1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
|
"""
These method recursively evaluate the SPARQL Algebra
evalQuery is the entry-point, it will setup context and
return the SPARQLResult object
evalPart is called on each level and will delegate to the right method
A rdflib.plugins.sparql.sparql.QueryContext is passed along, keeping
information needed for evaluation
A list of dicts (solution mappings) is returned, apart from GroupBy which may
also return a dict of list of dicts
"""
import collections
import itertools
import json as j
import re
from typing import Any, Deque, Dict, Generator, Iterable, List, Tuple, Union
from urllib.parse import urlencode
from urllib.request import Request, urlopen
from pyparsing import ParseException
from rdflib.graph import Graph
from rdflib.plugins.sparql import CUSTOM_EVALS, parser
from rdflib.plugins.sparql.aggregates import Aggregator
from rdflib.plugins.sparql.evalutils import (
_ebv,
_eval,
_fillTemplate,
_join,
_minus,
_val,
)
from rdflib.plugins.sparql.parserutils import CompValue, value
from rdflib.plugins.sparql.sparql import (
AlreadyBound,
FrozenBindings,
FrozenDict,
Query,
QueryContext,
SPARQLError,
)
from rdflib.term import BNode, Identifier, Literal, URIRef, Variable
_Triple = Tuple[Identifier, Identifier, Identifier]
def evalBGP(
ctx: QueryContext, bgp: List[_Triple]
) -> Generator[FrozenBindings, None, None]:
"""
A basic graph pattern
"""
if not bgp:
yield ctx.solution()
return
s, p, o = bgp[0]
_s = ctx[s]
_p = ctx[p]
_o = ctx[o]
# type error: Item "None" of "Optional[Graph]" has no attribute "triples"
for ss, sp, so in ctx.graph.triples((_s, _p, _o)): # type: ignore[union-attr]
if None in (_s, _p, _o):
c = ctx.push()
else:
c = ctx
if _s is None:
c[s] = ss
try:
if _p is None:
c[p] = sp
except AlreadyBound:
continue
try:
if _o is None:
c[o] = so
except AlreadyBound:
continue
for x in evalBGP(c, bgp[1:]):
yield x
def evalExtend(
ctx: QueryContext, extend: CompValue
) -> Generator[FrozenBindings, None, None]:
# TODO: Deal with dict returned from evalPart from GROUP BY
for c in evalPart(ctx, extend.p):
try:
e = _eval(extend.expr, c.forget(ctx, _except=extend._vars))
if isinstance(e, SPARQLError):
raise e
yield c.merge({extend.var: e})
except SPARQLError:
yield c
def evalLazyJoin(
ctx: QueryContext, join: CompValue
) -> Generator[FrozenBindings, None, None]:
"""
A lazy join will push the variables bound
in the first part to the second part,
essentially doing the join implicitly
hopefully evaluating much fewer triples
"""
for a in evalPart(ctx, join.p1):
c = ctx.thaw(a)
for b in evalPart(c, join.p2):
yield b.merge(a) # merge, as some bindings may have been forgotten
def evalJoin(ctx: QueryContext, join: CompValue) -> Generator[FrozenDict, None, None]:
# TODO: Deal with dict returned from evalPart from GROUP BY
# only ever for join.p1
if join.lazy:
return evalLazyJoin(ctx, join)
else:
a = evalPart(ctx, join.p1)
b = set(evalPart(ctx, join.p2))
return _join(a, b)
def evalUnion(ctx: QueryContext, union: CompValue) -> Iterable[FrozenBindings]:
branch1_branch2 = []
for x in evalPart(ctx, union.p1):
branch1_branch2.append(x)
for x in evalPart(ctx, union.p2):
branch1_branch2.append(x)
return branch1_branch2
def evalMinus(ctx: QueryContext, minus: CompValue) -> Generator[FrozenDict, None, None]:
a = evalPart(ctx, minus.p1)
b = set(evalPart(ctx, minus.p2))
return _minus(a, b)
def evalLeftJoin(
ctx: QueryContext, join: CompValue
) -> Generator[FrozenBindings, None, None]:
# import pdb; pdb.set_trace()
for a in evalPart(ctx, join.p1):
ok = False
c = ctx.thaw(a)
for b in evalPart(c, join.p2):
if _ebv(join.expr, b.forget(ctx)):
ok = True
yield b
if not ok:
# we've cheated, the ctx above may contain
# vars bound outside our scope
# before we yield a solution without the OPTIONAL part
# check that we would have had no OPTIONAL matches
# even without prior bindings...
p1_vars = join.p1._vars
if p1_vars is None or not any(
_ebv(join.expr, b)
for b in evalPart(ctx.thaw(a.remember(p1_vars)), join.p2)
):
yield a
def evalFilter(
ctx: QueryContext, part: CompValue
) -> Generator[FrozenBindings, None, None]:
# TODO: Deal with dict returned from evalPart!
for c in evalPart(ctx, part.p):
if _ebv(
part.expr,
c.forget(ctx, _except=part._vars) if not part.no_isolated_scope else c,
):
yield c
def evalGraph(
ctx: QueryContext, part: CompValue
) -> Generator[FrozenBindings, None, None]:
if ctx.dataset is None:
raise Exception(
"Non-conjunctive-graph doesn't know about "
+ "graphs. Try a query without GRAPH."
)
ctx = ctx.clone()
graph = ctx[part.term]
prev_graph = ctx.graph
if graph is None:
for graph in ctx.dataset.contexts():
# in SPARQL the default graph is NOT a named graph
if graph == ctx.dataset.default_context:
continue
c = ctx.pushGraph(graph)
c = c.push()
graphSolution = [{part.term: graph.identifier}]
for x in _join(evalPart(c, part.p), graphSolution):
x.ctx.graph = prev_graph
yield x
else:
c = ctx.pushGraph(ctx.dataset.get_context(graph))
for x in evalPart(c, part.p):
x.ctx.graph = prev_graph
yield x
def evalValues(
ctx: QueryContext, part: CompValue
) -> Generator[FrozenBindings, None, None]:
for r in part.p.res:
c = ctx.push()
try:
for k, v in r.items():
if v != "UNDEF":
c[k] = v
except AlreadyBound:
continue
yield c.solution()
def evalMultiset(ctx: QueryContext, part: CompValue):
if part.p.name == "values":
return evalValues(ctx, part)
return evalPart(ctx, part.p)
def evalPart(ctx: QueryContext, part: CompValue):
# try custom evaluation functions
for name, c in CUSTOM_EVALS.items():
try:
return c(ctx, part)
except NotImplementedError:
pass # the given custome-function did not handle this part
if part.name == "BGP":
# Reorder triples patterns by number of bound nodes in the current ctx
# Do patterns with more bound nodes first
triples = sorted(
part.triples, key=lambda t: len([n for n in t if ctx[n] is None])
)
return evalBGP(ctx, triples)
elif part.name == "Filter":
return evalFilter(ctx, part)
elif part.name == "Join":
return evalJoin(ctx, part)
elif part.name == "LeftJoin":
return evalLeftJoin(ctx, part)
elif part.name == "Graph":
return evalGraph(ctx, part)
elif part.name == "Union":
return evalUnion(ctx, part)
elif part.name == "ToMultiSet":
return evalMultiset(ctx, part)
elif part.name == "Extend":
return evalExtend(ctx, part)
elif part.name == "Minus":
return evalMinus(ctx, part)
elif part.name == "Project":
return evalProject(ctx, part)
elif part.name == "Slice":
return evalSlice(ctx, part)
elif part.name == "Distinct":
return evalDistinct(ctx, part)
elif part.name == "Reduced":
return evalReduced(ctx, part)
elif part.name == "OrderBy":
return evalOrderBy(ctx, part)
elif part.name == "Group":
return evalGroup(ctx, part)
elif part.name == "AggregateJoin":
return evalAggregateJoin(ctx, part)
elif part.name == "SelectQuery":
return evalSelectQuery(ctx, part)
elif part.name == "AskQuery":
return evalAskQuery(ctx, part)
elif part.name == "ConstructQuery":
return evalConstructQuery(ctx, part)
elif part.name == "ServiceGraphPattern":
return evalServiceQuery(ctx, part)
elif part.name == "DescribeQuery":
return evalDescribeQuery(ctx, part)
else:
raise Exception("I dont know: %s" % part.name)
def evalServiceQuery(ctx: QueryContext, part: CompValue):
res = {}
match = re.match(
"^service <(.*)>[ \n]*{(.*)}[ \n]*$",
part.get("service_string", ""),
re.DOTALL | re.I,
)
if match:
service_url = match.group(1)
service_query = _buildQueryStringForServiceCall(ctx, match.group(2))
query_settings = {"query": service_query, "output": "json"}
headers = {
"accept": "application/sparql-results+json",
"user-agent": "rdflibForAnUser",
}
# GET is easier to cache so prefer that if the query is not to long
if len(service_query) < 600:
response = urlopen(
Request(service_url + "?" + urlencode(query_settings), headers=headers)
)
else:
response = urlopen(
Request(
service_url,
data=urlencode(query_settings).encode(),
headers=headers,
)
)
if response.status == 200:
json = j.loads(response.read())
variables = res["vars_"] = json["head"]["vars"]
# or just return the bindings?
res = json["results"]["bindings"]
if len(res) > 0:
for r in res:
# type error: Argument 2 to "_yieldBindingsFromServiceCallResult" has incompatible type "str"; expected "Dict[str, Dict[str, str]]"
for bound in _yieldBindingsFromServiceCallResult(ctx, r, variables): # type: ignore[arg-type]
yield bound
else:
raise Exception(
"Service: %s responded with code: %s", service_url, response.status
)
"""
Build a query string to be used by the service call.
It is supposed to pass in the existing bound solutions.
Re-adds prefixes if added and sets the base.
Wraps it in select if needed.
"""
def _buildQueryStringForServiceCall(ctx: QueryContext, service_query: str) -> str:
try:
parser.parseQuery(service_query)
except ParseException:
# This could be because we don't have a select around the service call.
service_query = "SELECT REDUCED * WHERE {" + service_query + "}"
# type error: Item "None" of "Optional[Prologue]" has no attribute "namespace_manager"
for p in ctx.prologue.namespace_manager.store.namespaces(): # type: ignore[union-attr]
service_query = "PREFIX " + p[0] + ":" + p[1].n3() + " " + service_query
# re add the base if one was defined
# type error: Item "None" of "Optional[Prologue]" has no attribute "base" [union-attr]
base = ctx.prologue.base # type: ignore[union-attr]
if base is not None and len(base) > 0:
service_query = "BASE <" + base + "> " + service_query
sol = [v for v in ctx.solution() if isinstance(v, Variable)]
if len(sol) > 0:
variables = " ".join([v.n3() for v in sol])
variables_bound = " ".join([ctx.get(v).n3() for v in sol])
service_query = (
service_query + "VALUES (" + variables + ") {(" + variables_bound + ")}"
)
return service_query
def _yieldBindingsFromServiceCallResult(
ctx: QueryContext, r: Dict[str, Dict[str, str]], variables: List[str]
) -> Generator[FrozenBindings, None, None]:
res_dict: Dict[Variable, Identifier] = {}
for var in variables:
if var in r and r[var]:
var_binding = r[var]
var_type = var_binding["type"]
if var_type == "uri":
res_dict[Variable(var)] = URIRef(var_binding["value"])
elif var_type == "literal":
res_dict[Variable(var)] = Literal(
var_binding["value"],
datatype=var_binding.get("datatype"),
lang=var_binding.get("xml:lang"),
)
# This is here because of
# https://www.w3.org/TR/2006/NOTE-rdf-sparql-json-res-20061004/#variable-binding-results
elif var_type == "typed-literal":
res_dict[Variable(var)] = Literal(
var_binding["value"], datatype=URIRef(var_binding["datatype"])
)
elif var_type == "bnode":
res_dict[Variable(var)] = BNode(var_binding["value"])
else:
raise ValueError(f"invalid type {var_type!r} for variable {var!r}")
yield FrozenBindings(ctx, res_dict)
def evalGroup(ctx: QueryContext, group: CompValue):
"""
http://www.w3.org/TR/sparql11-query/#defn_algGroup
"""
# grouping should be implemented by evalAggregateJoin
return evalPart(ctx, group.p)
def evalAggregateJoin(
ctx: QueryContext, agg: CompValue
) -> Generator[FrozenBindings, None, None]:
# import pdb ; pdb.set_trace()
p = evalPart(ctx, agg.p)
# p is always a Group, we always get a dict back
group_expr = agg.p.expr
res: Dict[Any, Any] = collections.defaultdict(
lambda: Aggregator(aggregations=agg.A)
)
if group_expr is None:
# no grouping, just COUNT in SELECT clause
# get 1 aggregator for counting
aggregator = res[True]
for row in p:
aggregator.update(row)
else:
for row in p:
# determine right group aggregator for row
k = tuple(_eval(e, row, False) for e in group_expr)
res[k].update(row)
# all rows are done; yield aggregated values
for aggregator in res.values():
yield FrozenBindings(ctx, aggregator.get_bindings())
# there were no matches
if len(res) == 0:
yield FrozenBindings(ctx)
def evalOrderBy(
ctx: QueryContext, part: CompValue
) -> Generator[FrozenBindings, None, None]:
res = evalPart(ctx, part.p)
for e in reversed(part.expr):
reverse = bool(e.order and e.order == "DESC")
res = sorted(
res, key=lambda x: _val(value(x, e.expr, variables=True)), reverse=reverse
)
return res
def evalSlice(ctx: QueryContext, slice: CompValue):
res = evalPart(ctx, slice.p)
return itertools.islice(
res,
slice.start,
slice.start + slice.length if slice.length is not None else None,
)
def evalReduced(
ctx: QueryContext, part: CompValue
) -> Generator[FrozenBindings, None, None]:
"""apply REDUCED to result
REDUCED is not as strict as DISTINCT, but if the incoming rows were sorted
it should produce the same result with limited extra memory and time per
incoming row.
"""
# This implementation uses a most recently used strategy and a limited
# buffer size. It relates to a LRU caching algorithm:
# https://en.wikipedia.org/wiki/Cache_algorithms#Least_Recently_Used_.28LRU.29
MAX = 1
# TODO: add configuration or determine "best" size for most use cases
# 0: No reduction
# 1: compare only with the last row, almost no reduction with
# unordered incoming rows
# N: The greater the buffer size the greater the reduction but more
# memory and time are needed
# mixed data structure: set for lookup, deque for append/pop/remove
mru_set = set()
mru_queue: Deque[Any] = collections.deque()
for row in evalPart(ctx, part.p):
if row in mru_set:
# forget last position of row
mru_queue.remove(row)
else:
# row seems to be new
yield row
mru_set.add(row)
if len(mru_set) > MAX:
# drop the least recently used row from buffer
mru_set.remove(mru_queue.pop())
# put row to the front
mru_queue.appendleft(row)
def evalDistinct(
ctx: QueryContext, part: CompValue
) -> Generator[FrozenBindings, None, None]:
res = evalPart(ctx, part.p)
done = set()
for x in res:
if x not in done:
yield x
done.add(x)
def evalProject(ctx: QueryContext, project: CompValue):
res = evalPart(ctx, project.p)
return (row.project(project.PV) for row in res)
def evalSelectQuery(ctx: QueryContext, query: CompValue):
res = {}
res["type_"] = "SELECT"
res["bindings"] = evalPart(ctx, query.p)
res["vars_"] = query.PV
return res
def evalAskQuery(ctx: QueryContext, query: CompValue):
res: Dict[str, Union[bool, str]] = {}
res["type_"] = "ASK"
res["askAnswer"] = False
for x in evalPart(ctx, query.p):
res["askAnswer"] = True
break
return res
def evalConstructQuery(ctx: QueryContext, query) -> Dict[str, Union[str, Graph]]:
template = query.template
if not template:
# a construct-where query
template = query.p.p.triples # query->project->bgp ...
graph = Graph()
for c in evalPart(ctx, query.p):
graph += _fillTemplate(template, c)
res: Dict[str, Union[str, Graph]] = {}
res["type_"] = "CONSTRUCT"
res["graph"] = graph
return res
def evalDescribeQuery(ctx: QueryContext, query) -> Dict[str, Union[str, Graph]]:
# Create a result graph and bind namespaces from the graph being queried
graph = Graph()
# type error: Item "None" of "Optional[Graph]" has no attribute "namespaces"
for pfx, ns in ctx.graph.namespaces(): # type: ignore[union-attr]
graph.bind(pfx, ns)
to_describe = set()
# Explicit IRIs may be provided to a DESCRIBE query.
# If there is a WHERE clause, explicit IRIs may be provided in
# addition to projected variables. Find those explicit IRIs and
# prepare to describe them.
for iri in query.PV:
if isinstance(iri, URIRef):
to_describe.add(iri)
# If there is a WHERE clause, evaluate it then find the unique set of
# resources to describe across all bindings and projected variables
if query.p is not None:
bindings = evalPart(ctx, query.p)
to_describe.update(*(set(binding.values()) for binding in bindings))
# Get a CBD for all resources identified to describe
for resource in to_describe:
# type error: Item "None" of "Optional[Graph]" has no attribute "cbd"
graph += ctx.graph.cbd(resource) # type: ignore[union-attr]
res: Dict[str, Union[str, Graph]] = {}
res["type_"] = "DESCRIBE"
res["graph"] = graph
return res
def evalQuery(graph: Graph, query: Query, initBindings, base=None):
initBindings = dict((Variable(k), v) for k, v in initBindings.items())
ctx = QueryContext(graph, initBindings=initBindings)
ctx.prologue = query.prologue
main = query.algebra
if main.datasetClause:
if ctx.dataset is None:
raise Exception(
"Non-conjunctive-graph doesn't know about "
+ "graphs! Try a query without FROM (NAMED)."
)
ctx = ctx.clone() # or push/pop?
firstDefault = False
for d in main.datasetClause:
if d.default:
if firstDefault:
# replace current default graph
dg = ctx.dataset.get_context(BNode())
ctx = ctx.pushGraph(dg)
firstDefault = True
ctx.load(d.default, default=True)
elif d.named:
g = d.named
ctx.load(g, default=False)
return evalPart(ctx, main)
|