aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorMatthias Baumgartner <dev@igsor.net>2023-01-21 22:32:33 +0100
committerMatthias Baumgartner <dev@igsor.net>2023-01-21 22:32:33 +0100
commit9310610a7edf4dcbb934aedcecff1d11348197bb (patch)
treee7d342d8f25ae40e6bdaf33549f7145bb8697f23
parentc196d2ce73d8351a18c19bcddd4b06d224e644fc (diff)
downloadbsfs-9310610a7edf4dcbb934aedcecff1d11348197bb.tar.gz
bsfs-9310610a7edf4dcbb934aedcecff1d11348197bb.tar.bz2
bsfs-9310610a7edf4dcbb934aedcecff1d11348197bb.zip
nodes predicate walk sugar
-rw-r--r--bsfs/graph/nodes.py15
-rw-r--r--bsfs/graph/walk.py120
-rw-r--r--bsfs/schema/schema.py4
-rw-r--r--test/graph/test_nodes.py14
-rw-r--r--test/graph/test_walk.py173
5 files changed, 324 insertions, 2 deletions
diff --git a/bsfs/graph/nodes.py b/bsfs/graph/nodes.py
index a4ba45f..18ab30d 100644
--- a/bsfs/graph/nodes.py
+++ b/bsfs/graph/nodes.py
@@ -19,6 +19,7 @@ from bsfs.utils import errors, URI, typename
# inner-module imports
from . import ac
from . import result
+from . import walk
# exports
__all__: typing.Sequence[str] = (
@@ -87,6 +88,18 @@ class Nodes():
"""Return all node guids."""
return iter(self._guids)
+ @property
+ def schema(self) -> bsc.Schema:
+ """Return the store's local schema."""
+ return self._backend.schema
+
+ def __getattr__(self, name: str):
+ try:
+ return super().__getattr__(name) # type: ignore [misc] # parent has no getattr
+ except AttributeError:
+ pass
+ return walk.Walk(self, walk.Walk.step(self.schema, self.node_type, name))
+
def set(
self,
pred: URI, # FIXME: URI or bsc.Predicate?
@@ -141,7 +154,7 @@ class Nodes():
if view not in (dict, list):
raise ValueError(f'expected dict or list, found {view}')
# process paths: create fetch ast, build name mapping, and find unique paths
- schema = self._backend.schema
+ schema = self.schema
statements = set()
name2path = {}
unique_paths = set() # paths that result in a single (unique) value
diff --git a/bsfs/graph/walk.py b/bsfs/graph/walk.py
new file mode 100644
index 0000000..63ef5e9
--- /dev/null
+++ b/bsfs/graph/walk.py
@@ -0,0 +1,120 @@
+"""
+
+Part of the BlackStar filesystem (bsfs) module.
+A copy of the license is provided with the project.
+Author: Matthias Baumgartner, 2022
+"""
+# imports
+from collections import abc
+import typing
+
+# bsfs imports
+from bsfs import schema as bsc
+
+# inner-module imports
+# NOTE: circular import! OK as long as only used for type annotations.
+from . import nodes # pylint: disable=cyclic-import
+
+# exports
+__all__: typing.Sequence[str] = (
+ 'Walk',
+ )
+
+
+## code ##
+
+class Walk(abc.Hashable, abc.Callable): # type: ignore [misc] # invalid base class (Callable)
+ """Syntactic sugar for `Nodes` to build and act on predicate paths via members."""
+
+ # Link to Nodes instance.
+ _root: 'nodes.Nodes'
+
+ # Current predicate path.
+ _path: typing.Tuple[bsc.Predicate, ...]
+
+ def __init__(
+ self,
+ root: 'nodes.Nodes',
+ path: typing.Sequence[bsc.Predicate],
+ ):
+ self._root = root
+ self._path = tuple(path)
+
+ @property
+ def tail(self):
+ """Return the node type at the end of the path."""
+ return self._path[-1].range
+
+
+ ## comparison
+
+ def __hash__(self) -> int:
+ """Return an integer hash that identifies the instance."""
+ return hash((type(self), self._root, self._path))
+
+ def __eq__(self, other) -> bool:
+ """Compare against *other* backend."""
+ return isinstance(other, type(self)) \
+ and self._root == other._root \
+ and self._path == other._path
+
+
+ ## representation
+
+ def __repr__(self) -> str:
+ """Return a formal string representation."""
+ path = ', '.join(pred.uri for pred in self._path)
+ return f'Walk({self._root.node_type.uri}, ({path}))'
+
+ def __str__(self) -> str:
+ """Return an informal string representation."""
+ path = ', '.join(pred.uri for pred in self._path)
+ return f'Walk(@{self._root.node_type.uri}: {path})'
+
+
+ ## walk
+
+ @staticmethod
+ def step(
+ schema: bsc.Schema,
+ node: bsc.Node,
+ name: str,
+ ) -> typing.Tuple[bsc.Predicate]:
+ """Get an predicate at *node* whose fragment matches *name*."""
+ predicates = tuple(
+ pred
+ for pred
+ in schema.predicates_at(node)
+ if pred.uri.get('fragment', None) == name
+ )
+ if len(predicates) == 0: # no fragment found for name
+ raise ValueError(f'no available predicate matches {name}')
+ if len(predicates) > 1: # ambiguous name
+ raise ValueError(f'{name} matches multiple predicates')
+ # append predicate to walk
+ return predicates # type: ignore [return-value] # size is one
+
+ def __getattr__(self, name: str) -> 'Walk':
+ """Alias for `Walk.step(name)`."""
+ try:
+ return super().__getattr__(name)
+ except AttributeError:
+ pass
+ # get predicate
+ pred = self.step(self._root.schema, self.tail, name)
+ # append predicate to walk
+ return Walk(self._root, self._path + pred)
+
+
+ ## get paths ##
+
+ def get(self, **kwargs) -> typing.Any:
+ """Alias for `Nodes.get(..)`."""
+ return self._root.get(tuple(pred.uri for pred in self._path), **kwargs)
+
+ def __call__(self, **kwargs) -> typing.Any: # pylint: disable=arguments-differ
+ """Alias for `Walk.get(...)`."""
+ return self.get(**kwargs)
+
+
+## EOF ##
diff --git a/bsfs/schema/schema.py b/bsfs/schema/schema.py
index 8d9a821..1644926 100644
--- a/bsfs/schema/schema.py
+++ b/bsfs/schema/schema.py
@@ -312,4 +312,8 @@ class Schema():
"""Return the Literal matching the *uri*."""
return self._literals[uri]
+ def predicates_at(self, node: types.Node) -> typing.Iterator[types.Predicate]:
+ """Return predicates that have domain *node* (or superclass thereof)."""
+ return iter(pred for pred in self._predicates.values() if node <= pred.domain)
+
## EOF ##
diff --git a/test/graph/test_nodes.py b/test/graph/test_nodes.py
index a4e07ee..670df69 100644
--- a/test/graph/test_nodes.py
+++ b/test/graph/test_nodes.py
@@ -10,6 +10,7 @@ import unittest
# bsie imports
from bsfs import schema as bsc
+from bsfs.graph.walk import Walk
from bsfs.namespace import Namespace, ns
from bsfs.triple_store.sparql import SparqlStore
from bsfs.utils import errors, URI
@@ -108,7 +109,8 @@ class TestNodes(unittest.TestCase):
self.p_filesize = self.backend.schema.predicate(ns.bse.filesize)
self.p_author = self.backend.schema.predicate(ns.bse.author)
self.p_tag = self.backend.schema.predicate(ns.bse.tag)
- self.p_representative = self.backend.schema.predicate(URI('http://bsfs.ai/schema/Tag#representative'))
+ self.p_representative = self.backend.schema.predicate(bst.representative)
+ self.p_label = self.backend.schema.predicate(bst.label)
self.t_created = self.backend.schema.predicate(ns.bsm.t_created)
self.ent_ids = {
URI('http://example.com/me/entity#1234'),
@@ -458,6 +460,16 @@ class TestNodes(unittest.TestCase):
Nodes(self.backend, self.user, self.ent_type, {'http://example.com/me/entity#1234'}): {'hello world'},
})
+ def test_getattr(self):
+ nodes = Nodes(self.backend, self.user, self.ent_type, {'http://example.com/me/entity#1234'})
+ # can get walks to values
+ self.assertEqual(nodes.filesize, Walk(nodes, (self.p_filesize, )))
+ # can get walks to nodes
+ self.assertEqual(nodes.tag, Walk(nodes, (self.p_tag, )))
+ # can do multiple hops
+ self.assertEqual(nodes.tag.label, Walk(nodes, (self.p_tag, self.p_label)))
+ # invalid step raises an error
+ self.assertRaises(ValueError, getattr, nodes, 'foobar')
## main ##
diff --git a/test/graph/test_walk.py b/test/graph/test_walk.py
new file mode 100644
index 0000000..057ac85
--- /dev/null
+++ b/test/graph/test_walk.py
@@ -0,0 +1,173 @@
+"""
+
+Part of the bsfs test suite.
+A copy of the license is provided with the project.
+Author: Matthias Baumgartner, 2022
+"""
+# imports
+import unittest
+
+# bsfs imports
+from bsfs import schema as bsc
+from bsfs.graph import Graph
+from bsfs.namespace import Namespace, ns
+from bsfs.triple_store.sparql import SparqlStore
+from bsfs.utils import URI
+
+# symbol to test
+from bsfs.graph.walk import Walk
+
+## code ##
+
+bse = ns.bse
+bst = Namespace('http://bsfs.ai/schema/Tag')
+
+class TestWalk(unittest.TestCase):
+ def setUp(self):
+ # backend setup
+ 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#>
+ prefix bst: <http://bsfs.ai/schema/Tag#>
+
+ bsfs:Entity rdfs:subClassOf bsfs:Node .
+ bsfs:Tag rdfs:subClassOf bsfs:Node .
+ bsfs:User rdfs:subClassOf bsfs:Node .
+ xsd:string rdfs:subClassOf bsfs:Literal .
+
+ bse:author rdfs:subClassOf bsfs:Predicate ;
+ rdfs:domain bsfs:Node ;
+ rdfs:range bsfs:User .
+
+ bse:tag rdfs:subClassOf bsfs:Predicate ;
+ rdfs:domain bsfs:Entity ;
+ rdfs:range bsfs:Tag .
+
+ bst:label rdfs:subClassOf bsfs:Predicate ;
+ rdfs:domain bsfs:Tag ;
+ rdfs:range xsd:string ;
+ bsfs:unique "true"^^xsd:boolean .
+
+ bst:subTagOf rdfs:subClassOf bsfs:Predicate ;
+ rdfs:domain bsfs:Tag ;
+ rdfs:range bsfs:Tag .
+
+ bst:main rdfs:subClassOf bsfs:Predicate ;
+ rdfs:domain bsfs:Tag ;
+ rdfs:range bsfs:Entity .
+
+ bst:author rdfs:subClassOf bsfs:Predicate ;
+ rdfs:domain bsfs:Tag ;
+ rdfs:range xsd:string .
+
+ ''')
+ self.backend = SparqlStore.Open()
+ self.user = URI('http://example.com/me')
+ self.graph = Graph(self.backend, self.user)
+ self.graph.migrate(self.schema)
+
+ # nodes setup
+ self.ents = self.graph.nodes(ns.bsfs.Entity, {
+ URI('http://example.com/me/entity#1234'),
+ URI('http://example.com/me/entity#4321')})
+ self.tags = self.graph.nodes(ns.bsfs.Tag, {
+ URI('http://example.com/me/tag#1234'),
+ URI('http://example.com/me/tag#4321')})
+ # add some instances
+ self.ents.set(bse.tag, self.tags)
+ self.graph.node(ns.bsfs.Tag, URI('http://example.com/me/tag#1234')).set(bst.label, 'hello')
+ self.graph.node(ns.bsfs.Tag, URI('http://example.com/me/tag#4321')).set(bst.label, 'world')
+
+ def test_essentials(self): # __eq__, __hash__, __str__, __repr__
+ p_author = self.schema.predicate(bse.author)
+ p_tag = self.schema.predicate(bse.tag)
+ p_main = self.schema.predicate(bst.main)
+ # comparison
+ self.assertEqual(Walk(self.ents, [p_tag]), Walk(self.ents, [p_tag]))
+ self.assertEqual(hash(Walk(self.ents, [p_tag])), hash(Walk(self.ents, [p_tag])))
+ # comparison respects type
+ class Foo(Walk): pass
+ self.assertNotEqual(Walk(self.ents, [p_tag]), Foo(self.ents, [p_tag]))
+ self.assertNotEqual(hash(Walk(self.ents, [p_tag])), hash(Foo(self.ents, [p_tag])))
+ # comparison respects root
+ self.assertNotEqual(Walk(self.ents, [p_author]), Walk(self.tags, [p_author]))
+ self.assertNotEqual(hash(Walk(self.ents, [p_author])), hash(Walk(self.tags, [p_author])))
+ # comparison respects path
+ self.assertNotEqual(Walk(self.tags, [p_author]), Walk(self.tags, [p_main]))
+ self.assertNotEqual(hash(Walk(self.tags, [p_author])), hash(Walk(self.tags, [p_main])))
+ # string conversion
+ self.assertEqual(str(Walk(self.ents, [p_tag, p_main])),
+ 'Walk(@http://bsfs.ai/schema/Entity: http://bsfs.ai/schema/Entity#tag, http://bsfs.ai/schema/Tag#main)')
+ self.assertEqual(repr(Walk(self.ents, [p_tag, p_main])),
+ 'Walk(http://bsfs.ai/schema/Entity, (http://bsfs.ai/schema/Entity#tag, http://bsfs.ai/schema/Tag#main))')
+
+ def test_tail(self):
+ self.assertEqual(Walk(self.ents, (
+ self.schema.predicate(bse.tag),
+ )).tail,
+ self.schema.node(ns.bsfs.Tag))
+ self.assertEqual(Walk(self.ents, (
+ self.schema.predicate(bse.tag),
+ self.schema.predicate(bst.main),
+ )).tail,
+ self.schema.node(ns.bsfs.Entity))
+
+ def test_step(self):
+ tag_type = self.schema.node(ns.bsfs.Tag)
+ # step returns a predicate
+ self.assertEqual(Walk.step(self.schema, tag_type, 'subTagOf'),
+ (self.schema.predicate(bst.subTagOf), ))
+ # invalid step raises an error
+ self.assertRaises(ValueError, Walk.step, self.schema, tag_type, 'foobar')
+ # ambiguous step raises an error
+ self.assertRaises(ValueError, Walk.step, self.schema, tag_type, 'author')
+
+ def test_getattr(self): # __getattr__
+ walk = Walk(self.ents, (self.schema.predicate(bse.tag), ))
+ # first step
+ self.assertEqual(walk.subTagOf, Walk(self.ents, (
+ self.schema.predicate(bse.tag),
+ self.schema.predicate(bst.subTagOf),
+ )))
+ # second step
+ self.assertEqual(walk.subTagOf.main, Walk(self.ents, (
+ self.schema.predicate(bse.tag),
+ self.schema.predicate(bst.subTagOf),
+ self.schema.predicate(bst.main),
+ )))
+ # invalid step raises an error
+ self.assertRaises(ValueError, getattr, walk, 'foobar')
+ # ambiguous step raises an error
+ self.assertRaises(ValueError, getattr, walk, 'author')
+
+ def test_get(self): # get, __call__
+ walk = Walk(self.ents, (self.schema.predicate(bse.tag), ))
+ tags = {
+ self.graph.node(ns.bsfs.Tag, URI('http://example.com/me/tag#1234')),
+ self.graph.node(ns.bsfs.Tag, URI('http://example.com/me/tag#4321'))}
+ # get returns from Nodes.get
+ self.assertDictEqual(walk.get(), {
+ self.graph.node(ns.bsfs.Entity, URI('http://example.com/me/entity#1234')): tags,
+ self.graph.node(ns.bsfs.Entity, URI('http://example.com/me/entity#4321')): tags,
+ })
+ self.assertDictEqual(walk(), {
+ self.graph.node(ns.bsfs.Entity, URI('http://example.com/me/entity#1234')): tags,
+ self.graph.node(ns.bsfs.Entity, URI('http://example.com/me/entity#4321')): tags,
+ })
+ # get passes kwargs to Nodes.get
+ self.assertSetEqual(tags, walk.get(node=True))
+ self.assertSetEqual(tags, walk(node=True))
+ self.assertSetEqual(tags, set(walk.get(view=list, node=True)))
+ self.assertSetEqual(tags, set(walk(view=list, node=True)))
+ # get returns values if need be
+ self.assertSetEqual(walk.label(node=True), {'hello', 'world'})
+
+
+## main ##
+
+if __name__ == '__main__':
+ unittest.main()
+
+## EOF ##