns.xsd.integer), node.count) + self._parse_filter_expression(self.schema.literal(ns.bsfs.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_: T_VERTEX, node: ast.filter.Is): # pylint: disable=unused-argument # (node) + 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_: T_VERTEX, node: ast.filter._Value): # pylint: disable=unused-argument # (node) + 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_}') @@ -211,13 +210,16 @@ class Filter(): # 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_: T_VERTEX, node: ast.filter._Bounded): # pylint: disable=unused-argument # (node) + 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.bsfs.Number): + raise errors.ConsistencyError(f'expected a number type, found {type_}') # FIXME: Check if node.value corresponds to type_ diff --git a/bsfs/schema/__init__.py b/bsfs/schema/__init__.py index ad4d456..f53512e 100644 --- a/bsfs/schema/__init__.py +++ b/bsfs/schema/__init__.py @@ -9,7 +9,12 @@ import typing # inner-module imports from .schema import Schema -from .types import Literal, Node, Predicate +from .serialize import from_string, to_string +from .types import Literal, Node, Predicate, Vertex, Feature, \ + ROOT_VERTEX, ROOT_NODE, ROOT_LITERAL, \ + ROOT_NUMBER, ROOT_TIME, \ + ROOT_ARRAY, ROOT_FEATURE, \ + ROOT_PREDICATE # exports __all__: typing.Sequence[str] = ( @@ -17,6 +22,8 @@ __all__: typing.Sequence[str] = ( 'Node', 'Predicate', 'Schema', + 'from_string', + 'to_string', ) ## EOF ## diff --git a/bsfs/schema/schema.py b/bsfs/schema/schema.py index c5d4571..8d9a821 100644 --- a/bsfs/schema/schema.py +++ b/bsfs/schema/schema.py @@ -7,10 +7,8 @@ Author: Matthias Baumgartner, 2022 # imports from collections import abc, namedtuple import typing -import rdflib # bsfs imports -from bsfs.namespace import ns from bsfs.utils import errors, URI, typename # inner-module imports @@ -51,11 +49,13 @@ class Schema(): def __init__( self, - predicates: typing.Iterable[types.Predicate], + predicates: typing.Optional[typing.Iterable[types.Predicate]] = None, nodes: typing.Optional[typing.Iterable[types.Node]] = None, literals: typing.Optional[typing.Iterable[types.Literal]] = None, ): # materialize arguments + if predicates is None: + predicates = set() if nodes is None: nodes = set() if literals is None: @@ -63,24 +63,40 @@ class Schema(): nodes = set(nodes) literals = set(literals) predicates = set(predicates) + + # add root types to the schema + nodes.add(types.ROOT_NODE) + literals.add(types.ROOT_LITERAL) + predicates.add(types.ROOT_PREDICATE) + # add minimally necessary types to the schema + literals.add(types.ROOT_NUMBER) + literals.add(types.ROOT_TIME) + literals.add(types.ROOT_ARRAY) + literals.add(types.ROOT_FEATURE) + + # FIXME: ensure that types derive from the right root? + # include parents in predicates set # TODO: review type annotations and ignores for python >= 3.11 (parents is _Type but should be typing.Self) predicates |= {par for pred in predicates for par in pred.parents()} # type: ignore [misc] # include predicate domain in nodes set nodes |= {pred.domain for pred in predicates} # include predicate range in nodes and literals sets - prange = {pred.range for pred in predicates if pred.range is not None} + prange = {pred.range for pred in predicates} nodes |= {vert for vert in prange if isinstance(vert, types.Node)} literals |= {vert for vert in prange if isinstance(vert, types.Literal)} + # NOTE: ROOT_PREDICATE has a Vertex as range which is neither in nodes nor literals + # FIXME: with the ROOT_VERTEX missing, the schema is not complete anymore! + # include parents in nodes and literals sets - # NOTE: Must be done after predicate domain/range was handled - # so that their parents are included as well. + # NOTE: Must come after predicate domain/range was handled to have their parents as well. nodes |= {par for node in nodes for par in node.parents()} # type: ignore [misc] literals |= {par for lit in literals for par in lit.parents()} # type: ignore [misc] # assign members self._nodes = {node.uri: node for node in nodes} self._literals = {lit.uri: lit for lit in literals} self._predicates = {pred.uri: pred for pred in predicates} + # verify unique uris if len(nodes) != len(self._nodes): raise errors.ConsistencyError('inconsistent nodes') @@ -214,6 +230,7 @@ class Schema(): >>> Schema.Union([a, b, c]) """ + # FIXME: copy type annotations? if len(args) == 0: raise TypeError('Schema.Union requires at least one argument (Schema or Iterable)') if isinstance(args[0], cls): # args is sequence of Schema instances @@ -295,92 +312,4 @@ class Schema(): """Return the Literal matching the *uri*.""" return self._literals[uri] - - ## constructors ## - - - @classmethod - def Empty(cls) -> 'Schema': # pylint: disable=invalid-name # capitalized classmethod - """Return a minimal Schema.""" - node = types.Node(ns.bsfs.Node, None) - literal = types.Literal(ns.bsfs.Literal, None) - predicate = types.Predicate( - uri=ns.bsfs.Predicate, - parent=None, - domain=node, - range=None, - unique=False, - ) - return cls((predicate, ), (node, ), (literal, )) - - - @classmethod - def from_string(cls, schema: str) -> 'Schema': # pylint: disable=invalid-name # capitalized classmethod - """Load and return a Schema from a string.""" - # parse string into rdf graph - graph = rdflib.Graph() - graph.parse(data=schema, format='turtle') - - def _fetch_hierarchically(factory, curr): - # emit current node - yield curr - # walk through childs - for child in graph.subjects(rdflib.URIRef(ns.rdfs.subClassOf), rdflib.URIRef(curr.uri)): - # convert to URI - child = URI(child) - # check circular dependency - if child == curr.uri or child in {node.uri for node in curr.parents()}: - raise errors.ConsistencyError('circular dependency') - # recurse and emit (sub*)childs - yield from _fetch_hierarchically(factory, factory(child, curr)) - - # fetch nodes - nodes = set(_fetch_hierarchically(types.Node, types.Node(ns.bsfs.Node, None))) - nodes_lut = {node.uri: node for node in nodes} - if len(nodes_lut) != len(nodes): - raise errors.ConsistencyError('inconsistent nodes') - - # fetch literals - literals = set(_fetch_hierarchically(types.Literal, types.Literal(ns.bsfs.Literal, None))) - literals_lut = {lit.uri: lit for lit in literals} - if len(literals_lut) != len(literals): - raise errors.ConsistencyError('inconsistent literals') - - # fetch predicates - def build_predicate(uri, parent): - uri = rdflib.URIRef(uri) - # get domain - domains = set(graph.objects(uri, rdflib.RDFS.domain)) - if len(domains) != 1: - raise errors.ConsistencyError(f'inconsistent domain: {domains}') - dom = nodes_lut.get(next(iter(domains))) - if dom is None: - raise errors.ConsistencyError('missing domain') - # get range - ranges = set(graph.objects(uri, rdflib.RDFS.range)) - if len(ranges) != 1: - raise errors.ConsistencyError(f'inconsistent range: {ranges}') - rng = next(iter(ranges)) - rng = nodes_lut.get(rng, literals_lut.get(rng)) - if rng is None: - raise errors.ConsistencyError('missing range') - # get unique flag - uniques = set(graph.objects(uri, rdflib.URIRef(ns.bsfs.unique))) - if len(uniques) != 1: - raise errors.ConsistencyError(f'inconsistent unique flags: {uniques}') - unique = bool(next(iter(uniques))) - # build Predicate - return types.Predicate(URI(uri), parent, dom, rng, unique) - - root_predicate = types.Predicate( - uri=ns.bsfs.Predicate, - parent=None, - domain=nodes_lut[ns.bsfs.Node], - range=None, # FIXME: Unclear how to handle this! Can be either a Literal or a Node - unique=False, - ) - predicates = _fetch_hierarchically(build_predicate, root_predicate) - # return Schema - return cls(predicates, nodes, literals) - ## EOF ## diff --git a/bsfs/schema/serialize.py b/bsfs/schema/serialize.py new file mode 100644 index 0000000..acc009a --- /dev/null +++ b/bsfs/schema/serialize.py @@ -0,0 +1,259 @@ +""" + +Part of the BlackStar filesystem (bsfs) module. +A copy of the license is provided with the project. +Author: Matthias Baumgartner, 2022 +""" +# standard imports +import itertools +import typing + +# external imports +import rdflib + +# bsfs imports +from bsfs.namespace import ns +from bsfs.utils import errors, URI, typename + +# inner-module imports +from . import types +from . import schema + +# exports +__all__: typing.Sequence[str] = ( + 'to_string', + 'from_string', + ) + + +## code ## + +def from_string(schema_str: str) -> schema.Schema: + """Load and return a Schema from a string.""" + # parse string into rdf graph + graph = rdflib.Graph() + graph.parse(data=schema_str, format='turtle') + + # helper functions + def _fetch_value( + subject: URI, + predicate: rdflib.URIRef, + value_factory: typing.Callable[[typing.Any], typing.Any], + ) -> typing.Optional[typing.Any]: + """Fetch the object of a given subject and predicate. + Raises a `errors.ConsistencyError` if multiple objects match. + """ + values = list(graph.objects(rdflib.URIRef(subject), predicate)) + if len(values) == 0: + return None + if len(values) == 1: + return value_factory(values[0]) + raise errors.ConsistencyError( + f'{subject} has multiple values for predicate {str(predicate)}, expected zero or one') + + def _convert(value): + """Convert the subject type from rdflib to a bsfs native type.""" + if isinstance(value, rdflib.Literal): + return value.value + if isinstance(value, rdflib.URIRef): + return URI(value) + # value is neither a node nor a literal, but e.g. a blank node + raise errors.BackendError(f'expected Literal or URIRef, found {typename(value)}') + + def _fetch_hierarchically(factory, curr): + """Walk through a rdfs:subClassOf hierarchy, creating symbols along the way.""" + # emit current node + yield curr + # walk through childs + for child in graph.subjects(rdflib.URIRef(ns.rdfs.subClassOf), rdflib.URIRef(curr.uri)): + # fetch annotations + annotations = { + URI(pred): _convert(value) + for pred, value # FIXME: preserve datatype of value?! + in graph.predicate_objects(child) + if URI(pred) != ns.rdfs.subClassOf + } + # convert child to URI + child = URI(child) + # check circular dependency + if child == curr.uri or child in {node.uri for node in curr.parents()}: + raise errors.ConsistencyError('circular dependency') + # recurse and emit (sub*)childs + yield from _fetch_hierarchically(factory, factory(child, curr, **annotations)) + + # fetch nodes + nodes = set(_fetch_hierarchically(types.Node, types.ROOT_NODE)) + nodes_lut = {node.uri: node for node in nodes} + if len(nodes_lut) != len(nodes): + raise errors.ConsistencyError('inconsistent nodes') + + # fetch literals + def _build_literal(uri, parent, **annotations): + """Literal factory.""" + # break out on root feature type + if uri == types.ROOT_FEATURE.uri: + return types.ROOT_FEATURE + # handle feature types + if isinstance(parent, types.Feature): + # clean annotations + annotations.pop(ns.bsfs.dimension, None) + annotations.pop(ns.bsfs.dtype, None) + annotations.pop(ns.bsfs.distance, None) + # get dimension + dimension = _fetch_value(uri, rdflib.URIRef(ns.bsfs.dimension), int) + # get dtype + dtype = _fetch_value(uri, rdflib.URIRef(ns.bsfs.dtype), URI) + # get distance + distance = _fetch_value(uri, rdflib.URIRef(ns.bsfs.distance), URI) + # return feature + return parent.child(URI(uri), dtype=dtype, dimension=dimension, distance=distance, **annotations) + # handle non-feature types + return parent.child(URI(uri), **annotations) + + literals = set(_fetch_hierarchically(_build_literal, types.ROOT_LITERAL)) + literals_lut = {lit.uri: lit for lit in literals} + if len(literals_lut) != len(literals): + raise errors.ConsistencyError('inconsistent literals') + + # fetch predicates + def _build_predicate(uri, parent, **annotations): + """Predicate factory.""" + # clean annotations + annotations.pop(ns.rdfs.domain, None) + annotations.pop(ns.rdfs.range, None) + annotations.pop(ns.bsfs.unique, None) + # get domain + dom = _fetch_value(uri, rdflib.RDFS.domain, URI) + if dom is not None and dom not in nodes_lut: + raise errors.ConsistencyError(f'predicate {uri} has undefined domain {dom}') + if dom is not None: + dom = nodes_lut[dom] + # get range + rng = _fetch_value(uri, rdflib.RDFS.range, URI) + if rng is not None and rng not in nodes_lut and rng not in literals_lut: + raise errors.ConsistencyError(f'predicate {uri} has undefined range {rng}') + if rng is not None: + rng = nodes_lut.get(rng, literals_lut.get(rng)) + # get unique + unique = _fetch_value(uri, rdflib.URIRef(ns.bsfs.unique), bool) + # build predicate + return parent.child(URI(uri), domain=dom, range=rng, unique=unique, **annotations) + + predicates = _fetch_hierarchically(_build_predicate, types.ROOT_PREDICATE) + + return schema.Schema(predicates, nodes, literals) + + + +def to_string(schema_inst: schema.Schema, fmt: str = 'turtle') -> str: + """Serialize a `bsfs.schema.Schema` to a string. + See `rdflib.Graph.serialize` for viable formats (default: turtle). + """ + + # type of emitted triples. + T_TRIPLE = typing.Iterator[typing.Tuple[rdflib.URIRef, rdflib.URIRef, rdflib.term.Identifier]] + + def _type(tpe: types._Type) -> T_TRIPLE : + """Emit _Type properties (parent, annotations).""" + # emit parent + if tpe.parent is not None: + yield ( + rdflib.URIRef(tpe.uri), + rdflib.URIRef(ns.rdfs.subClassOf), + rdflib.URIRef(tpe.parent.uri), + ) + # emit annotations + for prop, value in tpe.annotations.items(): + yield ( + rdflib.URIRef(tpe.uri), + rdflib.URIRef(prop), + rdflib.Literal(value), # FIXME: datatype?! + ) + + def _predicate(pred: types.Predicate) -> T_TRIPLE: + """Emit Predicate properties (domain, range, unique).""" + # no need to emit anything for the root predicate + if pred == types.ROOT_PREDICATE: + return + # emit domain + if pred.domain != getattr(pred.parent, 'domain', None): + yield ( + rdflib.URIRef(pred.uri), + rdflib.URIRef(ns.rdfs.domain), + rdflib.URIRef(pred.domain.uri), + ) + # emit range + if pred.range != getattr(pred.parent, 'range', None): + yield ( + rdflib.URIRef(pred.uri), + rdflib.URIRef(ns.rdfs.range), + rdflib.URIRef(pred.range.uri), + ) + # emit cardinality + if pred.unique != getattr(pred.parent, 'unique', None): + yield ( + rdflib.URIRef(pred.uri), + rdflib.URIRef(ns.bsfs.unique), + rdflib.Literal(pred.unique, datatype=rdflib.XSD.boolean), + ) + + def _feature(feat: types.Feature) -> T_TRIPLE: + """Emit Feature properties (dimension, dtype, distance).""" + # emit size + if feat.dimension != getattr(feat.parent, 'dimension', None): + yield ( + rdflib.URIRef(feat.uri), + rdflib.URIRef(ns.bsfs.dimension), + rdflib.Literal(feat.dimension, datatype=rdflib.XSD.integer), + ) + # emit dtype + if feat.dtype != getattr(feat.parent, 'dtype', None): + yield ( + rdflib.URIRef(feat.uri), + rdflib.URIRef(ns.bsfs.dtype), + rdflib.URIRef(feat.dtype), + ) + # emit distance + if feat.distance != getattr(feat.parent, 'distance', None): + yield ( + rdflib.URIRef(feat.uri), + rdflib.URIRef(ns.bsfs.distance), + rdflib.URIRef(feat.distance), + ) + + def _parse(node: types._Type) -> T_TRIPLE: + """Emit all properties of a type.""" + # check arg + if not isinstance(node, types._Type): # pylint: disable=protected-access + raise TypeError(node) + # emit _Type essentials + yield from _type(node) + # emit properties of derived types + if isinstance(node, types.Predicate): + yield from _predicate(node) + if isinstance(node, types.Feature): + yield from _feature(node) + + # create graph + graph = rdflib.Graph() + # add triples to graph + nodes = itertools.chain( + schema_inst.nodes(), + schema_inst.literals(), + schema_inst.predicates()) + for node in nodes: + for triple in _parse(node): + graph.add(triple) + # add known namespaces for readability + # FIXME: more generically? + graph.bind('bse', rdflib.URIRef(ns.bse[''])) + graph.bind('bsfs', rdflib.URIRef(ns.bsfs[''])) + graph.bind('bsm', rdflib.URIRef(ns.bsm[''])) + graph.bind('rdf', rdflib.URIRef(ns.rdf[''])) + graph.bind('rdfs', rdflib.URIRef(ns.rdfs[''])) + graph.bind('schema', rdflib.URIRef(ns.schema[''])) + graph.bind('xsd', rdflib.URIRef(ns.xsd[''])) + # serialize to turtle + return graph.serialize(format=fmt) + +## EOF ## diff --git a/bsfs/schema/types.py b/bsfs/schema/types.py index 54a7e99..3a2e10c 100644 --- a/bsfs/schema/types.py +++ b/bsfs/schema/types.py @@ -8,6 +8,7 @@ Author: Matthias Baumgartner, 2022 import typing # bsfs imports +from bsfs.namespace import ns from bsfs.utils import errors, URI, typename # exports @@ -15,6 +16,7 @@ __all__: typing.Sequence[str] = ( 'Literal', 'Node', 'Predicate', + 'Feature', ) @@ -99,9 +101,11 @@ class _Type(): self, uri: URI, parent: typing.Optional['_Type'] = None, + **annotations: typing.Any, ): self.uri = uri self.parent = parent + self.annotations = annotations def parents(self) -> typing.Generator['_Type', None, None]: """Generate a list of parent nodes.""" @@ -110,9 +114,17 @@ class _Type(): yield curr curr = curr.parent - def get_child(self, uri: URI, **kwargs): + def child( + self, + uri: URI, + **kwargs, + ): """Return a child of the current class.""" - return type(self)(uri, self, **kwargs) + return type(self)( + uri=uri, + parent=self, + **kwargs + ) def __str__(self) -> str: return f'{typename(self)}({self.uri})' @@ -138,8 +150,10 @@ class _Type(): def __lt__(self, other: typing.Any) -> bool: """Return True iff *self* is a true subclass of *other*.""" - if not type(self) is type(other): # type mismatch # pylint: disable=unidiomatic-typecheck + if not isinstance(other, _Type): return NotImplemented + if not isinstance(other, type(self)): # FIXME: necessary? + return False if self.uri == other.uri: # equivalence return False if self in other.parents(): # superclass @@ -151,8 +165,10 @@ class _Type(): def __le__(self, other: typing.Any) -> bool: """Return True iff *self* is equivalent or a subclass of *other*.""" - if not type(self) is type(other): # type mismatch # pylint: disable=unidiomatic-typecheck + if not isinstance(other, _Type): return NotImplemented + if not isinstance(other, type(self)): # FIXME: necessary? + return False if self.uri == other.uri: # equivalence return True if self in other.parents(): # superclass @@ -164,8 +180,10 @@ class _Type(): def __gt__(self, other: typing.Any) -> bool: """Return True iff *self* is a true superclass of *other*.""" - if not type(self) is type(other): # type mismatch # pylint: disable=unidiomatic-typecheck + if not isinstance(other, _Type): return NotImplemented + if not isinstance(other, type(self)): # FIXME: necessary? + return False if self.uri == other.uri: # equivalence return False if self in other.parents(): # superclass @@ -177,8 +195,10 @@ class _Type(): def __ge__(self, other: typing.Any) -> bool: """Return True iff *self* is eqiuvalent or a superclass of *other*.""" - if not type(self) is type(other): # type mismatch # pylint: disable=unidiomatic-typecheck + if not isinstance(other, _Type): return NotImplemented + if not isinstance(other, type(self)): # FIXME: necessary? + return False if self.uri == other.uri: # equivalence return True if self in other.parents(): # superclass @@ -189,32 +209,95 @@ class _Type(): return False -class _Vertex(_Type): +class Vertex(_Type): """Graph vertex types. Can be a Node or a Literal.""" - def __init__(self, uri: URI, parent: typing.Optional['_Vertex']): - super().__init__(uri, parent) + parent: typing.Optional['Vertex'] + def __init__(self, uri: URI, parent: typing.Optional['Vertex'], **kwargs): + super().__init__(uri, parent, **kwargs) -class Node(_Vertex): +class Node(Vertex): """Node type.""" - def __init__(self, uri: URI, parent: typing.Optional['Node']): - super().__init__(uri, parent) + parent: typing.Optional['Node'] + def __init__(self, uri: URI, parent: typing.Optional['Node'], **kwargs): + super().__init__(uri, parent, **kwargs) -class Literal(_Vertex): +class Literal(Vertex): """Literal type.""" - def __init__(self, uri: URI, parent: typing.Optional['Literal']): - super().__init__(uri, parent) + parent: typing.Optional['Literal'] + def __init__(self, uri: URI, parent: typing.Optional['Literal'], **kwargs): + super().__init__(uri, parent, **kwargs) + + +class Feature(Literal): + """Feature type.""" + + # Number of feature vector dimensions. + dimension: int + + # Feature vector datatype. + dtype: URI + + # Distance measure to compare feature vectors. + distance: URI + + def __init__( + self, + # Type members + uri: URI, + parent: typing.Optional[Literal], + # Feature members + dimension: int, + dtype: URI, + distance: URI, + **kwargs, + ): + super().__init__(uri, parent, **kwargs) + self.dimension = int(dimension) + self.dtype = URI(dtype) + self.distance = URI(distance) + + def __hash__(self) -> int: + return hash((super().__hash__(), self.dimension, self.dtype, self.distance)) + + def __eq__(self, other: typing.Any) -> bool: + return super().__eq__(other) \ + and self.dimension == other.dimension \ + and self.dtype == other.dtype \ + and self.distance == other.distance + def child( + self, + uri: URI, + dimension: typing.Optional[int] = None, + dtype: typing.Optional[URI] = None, + distance: typing.Optional[URI] = None, + **kwargs, + ): + """Return a child of the current class.""" + if dimension is None: + dimension = self.dimension + if dtype is None: + dtype = self.dtype + if distance is None: + distance = self.distance + return super().child( + uri=uri, + dimension=dimension, + dtype=dtype, + distance=distance, + **kwargs, + ) class Predicate(_Type): - """Predicate type.""" + """Predicate base type.""" # source type. domain: Node # destination type. - range: typing.Optional[typing.Union[Node, Literal]] + range: Vertex # maximum cardinality of type. unique: bool @@ -226,22 +309,23 @@ class Predicate(_Type): parent: typing.Optional['Predicate'], # Predicate members domain: Node, - range: typing.Optional[typing.Union[Node, Literal]], # pylint: disable=redefined-builtin + range: Vertex, # pylint: disable=redefined-builtin unique: bool, + **kwargs, ): # check arguments if not isinstance(domain, Node): raise TypeError(domain) - if range is not None and not isinstance(range, Node) and not isinstance(range, Literal): + if range != ROOT_VERTEX and not isinstance(range, (Node, Literal)): raise TypeError(range) # initialize - super().__init__(uri, parent) + super().__init__(uri, parent, **kwargs) self.domain = domain self.range = range - self.unique = unique + self.unique = bool(unique) def __hash__(self) -> int: - return hash((super().__hash__(), self.domain, self.range, self.unique)) + return hash((super().__hash__(), self.domain, self.unique, self.range)) def __eq__(self, other: typing.Any) -> bool: return super().__eq__(other) \ @@ -249,11 +333,11 @@ class Predicate(_Type): and self.range == other.range \ and self.unique == other.unique - def get_child( + def child( self, uri: URI, domain: typing.Optional[Node] = None, - range: typing.Optional[_Vertex] = None, # pylint: disable=redefined-builtin + range: typing.Optional[Vertex] = None, # pylint: disable=redefined-builtin unique: typing.Optional[bool] = None, **kwargs, ): @@ -264,13 +348,68 @@ class Predicate(_Type): raise errors.ConsistencyError(f'{domain} must be a subclass of {self.domain}') if range is None: range = self.range - if range is None: # inherited range from ns.bsfs.Predicate - raise ValueError('range must be defined by the parent or argument') - if self.range is not None and not range <= self.range: + # NOTE: The root predicate has a Vertex as range, which is neither a parent of the root + # Node nor Literal. Hence, that test is skipped since a child should be allowed to + # specialize from Vertex to anything. + if self.range != ROOT_VERTEX and not range <= self.range: raise errors.ConsistencyError(f'{range} must be a subclass of {self.range}') if unique is None: unique = self.unique - return super().get_child(uri, domain=domain, range=range, unique=unique, **kwargs) + return super().child( + uri=uri, + domain=domain, + range=range, + unique=unique, + **kwargs + ) + + +# essential vertices +ROOT_VERTEX = Vertex( + uri=ns.bsfs.Vertex, + parent=None, + ) + +ROOT_NODE = Node( + uri=ns.bsfs.Node, + parent=None, + ) +ROOT_LITERAL = Literal( + uri=ns.bsfs.Literal, + parent=None, + ) + +ROOT_NUMBER = Literal( + uri=ns.bsfs.Number, + parent=ROOT_LITERAL, + ) + +ROOT_TIME = Literal( + uri=ns.bsfs.Time, + parent=ROOT_LITERAL, + ) + +ROOT_ARRAY = Literal( + uri=ns.bsfs.Array, + parent=ROOT_LITERAL, + ) + +ROOT_FEATURE = Feature( + uri=ns.bsfs.Feature, + parent=ROOT_ARRAY, + dimension=1, + dtype=ns.bsfs.f16, + distance=ns.bsfs.euclidean, + ) + +# essential predicates +ROOT_PREDICATE = Predicate( + uri=ns.bsfs.Predicate, + parent=None, + domain=ROOT_NODE, + range=ROOT_VERTEX, + unique=False, + ) ## EOF ## diff --git a/bsfs/triple_store/sparql/distance.py b/bsfs/triple_store/sparql/distance.py new file mode 100644 index 0000000..2f5387a --- /dev/null +++ b/bsfs/triple_store/sparql/distance.py @@ -0,0 +1,56 @@ +""" + +Part of the BlackStar filesystem (bsfs) module. +A copy of the license is provided with the project. +Author: Matthias Baumgartner, 2022 +""" +# standard imports +import typing + +# external imports +import numpy as np + +# bsfs imports +from bsfs.namespace import ns + +# constants +EPS = 1e-9 + +# exports +__all__: typing.Sequence[str] = ( + 'DISTANCE_FU', + ) + + +## code ## + +def euclid(fst, snd) -> float: + """Euclidean distance (l2 norm).""" + fst = np.array(fst) + snd = np.array(snd) + return float(np.linalg.norm(fst - snd)) + +def cosine(fst, snd) -> float: + """Cosine distance.""" + fst = np.array(fst) + snd = np.array(snd) + if (fst == snd).all(): + return 0.0 + nrm0 = np.linalg.norm(fst) + nrm1 = np.linalg.norm(snd) + return float(1.0 - np.dot(fst, snd) / (nrm0 * nrm1 + EPS)) + +def manhatten(fst, snd) -> float: + """Manhatten (cityblock) distance (l1 norm).""" + fst = np.array(fst) + snd = np.array(snd) + return float(np.abs(fst - snd).sum()) + +# Known distance functions. +DISTANCE_FU = { + ns.bsfs.euclidean: euclid, + ns.bsfs.cosine: cosine, + ns.bsfs.manhatten: manhatten, +} + +## EOF ## diff --git a/bsfs/triple_store/sparql/parse_filter.py b/bsfs/triple_store/sparql/parse_filter.py index d4db0aa..8b6b976 100644 --- a/bsfs/triple_store/sparql/parse_filter.py +++ b/bsfs/triple_store/sparql/parse_filter.py @@ -5,19 +5,29 @@ A copy of the license is provided with the project. Author: Matthias Baumgartner, 2022 """ # imports +import operator import typing +# external imports +import rdflib + # bsfs imports from bsfs import schema as bsc from bsfs.namespace import ns from bsfs.query import ast from bsfs.utils import URI, errors +# inner-module imports +from .distance import DISTANCE_FU + # exports __all__: typing.Sequence[str] = ( 'Filter', ) + +## code ## + class _GenHopName(): """Generator that produces a new unique symbol name with each iteration.""" @@ -46,10 +56,8 @@ class Filter(): # Generator that produces unique symbol names. ngen: _GenHopName - # Vertex type. - T_VERTEX = typing.Union[bsc.Node, bsc.Literal] - - def __init__(self, schema): + def __init__(self, graph, schema): + self.graph = graph self.schema = schema self.ngen = _GenHopName() @@ -79,7 +87,7 @@ class Filter(): }} ''' - def _parse_filter_expression(self, type_: T_VERTEX, node: ast.filter.FilterExpression, head: str) -> str: + def _parse_filter_expression(self, type_: bsc.Vertex, node: ast.filter.FilterExpression, head: str) -> str: """Route *node* to the handler of the respective FilterExpression subclass.""" if isinstance(node, ast.filter.Is): return self._is(type_, node, head) @@ -87,6 +95,8 @@ class Filter(): return self._not(type_, node, head) if isinstance(node, ast.filter.Has): return self._has(type_, node, head) + if isinstance(node, ast.filter.Distance): + return self._distance(type_, node, head) if isinstance(node, ast.filter.Any): return self._any(type_, node, head) if isinstance(node, ast.filter.All): @@ -112,9 +122,9 @@ class Filter(): def _parse_predicate_expression( self, - type_: T_VERTEX, + type_: bsc.Vertex, node: ast.filter.PredicateExpression - ) -> typing.Tuple[str, T_VERTEX]: + ) -> typing.Tuple[str, bsc.Vertex]: """Route *node* to the handler of the respective PredicateExpression subclass.""" if isinstance(node, ast.filter.Predicate): return self._predicate(type_, node) @@ -123,7 +133,7 @@ class Filter(): # invalid node raise errors.BackendError(f'expected predicate expression, found {node}') - def _one_of(self, node_type: T_VERTEX, node: ast.filter.OneOf) -> typing.Tuple[str, T_VERTEX]: + def _one_of(self, node_type: bsc.Vertex, node: ast.filter.OneOf) -> typing.Tuple[str, bsc.Vertex]: """ """ if not isinstance(node_type, bsc.Node): @@ -134,23 +144,18 @@ class Filter(): puri, subrng = self._parse_predicate_expression(node_type, pred) # track predicate uris suburi.add(puri) - try: - # check for more generic range - if rng is None or subrng > rng: - rng = subrng - # check range consistency - if not subrng <= rng and not subrng >= rng: - raise errors.ConsistencyError(f'ranges {subrng} and {rng} are not related') - except TypeError as err: # subrng and rng are not comparable - raise errors.ConsistencyError(f'ranges {subrng} and {rng} are not related') from err - if rng is None: - # for mypy to be certain of the rng type - # if rng were None, we'd have gotten a TypeError above (None > None) - raise errors.UnreachableError() + # check for more generic range + if rng is None or subrng > rng: + rng = subrng + # check range consistency + if not subrng <= rng and not subrng >= rng: + raise errors.ConsistencyError(f'ranges {subrng} and {rng} are not related') # return joint predicate expression and next range - return '|'.join(suburi), rng + # OneOf guarantees at least one expression, rng is always a bsc.Vertex. + # mypy does not realize this, hence we ignore the warning. + return '|'.join(suburi), rng # type: ignore [return-value] - def _predicate(self, node_type: T_VERTEX, node: ast.filter.Predicate) -> typing.Tuple[str, T_VERTEX]: + def _predicate(self, node_type: bsc.Vertex, node: ast.filter.Predicate) -> typing.Tuple[str, bsc.Vertex]: """ """ # check node_type @@ -162,9 +167,8 @@ class Filter(): if not self.schema.has_predicate(puri): raise errors.ConsistencyError(f'predicate {puri} is not in the schema') pred = self.schema.predicate(puri) - if pred.range is None: - # FIXME: It is a design error that Predicates can have a None range... - raise errors.BackendError(f'predicate {pred} has no range') + 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 # encapsulate predicate uri puri = f'<{puri}>' # type: ignore [assignment] # variable re-use confuses mypy @@ -178,7 +182,7 @@ class Filter(): # return predicate URI and next node type return puri, rng - def _any(self, node_type: T_VERTEX, node: ast.filter.Any, head: str) -> str: + def _any(self, node_type: bsc.Vertex, node: ast.filter.Any, head: str) -> str: """ """ if not isinstance(node_type, bsc.Node): @@ -191,7 +195,7 @@ class Filter(): # combine results return f'{head} {pred} {nexthead} . {expr}' - def _all(self, node_type: T_VERTEX, node: ast.filter.All, head: str) -> str: + def _all(self, node_type: bsc.Vertex, node: ast.filter.All, head: str) -> str: """ """ # NOTE: All(P, E) := Not(Any(P, Not(E))) and EXISTS(P, ?) @@ -208,13 +212,13 @@ class Filter(): # return existence and rewritten expression return f'FILTER EXISTS {{ {head} {pred} {temphead} }} . ' + expr - def _and(self, node_type: T_VERTEX, node: ast.filter.And, head: str) -> str: + def _and(self, node_type: bsc.Vertex, node: ast.filter.And, head: str) -> str: """ """ sub = [self._parse_filter_expression(node_type, expr, head) for expr in node] return ' . '.join(sub) - def _or(self, node_type: T_VERTEX, node: ast.filter.Or, head: str) -> str: + def _or(self, node_type: bsc.Vertex, node: ast.filter.Or, head: str) -> str: """ """ # potential special case optimization: @@ -224,7 +228,7 @@ class Filter(): sub = ['{' + expr + '}' for expr in sub] return ' UNION '.join(sub) - def _not(self, node_type: T_VERTEX, node: ast.filter.Not, head: str) -> str: + def _not(self, node_type: bsc.Vertex, node: ast.filter.Not, head: str) -> str: """ """ expr = self._parse_filter_expression(node_type, node.expr, head) @@ -235,7 +239,7 @@ class Filter(): # The simplest (and non-interfering) choice is a type statement. return f'MINUS {{ {head} <{ns.rdf.type}>/<{ns.rdfs.subClassOf}>* <{node_type.uri}> . {expr} }}' - def _has(self, node_type: T_VERTEX, node: ast.filter.Has, head: str) -> str: + def _has(self, node_type: bsc.Vertex, node: ast.filter.Has, head: str) -> str: """ """ if not isinstance(node_type, bsc.Node): @@ -248,47 +252,72 @@ class Filter(): # predicate count expression (fetch number of predicates at *head*) num_preds = f'{{ SELECT (COUNT(distinct {inner}) as {outer}) WHERE {{ {head} {pred} {inner} }} }}' # count expression - # FIXME: We have to ensure that ns.xsd.integer is always known in the schema! count_bounds = self._parse_filter_expression(self.schema.literal(ns.xsd.integer), node.count, outer) # combine return num_preds + ' . ' + count_bounds - def _is(self, node_type: T_VERTEX, node: ast.filter.Is, head: str) -> str: + def _distance(self, node_type: bsc.Vertex, node: ast.filter.Distance, head: str) -> str: + """ + """ + if not isinstance(node_type, bsc.Feature): + raise errors.BackendError(f'expected Feature, found {node_type}') + if len(node.reference) != node_type.dimension: + raise errors.ConsistencyError( + f'reference has dimension {len(node.reference)}, expected {node_type.dimension}') + # get distance metric + dist = DISTANCE_FU[node_type.distance] + # get operator + cmp = operator.lt if node.strict else operator.le + # get candidate values + candidates = { + f'"{cand}"^^<{node_type.uri}>' + for cand + in self.graph.objects() + if isinstance(cand, rdflib.Literal) + and cand.datatype == rdflib.URIRef(node_type.uri) + and cmp(dist(cand.value, node.reference), node.threshold) + } + # combine candidate values + values = ' '.join(candidates) if len(candidates) else f'"impossible value"^^<{ns.xsd.string}>' + # return sparql fragment + return f'VALUES {head} {{ {values} }}' + + def _is(self, node_type: bsc.Vertex, node: ast.filter.Is, head: str) -> str: """ """ if not isinstance(node_type, bsc.Node): raise errors.BackendError(f'expected Node, found {node_type}') return f'VALUES {head} {{ <{node.value}> }}' - def _equals(self, node_type: T_VERTEX, node: ast.filter.Equals, head: str) -> str: + def _equals(self, node_type: bsc.Vertex, node: ast.filter.Equals, head: str) -> str: """ """ if not isinstance(node_type, bsc.Literal): raise errors.BackendError(f'expected Literal, found {node}') return f'VALUES {head} {{ "{node.value}"^^<{node_type.uri}> }}' - def _substring(self, node_type: T_VERTEX, node: ast.filter.Substring, head: str) -> str: + def _substring(self, node_type: bsc.Vertex, node: ast.filter.Substring, head: str) -> str: """ """ if not isinstance(node_type, bsc.Literal): raise errors.BackendError(f'expected Literal, found {node_type}') return f'FILTER contains(str({head}), "{node.value}")' - def _starts_with(self, node_type: T_VERTEX, node: ast.filter.StartsWith, head: str) -> str: + def _starts_with(self, node_type: bsc.Vertex, node: ast.filter.StartsWith, head: str) -> str: """ """ if not isinstance(node_type, bsc.Literal): raise errors.BackendError(f'expected Literal, found {node_type}') return f'FILTER strstarts(str({head}), "{node.value}")' - def _ends_with(self, node_type: T_VERTEX, node: ast.filter.EndsWith, head: str) -> str: + def _ends_with(self, node_type: bsc.Vertex, node: ast.filter.EndsWith, head: str) -> str: """ """ if not isinstance(node_type, bsc.Literal): raise errors.BackendError(f'expected Literal, found {node_type}') return f'FILTER strends(str({head}), "{node.value}")' - def _less_than(self, node_type: T_VERTEX, node: ast.filter.LessThan, head: str) -> str: + def _less_than(self, node_type: bsc.Vertex, node: ast.filter.LessThan, head: str) -> str: """ """ if not isinstance(node_type, bsc.Literal): @@ -296,7 +325,7 @@ class Filter(): equality = '=' if not node.strict else '' return f'FILTER ({head} <{equality} {float(node.threshold)})' - def _greater_than(self, node_type: T_VERTEX, node: ast.filter.GreaterThan, head: str) -> str: + def _greater_than(self, node_type: bsc.Vertex, node: ast.filter.GreaterThan, head: str) -> str: """ """ if not isinstance(node_type, bsc.Literal): diff --git a/bsfs/triple_store/sparql/sparql.py b/bsfs/triple_store/sparql/sparql.py index c3cbff6..fedd227 100644 --- a/bsfs/triple_store/sparql/sparql.py +++ b/bsfs/triple_store/sparql/sparql.py @@ -11,12 +11,14 @@ import rdflib # bsfs imports from bsfs import schema as bsc +from bsfs.namespace import ns from bsfs.query import ast from bsfs.utils import errors, URI # inner-module imports from . import parse_filter from .. import base +from .distance import DISTANCE_FU # exports @@ -94,8 +96,9 @@ class SparqlStore(base.TripleStoreBase): super().__init__(None) self._graph = rdflib.Graph() self._transaction = _Transaction(self._graph) - self._schema = bsc.Schema.Empty() - self._filter_parser = parse_filter.Filter(self._schema) + # NOTE: parsing bsfs.query.ast.filter.Has requires xsd:integer. + self._schema = bsc.Schema(literals={bsc.ROOT_NUMBER.child(ns.xsd.integer)}) + self._filter_parser = parse_filter.Filter(self._graph, self._schema) # NOTE: mypy and pylint complain about the **kwargs not being listed (contrasting super) # However, not having it here is clearer since it's explicit that there are no arguments. @@ -121,6 +124,16 @@ class SparqlStore(base.TripleStoreBase): # check compatibility: No contradicting definitions if not self.schema.consistent_with(schema): raise errors.ConsistencyError(f'{schema} is inconsistent with {self.schema}') + # check distance functions of features + invalid = { + (cand.uri, cand.distance) + for cand + in schema.literals() + if isinstance(cand, bsc.Feature) and cand.distance not in DISTANCE_FU} + if len(invalid) > 0: + cand, dist = zip(*invalid) + raise errors.UnsupportedError( + f'unknown distance function {",".join(dist)} in feature {", ".join(cand)}') # commit the current transaction self.commit() @@ -137,7 +150,7 @@ class SparqlStore(base.TripleStoreBase): for src, trg in self._graph.subject_objects(rdflib.URIRef(pred.uri)): self._transaction.remove((src, rdflib.URIRef(pred.uri), trg)) # remove predicate definition - if pred.parent is not None: + if pred.parent is not None: # NOTE: there shouldn't be any predicate w/o parent self._transaction.remove(( rdflib.URIRef(pred.uri), rdflib.RDFS.subClassOf, @@ -157,7 +170,7 @@ class SparqlStore(base.TripleStoreBase): # remove instance self._transaction.remove((inst, rdflib.RDF.type, rdflib.URIRef(node.uri))) # remove node definition - if node.parent is not None: + if node.parent is not None: # NOTE: there shouldn't be any node w/o parent self._transaction.remove(( rdflib.URIRef(node.uri), rdflib.RDFS.subClassOf, @@ -166,7 +179,7 @@ class SparqlStore(base.TripleStoreBase): for lit in sub.literals: # remove literal definition - if lit.parent is not None: + if lit.parent is not None: # NOTE: there shouldn't be any literal w/o parent self._transaction.remove(( rdflib.URIRef(lit.uri), rdflib.RDFS.subClassOf, diff --git a/bsfs/utils/errors.py b/bsfs/utils/errors.py index be9d40e..6ae6484 100644 --- a/bsfs/utils/errors.py +++ b/bsfs/utils/errors.py @@ -41,4 +41,7 @@ class ConfigError(_BSFSError): class BackendError(_BSFSError): """Could not parse an AST structure.""" +class UnsupportedError(_BSFSError): + """Some requested functionality is not supported by an implementation.""" + ## EOF ## diff --git a/bsfs/utils/uuid.py b/bsfs/utils/uuid.py index 6366b18..ba5cf52 100644 --- a/bsfs/utils/uuid.py +++ b/bsfs/utils/uuid.py @@ -7,6 +7,7 @@ Author: Matthias Baumgartner, 2022 # imports from collections import abc import hashlib +import json import os import platform import random @@ -105,4 +106,10 @@ class UCID(): with open(path, 'rb') as ifile: return HASH(ifile.read()).hexdigest() + + @staticmethod + def from_dict(content: dict) -> str: + """Get the content from a dict.""" + return HASH(json.dumps(content).encode('ascii', 'ignore')).hexdigest() + ## EOF ## diff --git a/test/apps/schema-2.nt b/test/apps/schema-2.nt index 525ac99..4c5468f 100644 --- a/test/apps/schema-2.nt +++ b/test/apps/schema-2.nt @@ -10,7 +10,8 @@ prefix bse: <http://bsfs.ai/schema/Entity#> bsfs:Entity rdfs:subClassOf bsfs:Node . # common definitions -xsd:integer rdfs:subClassOf bsfs:Literal . +bsfs:Number rdfs:subClassOf bsfs:Literal . +xsd:integer rdfs:subClassOf bsfs:Number . bse:filesize rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Entity ; diff --git a/test/apps/test_migrate.py b/test/apps/test_migrate.py index 957509a..230c032 100644 --- a/test/apps/test_migrate.py +++ b/test/apps/test_migrate.py @@ -13,7 +13,7 @@ import unittest import unittest.mock # bsie imports -from bsfs.schema import Schema +from bsfs import schema # objects to test from bsfs.apps.migrate import main @@ -33,21 +33,21 @@ class TestMigrate(unittest.TestCase): # read schema from file with open(schema_1) as ifile: - target = Schema.from_string(ifile.read()) + target = schema.from_string(ifile.read()) graph = main([config, schema_1]) self.assertTrue(target <= graph.schema) # read schema from multiple files with open(schema_1) as ifile: - target = Schema.from_string(ifile.read()) + target = schema.from_string(ifile.read()) with open(schema_2) as ifile: - target = target + Schema.from_string(ifile.read()) + target = target + schema.from_string(ifile.read()) graph = main([config, schema_1, schema_2]) self.assertTrue(target <= graph.schema) # read schema from stdin with open(schema_1, 'rt') as ifile: - target = Schema.from_string(ifile.read()) + target = schema.from_string(ifile.read()) with open(schema_1, 'rt') as ifile: with unittest.mock.patch('sys.stdin', ifile): graph = main([config]) diff --git a/test/graph/ac/test_null.py b/test/graph/ac/test_null.py index c863943..e35852d 100644 --- a/test/graph/ac/test_null.py +++ b/test/graph/ac/test_null.py @@ -8,7 +8,7 @@ Author: Matthias Baumgartner, 2022 import unittest # bsie imports -from bsfs import schema as _schema +from bsfs import schema as bsc from bsfs.namespace import ns from bsfs.query import ast from bsfs.triple_store import SparqlStore @@ -23,7 +23,7 @@ from bsfs.graph.ac.null import NullAC class TestNullAC(unittest.TestCase): def setUp(self): self.backend = SparqlStore() - self.backend.schema = _schema.Schema.from_string(''' + self.backend.schema = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> @@ -34,7 +34,8 @@ class TestNullAC(unittest.TestCase): bsfs:Entity rdfs:subClassOf bsfs:Node . bsfs:Tag rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . # predicates mandated by Nodes bsm:t_created rdfs:subClassOf bsfs:Predicate ; diff --git a/test/graph/test_graph.py b/test/graph/test_graph.py index 8503d5b..f97783b 100644 --- a/test/graph/test_graph.py +++ b/test/graph/test_graph.py @@ -25,7 +25,7 @@ class TestGraph(unittest.TestCase): def setUp(self): self.user = URI('http://example.com/me') self.backend = SparqlStore.Open() - self.backend.schema = schema.Schema.from_string(''' + self.backend.schema = schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix bsfs: <http://bsfs.ai/schema/> bsfs:Entity rdfs:subClassOf bsfs:Node . @@ -118,14 +118,15 @@ class TestGraph(unittest.TestCase): schema.Node(ns.bsfs.Node, None)))}), append=False) # can migrate to compatible schema - target_1 = schema.Schema.from_string(''' + target_1 = schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> prefix bse: <http://bsfs.ai/schema/Entity#> bsfs:Entity rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . bse:filename rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Entity ; @@ -142,12 +143,13 @@ class TestGraph(unittest.TestCase): # new schema is applied self.assertLess(target_1, graph.schema) # graph appends its predicates - self.assertEqual(graph.schema, target_1 + schema.Schema.from_string(''' + self.assertEqual(graph.schema, target_1 + schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> prefix bsm: <http://bsfs.ai/schema/Meta#> - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . bsm:t_created rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Node ; rdfs:range xsd:integer ; @@ -155,14 +157,15 @@ class TestGraph(unittest.TestCase): ''')) # can overwrite the current schema - target_2 = schema.Schema.from_string(''' + target_2 = schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> prefix bse: <http://bsfs.ai/schema/Entity#> bsfs:Entity rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . bse:filename rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Entity ; @@ -181,12 +184,13 @@ class TestGraph(unittest.TestCase): # new schema is applied self.assertLess(target_2, graph.schema) # graph appends its predicates - self.assertEqual(graph.schema, target_2 + schema.Schema.from_string(''' + self.assertEqual(graph.schema, target_2 + schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> prefix bsm: <http://bsfs.ai/schema/Meta#> - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . bsm:t_created rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Node ; rdfs:range xsd:integer ; @@ -196,7 +200,7 @@ class TestGraph(unittest.TestCase): def test_get(self): # setup graph = Graph(self.backend, self.user) - graph.migrate(schema.Schema.from_string(''' + graph.migrate(schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> diff --git a/test/graph/test_nodes.py b/test/graph/test_nodes.py index 11ae46d..2870f35 100644 --- a/test/graph/test_nodes.py +++ b/test/graph/test_nodes.py @@ -24,7 +24,7 @@ class TestNodes(unittest.TestCase): def setUp(self): # initialize backend self.backend = SparqlStore() - self.backend.schema = _schema.Schema.from_string(''' + self.backend.schema = _schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> @@ -37,7 +37,8 @@ class TestNodes(unittest.TestCase): bsfs:Tag rdfs:subClassOf bsfs:Node . bsfs:User rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . # predicates mandated by Nodes bsm:t_created rdfs:subClassOf bsfs:Predicate ; @@ -78,7 +79,11 @@ class TestNodes(unittest.TestCase): (rdflib.URIRef(ns.bsfs.Tag), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Node)), (rdflib.URIRef(ns.bsfs.User), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Node)), (rdflib.URIRef(ns.xsd.string), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), - (rdflib.URIRef(ns.xsd.integer), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Array), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Feature), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Array)), + (rdflib.URIRef(ns.bsfs.Number), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Time), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.xsd.integer), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Number)), (rdflib.URIRef(ns.bsm.t_created), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), (rdflib.URIRef(ns.bse.comment), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), (rdflib.URIRef(ns.bse.filesize), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), diff --git a/test/graph/test_resolve.py b/test/graph/test_resolve.py index 5bc99e4..0918b02 100644 --- a/test/graph/test_resolve.py +++ b/test/graph/test_resolve.py @@ -31,7 +31,7 @@ class TestFilter(unittest.TestCase): """ def test_call(self): - schema = bsc.Schema.from_string(''' + schema = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> @@ -41,7 +41,17 @@ class TestFilter(unittest.TestCase): bsfs:Entity rdfs:subClassOf bsfs:Node . bsfs:Tag rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array . + xsd:integer rdfs:subClassOf bsfs:Number . + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "5"^^xsd:integer . + + bse:colors rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range bsfs:Colors . bse:comment rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Entity ; @@ -65,7 +75,7 @@ class TestFilter(unittest.TestCase): {URI('http://example.com/me/entity#1234'), URI('http://example.com/me/entity#4321')}) tags = graph.nodes(ns.bsfs.Tag, {URI('http://example.com/me/tag#1234'), URI('http://example.com/me/tag#4321')}) - invalid = nodes.Nodes(None, '', schema.node(ns.bsfs.Node).get_child(ns.bsfs.Invalid), + invalid = nodes.Nodes(None, '', schema.node(ns.bsfs.Node).child(ns.bsfs.Invalid), {'http://example.com/you/invalid#1234', 'http://example.com/you/invalid#4321'}) resolver = Filter(schema) @@ -144,12 +154,18 @@ class TestFilter(unittest.TestCase): self.assertEqual(resolver(schema.node(ns.bsfs.Entity), ast.filter.Has(ns.bse.comment)), ast.filter.Has(ns.bse.comment)) + # for sake of completeness: Distance + self.assertEqual(resolver(schema.node(ns.bsfs.Entity), + ast.filter.Any(ns.bse.colors, ast.filter.Distance([1,2,3,4,5], 1))), + ast.filter.Any(ns.bse.colors, ast.filter.Distance([1,2,3,4,5], 1))) # route errors self.assertRaises(errors.BackendError, resolver, schema.node(ns.bsfs.Tag), ast.filter.Predicate(ns.bse.comment)) self.assertRaises(errors.BackendError, resolver, schema.node(ns.bsfs.Tag), ast.filter.Any(ast.filter.PredicateExpression(), ast.filter.Equals('foo'))) - self.assertRaises(errors.UnreachableError, resolver._one_of, ast.filter.OneOf(ast.filter.Predicate(ns.bsfs.Predicate))) + self.assertRaises(errors.BackendError, resolver._one_of, ast.filter.OneOf(ast.filter.Predicate(ns.bsfs.Predicate))) + # for sake of coverage completeness: valid OneOf + self.assertIsNotNone(resolver._one_of(ast.filter.OneOf(ast.filter.Predicate(ns.bse.colors)))) # check schema consistency self.assertRaises(errors.ConsistencyError, resolver, schema.node(ns.bsfs.Tag), diff --git a/test/namespace/test_namespace.py b/test/namespace/test_namespace.py index f109653..2536203 100644 --- a/test/namespace/test_namespace.py +++ b/test/namespace/test_namespace.py @@ -20,15 +20,15 @@ from bsfs.namespace.namespace import Namespace, ClosedNamespace class TestNamespace(unittest.TestCase): def test_essentials(self): # string conversion - self.assertEqual(str(Namespace('http://example.org/')), 'Namespace(http://example.org)') - self.assertEqual(str(Namespace('http://example.org#')), 'Namespace(http://example.org)') + self.assertEqual(str(Namespace('http://example.org/')), 'http://example.org') + self.assertEqual(str(Namespace('http://example.org#')), 'http://example.org') self.assertEqual(repr(Namespace('http://example.org/')), 'Namespace(http://example.org, #, /)') self.assertEqual(repr(Namespace('http://example.org#')), 'Namespace(http://example.org, #, /)') self.assertEqual(repr(Namespace('http://example.org', fsep='.')), 'Namespace(http://example.org, ., /)') self.assertEqual(repr(Namespace('http://example.org', psep='.')), 'Namespace(http://example.org, #, .)') # repeated separators are truncated - self.assertEqual(str(Namespace('http://example.org////')), 'Namespace(http://example.org)') - self.assertEqual(str(Namespace('http://example.org####')), 'Namespace(http://example.org)') + self.assertEqual(str(Namespace('http://example.org////')), 'http://example.org') + self.assertEqual(str(Namespace('http://example.org####')), 'http://example.org') self.assertEqual(repr(Namespace('http://example.org///##')), 'Namespace(http://example.org, #, /)') # comparison class Foo(Namespace): pass @@ -83,8 +83,8 @@ class TestNamespace(unittest.TestCase): class TestClosedNamespace(unittest.TestCase): def test_essentials(self): # string conversion - self.assertEqual(str(ClosedNamespace('http://example.org/')), 'ClosedNamespace(http://example.org)') - self.assertEqual(str(ClosedNamespace('http://example.org#')), 'ClosedNamespace(http://example.org)') + self.assertEqual(str(ClosedNamespace('http://example.org/')), 'http://example.org') + self.assertEqual(str(ClosedNamespace('http://example.org#')), 'http://example.org') self.assertEqual(repr(ClosedNamespace('http://example.org/')), 'ClosedNamespace(http://example.org, #, /)') self.assertEqual(repr(ClosedNamespace('http://example.org#')), 'ClosedNamespace(http://example.org, #, /)') self.assertEqual(repr(ClosedNamespace('http://example.org', fsep='.')), 'ClosedNamespace(http://example.org, ., /)') diff --git a/test/query/ast/__init__.py b/test/query/ast_test/__init__.py index e69de29..e69de29 100644 --- a/test/query/ast/__init__.py +++ b/test/query/ast_test/__init__.py diff --git a/test/query/ast/test_filter_.py b/test/query/ast_test/test_filter_.py index 4f69bdc..9eb92e2 100644 --- a/test/query/ast/test_filter_.py +++ b/test/query/ast_test/test_filter_.py @@ -15,7 +15,7 @@ from bsfs.utils import URI from bsfs.query.ast.filter_ import _Expression, FilterExpression, PredicateExpression from bsfs.query.ast.filter_ import _Branch, Any, All from bsfs.query.ast.filter_ import _Agg, And, Or -from bsfs.query.ast.filter_ import Not, Has +from bsfs.query.ast.filter_ import Not, Has, Distance from bsfs.query.ast.filter_ import _Value, Is, Equals, Substring, StartsWith, EndsWith from bsfs.query.ast.filter_ import _Bounded, LessThan, GreaterThan from bsfs.query.ast.filter_ import Predicate, OneOf @@ -284,6 +284,39 @@ class TestValue(unittest.TestCase): self.assertEqual(cls(f).value, f) +class TestDistance(unittest.TestCase): + def test_essentials(self): + ref = (1,2,3) + # comparison + self.assertEqual(Distance(ref, 3), Distance(ref, 3)) + self.assertEqual(hash(Distance(ref, 3)), hash(Distance(ref, 3))) + # comparison respects type + self.assertNotEqual(Distance(ref, 3), FilterExpression()) + self.assertNotEqual(hash(Distance(ref, 3)), hash(FilterExpression())) + # comparison respects reference + self.assertNotEqual(Distance((1,2,3), 3, False), Distance((1,2), 3, False)) + self.assertNotEqual(hash(Distance((1,2,3), 3, False)), hash(Distance((1,2), 3, False))) + self.assertNotEqual(Distance((1,2,3), 3, False), Distance((1,5,3), 3, False)) + self.assertNotEqual(hash(Distance((1,2,3), 3, False)), hash(Distance((1,5,3), 3, False))) + # comparison respects threshold + self.assertNotEqual(Distance((1,2,3), 3, False), Distance((1,2,3), 3.1, False)) + self.assertNotEqual(hash(Distance((1,2,3), 3, False)), hash(Distance((1,2,3), 3.1, False))) + # comparison respects strict flag + self.assertNotEqual(Distance((1,2,3), 3, False), Distance((1,2,3), 3, True)) + self.assertNotEqual(hash(Distance((1,2,3), 3, False)), hash(Distance((1,2,3), 3, True))) + # string conversion + self.assertEqual(str(Distance(ref, 3, False)), 'Distance((1, 2, 3), 3.0, False)') + self.assertEqual(repr(Distance(ref, 3, False)), 'Distance((1, 2, 3), 3.0, False)') + + def test_members(self): + self.assertEqual(Distance((1,2,3), 3, False).reference, (1,2,3)) + self.assertEqual(Distance((3,2,1), 3, False).reference, (3,2,1)) + self.assertEqual(Distance((1,2,3), 3, False).threshold, 3.0) + self.assertEqual(Distance((1,2,3), 53.45, False).threshold, 53.45) + self.assertEqual(Distance((1,2,3), 3, False).strict, False) + self.assertEqual(Distance((1,2,3), 3, True).strict, True) + + class TestBounded(unittest.TestCase): def test_essentials(self): # comparison respects type diff --git a/test/query/test_validator.py b/test/query/test_validator.py index 4f8364a..dc9d913 100644 --- a/test/query/test_validator.py +++ b/test/query/test_validator.py @@ -21,7 +21,7 @@ from bsfs.query.validator import Filter class TestFilter(unittest.TestCase): def setUp(self): - self.schema = _schema.Schema.from_string(''' + self.schema = _schema.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> @@ -33,7 +33,19 @@ class TestFilter(unittest.TestCase): bsfs:Tag rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array . + xsd:integer rdfs:subClassOf bsfs:Number . + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "5"^^xsd:integer ; + bsfs:dtype bsfs:f32 . + + bse:color rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Node ; + rdfs:range bsfs:Colors ; + bsfs:unique "true"^^xsd:boolean . bse:comment rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Node ; @@ -69,8 +81,8 @@ class TestFilter(unittest.TestCase): self.assertRaises(TypeError, self.validate, '1234', None) self.assertRaises(TypeError, self.validate, self.schema.literal(ns.bsfs.URI), None) # root_type must exist in the schema - self.assertRaises(errors.ConsistencyError, self.validate, self.schema.node(ns.bsfs.Node).get_child(ns.bsfs.Image), None) - self.assertRaises(errors.ConsistencyError, self.validate, self.schema.node(ns.bsfs.Entity).get_child(ns.bsfs.Image), None) + self.assertRaises(errors.ConsistencyError, self.validate, self.schema.node(ns.bsfs.Node).child(ns.bsfs.Image), None) + self.assertRaises(errors.ConsistencyError, self.validate, self.schema.node(ns.bsfs.Entity).child(ns.bsfs.Image), None) # valid query returns true self.assertTrue(self.validate(self.schema.node(ns.bsfs.Entity), ast.filter.Any(ast.filter.OneOf(ns.bse.tag, ns.bse.buddy), @@ -85,6 +97,7 @@ class TestFilter(unittest.TestCase): ), ast.filter.Not(ast.filter.Any(ns.bse.comment, ast.filter.Not(ast.filter.Equals('hello world')))), + ast.filter.Any(ns.bse.color, ast.filter.Distance([1,2,3,4,5], 3)), ))))) # invalid paths raise consistency error self.assertRaises(errors.ConsistencyError, self.validate, self.schema.node(ns.bsfs.Entity), @@ -130,7 +143,7 @@ class TestFilter(unittest.TestCase): # type must be a node self.assertRaises(errors.ConsistencyError, self.validate._branch, self.schema.literal(ns.bsfs.Literal), None) # type must be in the schema - self.assertRaises(errors.ConsistencyError, self.validate._branch, self.schema.node(ns.bsfs.Node).get_child(ns.bsfs.Invalid), None) + self.assertRaises(errors.ConsistencyError, self.validate._branch, self.schema.node(ns.bsfs.Node).child(ns.bsfs.Invalid), None) # predicate is verified self.assertRaises(errors.ConsistencyError, self.validate._branch, self.schema.node(ns.bsfs.Entity), ast.filter.Any(ns.bsfs.Invalid, ast.filter.Is('http://example.com/entity#1234'))) @@ -187,7 +200,7 @@ class TestFilter(unittest.TestCase): self.assertRaises(errors.ConsistencyError, self.validate._has, self.schema.literal(ns.bsfs.Literal), ast.filter.Has(ns.bse.tag)) # type must be in the schema - self.assertRaises(errors.ConsistencyError, self.validate._has, self.schema.node(ns.bsfs.Node).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._has, self.schema.node(ns.bsfs.Node).child(ns.bsfs.Invalid), ast.filter.Has(ns.bse.tag)) # has checks predicate self.assertRaises(errors.ConsistencyError, self.validate._has, self.schema.node(ns.bsfs.Entity), @@ -206,7 +219,7 @@ class TestFilter(unittest.TestCase): self.assertRaises(errors.ConsistencyError, self.validate._is, self.schema.literal(ns.bsfs.Literal), ast.filter.Is('http://example.com/foo')) # type must be in the schema - self.assertRaises(errors.ConsistencyError, self.validate._is, self.schema.node(ns.bsfs.Node).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._is, self.schema.node(ns.bsfs.Node).child(ns.bsfs.Invalid), ast.filter.Is('http://example.com/foo')) # is accepts correct expressions self.assertIsNone(self.validate._is(self.schema.node(ns.bsfs.Entity), ast.filter.Is('http://example.com/entity#1234'))) @@ -222,13 +235,13 @@ class TestFilter(unittest.TestCase): self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.node(ns.bsfs.Node), ast.filter.EndsWith('hello world')) # type must be in the schema - self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).child(ns.bsfs.Invalid), ast.filter.Equals('hello world')) - self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).child(ns.bsfs.Invalid), ast.filter.Substring('hello world')) - self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).child(ns.bsfs.Invalid), ast.filter.StartsWith('hello world')) - self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._value, self.schema.literal(ns.bsfs.Literal).child(ns.bsfs.Invalid), ast.filter.EndsWith('hello world')) # value accepts correct expressions self.assertIsNone(self.validate._value(self.schema.literal(ns.xsd.string), ast.filter.Equals('hello world'))) @@ -243,14 +256,34 @@ class TestFilter(unittest.TestCase): self.assertRaises(errors.ConsistencyError, self.validate._bounded, self.schema.node(ns.bsfs.Node), ast.filter.LessThan(0)) # type must be in the schema - self.assertRaises(errors.ConsistencyError, self.validate._bounded, self.schema.literal(ns.bsfs.Literal).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._bounded, self.schema.literal(ns.bsfs.Literal).child(ns.bsfs.Invalid), ast.filter.GreaterThan(0)) - self.assertRaises(errors.ConsistencyError, self.validate._bounded, self.schema.literal(ns.bsfs.Literal).get_child(ns.bsfs.Invalid), + self.assertRaises(errors.ConsistencyError, self.validate._bounded, self.schema.literal(ns.bsfs.Literal).child(ns.bsfs.Invalid), + ast.filter.LessThan(0)) + # type must be a number + self.assertRaises(errors.ConsistencyError, self.validate._bounded, self.schema.literal(ns.xsd.string), ast.filter.LessThan(0)) # bounded accepts correct expressions self.assertIsNone(self.validate._bounded(self.schema.literal(ns.xsd.integer), ast.filter.LessThan(0))) self.assertIsNone(self.validate._bounded(self.schema.literal(ns.xsd.integer), ast.filter.GreaterThan(0))) + def test_distance(self): + # type must be a literal + self.assertRaises(errors.ConsistencyError, self.validate._distance, self.schema.node(ns.bsfs.Node), + ast.filter.Distance([1,2,3], 1, False)) + # type must be a feature + self.assertRaises(errors.ConsistencyError, self.validate._distance, self.schema.literal(ns.bsfs.Array), + ast.filter.Distance([1,2,3], 1, False)) + # type must be in the schema + self.assertRaises(errors.ConsistencyError, self.validate._distance, self.schema.literal(ns.bsfs.Feature).child(ns.bsfs.Invalid), + ast.filter.Distance([1,2,3], 1, False)) + # FIXME: reference must be a numpy array + # reference must have the correct dimension + self.assertRaises(errors.ConsistencyError, self.validate._distance, self.schema.literal(ns.bsfs.Colors), + ast.filter.Distance([1,2,3], 1, False)) + # FIXME: reference must have the correct dtype + # distance accepts correct expressions + self.assertIsNone(self.validate._distance(self.schema.literal(ns.bsfs.Colors), ast.filter.Distance([1,2,3,4,5], 1, False))) ## main ## diff --git a/test/schema/test_schema.py b/test/schema/test_schema.py index 888cdca..32dbc93 100644 --- a/test/schema/test_schema.py +++ b/test/schema/test_schema.py @@ -10,7 +10,7 @@ import unittest # bsfs imports from bsfs.namespace import ns -from bsfs.schema import types +from bsfs.schema import types, from_string from bsfs.utils import errors # objects to test @@ -35,7 +35,8 @@ class TestSchema(unittest.TestCase): bsfs:Unused rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . xsd:boolean rdfs:subClassOf bsfs:Literal . bse:tag rdfs:subClassOf bsfs:Predicate ; @@ -55,32 +56,42 @@ class TestSchema(unittest.TestCase): ''' # nodes - self.n_root = types.Node(ns.bsfs.Node, None) - self.n_ent = types.Node(ns.bsfs.Entity, types.Node(ns.bsfs.Node, None)) - self.n_img = types.Node(ns.bsfs.Image, types.Node(ns.bsfs.Entity, types.Node(ns.bsfs.Node, None))) - self.n_tag = types.Node(ns.bsfs.Tag, types.Node(ns.bsfs.Node, None)) - self.n_unused = types.Node(ns.bsfs.Unused, types.Node(ns.bsfs.Node, None)) + self.n_root = types.ROOT_NODE + self.n_ent = self.n_root.child(ns.bsfs.Entity) + self.n_img = self.n_ent.child(ns.bsfs.Image) + self.n_tag = self.n_root.child(ns.bsfs.Tag) + self.n_unused = self.n_root.child(ns.bsfs.Unused) self.nodes = [self.n_root, self.n_ent, self.n_img, self.n_tag, self.n_unused] # literals - self.l_root = types.Literal(ns.bsfs.Literal, None) - self.l_string = types.Literal(ns.xsd.string, types.Literal(ns.bsfs.Literal, None)) - self.l_integer = types.Literal(ns.xsd.integer, types.Literal(ns.bsfs.Literal, None)) - self.l_unused = types.Literal(ns.xsd.boolean, types.Literal(ns.bsfs.Literal, None)) - self.literals = [self.l_root, self.l_string, self.l_integer, self.l_unused] + self.l_root = types.ROOT_LITERAL + self.l_number = types.ROOT_NUMBER + self.l_array = types.ROOT_ARRAY + self.l_time = types.ROOT_TIME + self.l_string = self.l_root.child(ns.xsd.string) + self.l_integer = self.l_root.child(ns.xsd.integer) + self.l_unused = self.l_root.child(ns.xsd.boolean) + self.f_root = types.ROOT_FEATURE + self.literals = [self.l_root, self.l_array, self.f_root, self.l_number, self.l_time, self.l_string, self.l_integer, self.l_unused] # predicates - self.p_root = types.Predicate(ns.bsfs.Predicate, None, types.Node(ns.bsfs.Node, None), None, False) - self.p_tag = self.p_root.get_child(ns.bse.tag, self.n_ent, self.n_tag, False) - self.p_group = self.p_tag.get_child(ns.bse.group, self.n_img, self.n_tag, False) - self.p_comment = self.p_root.get_child(ns.bse.comment, self.n_root, self.l_string, True) + self.p_root = types.ROOT_PREDICATE + self.p_tag = self.p_root.child(ns.bse.tag, self.n_ent, self.n_tag, False) + self.p_group = self.p_tag.child(ns.bse.group, self.n_img, self.n_tag, False) + self.p_comment = self.p_root.child(ns.bse.comment, self.n_root, self.l_string, True) self.predicates = [self.p_root, self.p_tag, self.p_group, self.p_comment] def test_construction(self): + # no args yields a minimal schema + schema = Schema() + self.assertSetEqual(set(schema.nodes()), {self.n_root}) + self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_number, self.l_array, self.l_time, self.f_root}) + self.assertSetEqual(set(schema.predicates()), {self.p_root}) + # nodes and literals are optional schema = Schema(self.predicates) self.assertSetEqual(set(schema.nodes()), {self.n_root, self.n_ent, self.n_img, self.n_tag}) - self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string}) + self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string, self.l_number, self.l_time, self.l_array, self.f_root}) self.assertSetEqual(set(schema.predicates()), set(self.predicates)) # predicates, nodes, and literals are respected @@ -101,19 +112,19 @@ class TestSchema(unittest.TestCase): # literals are complete schema = Schema(self.predicates, self.nodes, None) - self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string}) + self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string, self.l_number, self.l_array, self.l_time, self.f_root}) schema = Schema(self.predicates, self.nodes, []) - self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string}) + self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string, self.l_number, self.l_array, self.l_time, self.f_root}) schema = Schema(self.predicates, self.nodes, [self.l_string]) - self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string}) + self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string, self.l_number, self.l_array, self.l_time, self.f_root}) schema = Schema(self.predicates, self.nodes, [self.l_integer]) - self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string, self.l_integer}) + self.assertSetEqual(set(schema.literals()), {self.l_root, self.l_string, self.l_integer, self.l_number, self.l_array, self.l_time, self.f_root}) schema = Schema(self.predicates, self.nodes, [self.l_integer, self.l_unused]) self.assertSetEqual(set(schema.literals()), set(self.literals)) # predicates are complete schema = Schema([], self.nodes, self.literals) - self.assertSetEqual(set(schema.predicates()), set()) + self.assertSetEqual(set(schema.predicates()), {self.p_root}) schema = Schema([self.p_group], self.nodes, self.literals) self.assertSetEqual(set(schema.predicates()), {self.p_root, self.p_tag, self.p_group}) schema = Schema([self.p_group, self.p_comment], self.nodes, self.literals) @@ -153,20 +164,27 @@ class TestSchema(unittest.TestCase): self.assertRaises(errors.ConsistencyError, Schema, {}, {types.Node(ns.bsfs.Foo, None)}, {types.Node(ns.bsfs.Foo, None)}) self.assertRaises(errors.ConsistencyError, Schema, - {types.Predicate(ns.bsfs.Foo, None, types.Node(ns.bsfs.Node, None), None, False)}, {}, {types.Node(ns.bsfs.Foo, None)}) + {types.Predicate(ns.bsfs.Foo, None, types.Node(ns.bsfs.Node, None), types.ROOT_VERTEX, False)}, {}, {types.Node(ns.bsfs.Foo, None)}) self.assertRaises(errors.ConsistencyError, Schema, - {types.Predicate(ns.bsfs.Foo, None, types.Node(ns.bsfs.Node, None), None, False)}, {types.Node(ns.bsfs.Foo, None)}, {}) + {types.Predicate(ns.bsfs.Foo, None, types.Node(ns.bsfs.Node, None), types.ROOT_VERTEX, False)}, {types.Node(ns.bsfs.Foo, None)}, {}) self.assertRaises(errors.ConsistencyError, Schema, - {types.Predicate(ns.bsfs.Foo, None, types.Node(ns.bsfs.Node, None), None, False)}, {types.Node(ns.bsfs.Foo, None)}, {types.Node(ns.bsfs.Foo, None)}) + {types.Predicate(ns.bsfs.Foo, None, types.Node(ns.bsfs.Node, None), types.ROOT_VERTEX, False)}, {types.Node(ns.bsfs.Foo, None)}, {types.Node(ns.bsfs.Foo, None)}) + def test_str(self): + # string conversion self.assertEqual(str(Schema([])), 'Schema()') self.assertEqual(str(Schema([], [], [])), 'Schema()') self.assertEqual(str(Schema(self.predicates, self.nodes, self.literals)), 'Schema()') - self.assertEqual(repr(Schema([])), 'Schema([], [], [])') - self.assertEqual(repr(Schema([], [], [])), 'Schema([], [], [])') + # repr conversion with only default nodes, literals, and predicates + n = [ns.bsfs.Node] + l = [ns.bsfs.Array, ns.bsfs.Feature, ns.bsfs.Literal, ns.bsfs.Number, ns.bsfs.Time] + p = [ns.bsfs.Predicate] + self.assertEqual(repr(Schema()), f'Schema({n}, {l}, {p})') + self.assertEqual(repr(Schema([], [], [])), f'Schema({n}, {l}, {p})') + # repr conversion n = [ns.bsfs.Entity, ns.bsfs.Image, ns.bsfs.Node, ns.bsfs.Tag, ns.bsfs.Unused] - l = [ns.bsfs.Literal, ns.xsd.boolean, ns.xsd.integer, ns.xsd.string] + l = [ns.bsfs.Array, ns.bsfs.Feature, ns.bsfs.Literal, ns.bsfs.Number, ns.bsfs.Time, ns.xsd.boolean, ns.xsd.integer, ns.xsd.string] p = [ns.bse.comment, ns.bse.group, ns.bse.tag, ns.bsfs.Predicate] self.assertEqual(repr(Schema(self.predicates, self.nodes, self.literals)), f'Schema({n}, {l}, {p})') @@ -202,16 +220,16 @@ class TestSchema(unittest.TestCase): self.assertNotEqual(hash(schema), hash(Schema([self.p_group, self.p_tag, self.p_root], self.nodes, self.literals))) self.assertNotEqual(schema, - Schema(self.predicates + [self.p_root.get_child(ns.bse.filesize, self.n_ent, self.l_integer)], self.nodes, self.literals)) + Schema(self.predicates + [self.p_root.child(ns.bse.filesize, self.n_ent, self.l_integer)], self.nodes, self.literals)) self.assertNotEqual(hash(schema), - hash(Schema(self.predicates + [self.p_root.get_child(ns.bse.filesize, self.n_ent, self.l_integer)], self.nodes, self.literals))) + hash(Schema(self.predicates + [self.p_root.child(ns.bse.filesize, self.n_ent, self.l_integer)], self.nodes, self.literals))) def test_order(self): # setup class Foo(): pass - p_foo = self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_string, True) - p_sub = p_foo.get_child(ns.bse.sub, self.n_ent, self.l_string, True) - p_bar = self.p_root.get_child(ns.bse.bar, self.n_ent, self.l_string, True) + p_foo = self.p_root.child(ns.bse.foo, self.n_ent, self.l_string, True) + p_sub = p_foo.child(ns.bse.sub, self.n_ent, self.l_string, True) + p_bar = self.p_root.child(ns.bse.bar, self.n_ent, self.l_string, True) # can only compare schema to other schema # < @@ -258,11 +276,11 @@ class TestSchema(unittest.TestCase): self.assertTrue(operator.lt(Schema({self.p_tag}), Schema({self.p_group}))) self.assertTrue(operator.le(Schema({self.p_tag}), Schema({self.p_group}))) # subset considers differences in predicates and literals - self.assertTrue(operator.lt(Schema.Empty(), Schema({self.p_comment}))) + self.assertTrue(operator.lt(Schema(), Schema({self.p_comment}))) # subset considers differences in predicates, nodes, and literals - self.assertTrue(operator.lt(Schema({}), Schema.Empty())) - self.assertTrue(operator.lt(Schema({self.p_tag}), Schema.from_string(self.schema_str))) - self.assertTrue(operator.le(Schema({self.p_tag}), Schema.from_string(self.schema_str))) + self.assertTrue(operator.le(Schema({}), Schema())) + self.assertTrue(operator.lt(Schema({self.p_tag}), from_string(self.schema_str))) + self.assertTrue(operator.le(Schema({self.p_tag}), from_string(self.schema_str))) self.assertFalse(operator.lt(Schema({self.p_comment}), Schema({self.p_tag}))) self.assertFalse(operator.le(Schema({self.p_comment}), Schema({self.p_tag}))) @@ -280,54 +298,54 @@ class TestSchema(unittest.TestCase): self.assertTrue(operator.gt(Schema({self.p_group}), Schema({self.p_tag}))) self.assertTrue(operator.ge(Schema({self.p_group}), Schema({self.p_tag}))) # superset considers differences in predicates and literals - self.assertTrue(operator.gt(Schema({self.p_comment}), Schema.Empty())) + self.assertTrue(operator.gt(Schema({self.p_comment}), Schema())) # superset considers differences in predicates, nodes, and literals - self.assertTrue(operator.gt(Schema.Empty(), Schema({}))) - self.assertTrue(operator.gt(Schema.from_string(self.schema_str), Schema({self.p_tag}))) - self.assertTrue(operator.ge(Schema.from_string(self.schema_str), Schema({self.p_tag}))) + self.assertTrue(operator.ge(Schema(), Schema({}))) + self.assertTrue(operator.gt(from_string(self.schema_str), Schema({self.p_tag}))) + self.assertTrue(operator.ge(from_string(self.schema_str), Schema({self.p_tag}))) self.assertFalse(operator.gt(Schema({self.p_tag}), Schema({self.p_comment}))) self.assertFalse(operator.ge(Schema({self.p_tag}), Schema({self.p_comment}))) # inconsistent schema cannot be a subset self.assertFalse(operator.le(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal + self.p_root.child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal self.assertFalse(operator.le(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node + self.p_root.child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node self.assertFalse(operator.le(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique + self.p_root.child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique self.assertFalse(operator.le(Schema({}, {self.n_img}), Schema({}, { types.Node(ns.bsfs.Image, types.Node(ns.bsfs.Node, None))}))) self.assertFalse(operator.le(Schema({}, {}, {self.l_integer}), Schema({}, {}, { types.Literal(ns.xsd.integer, types.Literal(ns.xsd.number, types.Literal(ns.bsfs.Literal, None)))}))) # inconsistent schema cannot be a true subset self.assertFalse(operator.lt(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal + self.p_root.child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal self.assertFalse(operator.lt(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node + self.p_root.child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node self.assertFalse(operator.lt(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique + self.p_root.child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique self.assertFalse(operator.lt(Schema({}, {self.n_img}), Schema({}, { types.Node(ns.bsfs.Image, types.Node(ns.bsfs.Node, None))}))) self.assertFalse(operator.lt(Schema({}, {}, {self.l_integer}), Schema({}, {}, { types.Literal(ns.xsd.integer, types.Literal(ns.xsd.number, types.Literal(ns.bsfs.Literal, None)))}))) # inconsistent schema cannot be a superset self.assertFalse(operator.ge(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal + self.p_root.child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal self.assertFalse(operator.ge(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node + self.p_root.child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node self.assertFalse(operator.ge(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique + self.p_root.child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique self.assertFalse(operator.ge(Schema({}, {self.n_img}), Schema({}, { types.Node(ns.bsfs.Image, types.Node(ns.bsfs.Node, None))}))) self.assertFalse(operator.ge(Schema({}, {}, {self.l_integer}), Schema({}, {}, { types.Literal(ns.xsd.integer, types.Literal(ns.xsd.number, types.Literal(ns.bsfs.Literal, None)))}))) # inconsistent schema cannot be a true superset self.assertFalse(operator.gt(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal + self.p_root.child(ns.bse.foo, self.n_ent, self.l_integer, True)}))) # inconsistent w.r.t. literal self.assertFalse(operator.gt(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node + self.p_root.child(ns.bse.foo, self.n_img, self.l_string, True)}))) # inconsistent w.r.t. node self.assertFalse(operator.gt(Schema({p_foo}), Schema({ - self.p_root.get_child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique + self.p_root.child(ns.bse.foo, self.n_ent, self.l_string, False)}))) # inconsistent w.r.t. unique self.assertFalse(operator.gt(Schema({}, {self.n_img}), Schema({}, { types.Node(ns.bsfs.Image, types.Node(ns.bsfs.Node, None))}))) self.assertFalse(operator.gt(Schema({}, {}, {self.l_integer}), Schema({}, {}, { @@ -351,26 +369,26 @@ class TestSchema(unittest.TestCase): # difference does not contain predicates from the RHS diff = Schema({self.p_tag, self.p_comment}).diff(Schema({self.p_group})) self.assertSetEqual(set(diff.nodes), set()) - self.assertSetEqual(set(diff.literals), {self.l_root, self.l_string}) + self.assertSetEqual(set(diff.literals), {self.l_string}) self.assertSetEqual(set(diff.predicates), {self.p_comment}) # difference considers extra nodes and literals diff = Schema({self.p_tag}, {self.n_unused}, {self.l_unused}).diff(Schema({self.p_tag})) self.assertSetEqual(set(diff.nodes), {self.n_unused}) - self.assertSetEqual(set(diff.literals), {self.l_root, self.l_unused}) + self.assertSetEqual(set(diff.literals), {self.l_unused}) self.assertSetEqual(set(diff.predicates), set()) # difference considers inconsistent types diff = Schema({self.p_tag}, {self.n_unused}, {self.l_unused}).diff( Schema({self.p_tag}, {types.Node(ns.bsfs.Unused, None)}, {types.Literal(ns.xsd.boolean, None)})) self.assertSetEqual(set(diff.nodes), {self.n_unused}) - self.assertSetEqual(set(diff.literals), {self.l_root, self.l_unused}) + self.assertSetEqual(set(diff.literals), {self.l_unused}) self.assertSetEqual(set(diff.predicates), set()) # __sub__ is an alias for diff diff = Schema({self.p_comment}, {self.n_unused}, {self.l_unused}) - Schema({self.p_group}) self.assertSetEqual(set(diff.nodes), {self.n_unused}) - self.assertSetEqual(set(diff.literals), {self.l_root, self.l_string, self.l_unused}) + self.assertSetEqual(set(diff.literals), {self.l_string, self.l_unused}) self.assertSetEqual(set(diff.predicates), {self.p_comment}) # __sub__ only accepts Schema instances class Foo(): pass @@ -547,196 +565,6 @@ class TestSchema(unittest.TestCase): self.assertFalse(schema.has_predicate(ns.bse.mimetype)) self.assertFalse(schema.has_predicate(self.p_root)) - def test_empty(self): - self.assertEqual(Schema.Empty(), Schema( - [types.Predicate(ns.bsfs.Predicate, None, types.Node(ns.bsfs.Node, None), None, False)], - [types.Node(ns.bsfs.Node, None)], - [types.Literal(ns.bsfs.Literal, None)], - )) - - def test_from_string(self): - # from_string creates a schema - self.assertEqual( - Schema(self.predicates, self.nodes, self.literals), - Schema.from_string(self.schema_str)) - - # schema contains at least the root types - self.assertEqual(Schema.from_string(''), Schema({self.p_root}, {self.n_root}, {self.l_root})) - - # custom example - self.assertEqual( - Schema({types.Predicate(ns.bsfs.Predicate, None, self.n_root, None, False).get_child( - ns.bse.filename, self.n_ent, self.l_string, False)}), - Schema.from_string(''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - xsd:string rdfs:subClassOf bsfs:Literal . - - bse:filename rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Entity ; - rdfs:range xsd:string ; - bsfs:unique "false"^^xsd:boolean . - ''')) - - # all nodes must be defined - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - xsd:string rdfs:subClassOf bsfs:Literal . - - bse:filename rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Entity ; - rdfs:range xsd:string ; - bsfs:unique "false"^^xsd:boolean . - ''') - - # all literals must be defined - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - - bse:filename rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Entity ; - rdfs:range xsd:string ; - bsfs:unique "false"^^xsd:boolean . - ''') - - # must not have circular dependencies - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix bsfs: <http://bsfs.ai/schema/> - bsfs:Entity rdfs:subClassOf bsfs:Node . - # ah, a nice circular dependency - bsfs:Entity rdfs:subClassOf bsfs:Document . - bsfs:Document rdfs:subClassOf bsfs:Entity . - bsfs:PDF rdfs:subClassOf bsfs:Document . - ''') - - # range must be a node or literal - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - - bse:filename rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Entity ; - rdfs:range xsd:string ; - bsfs:unique "false"^^xsd:boolean . - ''') - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - - bse:filename rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Entity ; - rdfs:range bsfs:Foo ; - bsfs:unique "false"^^xsd:boolean . - ''') - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - - bse:filename rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Entity ; - rdfs:range bsfs:Predicate ; - bsfs:unique "false"^^xsd:boolean . - ''') - - # must be consistent - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - bsfs:Document rdfs:subClassOf bsfs:Node . - bsfs:Document rdfs:subClassOf bsfs:Entity. - ''') - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - - xsd:string rdfs:subClassOf bsfs:Literal . - xsd:name rdfs:subClassOf bsfs:Literal . - xsd:name rdfs:subClassOf xsd:string . - ''') - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - - bse:foo rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Node ; - rdfs:range bsfs:Node ; - bsfs:unique "false"^^xsd:boolean . - - bse:foo rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Entity . - - ''') - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - - bse:foo rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Node ; - rdfs:range bsfs:Node ; - bsfs:unique "false"^^xsd:boolean . - - bse:foo rdfs:subClassOf bsfs:Predicate ; - rdfs:range bsfs:Entity . - - ''') - self.assertRaises(errors.ConsistencyError, Schema.from_string, ''' - prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> - prefix xsd: <http://www.w3.org/2001/XMLSchema#> - prefix bsfs: <http://bsfs.ai/schema/> - prefix bse: <http://bsfs.ai/schema/Entity#> - - bsfs:Entity rdfs:subClassOf bsfs:Node . - - bse:foo rdfs:subClassOf bsfs:Predicate ; - rdfs:domain bsfs:Node ; - rdfs:range bsfs:Node ; - bsfs:unique "false"^^xsd:boolean . - - bse:foo rdfs:subClassOf bsfs:Predicate ; - bsfs:unique "true"^^xsd:boolean . - - ''') - - - ## main ## if __name__ == '__main__': diff --git a/test/schema/test_serialize.py b/test/schema/test_serialize.py new file mode 100644 index 0000000..fc6b20a --- /dev/null +++ b/test/schema/test_serialize.py @@ -0,0 +1,1030 @@ +""" + +Part of the tagit test suite. +A copy of the license is provided with the project. +Author: Matthias Baumgartner, 2022 +""" +# imports +import re +import unittest + +# bsfs imports +from bsfs.namespace import ns +from bsfs.schema import Schema, types +from bsfs.utils import errors, URI + +# objects to test +from bsfs.schema.serialize import from_string, to_string + + +## code ## + +class TestFromString(unittest.TestCase): + + def test_empty(self): + # schema contains at least the root types + self.assertEqual(from_string(''), Schema()) + + + def test_circular_dependency(self): + # must not have circular dependencies + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix bsfs: <http://bsfs.ai/schema/> + bsfs:Entity rdfs:subClassOf bsfs:Node . + # ah, a nice circular dependency + bsfs:Entity rdfs:subClassOf bsfs:Document . + bsfs:Document rdfs:subClassOf bsfs:Entity . + bsfs:PDF rdfs:subClassOf bsfs:Document . + ''') + + + def test_node(self): + # all nodes must be defined + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + xsd:string rdfs:subClassOf bsfs:Literal . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + ''') + + # node definitions must be consistent (cannot re-use a node uri) + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + bsfs:Document rdfs:subClassOf bsfs:Node . + bsfs:Document rdfs:subClassOf bsfs:Entity . # conflicting parent + ''') + + # additional nodes can be defined + n_unused = types.ROOT_NODE.child(ns.bsfs.unused) + self.assertEqual(Schema({}, {n_unused}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:unused rdfs:subClassOf bsfs:Node . # unused symbol + ''')) + + # a node can have multiple children + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + n_tag = types.ROOT_NODE.child(ns.bsfs.Tag) + n_doc = n_ent.child(ns.bsfs.Document) + n_image = n_ent.child(ns.bsfs.Image) + self.assertEqual(Schema({}, {n_ent, n_tag, n_doc, n_image}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + # nodes inherit from same parent + bsfs:Entity rdfs:subClassOf bsfs:Node . + bsfs:Tag rdfs:subClassOf bsfs:Node . + + # nodes inherit from same parent + bsfs:Document rdfs:subClassOf bsfs:Entity . + bsfs:Image rdfs:subClassOf bsfs:Entity . + ''')) + + # additional nodes can be defined and used + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + l_string = types.ROOT_LITERAL.child(ns.xsd.string) + p_filename = types.ROOT_PREDICATE.child(ns.bse.filename, + n_ent, l_string, False) + self.assertEqual(Schema({p_filename}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + xsd:string rdfs:subClassOf bsfs:Literal . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + ''')) + + # nodes can have annotations + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + ''').node(ns.bsfs.Entity).annotations, {}) + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + + bsfs:Entity rdfs:subClassOf bsfs:Node ; + rdfs:label "hello world"^^xsd:string ; + bsfs:foo "1234"^^xsd:integer . + + ''').node(ns.bsfs.Entity).annotations, { + ns.rdfs.label: 'hello world', + ns.bsfs.foo: 1234, + }) + + + def test_literal(self): + # all literals must be defined + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range xsd:string ; # undefined symbol + bsfs:unique "false"^^xsd:boolean . + ''') + + # literal definitions must be consistent (cannot re-use a literal uri) + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + + xsd:string rdfs:subClassOf bsfs:Literal . + xsd:name rdfs:subClassOf bsfs:Literal . + xsd:name rdfs:subClassOf xsd:string . # conflicting parent + ''') + + # additional literals can be defined + l_unused = types.ROOT_LITERAL.child(ns.xsd.unused) + self.assertEqual(Schema({}, {}, {l_unused}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + xsd:unused rdfs:subClassOf bsfs:Literal . # unused symbol + ''')) + + # a literal can have multiple children + l_string = types.ROOT_LITERAL.child(ns.xsd.string) + l_integer = types.ROOT_NUMBER.child(ns.xsd.integer) + l_unsigned = l_integer.child(ns.xsd.unsigned) + l_signed = l_integer.child(ns.xsd.signed) + self.assertEqual(Schema({}, {}, {l_string, l_integer, l_unsigned, l_signed}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + # literals inherit from same parent + xsd:string rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . + + # literals inherit from same parent + xsd:unsigned rdfs:subClassOf xsd:integer . + xsd:signed rdfs:subClassOf xsd:integer . + ''')) + + # additional literals can be defined and used + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + l_string = types.ROOT_LITERAL.child(ns.xsd.string) + p_filename = types.ROOT_PREDICATE.child(ns.bse.filename, + n_ent, l_string, False) + self.assertEqual(Schema({p_filename}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + xsd:string rdfs:subClassOf bsfs:Literal . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + ''')) + + # literals can have annotations + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + + xsd:string rdfs:subClassOf bsfs:Literal . + + ''').literal(ns.xsd.string).annotations, {}) + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + + xsd:string rdfs:subClassOf bsfs:Literal ; + rdfs:label "hello world"^^xsd:string ; + bsfs:foo "1234"^^xsd:integer . + + ''').literal(ns.xsd.string).annotations, { + ns.rdfs.label: 'hello world', + ns.bsfs.foo: 1234, + }) + + + def test_predicate(self): + # domain must be defined + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + xsd:string rdfs:subClassOf bsfs:Literal . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; # undefined symbol + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + ''') + # domain cannot be a literal + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Literal . + xsd:string rdfs:subClassOf bsfs:Literal . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; # literal instead of node + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + ''') + + # range must be defined + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range xsd:string ; # undefined symbol + bsfs:unique "false"^^xsd:boolean . + ''') + # range must be defined + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range bsfs:Foo ; # undefined symbol + bsfs:unique "false"^^xsd:boolean . + ''') + # range must be a node or a literal + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:filename rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range bsfs:Predicate ; # invalid symbol + bsfs:unique "false"^^xsd:boolean . + ''') + + # additional predicates can be defined + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + l_string = types.ROOT_LITERAL.child(ns.xsd.string) + p_comment = types.ROOT_PREDICATE.child(ns.bse.comment, domain=n_ent, range=l_string, unique=False) + self.assertEqual(Schema({p_comment}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + xsd:string rdfs:subClassOf bsfs:Literal . + + bse:comment rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + ''')) + + # predicates inherit properties from parents + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + l_string = types.ROOT_LITERAL.child(ns.xsd.string) + p_annotation = types.ROOT_PREDICATE.child(ns.bsfs.Annotation, domain=n_ent, range=l_string) + p_comment = p_annotation.child(ns.bse.comment, unique=True) + self.assertEqual(Schema({p_comment}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + xsd:string rdfs:subClassOf bsfs:Literal . + + bsfs:Annotation rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range xsd:string . + + bse:comment rdfs:subClassOf bsfs:Annotation ; # inherits domain/range from bsfs:Annotation + bsfs:unique "true"^^xsd:boolean . + ''')) + + # we can define partial predicates (w/o specifying a usable range) + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + l_string = types.ROOT_LITERAL.child(ns.xsd.string) + p_annotation = types.ROOT_PREDICATE.child(ns.bsfs.Annotation, domain=n_ent) + p_comment = p_annotation.child(ns.bse.comment, range=l_string, unique=False) + self.assertEqual(Schema({p_comment}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + xsd:string rdfs:subClassOf bsfs:Literal . + + bsfs:Annotation rdfs:subClassOf bsfs:Predicate ; # derive predicate w/o setting range + rdfs:domain bsfs:Entity . + + bse:comment rdfs:subClassOf bsfs:Annotation ; # derived predicate w/ setting range + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + ''')) + + # predicate definition can be split across multiple statements. + # statements can be repeated + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + p_foo = types.ROOT_PREDICATE.child(ns.bse.foo, domain=n_ent, range=types.ROOT_NODE, unique=True) + self.assertEqual(Schema({p_foo}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:range bsfs:Node ; + bsfs:unique "true"^^xsd:boolean . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity. + ''')) + + # domain must be a subtype of parent's domain + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + n_image = n_ent.child(ns.bsfs.Image) + p_foo = types.ROOT_PREDICATE.child(ns.bse.foo, domain=types.ROOT_NODE) + p_bar = p_foo.child(ns.bse.bar, domain=n_ent) + p_foobar = p_bar.child(ns.bse.foobar, domain=n_image) + self.assertEqual(Schema({p_foobar}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + bsfs:Image rdfs:subClassOf bsfs:Entity . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Node . + bse:bar rdfs:subClassOf bse:foo ; + rdfs:domain bsfs:Entity . + bse:foobar rdfs:subClassOf bse:bar ; + rdfs:domain bsfs:Image . + ''')) + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + bsfs:Image rdfs:subClassOf bsfs:Entity . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Image . + bse:bar rdfs:subClassOf bse:foo ; + rdfs:domain bsfs:Entity . + bse:foobar rdfs:subClassOf bse:bar ; + rdfs:domain bsfs:Node . + ''') + + # range must be a subtype of parent's range + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + n_image = n_ent.child(ns.bsfs.Image) + p_foo = types.ROOT_PREDICATE.child(ns.bse.foo, range=types.ROOT_NODE) + p_bar = p_foo.child(ns.bse.bar, range=n_ent) + p_foobar = p_bar.child(ns.bse.foobar, range=n_image) + self.assertEqual(Schema({p_foobar}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + bsfs:Image rdfs:subClassOf bsfs:Entity . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:range bsfs:Node . + bse:bar rdfs:subClassOf bse:foo ; + rdfs:range bsfs:Entity . + bse:foobar rdfs:subClassOf bse:bar ; + rdfs:range bsfs:Image . + ''')) + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + bsfs:Image rdfs:subClassOf bsfs:Entity . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:range bsfs:Image . + bse:bar rdfs:subClassOf bse:foo ; + rdfs:range bsfs:Entity . + bse:foobar rdfs:subClassOf bse:bar ; + rdfs:range bsfs:Node . + ''') + + # cannot define the same predicate from multiple parents + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + bsfs:Annotation rdfs:subClassOf bsfs:Predicate . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Node ; + rdfs:range bsfs:Node ; + bsfs:unique "false"^^xsd:boolean . + + bse:foo rdfs:subClassOf bsfs:Annotation ; + rdfs:domain bsfs:Node ; + rdfs:range bsfs:Node ; + bsfs:unique "false"^^xsd:boolean . + + ''') + # cannot assign multiple conflicting domains to the same predicate + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Node ; + rdfs:range bsfs:Node ; + bsfs:unique "false"^^xsd:boolean . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity . # conflicting domain + ''') + # cannot assign multiple conflicting ranges to the same predicate + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Node ; + rdfs:range bsfs:Node ; + bsfs:unique "false"^^xsd:boolean . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:range bsfs:Entity . # conflicting range + ''') + # cannot assign multiple conflicting uniques to the same predicate + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Entity rdfs:subClassOf bsfs:Node . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Node ; + rdfs:range bsfs:Node ; + bsfs:unique "false"^^xsd:boolean . + + bse:foo rdfs:subClassOf bsfs:Predicate ; + bsfs:unique "true"^^xsd:boolean . # conflicting unique + ''') + + # predicates can have annotations + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bse:comment rdfs:subClassOf bsfs:Predicate ; + rdfs:range bsfs:Node . + + ''').predicate(ns.bse.comment).annotations, {}) + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bse:comment rdfs:subClassOf bsfs:Predicate ; + rdfs:range bsfs:Node ; + rdfs:label "hello world"^^xsd:string ; + bsfs:foo "1234"^^xsd:integer . + + ''').predicate(ns.bse.comment).annotations, { + ns.rdfs.label: 'hello world', + ns.bsfs.foo: 1234, + }) + + + def test_feature(self): + # additional features can be defined + f_colors = types.ROOT_FEATURE.child(ns.bsfs.Colors) + self.assertEqual(Schema(literals={f_colors}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature . + + ''')) + + # features inherit properties from parents + f_colors = types.ROOT_FEATURE.child(ns.bsfs.Colors, dimension=1234, dtype=ns.bsfs.i32) + f_main_colors = f_colors.child(ns.bsfs.MainColor, distance=ns.bsfs.cosine, dtype=ns.bsfs.f16) + self.assertEqual(Schema(literals={f_colors, f_main_colors}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; # inherits distance from bsfs:Feature + bsfs:dimension "1234"^^xsd:integer ; # overwrites bsfs:Feature + bsfs:dtype bsfs:i32 . # overwrites bsfs:Feature + + bsfs:MainColor rdfs:subClassOf bsfs:Colors ; # inherits dimension from bsfs:Colors + bsfs:distance bsfs:cosine ; # overwrites bsfs:Feature + bsfs:dtype bsfs:f16 . # overwrites bsfs:Colors + + ''')) + + # feature definition can be split across multiple statements. + # statements can be repeated + f_colors = types.ROOT_FEATURE.child(ns.bsfs.Colors, dimension=1234, dtype=ns.bsfs.f32) + self.assertEqual(Schema(literals={f_colors}), from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "1234"^^xsd:integer . + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "1234"^^xsd:integer ; # non-conflicting repetition + bsfs:dtype bsfs:f32 . + ''')) + + # cannot define the same feature from multiple parents + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + bsfs:ColorSpace rdfs:subClassOf bsfs:Feature . + + bsfs:Colors rdfs:subClassOf bsfs:Feature . + bsfs:Colors rdfs:subClassOf bsfs:ColorSpace . + + ''') + # cannot assign multiple conflicting dimensions to the same feature + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "1234"^^xsd:integer . + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "4321"^^xsd:integer . # conflicting dimension + + ''') + # cannot assign multiple conflicting dtypes to the same feature + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dtype bsfs:f32 . + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dtype bsfs:f16 . # conflicting dtype + ''') + # cannot assign multiple conflicting distance metrics to the same feature + self.assertRaises(errors.ConsistencyError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:distance bsfs:euclidean . + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:distance bsfs:cosine . # conflicting distance + ''') + + # features can have annotations + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "1234"^^xsd:integer . + + ''').literal(ns.bsfs.Colors).annotations, {}) + self.assertDictEqual(from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "1234"^^xsd:integer ; + rdfs:label "hello world"^^xsd:string ; + bsfs:foo "1234"^^xsd:integer . + + ''').literal(ns.bsfs.Colors).annotations, { + ns.rdfs.label: 'hello world', + ns.bsfs.foo: 1234, + }) + + + def test_integration(self): + # nodes + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + n_tag = types.ROOT_NODE.child(ns.bsfs.Tag) + n_image = n_ent.child(ns.bsfs.Image) + # literals + l_string = types.ROOT_LITERAL.child(ns.xsd.string) + l_array = types.ROOT_LITERAL.child(ns.bsfs.array) + l_integer = types.ROOT_NUMBER.child(ns.xsd.integer) + l_boolean = types.ROOT_LITERAL.child(ns.xsd.boolean) + # predicates + p_annotation = types.ROOT_PREDICATE.child(ns.bsfs.Annotation) + p_tag = types.ROOT_PREDICATE.child(ns.bse.tag, domain=n_ent, range=n_tag) + p_group = p_tag.child(ns.bse.group, domain=n_image, unique=True) + p_comment = p_annotation.child(ns.bse.comment, range=l_string) + # features + f_colors = types.ROOT_FEATURE.child(URI('http://bsfs.ai/schema/Feature/colors_spatial'), + dtype=ns.bsfs.f16, distance=ns.bsfs.euclidean) + f_colors1234 = f_colors.child(URI('http://bsfs.ai/schema/Feature/colors_spatial#1234'), dimension=1024) + f_colors4321 = f_colors.child(URI('http://bsfs.ai/schema/Feature/colors_spatial#4321'), dimension=2048) + # schema + ref = Schema( + {p_annotation, p_tag, p_group, p_comment}, + {n_ent, n_tag, n_image}, + {l_string, l_integer, l_boolean, f_colors, f_colors1234, f_colors4321}) + # load from string + gen = from_string(''' + # generic prefixes + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + + # bsfs prefixes + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + # nodes + bsfs:Entity rdfs:subClassOf bsfs:Node ; + rdfs:label "Principal node"^^xsd:string . + bsfs:Tag rdfs:subClassOf bsfs:Node ; + rdfs:label "Tag"^^xsd:string . + bsfs:Image rdfs:subClassOf bsfs:Entity . + + # literals + xsd:string rdfs:subClassOf bsfs:Literal ; + rdfs:label "A sequence of characters"^^xsd:string . + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array. + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . + xsd:boolean rdfs:subClassOf bsfs:Literal . + + + # abstract predicates + bsfs:Annotation rdfs:subClassOf bsfs:Predicate ; + rdfs:label "node annotation"^^xsd:string . + + # feature instances + <http://bsfs.ai/schema/Feature/colors_spatial> rdfs:subClassOf bsfs:Feature ; + bsfs:dtype bsfs:f16 ; + bsfs:distance bsfs:euclidean ; + # annotations + rdfs:label "ColorsSpatial instances. Dimension depends on instance."^^xsd:string ; + bsfs:first_arg "1234"^^xsd:integer ; + bsfs:second_arg "hello world"^^xsd:string . + + <http://bsfs.ai/schema/Feature/colors_spatial#1234> rdfs:subClassOf <http://bsfs.ai/schema/Feature/colors_spatial> ; + bsfs:dimension "1024"^^xsd:integer ; + rdfs:label "Main colors spatial instance"^^xsd:string . + + <http://bsfs.ai/schema/Feature/colors_spatial#4321> rdfs:subClassOf <http://bsfs.ai/schema/Feature/colors_spatial> ; + bsfs:dimension "2048"^^xsd:integer . + + # predicate instances + bse:tag rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range bsfs:Tag ; + bsfs:unique "false"^^xsd:boolean ; + # annotations + rdfs:label "connect entity to a tag"^^xsd:string . + + bse:group rdfs:subClassOf bse:tag ; # subtype of another predicate + rdfs:domain bsfs:Image ; + bsfs:unique "true"^^xsd:boolean . + + bse:comment rdfs:subClassOf bsfs:Annotation ; # subtype of abstract predicate + rdfs:domain bsfs:Node ; + rdfs:range xsd:string ; + bsfs:unique "false"^^xsd:boolean . + + ''') + # schemas are equal + self.assertEqual(ref, gen) + # check annotations + self.assertDictEqual(gen.node(ns.bsfs.Entity).annotations, {ns.rdfs.label: 'Principal node'}) + self.assertDictEqual(gen.node(ns.bsfs.Tag).annotations, {ns.rdfs.label: 'Tag'}) + self.assertDictEqual(gen.literal(ns.xsd.string).annotations, {ns.rdfs.label: 'A sequence of characters'}) + self.assertDictEqual(gen.predicate(ns.bsfs.Annotation).annotations, {ns.rdfs.label: 'node annotation'}) + self.assertDictEqual(gen.literal(URI('http://bsfs.ai/schema/Feature/colors_spatial')).annotations, { + ns.rdfs.label: 'ColorsSpatial instances. Dimension depends on instance.', + ns.bsfs.first_arg: 1234, + ns.bsfs.second_arg: 'hello world', + }) + self.assertDictEqual(gen.literal(URI('http://bsfs.ai/schema/Feature/colors_spatial#1234')).annotations, { + ns.rdfs.label: 'Main colors spatial instance'}) + self.assertDictEqual(gen.predicate(ns.bse.tag).annotations, {ns.rdfs.label: 'connect entity to a tag'}) + + # blank nodes result in an error + self.assertRaises(errors.BackendError, from_string, ''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix bsfs: <http://bsfs.ai/schema/> + bsfs:Entity rdfs:subClassOf bsfs:Node ; + bsfs:foo _:bar . + ''') + + +class TestToString(unittest.TestCase): + + def test_empty(self): + self.assertEqual(Schema(), from_string(to_string(Schema()))) + + def test_parse(self): + schema = Schema() + schema._nodes[ns.bsfs.Invalid] = 123 # NOTE: Access protected to force an invalid schema + self.assertRaises(TypeError, to_string, schema) + + def test_literal(self): + # root literals + l_str = types.ROOT_LITERAL.child(ns.xsd.string) + # derived literals + l_int = types.ROOT_NUMBER.child(ns.xsd.integer) + l_unsigned = l_int.child(ns.xsd.unsigned) + # create schema + schema = Schema(literals={l_int, l_str, l_unsigned}) + + schema_str = to_string(schema) + # all symbols are serialized + self.assertIn('xsd:string', schema_str) + self.assertIn('xsd:integer', schema_str) + self.assertIn('xsd:unsigned', schema_str) + # unserialize yields the original schema + self.assertEqual(schema, from_string(schema_str)) + + # literals that have no parent are ignored + schema = Schema(literals={types.Literal(ns.bsfs.Invalid, None)}) + self.assertEqual(Schema(), from_string(to_string(schema))) + self.assertNotIn('Invalid', to_string(schema)) + + # literal annotations are serialized + annotations = { + ns.rdfs.label: 'hello world', + ns.schema.description: 'some text', + ns.bsfs.foo: 1234, + ns.bsfs.bar: True, + } + l_str = types.ROOT_LITERAL.child(ns.xsd.string, **annotations) + self.assertDictEqual( + annotations, + from_string(to_string(Schema(literals={l_str}))).literal(ns.xsd.string).annotations) + + + def test_node(self): + # root nodes + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + n_tag = types.ROOT_NODE.child(ns.bsfs.Tag) + # derived nodes + n_img = n_ent.child(ns.bsfs.Image) + n_doc = n_ent.child(ns.bsfs.Document) + n_grp = n_tag.child(ns.bsfs.Group) + # create schema + schema = Schema(nodes={n_ent, n_img, n_doc, n_tag, n_grp}) + + schema_str = to_string(schema) + # all symbols are serialized + self.assertIn('bsfs:Entity', schema_str) + self.assertIn('bsfs:Tag', schema_str) + self.assertIn('bsfs:Image', schema_str) + self.assertIn('bsfs:Document', schema_str) + self.assertIn('bsfs:Group', schema_str) + # unserialize yields the original schema + self.assertEqual(schema, from_string(schema_str)) + + # nodes that have no parent are ignored + schema = Schema(nodes={types.Node(ns.bsfs.Invalid, None)}) + self.assertEqual(Schema(), from_string(to_string(schema))) + self.assertNotIn('Invalid', to_string(schema)) + + # node annotations are serialized + annotations = { + ns.rdfs.label: 'hello world', + ns.schema.description: 'some text', + ns.bsfs.foo: 1234, + ns.bsfs.bar: True, + } + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity, **annotations) + self.assertDictEqual( + annotations, + from_string(to_string(Schema(nodes={n_ent}))).node(ns.bsfs.Entity).annotations) + + + def test_predicate(self): + # auxiliary types + n_ent = types.ROOT_NODE.child(ns.bsfs.Entity) + l_str = types.ROOT_LITERAL.child(ns.xsd.string) + # root predicates + p_annotation = types.ROOT_PREDICATE.child(ns.bsfs.Annotation, domain=n_ent) + p_owner = types.ROOT_PREDICATE.child(ns.bse.owner, range=l_str, unique=True) + # derived predicates + p_comment = p_annotation.child(ns.bse.comment, range=l_str) # inherits domain + p_note = p_comment.child(ns.bse.note, unique=True) # inherits domain/range + # create schema + schema = Schema({p_owner, p_comment, p_note}) + + schema_str = to_string(schema) + # all symbols are serialized + self.assertIn('bsfs:Entity', schema_str) + self.assertIn('xsd:string', schema_str) + self.assertIn('bsfs:Annotation', schema_str) + self.assertIn('bse:comment', schema_str) + self.assertIn('bse:owner', schema_str) + self.assertIn('bse:note', schema_str) + # inherited properties are not serialized + self.assertIsNotNone(re.search(r'bse:comment[^\.]*rdfs:range[^\.]', schema_str)) + self.assertIsNone(re.search(r'bse:comment[^\.]*rdfs:domain[^\.]', schema_str)) + #p_note has no domain/range + self.assertIsNone(re.search(r'bse:note[^\.]*rdfs:domain[^\.]', schema_str)) + self.assertIsNone(re.search(r'bse:note[^\.]*rdfs:range[^\.]', schema_str)) + # unserialize yields the original schema + self.assertEqual(schema, from_string(schema_str)) + + # predicate annotations are serialized + annotations = { + ns.rdfs.label: 'hello world', + ns.schema.description: 'some text', + ns.bsfs.foo: 1234, + ns.bsfs.bar: False, + } + p_annotation = types.ROOT_PREDICATE.child(ns.bsfs.Annotation, **annotations) + self.assertDictEqual( + annotations, + from_string(to_string(Schema({p_annotation}))).predicate(ns.bsfs.Annotation).annotations) + + + def test_feature(self): + # root features + f_colors = types.ROOT_FEATURE.child(URI('http://bsfs.ai/schema/Feature/colors'), + distance=ns.bsfs.cosine) + # derived features + f_colors1234 = f_colors.child(URI('http://bsfs.ai/schema/Feature/colors#1234'), + dimension=1024) # inherits dtype, distance + f_colors4321 = f_colors.child(URI('http://bsfs.ai/schema/Feature/colors#4321'), + dimension=2048, distance=ns.bsfs.euclidean) # inherits dtype + # create schema + schema = Schema(literals={f_colors, f_colors1234, f_colors4321}) + + schema_str = to_string(schema) + # all symbols are serialized + self.assertIn('bsfs:Array', schema_str) + self.assertIn('<http://bsfs.ai/schema/Feature/colors', schema_str) + self.assertIn('<http://bsfs.ai/schema/Feature/colors#1234', schema_str) + self.assertIn('<http://bsfs.ai/schema/Feature/colors#4321', schema_str) + # inherited properties are not serialized + self.assertIsNotNone(re.search(r'<http://bsfs\.ai/schema/Feature/colors#1234>[^\.]*bsfs:dimension[^\.]', schema_str)) + self.assertIsNone(re.search(r'<http://bsfs\.ai/schema/Feature/colors#1234>[^\.]*bsfs:dtype[^\.]', schema_str)) + self.assertIsNone(re.search(r'<http://bsfs\.ai/schema/Feature/colors#1234>[^\.]*bsfs:distance[^\.]', schema_str)) + self.assertIsNotNone(re.search(r'<http://bsfs\.ai/schema/Feature/colors#4321>[^\.]*bsfs:dimension[^\.]', schema_str)) + self.assertIsNotNone(re.search(r'<http://bsfs\.ai/schema/Feature/colors#4321>[^\.]*bsfs:distance[^\.]', schema_str)) + self.assertIsNone(re.search(r'<http://bsfs\.ai/schema/Feature/colors#4321>[^\.]*bsfs:dtype[^\.]', schema_str)) + # unserialize yields the original schema + self.assertEqual(schema, from_string(schema_str)) + + # predicate annotations are serialized + annotations = { + ns.rdfs.label: 'hello world', + ns.schema.description: 'some text', + ns.bsfs.foo: 1234, + ns.bsfs.bar: False, + } + f_colors = types.ROOT_FEATURE.child(URI('http://bsfs.ai/schema/Feature/colors'), + dtype=ns.bsfs.f16, distance=ns.bsfs.euclidean, + **annotations) + self.assertDictEqual( + annotations, + from_string(to_string(Schema(literals={f_colors}))).literal(URI('http://bsfs.ai/schema/Feature/colors')).annotations) + + +## main ## + +if __name__ == '__main__': + unittest.main() + +## EOF ## diff --git a/test/schema/test_types.py b/test/schema/test_types.py index 4a49e6e..c5895d2 100644 --- a/test/schema/test_types.py +++ b/test/schema/test_types.py @@ -10,15 +10,17 @@ import unittest # bsfs imports from bsfs.namespace import ns +from bsfs.schema.types import ROOT_PREDICATE, ROOT_VERTEX, ROOT_FEATURE from bsfs.utils import errors # objects to test -from bsfs.schema.types import _Type, _Vertex, Node, Literal, Predicate +from bsfs.schema.types import _Type, Vertex, Node, Literal, Predicate, Feature ## code ## class TestType(unittest.TestCase): + def test_parents(self): # create some types fst = _Type('First') @@ -31,7 +33,25 @@ class TestType(unittest.TestCase): self.assertListEqual(list(trd.parents()), [snd, fst]) self.assertListEqual(list(frd.parents()), [trd, snd, fst]) - def test_essentials(self): + def test_annotations(self): + # annotations can be empty + self.assertDictEqual(_Type('Foo', None).annotations, {}) + # annotations are stored + self.assertDictEqual(_Type('Foo', None, foo='bar', bar=123).annotations, { + 'foo': 'bar', + 'bar': 123}) + # comparison ignores annotations + self.assertEqual( + _Type('Foo', None, foo='bar', bar='foo'), + _Type('Foo', None, hello='world', foobar=1234)) + self.assertEqual( + hash(_Type('Foo', None, foo='bar', bar='foo')), + hash(_Type('Foo', None, hello='world', foobar=1234))) + # annotations can be passed to child + self.assertDictEqual(_Type('First', foo='bar').child('Second', bar='foo').annotations, { + 'bar': 'foo'}) + + def test_string_conversion(self): # type w/o parent self.assertEqual(str(_Type('Foo')), '_Type(Foo)') self.assertEqual(repr(_Type('Foo')), '_Type(Foo, None)') @@ -51,14 +71,17 @@ class TestType(unittest.TestCase): self.assertEqual(str(_Type('Foo', SubType('Bar'))), '_Type(Foo)') self.assertEqual(repr(_Type('Foo', SubType('Bar'))), '_Type(Foo, SubType(Bar, None))') - def test_get_child(self): + def test_child(self): # callee is used as parent - self.assertEqual(_Type('First').get_child('Second'), _Type('Second', _Type('First'))) + self.assertEqual(_Type('First').child('Second'), _Type('Second', _Type('First'))) # works with multiple parents - self.assertEqual(_Type('First').get_child('Second').get_child('Third'), _Type('Third', _Type('Second', _Type('First')))) + self.assertEqual(_Type('First').child('Second').child('Third'), _Type('Third', _Type('Second', _Type('First')))) # type persists class Foo(_Type): pass - self.assertEqual(Foo('First').get_child('Second'), Foo('Second', Foo('First'))) + self.assertEqual(Foo('First').child('Second'), Foo('Second', Foo('First'))) + # annotations are respected + self.assertDictEqual(_Type('First', foo='bar').child('Second', bar='foo').annotations, { + 'bar': 'foo'}) def test_equality(self): # equality depends on uri @@ -76,6 +99,13 @@ class TestType(unittest.TestCase): # comparison respects parent self.assertNotEqual(_Type('Foo', _Type('Bar')), _Type('Foo')) self.assertNotEqual(hash(_Type('Foo', _Type('Bar'))), hash(_Type('Foo'))) + # comparison ignores annotations + self.assertEqual( + _Type('Foo', None, foo='bar', bar='foo'), + _Type('Foo', None, hello='world', foobar=1234)) + self.assertEqual( + hash(_Type('Foo', None, foo='bar', bar='foo')), + hash(_Type('Foo', None, hello='world', foobar=1234))) def test_order(self): # create some types. @@ -109,27 +139,43 @@ class TestType(unittest.TestCase): self.assertFalse(bike > bicycle) self.assertFalse(bike >= bicycle) self.assertFalse(bike == bicycle) + + # comparing different classes returns False ... + # ... when classes are hierarchically related class Foo(_Type): pass - foo = Foo(bike.uri, bike.parent) - # cannot compare different types - self.assertRaises(TypeError, operator.lt, foo, bike) - self.assertRaises(TypeError, operator.le, foo, bike) - self.assertRaises(TypeError, operator.gt, foo, bike) - self.assertRaises(TypeError, operator.ge, foo, bike) + foo = Foo('Foo', bike) + self.assertFalse(foo < bike) + self.assertFalse(foo <= bike) + self.assertFalse(foo > bike) + self.assertFalse(foo >= bike) + # goes both ways + self.assertFalse(bike < foo) + self.assertFalse(bike <= foo) + self.assertFalse(bike > foo) + self.assertFalse(bike >= foo) + # ... when classes are unrelated + class Bar(_Type): pass + bar = Bar('Bar', bike) + self.assertFalse(foo < bar) + self.assertFalse(foo <= bar) + self.assertFalse(foo > bar) + self.assertFalse(foo >= bar) # goes both ways - self.assertRaises(TypeError, operator.lt, bike, foo) - self.assertRaises(TypeError, operator.le, bike, foo) - self.assertRaises(TypeError, operator.gt, bike, foo) - self.assertRaises(TypeError, operator.ge, bike, foo) + self.assertFalse(bar < foo) + self.assertFalse(bar <= foo) + self.assertFalse(bar > foo) + self.assertFalse(bar >= foo) + class TestPredicate(unittest.TestCase): def test_construction(self): # domain must be a node self.assertRaises(TypeError, Predicate, ns.bse.foo, 1234, None, True) self.assertRaises(TypeError, Predicate, ns.bse.foo, None, Literal(ns.bsfs.Foo, None), None, True) - # range must be None, a Literal, or a Node + # range must be a Literal, a Node, or the root Vertex + self.assertRaises(TypeError, Predicate, ns.bse.foo, None, Node(ns.bsfs.Node, None), None, True) self.assertRaises(TypeError, Predicate, ns.bse.foo, None, Node(ns.bsfs.Node, None), 1234, True) - self.assertRaises(TypeError, Predicate, ns.bse.foo, None, Node(ns.bsfs.Node, None), _Vertex(ns.bsfs.Foo, None), True) + self.assertRaises(TypeError, Predicate, ns.bse.foo, None, Node(ns.bsfs.Node, None), Vertex(ns.bsfs.Foo, None), True) self.assertRaises(TypeError, Predicate, ns.bse.foo, None, Node(ns.bsfs.Node, None), _Type(ns.bsfs.Foo, None), True) class Foo(): pass self.assertRaises(TypeError, Predicate, ns.bse.foo, None, Node(ns.bsfs.Node, None), Foo(), True) @@ -138,82 +184,160 @@ class TestPredicate(unittest.TestCase): n_root = Node(ns.bsfs.Node, None) n_ent = Node(ns.bsfs.Entity, Node(ns.bsfs.Node, None)) n_tag = Node(ns.bsfs.Tag, Node(ns.bsfs.Tag, None)) - root = Predicate( - uri=ns.bsfs.Predicate, - parent=None, + root = ROOT_PREDICATE + tag = Predicate( + uri=ns.bse.tag, + parent=root, domain=n_root, - range=None, + range=n_tag, unique=False, ) # instance is equal to itself - self.assertEqual(root, root) - self.assertEqual(hash(root), hash(root)) + self.assertEqual(tag, tag) + self.assertEqual(hash(tag), hash(tag)) # instance is equal to a clone - self.assertEqual(root, Predicate(ns.bsfs.Predicate, None, n_root, None, False)) - self.assertEqual(hash(root), hash(Predicate(ns.bsfs.Predicate, None, n_root, None, False))) + self.assertEqual(tag, Predicate(ns.bse.tag, root, n_root, n_tag, False)) + self.assertEqual(hash(tag), hash(Predicate(ns.bse.tag, root, n_root, n_tag, False))) # equality respects uri - self.assertNotEqual(root, Predicate(ns.bsfs.Alternative, None, n_root, None, False)) - self.assertNotEqual(hash(root), hash(Predicate(ns.bsfs.Alternative, None, n_root, None, False))) + self.assertNotEqual(tag, Predicate(ns.bsfs.Alternative, root, n_root, n_tag, False)) + self.assertNotEqual(hash(tag), hash(Predicate(ns.bsfs.Alternative, root, n_root, n_tag, False))) # equality respects parent - self.assertNotEqual(root, Predicate(ns.bsfs.Predicate, n_root, n_root, None, False)) - self.assertNotEqual(hash(root), hash(Predicate(ns.bsfs.Predicate, n_root, n_root, None, False))) + self.assertNotEqual(tag, Predicate(ns.bse.tag, n_root, n_root, n_tag, False)) + self.assertNotEqual(hash(tag), hash(Predicate(ns.bse.tag, n_root, n_root, n_tag, False))) # equality respects domain - self.assertNotEqual(root, Predicate(ns.bsfs.Predicate, None, n_ent, None, False)) - self.assertNotEqual(hash(root), hash(Predicate(ns.bsfs.Predicate, None, n_ent, None, False))) + self.assertNotEqual(tag, Predicate(ns.bse.tag, root, n_ent, n_tag, False)) + self.assertNotEqual(hash(tag), hash(Predicate(ns.bse.tag, root, n_ent, n_tag, False))) # equality respects range - self.assertNotEqual(root, Predicate(ns.bsfs.Predicate, None, n_root, n_root, False)) - self.assertNotEqual(hash(root), hash(Predicate(ns.bsfs.Predicate, None, n_root, n_root, False))) + self.assertNotEqual(tag, Predicate(ns.bse.tag, root, n_root, n_root, False)) + self.assertNotEqual(hash(tag), hash(Predicate(ns.bse.tag, root, n_root, n_root, False))) # equality respects unique - self.assertNotEqual(root, Predicate(ns.bsfs.Predicate, None, n_root, None, True)) - self.assertNotEqual(hash(root), hash(Predicate(ns.bsfs.Predicate, None, n_root, None, True))) + self.assertNotEqual(tag, Predicate(ns.bse.tag, root, n_root, n_tag, True)) + self.assertNotEqual(hash(tag), hash(Predicate(ns.bse.tag, root, n_root, n_tag, True))) - def test_get_child(self): + def test_child(self): n_root = Node(ns.bsfs.Node, None) + l_root = Literal(ns.bsfs.Literal, None) n_ent = Node(ns.bsfs.Entity, Node(ns.bsfs.Node, None)) n_tag = Node(ns.bsfs.Tag, Node(ns.bsfs.Tag, None)) - root = Predicate( - uri=ns.bsfs.Predicate, - parent=None, - domain=n_root, - range=None, - unique=False, - ) + root = ROOT_PREDICATE tag = Predicate( - uri=ns.bsfs.Entity, + uri=ns.bse.tag, parent=root, domain=n_ent, range=n_tag, unique=False, ) + # child returns Predicate + self.assertIsInstance(tag.child(ns.bse.foo), Predicate) # uri is respected - self.assertEqual(ns.bse.foo, tag.get_child(ns.bse.foo).uri) + self.assertEqual(ns.bse.foo, tag.child(ns.bse.foo).uri) # domain is respected dom = Node(ns.bsfs.Image, n_ent) - self.assertEqual(dom, tag.get_child(ns.bse.foo, domain=dom).domain) + self.assertEqual(dom, tag.child(ns.bse.foo, domain=dom).domain) # range is respected rng = Node(ns.bsfs.Group, n_tag) - self.assertEqual(rng, tag.get_child(ns.bse.foo, range=rng).range) + self.assertEqual(rng, tag.child(ns.bse.foo, range=rng).range) # cannot set range to None - self.assertEqual(n_tag, tag.get_child(ns.bse.foo, range=None).range) + self.assertEqual(n_tag, tag.child(ns.bse.foo, range=None).range) # unique is respected - self.assertTrue(tag.get_child(ns.bse.foo, unique=True).unique) + self.assertTrue(tag.child(ns.bse.foo, unique=True).unique) + # annotations are respected + self.assertDictEqual(tag.child(ns.bse.foo, foo='bar', bar=123).annotations, { + 'foo': 'bar', + 'bar': 123, + }) # domain is inherited from parent - self.assertEqual(n_ent, tag.get_child(ns.bse.foo).domain) + self.assertEqual(n_root, root.child(ns.bse.foo).domain) + self.assertEqual(n_ent, tag.child(ns.bse.foo).domain) # range is inherited from parent - self.assertEqual(n_tag, tag.get_child(ns.bse.foo).range) + self.assertEqual(ROOT_VERTEX, root.child(ns.bse.foo).range) + self.assertEqual(n_tag, tag.child(ns.bse.foo).range) # uniqueness is inherited from parent - self.assertFalse(tag.get_child(ns.bse.foo).unique) + self.assertFalse(tag.child(ns.bse.foo).unique) # domain must be subtype of parent's domain - self.assertRaises(errors.ConsistencyError, tag.get_child, ns.bse.foo, domain=n_root) - self.assertRaises(errors.ConsistencyError, tag.get_child, ns.bse.foo, domain=Node(ns.bsfs.Image, n_root)) - # range cannot be None - self.assertRaises(ValueError, root.get_child, ns.bse.foo) + self.assertRaises(errors.ConsistencyError, tag.child, ns.bse.foo, domain=n_root) + self.assertRaises(errors.ConsistencyError, tag.child, ns.bse.foo, domain=Node(ns.bsfs.Image, n_root)) # range must be subtype of parent's range - self.assertRaises(errors.ConsistencyError, tag.get_child, ns.bse.foo, range=n_root) - self.assertRaises(errors.ConsistencyError, tag.get_child, ns.bse.foo, range=Node(ns.bsfs.Image, n_root)) + self.assertRaises(errors.ConsistencyError, tag.child, ns.bse.foo, range=n_root) + self.assertRaises(errors.ConsistencyError, tag.child, ns.bse.foo, range=Node(ns.bsfs.Image, n_root)) + self.assertRaises(errors.ConsistencyError, tag.child, ns.bse.foo, range=Literal(ns.bsfs.Tag, l_root)) + # range can be subtyped from ROOT_VERTEX to Node or Literal + self.assertEqual(n_root, root.child(ns.bse.foo, range=n_root).range) + self.assertEqual(l_root, root.child(ns.bse.foo, range=l_root).range) + + +class TestFeature(unittest.TestCase): + def test_construction(self): + n_root = Node(ns.bsfs.Node, None) + l_root = Literal(ns.bsfs.Literal, None) + # dimension, dtype, and distance are respected + feat = Feature(ns.bsfs.Feature, None, 1234, ns.bsfs.float, ns.bsfs.euclidean) + self.assertEqual(1234, feat.dimension) + self.assertEqual(ns.bsfs.float, feat.dtype) + self.assertEqual(ns.bsfs.euclidean, feat.distance) + + def test_equality(self): + n_ent = Node(ns.bsfs.Entity, Node(ns.bsfs.Node, None)) + colors = Feature( + uri=ns.bse.colors, + parent=ROOT_FEATURE, + dimension=1234, + dtype=ns.bsfs.float, + distance=ns.bsfs.euclidean, + ) + # instance is equal to itself + self.assertEqual(colors, colors) + self.assertEqual(hash(colors), hash(colors)) + # instance is equal to a clone + self.assertEqual(colors, Feature(ns.bse.colors, ROOT_FEATURE, 1234, ns.bsfs.float, ns.bsfs.euclidean)) + self.assertEqual(hash(colors), hash(Feature(ns.bse.colors, ROOT_FEATURE, 1234, ns.bsfs.float, ns.bsfs.euclidean))) + # equality respects dimension + self.assertNotEqual(colors, Feature(ns.bse.colors, ROOT_FEATURE, 4321, ns.bsfs.float, ns.bsfs.euclidean)) + self.assertNotEqual(hash(colors), hash(Feature(ns.bse.colors, ROOT_FEATURE, 4321, ns.bsfs.float, ns.bsfs.euclidean))) + # equality respects dtype + self.assertNotEqual(colors, Feature(ns.bse.colors, ROOT_FEATURE, 1234, ns.bsfs.integer, ns.bsfs.euclidean)) + self.assertNotEqual(hash(colors), hash(Feature(ns.bse.colors, ROOT_FEATURE, 1234, ns.bsfs.integer, ns.bsfs.euclidean))) + # equality respects distance + self.assertNotEqual(colors, Feature(ns.bse.colors, ROOT_FEATURE, 1234, ns.bsfs.float, ns.bsfs.cosine)) + self.assertNotEqual(hash(colors), hash(Feature(ns.bse.colors, ROOT_FEATURE, 1234, ns.bsfs.float, ns.bsfs.cosine))) + + def test_child(self): + n_root = Node(ns.bsfs.Node, None) + n_ent = Node(ns.bsfs.Entity, n_root) + l_root = Literal(ns.bsfs.Literal, None) + colors = Feature( + uri=ns.bse.colors, + parent=ROOT_FEATURE, + dimension=1234, + dtype=ns.bsfs.float, + distance=ns.bsfs.euclidean, + ) + + # child returns Feature + self.assertIsInstance(colors.child(ns.bse.foo), Feature) + # uri is respected + self.assertEqual(ns.bse.foo, colors.child(ns.bse.foo).uri) + # dimension is respected + self.assertEqual(4321, colors.child(ns.bse.foo, dimension=4321).dimension) + # dtype is respected + self.assertEqual(ns.bsfs.integer, colors.child(ns.bse.foo, dtype=ns.bsfs.integer).dtype) + # distance is respected + self.assertEqual(ns.bsfs.cosine, colors.child(ns.bse.foo, distance=ns.bsfs.cosine).distance) + # annotations are respected + self.assertDictEqual(colors.child(ns.bse.foo, foo='bar', bar=123).annotations, { + 'foo': 'bar', + 'bar': 123, + }) + + # dimension is inherited from parent + self.assertEqual(1234, colors.child(ns.bse.foo).dimension) + # dtype is inherited from parent + self.assertEqual(ns.bsfs.float, colors.child(ns.bse.foo).dtype) + # distance is inherited from parent + self.assertEqual(ns.bsfs.euclidean, colors.child(ns.bse.foo).distance) ## main ## @@ -222,4 +346,3 @@ if __name__ == '__main__': unittest.main() ## EOF ## - diff --git a/test/triple_store/sparql/test_distance.py b/test/triple_store/sparql/test_distance.py new file mode 100644 index 0000000..0659459 --- /dev/null +++ b/test/triple_store/sparql/test_distance.py @@ -0,0 +1,61 @@ +""" + +Part of the bsfs test suite. +A copy of the license is provided with the project. +Author: Matthias Baumgartner, 2022 +""" +# imports +import numpy as np +import unittest + +# objects to test +from bsfs.triple_store.sparql import distance + + +## code ## + +class TestDistance(unittest.TestCase): + + def test_euclid(self): + # self-distance is zero + self.assertEqual(distance.euclid([1,2,3,4], [1,2,3,4]), 0.0) + # accepts list-like arguments + self.assertAlmostEqual(distance.euclid([1,2,3,4], [2,3,4,5]), 2.0, 3) + self.assertAlmostEqual(distance.euclid((1,2,3,4), (2,3,4,5)), 2.0, 3) + # dimension can vary + self.assertAlmostEqual(distance.euclid([1,2,3], [2,3,4]), 1.732, 3) + self.assertAlmostEqual(distance.euclid([1,2,3,4,5], [2,3,4,5,6]), 2.236, 3) + # vector can be zero + self.assertAlmostEqual(distance.euclid([0,0,0], [1,2,3]), 3.742, 3) + + def test_cosine(self): + # self-distance is zero + self.assertEqual(distance.cosine([1,2,3,4], [1,2,3,4]), 0.0) + # accepts list-like arguments + self.assertAlmostEqual(distance.cosine([1,2,3,4], [4,3,2,1]), 0.333, 3) + self.assertAlmostEqual(distance.cosine((1,2,3,4), (4,3,2,1)), 0.333, 3) + # dimension can vary + self.assertAlmostEqual(distance.cosine([1,2,3], [3,2,1]), 0.286, 3) + self.assertAlmostEqual(distance.cosine([1,2,3,4,5], [5,4,3,2,1]), 0.364, 3) + # vector can be zero + self.assertAlmostEqual(distance.cosine([0,0,0], [1,2,3]), 1.0, 3) + + def test_manhatten(self): + # self-distance is zero + self.assertEqual(distance.manhatten([1,2,3,4], [1,2,3,4]), 0.0) + # accepts list-like arguments + self.assertAlmostEqual(distance.manhatten([1,2,3,4], [2,3,4,5]), 4.0, 3) + self.assertAlmostEqual(distance.manhatten((1,2,3,4), (2,3,4,5)), 4.0, 3) + # dimension can vary + self.assertAlmostEqual(distance.manhatten([1,2,3], [2,3,4]), 3.0, 3) + self.assertAlmostEqual(distance.manhatten([1,2,3,4,5], [2,3,4,5,6]), 5.0, 3) + # vector can be zero + self.assertAlmostEqual(distance.manhatten([0,0,0], [1,2,3]), 6.0, 3) + + +## main ## + +if __name__ == '__main__': + unittest.main() + +## EOF ## diff --git a/test/triple_store/sparql/test_parse_filter.py b/test/triple_store/sparql/test_parse_filter.py index bd19803..8764535 100644 --- a/test/triple_store/sparql/test_parse_filter.py +++ b/test/triple_store/sparql/test_parse_filter.py @@ -9,7 +9,7 @@ import rdflib import unittest # bsie imports -from bsfs import schema as _schema +from bsfs import schema as bsc from bsfs.namespace import ns from bsfs.query import ast from bsfs.utils import errors @@ -23,21 +23,34 @@ from bsfs.triple_store.sparql.parse_filter import Filter class TestParseFilter(unittest.TestCase): def setUp(self): # schema - self.schema = _schema.Schema.from_string(''' + self.schema = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> prefix bse: <http://bsfs.ai/schema/Entity#> + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array . + bsfs:Number rdfs:subClassOf bsfs:Literal . + bsfs:Entity rdfs:subClassOf bsfs:Node . bsfs:Image rdfs:subClassOf bsfs:Entity . bsfs:Tag rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . bsfs:URI rdfs:subClassOf bsfs:Literal . + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "4"^^xsd:integer ; + bsfs:dtype xsd:integer ; + bsfs:distance bsfs:euclidean . + + bse:colors rdfs:subClassOf bsfs:Predicate ; + rdfs:domain bsfs:Entity ; + rdfs:range bsfs:Colors . + bse:comment rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Node ; rdfs:range xsd:string ; @@ -70,9 +83,6 @@ class TestParseFilter(unittest.TestCase): ''') - # parser instance - self.parser = Filter(self.schema) - # graph to test queries self.graph = rdflib.Graph() # schema hierarchies @@ -113,6 +123,13 @@ class TestParseFilter(unittest.TestCase): # image iso self.graph.add((rdflib.URIRef('http://example.com/image#1234'), rdflib.URIRef(ns.bse.iso), rdflib.Literal(1234, datatype=rdflib.XSD.integer))) self.graph.add((rdflib.URIRef('http://example.com/image#4321'), rdflib.URIRef(ns.bse.iso), rdflib.Literal(4321, datatype=rdflib.XSD.integer))) + # color features + self.graph.add((rdflib.URIRef('http://example.com/entity#1234'), rdflib.URIRef(ns.bse.colors), rdflib.Literal([1,2,3,4], datatype=rdflib.URIRef(ns.bsfs.Colors)))) + self.graph.add((rdflib.URIRef('http://example.com/entity#4321'), rdflib.URIRef(ns.bse.colors), rdflib.Literal([4,3,2,1], datatype=rdflib.URIRef(ns.bsfs.Colors)))) + self.graph.add((rdflib.URIRef('http://example.com/image#1234'), rdflib.URIRef(ns.bse.colors), rdflib.Literal([3,4,2,1], datatype=rdflib.URIRef(ns.bsfs.Colors)))) + + # parser instance + self.parser = Filter(self.graph, self.schema) def test_routing(self): @@ -124,7 +141,7 @@ class TestParseFilter(unittest.TestCase): # __call__ requires a valid root type self.assertRaises(errors.BackendError, self.parser, self.schema.literal(ns.bsfs.Literal), None) - self.assertRaises(errors.ConsistencyError, self.parser, self.schema.node(ns.bsfs.Node).get_child(ns.bsfs.Invalid), None) + self.assertRaises(errors.ConsistencyError, self.parser, self.schema.node(ns.bsfs.Node).child(ns.bsfs.Invalid), None) # __call__ requires a parseable root self.assertRaises(errors.BackendError, self.parser, self.schema.node(ns.bsfs.Entity), ast.filter.FilterExpression()) # __call__ returns an executable query @@ -613,6 +630,37 @@ class TestParseFilter(unittest.TestCase): {'http://example.com/tag#1234'}) + def test_distance(self): + # node colors distance to [2,4,3,1] + # entity#1234 [1,2,3,4] 3.742 + # entity#4321 [4,3,2,1] 2.449 + # image#1234 [3,4,2,1] 1.414 + + # _distance expects a feature + self.assertRaises(errors.BackendError, self.parser._distance, self.schema.node(ns.bsfs.Entity), ast.filter.Distance([1,2,3,4], 1), '') + # reference must have the correct dimension + self.assertRaises(errors.ConsistencyError, self.parser._distance, self.schema.literal(ns.bsfs.Colors), ast.filter.Distance([1,2,3], 1), '') + self.assertRaises(errors.ConsistencyError, self.parser._distance, self.schema.literal(ns.bsfs.Colors), ast.filter.Distance([1,2,3,4,5], 1), '') + # _distance respects threshold + q = self.parser(self.schema.node(ns.bsfs.Entity), ast.filter.Any(ns.bse.colors, ast.filter.Distance([2,4,3,1], 4))) + self.assertSetEqual({str(guid) for guid, in self.graph.query(q)}, + {'http://example.com/entity#1234', 'http://example.com/entity#4321', 'http://example.com/image#1234'}) + q = self.parser(self.schema.node(ns.bsfs.Entity), ast.filter.Any(ns.bse.colors, ast.filter.Distance([2,4,3,1], 3))) + self.assertSetEqual({str(guid) for guid, in self.graph.query(q)}, + {'http://example.com/entity#4321', 'http://example.com/image#1234'}) + q = self.parser(self.schema.node(ns.bsfs.Entity), ast.filter.Any(ns.bse.colors, ast.filter.Distance([2,4,3,1], 2))) + self.assertSetEqual({str(guid) for guid, in self.graph.query(q)}, + {'http://example.com/image#1234'}) + # result set can be empty + q = self.parser(self.schema.node(ns.bsfs.Entity), ast.filter.Any(ns.bse.colors, ast.filter.Distance([2,4,3,1], 1))) + self.assertSetEqual({str(guid) for guid, in self.graph.query(q)}, set()) + # _distance respects strict + q = self.parser(self.schema.node(ns.bsfs.Entity), ast.filter.Any(ns.bse.colors, ast.filter.Distance([1,2,3,4], 0, False))) + self.assertSetEqual({str(guid) for guid, in self.graph.query(q)}, + {'http://example.com/entity#1234'}) + q = self.parser(self.schema.node(ns.bsfs.Entity), ast.filter.Any(ns.bse.colors, ast.filter.Distance([1,2,3,4], 0, True))) + self.assertSetEqual({str(guid) for guid, in self.graph.query(q)}, set()) + def test_one_of(self): # _one_of expects a node self.assertRaises(errors.BackendError, self.parser._one_of, diff --git a/test/triple_store/sparql/test_sparql.py b/test/triple_store/sparql/test_sparql.py index 3d81de1..7fbfb65 100644 --- a/test/triple_store/sparql/test_sparql.py +++ b/test/triple_store/sparql/test_sparql.py @@ -9,7 +9,7 @@ import rdflib import unittest # bsie imports -from bsfs import schema as _schema +from bsfs import schema as bsc from bsfs.namespace import ns from bsfs.query import ast from bsfs.utils import errors, URI @@ -22,7 +22,7 @@ from bsfs.triple_store.sparql.sparql import SparqlStore class TestSparqlStore(unittest.TestCase): def setUp(self): - self.schema = _schema.Schema.from_string(''' + self.schema = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> @@ -33,7 +33,8 @@ class TestSparqlStore(unittest.TestCase): bsfs:Tag rdfs:subClassOf bsfs:Node . bsfs:User rdfs:subClassOf bsfs:Node . xsd:string rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . # non-unique literal bse:comment rdfs:subClassOf bsfs:Predicate ; @@ -66,7 +67,11 @@ class TestSparqlStore(unittest.TestCase): (rdflib.URIRef(ns.bsfs.Tag), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Node)), (rdflib.URIRef(ns.bsfs.User), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Node)), (rdflib.URIRef(ns.xsd.string), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), - (rdflib.URIRef(ns.xsd.integer), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Array), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Feature), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Array)), + (rdflib.URIRef(ns.bsfs.Number), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Time), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.xsd.integer), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Number)), (rdflib.URIRef(ns.bse.comment), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), (rdflib.URIRef(ns.bse.filesize), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), (rdflib.URIRef(ns.bse.tag), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), @@ -90,7 +95,7 @@ class TestSparqlStore(unittest.TestCase): def test__has_type(self): # setup store store = SparqlStore.Open() - store.schema = _schema.Schema.from_string(''' + store.schema = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> @@ -108,7 +113,7 @@ class TestSparqlStore(unittest.TestCase): store.create(store.schema.node(ns.bsfs.PDF), {URI('http://example.com/me/pdf#1234')}) # node_type must be in the schema - self.assertRaises(errors.ConsistencyError, store._has_type, URI('http://example.com/me/entity#1234'), store.schema.node(ns.bsfs.Node).get_child(ns.bsfs.invalid)) + self.assertRaises(errors.ConsistencyError, store._has_type, URI('http://example.com/me/entity#1234'), store.schema.node(ns.bsfs.Node).child(ns.bsfs.invalid)) # returns False on inexistent nodes self.assertFalse(store._has_type(URI('http://example.com/me/entity#4321'), store.schema.node(ns.bsfs.Entity))) @@ -195,7 +200,7 @@ class TestSparqlStore(unittest.TestCase): self.assertSetEqual(set(store._graph), instances) # add some classes to the schema - curr = curr + _schema.Schema.from_string(''' + curr = curr + bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> @@ -298,7 +303,7 @@ class TestSparqlStore(unittest.TestCase): # remove some classes from the schema - curr = _schema.Schema.from_string(''' + curr = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> @@ -310,7 +315,8 @@ class TestSparqlStore(unittest.TestCase): bsfs:User rdfs:subClassOf bsfs:Node . xsd:boolean rdfs:subClassOf bsfs:Literal . - xsd:integer rdfs:subClassOf bsfs:Literal . + bsfs:Number rdfs:subClassOf bsfs:Literal . + xsd:integer rdfs:subClassOf bsfs:Number . bse:filesize rdfs:subClassOf bsfs:Predicate ; rdfs:domain bsfs:Entity ; @@ -351,7 +357,11 @@ class TestSparqlStore(unittest.TestCase): (rdflib.URIRef(ns.bsfs.Tag), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Node)), (rdflib.URIRef(ns.bsfs.User), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Node)), (rdflib.URIRef(ns.xsd.boolean), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), - (rdflib.URIRef(ns.xsd.integer), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Array), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Feature), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Array)), + (rdflib.URIRef(ns.bsfs.Number), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.bsfs.Time), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Literal)), + (rdflib.URIRef(ns.xsd.integer), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Number)), (rdflib.URIRef(ns.bse.shared), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), (rdflib.URIRef(ns.bse.tag), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), (rdflib.URIRef(ns.bse.filesize), rdflib.RDFS.subClassOf, rdflib.URIRef(ns.bsfs.Predicate)), @@ -382,8 +392,25 @@ class TestSparqlStore(unittest.TestCase): class Foo(): pass self.assertRaises(TypeError, setattr, store, 'schema', Foo()) + # cannot define features w/o known distance function + invalid = bsc.from_string(''' + prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> + prefix xsd: <http://www.w3.org/2001/XMLSchema#> + prefix bsfs: <http://bsfs.ai/schema/> + prefix bse: <http://bsfs.ai/schema/Entity#> + + bsfs:Array rdfs:subClassOf bsfs:Literal . + bsfs:Feature rdfs:subClassOf bsfs:Array . + + bsfs:Colors rdfs:subClassOf bsfs:Feature ; + bsfs:dimension "4"^^xsd:integer ; + bsfs:distance bsfs:foobar . + + ''') + self.assertRaises(errors.UnsupportedError, setattr, store, 'schema', invalid) + # cannot migrate to incompatible schema - invalid = _schema.Schema.from_string(''' + invalid = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> @@ -399,7 +426,7 @@ class TestSparqlStore(unittest.TestCase): ''') self.assertRaises(errors.ConsistencyError, setattr, store, 'schema', invalid) - invalid = _schema.Schema.from_string(''' + invalid = bsc.from_string(''' prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> prefix xsd: <http://www.w3.org/2001/XMLSchema#> prefix bsfs: <http://bsfs.ai/schema/> @@ -509,7 +536,7 @@ class TestSparqlStore(unittest.TestCase): store.set(ent_type, {URI('http://example.com/me/entity#1234')}, self.schema.predicate(ns.bse.filesize), {1234}) store.set(ent_type, {URI('http://example.com/me/entity#4321')}, self.schema.predicate(ns.bse.filesize), {4321}) # node_type must be in the schema - self.assertRaises(errors.ConsistencyError, set, store.get(self.schema.node(ns.bsfs.Node).get_child(ns.bsfs.Invalid), ast.filter.IsIn(ent_ids))) + self.assertRaises(errors.ConsistencyError, set, store.get(self.schema.node(ns.bsfs.Node).child(ns.bsfs.Invalid), ast.filter.IsIn(ent_ids))) # query must be a filter expression class Foo(): pass self.assertRaises(TypeError, set, store.get(ent_type, 1234)) @@ -574,7 +601,7 @@ class TestSparqlStore(unittest.TestCase): store.schema = self.schema # node type must be valid - self.assertRaises(errors.ConsistencyError, store.create, self.schema.node(ns.bsfs.Entity).get_child(ns.bsfs.invalid), { + self.assertRaises(errors.ConsistencyError, store.create, self.schema.node(ns.bsfs.Entity).child(ns.bsfs.invalid), { URI('http://example.com/me/entity#1234'), URI('http://example.com/me/entity#4321')}) # can create some nodes @@ -636,7 +663,7 @@ class TestSparqlStore(unittest.TestCase): p_comment = store.schema.predicate(ns.bse.comment) p_author = store.schema.predicate(ns.bse.author) p_tag = store.schema.predicate(ns.bse.tag) - p_invalid = store.schema.predicate(ns.bsfs.Predicate).get_child(ns.bsfs.foo, range=store.schema.node(ns.bsfs.Tag)) + p_invalid = store.schema.predicate(ns.bsfs.Predicate).child(ns.bsfs.foo, range=store.schema.node(ns.bsfs.Tag)) # create node instances ent_ids = { URI('http://example.com/me/entity#1234'), @@ -659,7 +686,7 @@ class TestSparqlStore(unittest.TestCase): store.create(user_type, user_ids) # invalid node_type is not permitted - self.assertRaises(errors.ConsistencyError, store.set, self.schema.node(ns.bsfs.Node).get_child(ns.bse.foo), + self.assertRaises(errors.ConsistencyError, store.set, self.schema.node(ns.bsfs.Node).child(ns.bse.foo), ent_ids, p_comment, {'hello world'}) # invalid predicate is not permitted diff --git a/test/utils/test_uuid.py b/test/utils/test_uuid.py index 49176d4..0de96ed 100644 --- a/test/utils/test_uuid.py +++ b/test/utils/test_uuid.py @@ -83,6 +83,10 @@ class TestUCID(unittest.TestCase): def test_from_path(self): self.assertEqual(UCID.from_path(self._path), self._checksum) + def test_from_dict(self): + self.assertEqual(UCID.from_dict({'hello': 'world', 'foo': 1234, 'bar': False}), + '8d2544395a0d2827e3d9ce8cd619d5e3f801e8126bf3f93ee5abd38158959585') + ## main ## |