Source

whoosh / src / whoosh / query / nary.py

Full commit
  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
# Copyright 2007 Matt Chaput. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
#    1. Redistributions of source code must retain the above copyright notice,
#       this list of conditions and the following disclaimer.
#
#    2. Redistributions in binary form must reproduce the above copyright
#       notice, this list of conditions and the following disclaimer in the
#       documentation and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY MATT CHAPUT ``AS IS'' AND ANY EXPRESS OR
# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
# MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO
# EVENT SHALL MATT CHAPUT OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA,
# OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
# EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
# The views and conclusions contained in the software and documentation are
# those of the authors and should not be interpreted as representing official
# policies, either expressed or implied, of Matt Chaput.

from __future__ import division

from whoosh import matching
from whoosh.compat import text_type, u
from whoosh.compat import xrange
from whoosh.query import qcore
from whoosh.util import make_binary_tree, make_weighted_tree


class CompoundQuery(qcore.Query):
    """Abstract base class for queries that combine or manipulate the results
    of multiple sub-queries .
    """

    def __init__(self, subqueries, boost=1.0):
        self.subqueries = subqueries
        self.boost = boost

    def __repr__(self):
        r = "%s(%r" % (self.__class__.__name__, self.subqueries)
        if hasattr(self, "boost") and self.boost != 1:
            r += ", boost=%s" % self.boost
        r += ")"
        return r

    def __unicode__(self):
        r = u("(")
        r += (self.JOINT).join([text_type(s) for s in self.subqueries])
        r += u(")")
        return r

    __str__ = __unicode__

    def __eq__(self, other):
        return other and self.__class__ is other.__class__ and\
        self.subqueries == other.subqueries and\
        self.boost == other.boost

    def __getitem__(self, i):
        return self.subqueries.__getitem__(i)

    def __len__(self):
        return len(self.subqueries)

    def __iter__(self):
        return iter(self.subqueries)

    def __hash__(self):
        h = hash(self.__class__.__name__) ^ hash(self.boost)
        for q in self.subqueries:
            h ^= hash(q)
        return h

    def is_leaf(self):
        return False

    def children(self):
        return iter(self.subqueries)

    def apply(self, fn):
        return self.__class__([fn(q) for q in self.subqueries],
                              boost=self.boost)

    def field(self):
        if self.subqueries:
            f = self.subqueries[0].field()
            if all(q.field() == f for q in self.subqueries[1:]):
                return f

    def estimate_size(self, ixreader):
        return sum(q.estimate_size(ixreader) for q in self.subqueries)

    def estimate_min_size(self, ixreader):
        from whoosh.query import Not

        subs = self.subqueries
        for sub in subs:
            if isinstance(sub, Not):
                return 0

        return min(q.estimate_min_size(ixreader) for q in subs)

    def normalize(self):
        from whoosh.query import Every, TermRange, NumericRange

        # Normalize subqueries and merge nested instances of this class
        subqueries = []
        for s in self.subqueries:
            s = s.normalize()
            if isinstance(s, self.__class__):
                subqueries += [ss.with_boost(ss.boost * s.boost) for ss in s]
            else:
                subqueries.append(s)

        # If every subquery is Null, this query is Null
        if all(q is qcore.NullQuery for q in subqueries):
            return qcore.NullQuery

        # If there's an unfielded Every inside, then this query is Every
        if any((isinstance(q, Every) and q.fieldname is None)
               for q in subqueries):
            return Every()

        # Merge ranges and Everys
        everyfields = set()
        i = 0
        while i < len(subqueries):
            q = subqueries[i]
            f = q.field()
            if f in everyfields:
                subqueries.pop(i)
                continue

            if isinstance(q, (TermRange, NumericRange)):
                j = i + 1
                while j < len(subqueries):
                    if q.overlaps(subqueries[j]):
                        qq = subqueries.pop(j)
                        q = q.merge(qq, intersect=self.intersect_merge)
                    else:
                        j += 1
                q = subqueries[i] = q.normalize()

            if isinstance(q, Every):
                everyfields.add(q.fieldname)
            i += 1

        # Eliminate duplicate queries
        subqs = []
        seenqs = set()
        for s in subqueries:
            if (not isinstance(s, Every) and s.field() in everyfields):
                continue
            if s in seenqs:
                continue
            seenqs.add(s)
            subqs.append(s)

        # Remove NullQuerys
        subqs = [q for q in subqs if q is not qcore.NullQuery]

        if not subqs:
            return qcore.NullQuery

        if len(subqs) == 1:
            sub = subqs[0]
            if not (self.boost == 1.0 and sub.boost == 1.0):
                sub = sub.with_boost(sub.boost * self.boost)
            return sub

        return self.__class__(subqs, boost=self.boost)

    def simplify(self, ixreader):
        subs = self.subqueries
        if subs:
            q = self.__class__([subq.simplify(ixreader) for subq in subs],
                                boost=self.boost).normalize()
        else:
            q = qcore.NullQuery
        return q

    def matcher(self, searcher, context=None):
        # This method does a little sanity checking and then passes the info
        # down to the _matcher() method which subclasses must implement

        subs = self.subqueries
        if not subs:
            return matching.NullMatcher()

        if len(subs) == 1:
            m = subs[0].matcher(searcher, context)
        else:
            m = self._matcher(subs, searcher, context)
        return m

    def _matcher(self, subs, searcher, context):
        # Subclasses must implement this method

        raise NotImplementedError

    def _tree_matcher(self, subs, mcls, searcher, context, q_weight_fn,
                      **kwargs):
        # q_weight_fn is a function which is called on each query and returns a
        # "weight" value which is used to build a huffman-like matcher tree. If
        # q_weight_fn is None, an order-preserving binary tree is used instead.

        # Create a matcher from the list of subqueries
        subms = [q.matcher(searcher, context) for q in subs]
        if len(subms) == 1:
            m = subms[0]
        elif q_weight_fn is None:
            m = make_binary_tree(mcls, subms, **kwargs)
        else:
            w_subms = [(q_weight_fn(q), m) for q, m in zip(subs, subms)]
            m = make_weighted_tree(mcls, w_subms, **kwargs)

        # If this query had a boost, add a wrapping matcher to apply the boost
        if self.boost != 1.0:
            m = matching.WrappingMatcher(m, self.boost)

        return m


class And(CompoundQuery):
    """Matches documents that match ALL of the subqueries.

    >>> And([Term("content", u"render"),
    ...      Term("content", u"shade"),
    ...      Not(Term("content", u"texture"))])
    >>> # You can also do this
    >>> Term("content", u"render") & Term("content", u"shade")
    """

    # This is used by the superclass's __unicode__ method.
    JOINT = " AND "
    intersect_merge = True

    def requires(self):
        s = set()
        for q in self.subqueries:
            s |= q.requires()
        return s

    def estimate_size(self, ixreader):
        return min(q.estimate_size(ixreader) for q in self.subqueries)

    def _matcher(self, subs, searcher, context):
        r = searcher.reader()
        q_weight_fn = lambda q: 0 - q.estimate_size(r)
        return self._tree_matcher(subs, matching.IntersectionMatcher, searcher,
                                  context, q_weight_fn)


class Or(CompoundQuery):
    """Matches documents that match ANY of the subqueries.

    >>> Or([Term("content", u"render"),
    ...     And([Term("content", u"shade"), Term("content", u"texture")]),
    ...     Not(Term("content", u"network"))])
    >>> # You can also do this
    >>> Term("content", u"render") | Term("content", u"shade")
    """

    # This is used by the superclass's __unicode__ method.
    JOINT = " OR "
    intersect_merge = False
    binary_matcher = False

    def __init__(self, subqueries, boost=1.0, minmatch=0, scale=None):
        """
        :param subqueries: a list of :class:`Query` objects to search for.
        :param boost: a boost factor to apply to the scores of all matching
            documents.
        :param minmatch: not yet implemented.
        :param scale: a scaling factor for a "coordination bonus". If this
            value is not None, it should be a floating point number greater
            than 0 and less than 1. The scores of the matching documents are
            boosted/penalized based on the number of query terms that matched
            in the document. This number scales the effect of the bonuses.
        """

        CompoundQuery.__init__(self, subqueries, boost=boost)
        self.minmatch = minmatch
        self.scale = scale

    def __unicode__(self):
        r = u("(")
        r += (self.JOINT).join([text_type(s) for s in self.subqueries])
        r += u(")")
        if self.minmatch:
            r += u(">%s") % self.minmatch
        return r

    __str__ = __unicode__

    def normalize(self):
        norm = CompoundQuery.normalize(self)
        if norm.__class__ is self.__class__:
            norm.minmatch = self.minmatch
            norm.scale = self.scale
        return norm

    def requires(self):
        if len(self.subqueries) == 1:
            return self.subqueries[0].requires()
        else:
            return set()

    def _matcher(self, subs, searcher, context):
        needs_current = context.needs_current if context else True
        scored = context.weighting is not None if context else True

        # A binary tree of UnionMatchers is usually slower than
        # ArrayUnionMatcher, but in certain circumstances the binary tree is
        # necessary
        usebin = (self.binary_matcher
                  or self.scale
                  or needs_current
                  or len(subs) <= 2)

        if usebin:
            # Make a binary tree of UnionMatcher objects
            r = searcher.reader()
            q_weight_fn = lambda q: q.estimate_size(r)
            m = self._tree_matcher(subs, matching.UnionMatcher, searcher,
                                   context, q_weight_fn)
        else:
            # Get submatchers
            ms = [subq.matcher(searcher, context) for subq in subs]
            # Remove null matchers
            ms = [m for m in ms if m.is_active()]
            if not ms:
                return matching.NullMatcher()

            doccount = searcher.doc_count_all()
            # Hackish optimization: if all the sub-matchers are leaf matchers,
            # preload all scores for speed
            if all(m.is_leaf() for m in ms):
                m = matching.PreloadedUnionMatcher(ms, doccount,
                                                   boost=self.boost,
                                                   scored=scored)
            else:
                m = matching.ArrayUnionMatcher(ms, doccount, boost=self.boost,
                                               scored=scored)

        # If a scaling factor was given, wrap the matcher in a CoordMatcher
        # to alter scores based on term coordination
        if self.scale and any(m.term_matchers()):
            m = matching.CoordMatcher(m, scale=self.scale)
        return m


class DisjunctionMax(CompoundQuery):
    """Matches all documents that match any of the subqueries, but scores each
    document using the maximum score from the subqueries.
    """

    def __init__(self, subqueries, boost=1.0, tiebreak=0.0):
        CompoundQuery.__init__(self, subqueries, boost=boost)
        self.tiebreak = tiebreak

    def __unicode__(self):
        r = u("DisMax(")
        r += " ".join(sorted(text_type(s) for s in self.subqueries))
        r += u(")")
        if self.tiebreak:
            r += u("~") + text_type(self.tiebreak)
        return r

    __str__ = __unicode__

    def normalize(self):
        norm = CompoundQuery.normalize(self)
        if norm.__class__ is self.__class__:
            norm.tiebreak = self.tiebreak
        return norm

    def requires(self):
        if len(self.subqueries) == 1:
            return self.subqueries[0].requires()
        else:
            return set()

    def _matcher(self, subs, searcher, context):
        r = searcher.reader()
        q_weight_fn = lambda q: q.estimate_size(r)
        return self._tree_matcher(subs, matching.DisjunctionMaxMatcher,
                                  searcher, context, q_weight_fn,
                                  tiebreak=self.tiebreak)


# Boolean queries

class BinaryQuery(CompoundQuery):
    """Base class for binary queries (queries which are composed of two
    sub-queries). Subclasses should set the ``matcherclass`` attribute or
    override ``matcher()``, and may also need to override ``normalize()``,
    ``estimate_size()``, and/or ``estimate_min_size()``.
    """

    boost = 1.0

    def __init__(self, a, b):
        self.a = a
        self.b = b
        self.subqueries = (a, b)

    def __eq__(self, other):
        return (other and self.__class__ is other.__class__
                and self.a == other.a and self.b == other.b)

    def __hash__(self):
        return (hash(self.__class__.__name__) ^ hash(self.a) ^ hash(self.b))

    def apply(self, fn):
        return self.__class__(fn(self.a), fn(self.b))

    def field(self):
        f = self.a.field()
        if self.b.field() == f:
            return f

    def with_boost(self, boost):
        return self.__class__(self.a.with_boost(boost),
                              self.b.with_boost(boost))

    def normalize(self):
        a = self.a.normalize()
        b = self.b.normalize()
        if a is qcore.NullQuery and b is qcore.NullQuery:
            return qcore.NullQuery
        elif a is qcore.NullQuery:
            return b
        elif b is qcore.NullQuery:
            return a

        return self.__class__(a, b)

    def matcher(self, searcher, context=None):
        return self.matcherclass(self.a.matcher(searcher, context),
                                 self.b.matcher(searcher, context))


class AndNot(BinaryQuery):
    """Binary boolean query of the form 'a ANDNOT b', where documents that
    match b are removed from the matches for a.
    """

    JOINT = " ANDNOT "

    def with_boost(self, boost):
        return self.__class__(self.a.with_boost(boost), self.b)

    def normalize(self):
        a = self.a.normalize()
        b = self.b.normalize()

        if a is qcore.NullQuery:
            return qcore.NullQuery
        elif b is qcore.NullQuery:
            return a

        return self.__class__(a, b)

    def requires(self):
        return self.a.requires()

    def matcher(self, searcher, context=None):
        scoredm = self.a.matcher(searcher, context)
        notm = self.b.matcher(searcher, searcher.boolean_context())
        return matching.AndNotMatcher(scoredm, notm)


class Otherwise(BinaryQuery):
    """A binary query that only matches the second clause if the first clause
    doesn't match any documents.
    """

    JOINT = " OTHERWISE "

    def matcher(self, searcher, context=None):
        m = self.a.matcher(searcher, context)
        if not m.is_active():
            m = self.b.matcher(searcher, context)
        return m


class Require(BinaryQuery):
    """Binary query returns results from the first query that also appear in
    the second query, but only uses the scores from the first query. This lets
    you filter results without affecting scores.
    """

    JOINT = " REQUIRE "
    matcherclass = matching.RequireMatcher

    def requires(self):
        return self.a.requires() | self.b.requires()

    def estimate_size(self, ixreader):
        return self.b.estimate_size(ixreader)

    def estimate_min_size(self, ixreader):
        return self.b.estimate_min_size(ixreader)

    def with_boost(self, boost):
        return self.__class__(self.a.with_boost(boost), self.b)

    def normalize(self):
        a = self.a.normalize()
        b = self.b.normalize()
        if a is qcore.NullQuery or b is qcore.NullQuery:
            return qcore.NullQuery
        return self.__class__(a, b)

    def docs(self, searcher):
        return And(self.subqueries).docs(searcher)

    def matcher(self, searcher, context=None):
        scoredm = self.a.matcher(searcher, context)
        requiredm = self.b.matcher(searcher, searcher.boolean_context())
        return matching.AndNotMatcher(scoredm, requiredm)


class AndMaybe(BinaryQuery):
    """Binary query takes results from the first query. If and only if the
    same document also appears in the results from the second query, the score
    from the second query will be added to the score from the first query.
    """

    JOINT = " ANDMAYBE "
    matcherclass = matching.AndMaybeMatcher

    def normalize(self):
        a = self.a.normalize()
        b = self.b.normalize()
        if a is qcore.NullQuery:
            return qcore.NullQuery
        if b is qcore.NullQuery:
            return a
        return self.__class__(a, b)

    def requires(self):
        return self.a.requires()

    def estimate_min_size(self, ixreader):
        return self.subqueries[0].estimate_min_size(ixreader)

    def docs(self, searcher):
        return self.subqueries[0].docs(searcher)


def BooleanQuery(required, should, prohibited):
    return AndNot(AndMaybe(And(required), Or(should)),
                  Or(prohibited)).normalize()