Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Fetching contributors…

Octocat-spinner-32-eaf2f5

Cannot retrieve contributors at this time

executable file 1278 lines (1061 sloc) 50.445 kb
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277
# Copyright (C) 2009, 2010, 2011 David Sauve
# Copyright (C) 2009, 2010 Trapeze

__author__ = 'David Sauve'
__version__ = (1, 1, 6, 'beta')

import time
import datetime
import cPickle as pickle
import os
import re
import shutil
import sys
import warnings

from django.conf import settings
from django.core.exceptions import ImproperlyConfigured
from django.utils.encoding import smart_unicode, force_unicode

from haystack.backends import BaseSearchBackend, BaseSearchQuery, SearchNode, log_query
from haystack.constants import ID, DJANGO_CT, DJANGO_ID
from haystack.exceptions import HaystackError, MissingDependency, MoreLikeThisError
from haystack.fields import DateField, DateTimeField, IntegerField, FloatField, BooleanField, MultiValueField
from haystack.models import SearchResult
from haystack.utils import get_identifier

try:
    import xapian
except ImportError:
    raise MissingDependency("The 'xapian' backend requires the installation of 'xapian'. Please refer to the documentation.")


DOCUMENT_ID_TERM_PREFIX = 'Q'
DOCUMENT_CUSTOM_TERM_PREFIX = 'X'
DOCUMENT_CT_TERM_PREFIX = DOCUMENT_CUSTOM_TERM_PREFIX + 'CONTENTTYPE'

MEMORY_DB_NAME = ':memory:'

BACKEND_NAME = 'xapian'

DEFAULT_XAPIAN_FLAGS = (
    xapian.QueryParser.FLAG_PHRASE |
    xapian.QueryParser.FLAG_BOOLEAN |
    xapian.QueryParser.FLAG_LOVEHATE |
    xapian.QueryParser.FLAG_WILDCARD |
    xapian.QueryParser.FLAG_PURE_NOT
)


class InvalidIndexError(HaystackError):
    """Raised when an index can not be opened."""
    pass


class XHValueRangeProcessor(xapian.ValueRangeProcessor):
    def __init__(self, backend):
        self.backend = backend or SearchBackend()
        xapian.ValueRangeProcessor.__init__(self)
    
    def __call__(self, begin, end):
        """
Construct a tuple for value range processing.
`begin` -- a string in the format '<field_name>:[low_range]'
If 'low_range' is omitted, assume the smallest possible value.
`end` -- a string in the the format '[high_range|*]'. If '*', assume
the highest possible value.
Return a tuple of three strings: (column, low, high)
"""
        colon = begin.find(':')
        field_name = begin[:colon]
        begin = begin[colon + 1:len(begin)]
        for field_dict in self.backend.schema:
            if field_dict['field_name'] == field_name:
                if not begin:
                    if field_dict['type'] == 'text':
                        begin = u'a' # TODO: A better way of getting a min text value?
                    elif field_dict['type'] == 'long':
                        begin = -sys.maxint - 1
                    elif field_dict['type'] == 'float':
                        begin = float('-inf')
                    elif field_dict['type'] == 'date' or field_dict['type'] == 'datetime':
                        begin = u'00010101000000'
                elif end == '*':
                    if field_dict['type'] == 'text':
                        end = u'z' * 100 # TODO: A better way of getting a max text value?
                    elif field_dict['type'] == 'long':
                        end = sys.maxint
                    elif field_dict['type'] == 'float':
                        end = float('inf')
                    elif field_dict['type'] == 'date' or field_dict['type'] == 'datetime':
                        end = u'99990101000000'
                if field_dict['type'] == 'float':
                    begin = _marshal_value(float(begin))
                    end = _marshal_value(float(end))
                elif field_dict['type'] == 'long':
                    begin = _marshal_value(long(begin))
                    end = _marshal_value(long(end))
                return field_dict['column'], str(begin), str(end)


class XHExpandDecider(xapian.ExpandDecider):
    def __call__(self, term):
        """
Return True if the term should be used for expanding the search
query, False otherwise.
Currently, we only want to ignore terms beginning with `DOCUMENT_CT_TERM_PREFIX`
"""
        if term.startswith(DOCUMENT_CT_TERM_PREFIX):
            return False
        return True


class SearchBackend(BaseSearchBackend):
    """
`SearchBackend` defines the Xapian search backend for use with the Haystack
API for Django search.
It uses the Xapian Python bindings to interface with Xapian, and as
such is subject to this bug: <http://trac.xapian.org/ticket/364> when
Django is running with mod_python or mod_wsgi under Apache.
Until this issue has been fixed by Xapian, it is neccessary to set
`WSGIApplicationGroup to %{GLOBAL}` when using mod_wsgi, or
`PythonInterpreter main_interpreter` when using mod_python.
In order to use this backend, `HAYSTACK_XAPIAN_PATH` must be set in
your settings. This should point to a location where you would your
indexes to reside.
"""

    inmemory_db = None

    def __init__(self, site=None, language=None):
        """
Instantiates an instance of `SearchBackend`.
Optional arguments:
`site` -- The site to associate the backend with (default = None)

"""
        super(SearchBackend, self).__init__(site)
        
        if not hasattr(settings, 'HAYSTACK_XAPIAN_PATH'):
            raise ImproperlyConfigured('You must specify a HAYSTACK_XAPIAN_PATH in your settings.')
            
        if language:
            raise AttributeError('Language arg is now deprecated. Please use settings.HAYSTACK_XAPIAN_LANGUAGE instead.')

        if settings.HAYSTACK_XAPIAN_PATH != MEMORY_DB_NAME and \
           not os.path.exists(settings.HAYSTACK_XAPIAN_PATH):
            os.makedirs(settings.HAYSTACK_XAPIAN_PATH)
        
        self.language = getattr(settings, 'HAYSTACK_XAPIAN_LANGUAGE', 'english')
        self._schema = None
        self._content_field_name = None
        
    @property
    def schema(self):
        if not self._schema:
            self._content_field_name, self._schema = self.build_schema(self.site.all_searchfields())
        return self._schema

    @property
    def content_field_name(self):
        if not self._content_field_name:
            self._content_field_name, self._schema = self.build_schema(self.site.all_searchfields())
        return self._content_field_name
    
    def update(self, index, iterable):
        """
Updates the `index` with any objects in `iterable` by adding/updating
the database as needed.
Required arguments:
`index` -- The `SearchIndex` to process
`iterable` -- An iterable of model instances to index
For each object in `iterable`, a document is created containing all
of the terms extracted from `index.full_prepare(obj)` with field prefixes,
and 'as-is' as needed. Also, if the field type is 'text' it will be
stemmed and stored with the 'Z' prefix as well.
eg. `content:Testing` ==> `testing, Ztest, ZXCONTENTtest, XCONTENTtest`
Each document also contains an extra term in the format:
`XCONTENTTYPE<app_name>.<model_name>`
As well as a unique identifier in the the format:
`Q<app_name>.<model_name>.<pk>`
eg.: foo.bar (pk=1) ==> `Qfoo.bar.1`, `XCONTENTTYPEfoo.bar`
This is useful for querying for a specific document corresponding to
a model instance.
The document also contains a pickled version of the object itself and
the document ID in the document data field.
Finally, we also store field values to be used for sorting data. We
store these in the document value slots (position zero is reserver
for the document ID). All values are stored as unicode strings with
conversion of float, int, double, values being done by Xapian itself
through the use of the :method:xapian.sortable_serialise method.
"""
        database = self._database(writable=True)
        try:
            for obj in iterable:
                document = xapian.Document()
                
                term_generator = xapian.TermGenerator()
                term_generator.set_database(database)
                term_generator.set_stemmer(xapian.Stem(self.language))
                if getattr(settings, 'HAYSTACK_INCLUDE_SPELLING', False) is True:
                    term_generator.set_flags(xapian.TermGenerator.FLAG_SPELLING)
                term_generator.set_document(document)
                
                document_id = DOCUMENT_ID_TERM_PREFIX + get_identifier(obj)
                data = index.full_prepare(obj)
                weights = index.get_field_weights()
                for field in self.schema:
                    if field['field_name'] in data.keys():
                        prefix = DOCUMENT_CUSTOM_TERM_PREFIX + field['field_name'].upper()
                        value = data[field['field_name']]
                        try:
                            weight = int(weights[field['field_name']])
                        except KeyError:
                            weight = 1
                        if field['type'] == 'text':
                            if field['multi_valued'] == 'false':
                                term = _marshal_term(value)
                                term_generator.index_text(term, weight)
                                term_generator.index_text(term, weight, prefix)
                                if len(term.split()) == 1:
                                    document.add_term(term, weight)
                                    document.add_term(prefix + term, weight)
                                document.add_value(field['column'], _marshal_value(value))
                            else:
                                for term in value:
                                    term = _marshal_term(term)
                                    term_generator.index_text(term, weight)
                                    term_generator.index_text(term, weight, prefix)
                                    if len(term.split()) == 1:
                                        document.add_term(term, weight)
                                        document.add_term(prefix + term, weight)
                        else:
                            if field['multi_valued'] == 'false':
                                term = _marshal_term(value)
                                if len(term.split()) == 1:
                                    document.add_term(term, weight)
                                    document.add_term(prefix + term, weight)
                                    document.add_value(field['column'], _marshal_value(value))
                            else:
                                for term in value:
                                    term = _marshal_term(term)
                                    if len(term.split()) == 1:
                                        document.add_term(term, weight)
                                        document.add_term(prefix + term, weight)
                
                document.set_data(pickle.dumps(
                    (obj._meta.app_label, obj._meta.module_name, obj.pk, data),
                    pickle.HIGHEST_PROTOCOL
                ))
                document.add_term(document_id)
                document.add_term(
                    DOCUMENT_CT_TERM_PREFIX + u'%s.%s' %
                    (obj._meta.app_label, obj._meta.module_name)
                )
                database.replace_document(document_id, document)
        
        except UnicodeDecodeError:
            sys.stderr.write('Chunk failed.\n')
            pass
        
        finally:
            database = None
    
    def remove(self, obj):
        """
Remove indexes for `obj` from the database.
We delete all instances of `Q<app_name>.<model_name>.<pk>` which
should be unique to this object.
"""
        database = self._database(writable=True)
        database.delete_document(DOCUMENT_ID_TERM_PREFIX + get_identifier(obj))
    
    def clear(self, models=[]):
        """
Clear all instances of `models` from the database or all models, if
not specified.
Optional Arguments:
`models` -- Models to clear from the database (default = [])
If `models` is empty, an empty query is executed which matches all
documents in the database. Afterwards, each match is deleted.
Otherwise, for each model, a `delete_document` call is issued with
the term `XCONTENTTYPE<app_name>.<model_name>`. This will delete
all documents with the specified model type.
"""
        database = self._database(writable=True)
        if not models:
            # Because there does not appear to be a "clear all" method,
            # it's much quicker to remove the contents of the `HAYSTACK_XAPIAN_PATH`
            # folder than it is to remove each document one at a time.
            if os.path.exists(settings.HAYSTACK_XAPIAN_PATH):
                shutil.rmtree(settings.HAYSTACK_XAPIAN_PATH)
        else:
            for model in models:
                database.delete_document(
                    DOCUMENT_CT_TERM_PREFIX + '%s.%s' %
                    (model._meta.app_label, model._meta.module_name)
                )

    def document_count(self):
        try:
            return self._database().get_doccount()
        except InvalidIndexError:
            return 0

    @log_query
    def search(self, query, sort_by=None, start_offset=0, end_offset=None,
               fields='', highlight=False, facets=None, date_facets=None,
               query_facets=None, narrow_queries=None, spelling_query=None,
               limit_to_registered_models=True, result_class=None, **kwargs):
        """
Executes the Xapian::query as defined in `query`.
Required arguments:
`query` -- Search query to execute
Optional arguments:
`sort_by` -- Sort results by specified field (default = None)
`start_offset` -- Slice results from `start_offset` (default = 0)
`end_offset` -- Slice results at `end_offset` (default = None), if None, then all documents
`fields` -- Filter results on `fields` (default = '')
`highlight` -- Highlight terms in results (default = False)
`facets` -- Facet results on fields (default = None)
`date_facets` -- Facet results on date ranges (default = None)
`query_facets` -- Facet results on queries (default = None)
`narrow_queries` -- Narrow queries (default = None)
`spelling_query` -- An optional query to execute spelling suggestion on
`limit_to_registered_models` -- Limit returned results to models registered in the current `SearchSite` (default = True)
Returns:
A dictionary with the following keys:
`results` -- A list of `SearchResult`
`hits` -- The total available results
`facets` - A dictionary of facets with the following keys:
`fields` -- A list of field facets
`dates` -- A list of date facets
`queries` -- A list of query facets
If faceting was not used, the `facets` key will not be present
If `query` is None, returns no results.
If `HAYSTACK_INCLUDE_SPELLING` was enabled in `settings.py`, the
extra flag `FLAG_SPELLING_CORRECTION` will be passed to the query parser
and any suggestions for spell correction will be returned as well as
the results.
"""
        if not self.site:
            from haystack import site
        else:
            site = self.site
        
        if xapian.Query.empty(query):
            return {
                'results': [],
                'hits': 0,
            }
        
        database = self._database()
        
        if result_class is None:
            result_class = SearchResult
        
        if getattr(settings, 'HAYSTACK_INCLUDE_SPELLING', False) is True:
            spelling_suggestion = self._do_spelling_suggestion(database, query, spelling_query)
        else:
            spelling_suggestion = ''
        
        if narrow_queries is not None:
            query = xapian.Query(
                xapian.Query.OP_AND, query, xapian.Query(
                    xapian.Query.OP_OR, [self.parse_query(narrow_query) for narrow_query in narrow_queries]
                )
            )
        
        if limit_to_registered_models:
            registered_models = self.build_registered_models_list()
            
            if len(registered_models) > 0:
                query = xapian.Query(
                    xapian.Query.OP_AND, query,
                    xapian.Query(
                        xapian.Query.OP_OR, [
                            xapian.Query('%s%s' % (DOCUMENT_CT_TERM_PREFIX, model)) for model in registered_models
                        ]
                    )
                )
        
        enquire = xapian.Enquire(database)
        if hasattr(settings, 'HAYSTACK_XAPIAN_WEIGHTING_SCHEME'):
            enquire.set_weighting_scheme(xapian.BM25Weight(*settings.HAYSTACK_XAPIAN_WEIGHTING_SCHEME))
        enquire.set_query(query)
        
        if sort_by:
            sorter = xapian.MultiValueSorter()
            
            for sort_field in sort_by:
                if sort_field.startswith('-'):
                    reverse = True
                    sort_field = sort_field[1:] # Strip the '-'
                else:
                    reverse = False # Reverse is inverted in Xapian -- http://trac.xapian.org/ticket/311
                sorter.add(self._value_column(sort_field), reverse)
            
            enquire.set_sort_by_key_then_relevance(sorter, True)
        
        results = []
        facets_dict = {
            'fields': {},
            'dates': {},
            'queries': {},
        }
        
        if not end_offset:
            end_offset = database.get_doccount() - start_offset
        
        matches = self._get_enquire_mset(database, enquire, start_offset, end_offset)
        
        for match in matches:
            app_label, module_name, pk, model_data = pickle.loads(self._get_document_data(database, match.document))
            if highlight:
                model_data['highlighted'] = {
                    self.content_field_name: self._do_highlight(
                        model_data.get(self.content_field_name), query
                    )
                }
            results.append(
                result_class(app_label, module_name, pk, match.percent, searchsite=site, **model_data)
            )
        
        if facets:
            facets_dict['fields'] = self._do_field_facets(results, facets)
        if date_facets:
            facets_dict['dates'] = self._do_date_facets(results, date_facets)
        if query_facets:
            facets_dict['queries'] = self._do_query_facets(results, query_facets)
        
        return {
            'results': results,
            'hits': self._get_hit_count(database, enquire),
            'facets': facets_dict,
            'spelling_suggestion': spelling_suggestion,
        }
    
    def more_like_this(self, model_instance, additional_query=None,
                       start_offset=0, end_offset=None,
                       limit_to_registered_models=True, result_class=None, **kwargs):
        """
Given a model instance, returns a result set of similar documents.
Required arguments:
`model_instance` -- The model instance to use as a basis for
retrieving similar documents.
Optional arguments:
`additional_query` -- An additional query to narrow results
`start_offset` -- The starting offset (default=0)
`end_offset` -- The ending offset (default=None), if None, then all documents
`limit_to_registered_models` -- Limit returned results to models registered in the current `SearchSite` (default = True)
Returns:
A dictionary with the following keys:
`results` -- A list of `SearchResult`
`hits` -- The total available results
Opens a database connection, then builds a simple query using the
`model_instance` to build the unique identifier.
For each document retrieved(should always be one), adds an entry into
an RSet (relevance set) with the document id, then, uses the RSet
to query for an ESet (A set of terms that can be used to suggest
expansions to the original query), omitting any document that was in
the original query.
Finally, processes the resulting matches and returns.
"""
        if not self.site:
            from haystack import site
        else:
            site = self.site
        
        database = self._database()
        
        if result_class is None:
            result_class = SearchResult
        
        query = xapian.Query(DOCUMENT_ID_TERM_PREFIX + get_identifier(model_instance))
        
        enquire = xapian.Enquire(database)
        enquire.set_query(query)
        
        rset = xapian.RSet()
        
        if not end_offset:
            end_offset = database.get_doccount()
        
        for match in self._get_enquire_mset(database, enquire, 0, end_offset):
            rset.add_document(match.docid)
                
        query = xapian.Query(
            xapian.Query.OP_ELITE_SET,
            [expand.term for expand in enquire.get_eset(match.document.termlist_count(), rset, XHExpandDecider())],
            match.document.termlist_count()
        )
        query = xapian.Query(
            xapian.Query.OP_AND_NOT, [query, DOCUMENT_ID_TERM_PREFIX + get_identifier(model_instance)]
        )
        if limit_to_registered_models:
            registered_models = self.build_registered_models_list()
            
            if len(registered_models) > 0:
                query = xapian.Query(
                    xapian.Query.OP_AND, query,
                    xapian.Query(
                        xapian.Query.OP_OR, [
                            xapian.Query('%s%s' % (DOCUMENT_CT_TERM_PREFIX, model)) for model in registered_models
                        ]
                    )
                )
        if additional_query:
            query = xapian.Query(
                xapian.Query.OP_AND, query, additional_query
            )
        
        enquire.set_query(query)
        
        results = []
        matches = self._get_enquire_mset(database, enquire, start_offset, end_offset)
        
        for match in matches:
            app_label, module_name, pk, model_data = pickle.loads(self._get_document_data(database, match.document))
            results.append(
                result_class(app_label, module_name, pk, match.percent, searchsite=site, **model_data)
            )

        return {
            'results': results,
            'hits': self._get_hit_count(database, enquire),
            'facets': {
                'fields': {},
                'dates': {},
                'queries': {},
            },
            'spelling_suggestion': None,
        }
    
    def parse_query(self, query_string):
        """
Given a `query_string`, will attempt to return a xapian.Query
Required arguments:
``query_string`` -- A query string to parse
Returns a xapian.Query
"""
        if query_string == '*':
            return xapian.Query('') # Match everything
        elif query_string == '':
            return xapian.Query() # Match nothing
        
        flags = getattr(settings, 'HAYSTACK_XAPIAN_FLAGS', DEFAULT_XAPIAN_FLAGS)
        qp = xapian.QueryParser()
        qp.set_database(self._database())
        qp.set_stemmer(xapian.Stem(self.language))
        qp.set_stemming_strategy(xapian.QueryParser.STEM_SOME)
        qp.add_boolean_prefix('django_ct', DOCUMENT_CT_TERM_PREFIX)

        for field_dict in self.schema:
            qp.add_prefix(
                field_dict['field_name'],
                DOCUMENT_CUSTOM_TERM_PREFIX + field_dict['field_name'].upper()
            )
        
        vrp = XHValueRangeProcessor(self)
        qp.add_valuerangeprocessor(vrp)
        
        return qp.parse_query(query_string, flags)
    
    def build_schema(self, fields):
        """
Build the schema from fields.
Required arguments:
``fields`` -- A list of fields in the index
Returns a list of fields in dictionary format ready for inclusion in
an indexed meta-data.
"""
        content_field_name = ''
        schema_fields = [
            {'field_name': ID, 'type': 'text', 'multi_valued': 'false', 'column': 0},
        ]
        column = len(schema_fields)
        
        for field_name, field_class in sorted(fields.items(), key=lambda n: n[0]):
            if field_class.document is True:
                content_field_name = field_class.index_fieldname
            
            if field_class.indexed is True:
                field_data = {
                    'field_name': field_class.index_fieldname,
                    'type': 'text',
                    'multi_valued': 'false',
                    'column': column,
                }
                
                if field_class.field_type in ['date', 'datetime']:
                    field_data['type'] = 'date'
                elif field_class.field_type == 'integer':
                    field_data['type'] = 'long'
                elif field_class.field_type == 'float':
                    field_data['type'] = 'float'
                elif field_class.field_type == 'boolean':
                    field_data['type'] = 'boolean'
                
                if field_class.is_multivalued:
                    field_data['multi_valued'] = 'true'
                
                schema_fields.append(field_data)
                column += 1

        return (content_field_name, schema_fields)
    
    def _do_highlight(self, content, query, tag='em'):
        """
Highlight `query` terms in `content` with html `tag`.
This method assumes that the input text (`content`) does not contain
any special formatting. That is, it does not contain any html tags
or similar markup that could be screwed up by the highlighting.
Required arguments:
`content` -- Content to search for instances of `text`
`text` -- The text to be highlighted
"""
        for term in query:
            for match in re.findall('[^A-Z]+', term): # Ignore field identifiers
                match_re = re.compile(match, re.I)
                content = match_re.sub('<%s>%s</%s>' % (tag, term, tag), content)
        
        return content
    
    def _do_field_facets(self, results, field_facets):
        """
Private method that facets a document by field name.
Fields of type MultiValueField will be faceted on each item in the
(containing) list.
Required arguments:
`results` -- A list SearchResults to facet
`field_facets` -- A list of fields to facet on
"""
        facet_dict = {}
        
        # DS_TODO: Improve this algorithm. Currently, runs in O(N^2), ouch.
        for field in field_facets:
            facet_list = {}
            
            for result in results:
                field_value = getattr(result, field)
                if self._multi_value_field(field):
                    for item in field_value: # Facet each item in a MultiValueField
                        facet_list[item] = facet_list.get(item, 0) + 1
                else:
                    facet_list[field_value] = facet_list.get(field_value, 0) + 1
            
            facet_dict[field] = facet_list.items()
        
        return facet_dict
    
    def _do_date_facets(self, results, date_facets):
        """
Private method that facets a document by date ranges
Required arguments:
`results` -- A list SearchResults to facet
`date_facets` -- A dictionary containing facet parameters:
{'field': {'start_date': ..., 'end_date': ...: 'gap_by': '...', 'gap_amount': n}}
nb., gap must be one of the following:
year|month|day|hour|minute|second
For each date facet field in `date_facets`, generates a list
of date ranges (from `start_date` to `end_date` by `gap_by`) then
iterates through `results` and tallies the count for each date_facet.
Returns a dictionary of date facets (fields) containing a list with
entries for each range and a count of documents matching the range.
eg. {
'pub_date': [
('2009-01-01T00:00:00Z', 5),
('2009-02-01T00:00:00Z', 0),
('2009-03-01T00:00:00Z', 0),
('2009-04-01T00:00:00Z', 1),
('2009-05-01T00:00:00Z', 2),
],
}
"""
        facet_dict = {}
        
        for date_facet, facet_params in date_facets.iteritems():
            gap_type = facet_params.get('gap_by')
            gap_value = facet_params.get('gap_amount', 1)
            date_range = facet_params['start_date']
            facet_list = []
            while date_range < facet_params['end_date']:
                facet_list.append((date_range.isoformat(), 0))
                if gap_type == 'year':
                    date_range = date_range.replace(
                        year=date_range.year + int(gap_value)
                    )
                elif gap_type == 'month':
                    if date_range.month + int(gap_value) > 12:
                        date_range = date_range.replace(
                            month=((date_range.month + int(gap_value)) % 12),
                            year=(date_range.year + (date_range.month + int(gap_value)) / 12)
                        )
                    else:
                        date_range = date_range.replace(
                            month=date_range.month + int(gap_value)
                        )
                elif gap_type == 'day':
                    date_range += datetime.timedelta(days=int(gap_value))
                elif gap_type == 'hour':
                    date_range += datetime.timedelta(hours=int(gap_value))
                elif gap_type == 'minute':
                    date_range += datetime.timedelta(minutes=int(gap_value))
                elif gap_type == 'second':
                    date_range += datetime.timedelta(seconds=int(gap_value))
            
            facet_list = sorted(facet_list, key=lambda n:n[0], reverse=True)
            
            for result in results:
                result_date = getattr(result, date_facet)
                if result_date:
                    if not isinstance(result_date, datetime.datetime):
                        result_date = datetime.datetime(
                            year=result_date.year,
                            month=result_date.month,
                            day=result_date.day,
                        )
                    for n, facet_date in enumerate(facet_list):
                        if result_date > datetime.datetime(*(time.strptime(facet_date[0], '%Y-%m-%dT%H:%M:%S')[0:6])):
                            facet_list[n] = (facet_list[n][0], (facet_list[n][1] + 1))
                            break
            
            facet_dict[date_facet] = facet_list
        
        return facet_dict
    
    def _do_query_facets(self, results, query_facets):
        """
Private method that facets a document by query
Required arguments:
`results` -- A list SearchResults to facet
`query_facets` -- A dictionary containing facet parameters:
{'field': 'query', [...]}
For each query in `query_facets`, generates a dictionary entry with
the field name as the key and a tuple with the query and result count
as the value.
eg. {'name': ('a*', 5)}
"""
        facet_dict = {}
        
        for field, query in query_facets.iteritems():
            facet_dict[field] = (query, self.search(self.parse_query(query))['hits'])

        return facet_dict
    
    def _do_spelling_suggestion(self, database, query, spelling_query):
        """
Private method that returns a single spelling suggestion based on
`spelling_query` or `query`.
Required arguments:
`database` -- The database to check spelling against
`query` -- The query to check
`spelling_query` -- If not None, this will be checked instead of `query`
Returns a string with a suggested spelling
"""
        if spelling_query:
            if ' ' in spelling_query:
                return ' '.join([database.get_spelling_suggestion(term) for term in spelling_query.split()])
            else:
                return database.get_spelling_suggestion(spelling_query)
        
        term_set = set()
        for term in query:
            for match in re.findall('[^A-Z]+', term): # Ignore field identifiers
                term_set.add(database.get_spelling_suggestion(match))
        
        return ' '.join(term_set)
    
    def _database(self, writable=False):
        """
Private method that returns a xapian.Database for use.

Optional arguments:
``writable`` -- Open the database in read/write mode (default=False)

Returns an instance of a xapian.Database or xapian.WritableDatabase
"""
        if settings.HAYSTACK_XAPIAN_PATH == MEMORY_DB_NAME:
            if not SearchBackend.inmemory_db:
                SearchBackend.inmemory_db = xapian.inmemory_open()
            return SearchBackend.inmemory_db
        if writable:
            database = xapian.WritableDatabase(settings.HAYSTACK_XAPIAN_PATH, xapian.DB_CREATE_OR_OPEN)
        else:
            try:
                database = xapian.Database(settings.HAYSTACK_XAPIAN_PATH)
            except xapian.DatabaseOpeningError:
                raise InvalidIndexError(u'Unable to open index at %s' % settings.HAYSTACK_XAPIAN_PATH)

        return database

    def _get_enquire_mset(self, database, enquire, start_offset, end_offset):
        """
A safer version of Xapian.enquire.get_mset

Simply wraps the Xapian version and catches any `Xapian.DatabaseModifiedError`,
attempting a `database.reopen` as needed.

Required arguments:
`database` -- The database to be read
`enquire` -- An instance of an Xapian.enquire object
`start_offset` -- The start offset to pass to `enquire.get_mset`
`end_offset` -- The end offset to pass to `enquire.get_mset`
"""
        try:
            return enquire.get_mset(start_offset, end_offset)
        except xapian.DatabaseModifiedError:
            database.reopen()
            return enquire.get_mset(start_offset, end_offset)

    def _get_document_data(self, database, document):
        """
A safer version of Xapian.document.get_data

Simply wraps the Xapian version and catches any `Xapian.DatabaseModifiedError`,
attempting a `database.reopen` as needed.

Required arguments:
`database` -- The database to be read
`document` -- An instance of an Xapian.document object
"""
        try:
            return document.get_data()
        except xapian.DatabaseModifiedError:
            database.reopen()
            return document.get_data()

    def _get_hit_count(self, database, enquire):
        """
Given a database and enquire instance, returns the estimated number
of matches.
Required arguments:
`database` -- The database to be queried
`enquire` -- The enquire instance
"""
        return self._get_enquire_mset(
            database, enquire, 0, database.get_doccount()
        ).size()

    def _value_column(self, field):
        """
Private method that returns the column value slot in the database
for a given field.
Required arguemnts:
`field` -- The field to lookup
Returns an integer with the column location (0 indexed).
"""
        for field_dict in self.schema:
            if field_dict['field_name'] == field:
                return field_dict['column']
        return 0
    
    def _multi_value_field(self, field):
        """
Private method that returns `True` if a field is multi-valued, else
`False`.
Required arguemnts:
`field` -- The field to lookup
Returns a boolean value indicating whether the field is multi-valued.
"""
        for field_dict in self.schema:
            if field_dict['field_name'] == field:
                return field_dict['multi_valued'] == 'true'
        return False


class SearchQuery(BaseSearchQuery):
    """
This class is the Xapian specific version of the SearchQuery class.
It acts as an intermediary between the ``SearchQuerySet`` and the
``SearchBackend`` itself.
"""
    def __init__(self, backend=None, site=None):
        """
Create a new instance of the SearchQuery setting the backend as
specified. If no backend is set, will use the Xapian `SearchBackend`.
Optional arguments:
``backend`` -- The ``SearchBackend`` to use (default = None)
``site`` -- The site to use (default = None)
"""
        super(SearchQuery, self).__init__(backend=backend)
        self.backend = backend or SearchBackend(site=site)
    
    def build_params(self, *args, **kwargs):
        kwargs = super(SearchQuery, self).build_params(*args, **kwargs)

        if self.end_offset is not None:
            kwargs['end_offset'] = self.end_offset - self.start_offset
        
        return kwargs

    def build_query(self):
        if not self.query_filter:
            query = xapian.Query('')
        else:
            query = self._query_from_search_node(self.query_filter)
        
        if self.models:
            subqueries = [
                xapian.Query(
                    xapian.Query.OP_SCALE_WEIGHT, xapian.Query('%s%s.%s' % (
                            DOCUMENT_CT_TERM_PREFIX,
                            model._meta.app_label, model._meta.module_name
                        )
                    ), 0 # Pure boolean sub-query
                ) for model in self.models
            ]
            query = xapian.Query(
                xapian.Query.OP_AND, query,
                xapian.Query(xapian.Query.OP_OR, subqueries)
            )
        
        if self.boost:
            subqueries = [
                xapian.Query(
                    xapian.Query.OP_SCALE_WEIGHT, self._content_field(term, False), value
                ) for term, value in self.boost.iteritems()
            ]
            query = xapian.Query(
                xapian.Query.OP_AND_MAYBE, query,
                xapian.Query(xapian.Query.OP_OR, subqueries)
            )
        
        return query
    
    def _query_from_search_node(self, search_node, is_not=False):
        query_list = []
        
        for child in search_node.children:
            if isinstance(child, SearchNode):
                query_list.append(
                    self._query_from_search_node(child, child.negated)
                )
            else:
                expression, term = child
                field, filter_type = search_node.split_expression(expression)
                
                # Handle when we've got a ``ValuesListQuerySet``...
                if hasattr(term, 'values_list'):
                    term = list(term)
                
                if isinstance(term, (list, tuple)):
                    term = [_marshal_term(t) for t in term]
                else:
                    term = _marshal_term(term)
                
                if field == 'content':
                    query_list.append(self._content_field(term, is_not))
                else:
                    if filter_type == 'exact':
                        query_list.append(self._filter_exact(term, field, is_not))
                    elif filter_type == 'gt':
                        query_list.append(self._filter_gt(term, field, is_not))
                    elif filter_type == 'gte':
                        query_list.append(self._filter_gte(term, field, is_not))
                    elif filter_type == 'lt':
                        query_list.append(self._filter_lt(term, field, is_not))
                    elif filter_type == 'lte':
                        query_list.append(self._filter_lte(term, field, is_not))
                    elif filter_type == 'startswith':
                        query_list.append(self._filter_startswith(term, field, is_not))
                    elif filter_type == 'in':
                        query_list.append(self._filter_in(term, field, is_not))
        
        if search_node.connector == 'OR':
            return xapian.Query(xapian.Query.OP_OR, query_list)
        else:
            return xapian.Query(xapian.Query.OP_AND, query_list)
    
    def _content_field(self, term, is_not):
        """
Private method that returns a xapian.Query that searches for `value`
in all fields.
Required arguments:
``term`` -- The term to search for
``is_not`` -- Invert the search results
Returns:
A xapian.Query
"""
        if ' ' in term:
            if is_not:
                return xapian.Query(
                    xapian.Query.OP_AND_NOT, self._all_query(), self._phrase_query(
                        term.split(), self.backend.content_field_name
                    )
                )
            else:
                return self._phrase_query(term.split(), self.backend.content_field_name)
        else:
            if is_not:
                return xapian.Query(xapian.Query.OP_AND_NOT, self._all_query(), self._term_query(term))
            else:
                return self._term_query(term)
    
    def _filter_exact(self, term, field, is_not):
        """
Private method that returns a xapian.Query that searches for `term`
in a specified `field`.
Required arguments:
``term`` -- The term to search for
``field`` -- The field to search
``is_not`` -- Invert the search results
Returns:
A xapian.Query
"""
        if ' ' in term:
            if is_not:
                return xapian.Query(
                    xapian.Query.OP_AND_NOT, self._all_query(), self._phrase_query(term.split(), field)
                )
            else:
                return self._phrase_query(term.split(), field)
        else:
            if is_not:
                return xapian.Query(xapian.Query.OP_AND_NOT, self._all_query(), self._term_query(term, field))
            else:
                return self._term_query(term, field)
    
    def _filter_in(self, term_list, field, is_not):
        """
Private method that returns a xapian.Query that searches for any term
of `value_list` in a specified `field`.
Required arguments:
``term_list`` -- The terms to search for
``field`` -- The field to search
``is_not`` -- Invert the search results
Returns:
A xapian.Query
"""
        query_list = []
        for term in term_list:
            if ' ' in term:
                query_list.append(
                    self._phrase_query(term.split(), field)
                )
            else:
                query_list.append(
                    self._term_query(term, field)
                )
        if is_not:
            return xapian.Query(xapian.Query.OP_AND_NOT, self._all_query(), xapian.Query(xapian.Query.OP_OR, query_list))
        else:
            return xapian.Query(xapian.Query.OP_OR, query_list)
    
    def _filter_startswith(self, term, field, is_not):
        """
Private method that returns a xapian.Query that searches for any term
that begins with `term` in a specified `field`.
Required arguments:
``term`` -- The terms to search for
``field`` -- The field to search
``is_not`` -- Invert the search results
Returns:
A xapian.Query
"""
        if is_not:
            return xapian.Query(
                xapian.Query.OP_AND_NOT,
                self._all_query(),
                self.backend.parse_query('%s:%s*' % (field, term)),
            )
        return self.backend.parse_query('%s:%s*' % (field, term))
    
    def _filter_gt(self, term, field, is_not):
        return self._filter_lte(term, field, is_not=(is_not != True))
    
    def _filter_lt(self, term, field, is_not):
        return self._filter_gte(term, field, is_not=(is_not != True))

    def _filter_gte(self, term, field, is_not):
        """
Private method that returns a xapian.Query that searches for any term
that is greater than `term` in a specified `field`.
"""
        vrp = XHValueRangeProcessor(self.backend)
        pos, begin, end = vrp('%s:%s' % (field, _marshal_value(term)), '*')
        if is_not:
            return xapian.Query(xapian.Query.OP_AND_NOT,
                self._all_query(),
                xapian.Query(xapian.Query.OP_VALUE_RANGE, pos, begin, end)
            )
        return xapian.Query(xapian.Query.OP_VALUE_RANGE, pos, begin, end)
    
    def _filter_lte(self, term, field, is_not):
        """
Private method that returns a xapian.Query that searches for any term
that is less than `term` in a specified `field`.
"""
        vrp = XHValueRangeProcessor(self.backend)
        pos, begin, end = vrp('%s:' % field, '%s' % _marshal_value(term))
        if is_not:
            return xapian.Query(xapian.Query.OP_AND_NOT,
                self._all_query(),
                xapian.Query(xapian.Query.OP_VALUE_RANGE, pos, begin, end)
            )
        return xapian.Query(xapian.Query.OP_VALUE_RANGE, pos, begin, end)

    def _all_query(self):
        """
Private method that returns a xapian.Query that returns all documents,
Returns:
A xapian.Query
"""
        return xapian.Query('')
    
    def _term_query(self, term, field=None):
        """
Private method that returns a term based xapian.Query that searches
for `term`.
Required arguments:
``term`` -- The term to search for
``field`` -- The field to search (If `None`, all fields)
Returns:
A xapian.Query
"""
        stem = xapian.Stem(self.backend.language)
        
        if field == 'id':
            return xapian.Query('%s%s' % (DOCUMENT_ID_TERM_PREFIX, term))
        elif field == 'django_ct':
            return xapian.Query('%s%s' % (DOCUMENT_CT_TERM_PREFIX, term))
        elif field:
            stemmed = 'Z%s%s%s' % (
                DOCUMENT_CUSTOM_TERM_PREFIX, field.upper(), stem(term)
            )
            unstemmed = '%s%s%s' % (
                DOCUMENT_CUSTOM_TERM_PREFIX, field.upper(), term
            )
        else:
            stemmed = 'Z%s' % stem(term)
            unstemmed = term
            
        return xapian.Query(
            xapian.Query.OP_OR,
            xapian.Query(stemmed),
            xapian.Query(unstemmed)
        )
    
    def _phrase_query(self, term_list, field=None):
        """
Private method that returns a phrase based xapian.Query that searches
for terms in `term_list.
Required arguments:
``term_list`` -- The terms to search for
``field`` -- The field to search (If `None`, all fields)
Returns:
A xapian.Query
"""
        if field:
            return xapian.Query(
                xapian.Query.OP_PHRASE, [
                    '%s%s%s' % (
                        DOCUMENT_CUSTOM_TERM_PREFIX, field.upper(), term
                    ) for term in term_list
                ]
            )
        else:
            return xapian.Query(xapian.Query.OP_PHRASE, term_list)


def _marshal_value(value):
    """
Private utility method that converts Python values to a string for Xapian values.
"""
    if isinstance(value, datetime.datetime):
        value = _marshal_datetime(value)
    elif isinstance(value, datetime.date):
        value = _marshal_date(value)
    elif isinstance(value, bool):
        if value:
            value = u't'
        else:
            value = u'f'
    elif isinstance(value, float):
        value = xapian.sortable_serialise(value)
    elif isinstance(value, (int, long)):
        value = u'%012d' % value
    else:
        value = force_unicode(value).lower()
    return value


def _marshal_term(term):
    """
Private utility method that converts Python terms to a string for Xapian terms.
"""
    if isinstance(term, datetime.datetime):
        term = _marshal_datetime(term)
    elif isinstance(term, datetime.date):
        term = _marshal_date(term)
    else:
        term = force_unicode(term).lower()
    return term


def _marshal_date(d):
    return u'%04d%02d%02d000000' % (d.year, d.month, d.day)


def _marshal_datetime(dt):
    if dt.microsecond:
        return u'%04d%02d%02d%02d%02d%02d%06d' % (
            dt.year, dt.month, dt.day, dt.hour,
            dt.minute, dt.second, dt.microsecond
        )
    else:
        return u'%04d%02d%02d%02d%02d%02d' % (
            dt.year, dt.month, dt.day, dt.hour,
            dt.minute, dt.second
        )
Something went wrong with that request. Please try again.