def more_like_this(self, model_instance, additional_query_string=None, start_offset=0, end_offset=None, models=None, limit_to_registered_models=None, result_class=None, **kwargs): if not self.setup_complete: self.setup() field_name = self.content_field_name narrow_queries = set() narrowed_results = None self.index = self.index.refresh() if limit_to_registered_models is None: limit_to_registered_models = getattr( settings, 'HAYSTACK_LIMIT_TO_REGISTERED_MODELS', True) if models and len(models): model_choices = sorted(get_model_ct(model) for model in models) elif limit_to_registered_models: # Using narrow queries, limit the results to only models handled # with the current routers. model_choices = self.build_models_list() else: model_choices = [] if len(model_choices) > 0: if narrow_queries is None: narrow_queries = set() narrow_queries.add(' OR '.join( ['%s:%s' % (DJANGO_CT, rm) for rm in model_choices])) if additional_query_string and additional_query_string != '*': narrow_queries.add(additional_query_string) narrow_searcher = None if narrow_queries is not None: # Potentially expensive? I don't see another way to do it in Whoosh... narrow_searcher = self.index.searcher() for nq in narrow_queries: recent_narrowed_results = narrow_searcher.search( self.parser.parse(force_text(nq)), limit=None) if len(recent_narrowed_results) <= 0: return { 'results': [], 'hits': 0, } if narrowed_results: narrowed_results.filter(recent_narrowed_results) else: narrowed_results = recent_narrowed_results page_num, page_length = self.calculate_page(start_offset, end_offset) self.index = self.index.refresh() raw_results = EmptyResults() searcher = None if self.index.doc_count(): query = "%s:%s" % (ID, get_identifier(model_instance)) searcher = self.index.searcher() parsed_query = self.parser.parse(query) results = searcher.search(parsed_query) if len(results): raw_results = results[0].more_like_this(field_name, top=end_offset) # Handle the case where the results have been narrowed. if narrowed_results is not None and hasattr(raw_results, 'filter'): raw_results.filter(narrowed_results) try: raw_page = ResultsPage(raw_results, page_num, page_length) except ValueError: if not self.silently_fail: raise return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } # Because as of Whoosh 2.5.1, it will return the wrong page of # results if you request something too high. :( if raw_page.pagenum < page_num: return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } results = self._process_results(raw_page, result_class=result_class) if searcher: searcher.close() if hasattr(narrow_searcher, 'close'): narrow_searcher.close() return results
def more_like_this(self, model_instance, additional_query_string=None, start_offset=0, end_offset=None, models=None, limit_to_registered_models=None, result_class=None, **kwargs): if not self.setup_complete: self.setup() # Deferred models will have a different class ("RealClass_Deferred_fieldname") # which won't be in our registry: model_klass = model_instance._meta.concrete_model field_name = self.content_field_name narrow_queries = set() narrowed_results = None self.index = self.index.refresh() if limit_to_registered_models is None: limit_to_registered_models = getattr(settings, 'HAYSTACK_LIMIT_TO_REGISTERED_MODELS', True) if models and len(models): model_choices = sorted(get_model_ct(model) for model in models) elif limit_to_registered_models: # Using narrow queries, limit the results to only models handled # with the current routers. model_choices = self.build_models_list() else: model_choices = [] if len(model_choices) > 0: if narrow_queries is None: narrow_queries = set() narrow_queries.add(' OR '.join(['%s:%s' % (DJANGO_CT, rm) for rm in model_choices])) if additional_query_string and additional_query_string != '*': narrow_queries.add(additional_query_string) narrow_searcher = None if narrow_queries is not None: # Potentially expensive? I don't see another way to do it in Whoosh... narrow_searcher = self.index.searcher() for nq in narrow_queries: recent_narrowed_results = narrow_searcher.search(self.parser.parse(force_text(nq)), limit=None) if len(recent_narrowed_results) <= 0: return { 'results': [], 'hits': 0, } if narrowed_results: narrowed_results.filter(recent_narrowed_results) else: narrowed_results = recent_narrowed_results page_num, page_length = self.calculate_page(start_offset, end_offset) self.index = self.index.refresh() raw_results = EmptyResults() if self.index.doc_count(): query = "%s:%s" % (ID, get_identifier(model_instance)) searcher = self.index.searcher() parsed_query = self.parser.parse(query) results = searcher.search(parsed_query) if len(results): raw_results = results[0].more_like_this(field_name, top=end_offset) # Handle the case where the results have been narrowed. if narrowed_results is not None and hasattr(raw_results, 'filter'): raw_results.filter(narrowed_results) try: raw_page = ResultsPage(raw_results, page_num, page_length) except ValueError: if not self.silently_fail: raise return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } # Because as of Whoosh 2.5.1, it will return the wrong page of # results if you request something too high. :( if raw_page.pagenum < page_num: return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } results = self._process_results(raw_page, result_class=result_class) searcher.close() if hasattr(narrow_searcher, 'close'): narrow_searcher.close() return results
def more_like_this(self, model_instance, additional_query_string=None, start_offset=0, end_offset=None, limit_to_registered_models=None, result_class=None, **kwargs): if not self.setup_complete: self.setup() # Handle deferred models. if get_proxied_model and hasattr(model_instance, '_deferred') and model_instance._deferred: model_klass = get_proxied_model(model_instance._meta) else: model_klass = type(model_instance) field_name = self.content_field_name narrow_queries = set() narrowed_results = None self.index = self.index.refresh() if limit_to_registered_models is None: limit_to_registered_models = getattr(settings, 'HAYSTACK_LIMIT_TO_REGISTERED_MODELS', True) if limit_to_registered_models: # Using narrow queries, limit the results to only models registered # with the current site. if narrow_queries is None: narrow_queries = set() registered_models = self.build_models_list() if len(registered_models) > 0: narrow_queries.add(' OR '.join(['%s:%s' % (DJANGO_CT, rm) for rm in registered_models])) if additional_query_string and additional_query_string != '*': narrow_queries.add(additional_query_string) narrow_searcher = None if narrow_queries is not None: # Potentially expensive? I don't see another way to do it in Whoosh... narrow_searcher = self.index.searcher() for nq in narrow_queries: recent_narrowed_results = narrow_searcher.search(self.parser.parse(force_unicode(nq))) if narrowed_results: narrowed_results.filter(recent_narrowed_results) else: narrowed_results = recent_narrowed_results # Prevent against Whoosh throwing an error. Requires an end_offset # greater than 0. if not end_offset is None and end_offset <= 0: end_offset = 1 # Determine the page. page_num = 0 if end_offset is None: end_offset = 1000000 if start_offset is None: start_offset = 0 page_length = end_offset - start_offset if page_length and page_length > 0: page_num = start_offset / page_length # Increment because Whoosh uses 1-based page numbers. page_num += 1 self.index = self.index.refresh() raw_results = EmptyResults() if self.index.doc_count(): query = "%s:%s" % (ID, get_identifier(model_instance)) searcher = self.index.searcher() parsed_query = self.parser.parse(query) results = searcher.search(parsed_query) if len(results): raw_results = results[0].more_like_this(field_name, top=end_offset) # Handle the case where the results have been narrowed. if narrowed_results and hasattr(raw_results, 'filter'): raw_results.filter(narrowed_results) try: raw_page = ResultsPage(raw_results, page_num, page_length) except ValueError: if not self.silently_fail: raise return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } results = self._process_results(raw_page, result_class=result_class) searcher.close() if hasattr(narrow_searcher, 'close'): narrow_searcher.close() return results
def more_like_this(self, model_instance, additional_query_string=None, start_offset=0, end_offset=None, models=None, limit_to_registered_models=None, result_class=None, **kwargs): if not self.setup_complete: self.setup() # Handle deferred models. if get_proxied_model and hasattr( model_instance, '_deferred') and model_instance._deferred: model_klass = get_proxied_model(model_instance._meta) else: model_klass = type(model_instance) field_name = self.content_field_name narrow_queries = set() narrowed_results = None self.index = self.index.refresh() if limit_to_registered_models is None: limit_to_registered_models = getattr( settings, 'HAYSTACK_LIMIT_TO_REGISTERED_MODELS', True) if models and len(models): model_choices = sorted([ '%s.%s' % (model._meta.app_label, model._meta.module_name) for model in models ]) elif limit_to_registered_models: # Using narrow queries, limit the results to only models handled # with the current routers. model_choices = self.build_models_list() else: model_choices = [] if len(model_choices) > 0: if narrow_queries is None: narrow_queries = set() narrow_queries.add(' OR '.join( ['%s:%s' % (DJANGO_CT, rm) for rm in model_choices])) if additional_query_string and additional_query_string != '*': narrow_queries.add(additional_query_string) narrow_searcher = None if narrow_queries is not None: # Potentially expensive? I don't see another way to do it in Whoosh... narrow_searcher = self.index.searcher() for nq in narrow_queries: recent_narrowed_results = narrow_searcher.search( self.parser.parse(force_unicode(nq))) if len(recent_narrowed_results) <= 0: return { 'results': [], 'hits': 0, } if narrowed_results: narrowed_results.filter(recent_narrowed_results) else: narrowed_results = recent_narrowed_results # Prevent against Whoosh throwing an error. Requires an end_offset # greater than 0. if not end_offset is None and end_offset <= 0: end_offset = 1 # Determine the page. page_num = 0 if end_offset is None: end_offset = 1000000 if start_offset is None: start_offset = 0 page_length = end_offset - start_offset if page_length and page_length > 0: page_num = start_offset / page_length # Increment because Whoosh uses 1-based page numbers. page_num += 1 self.index = self.index.refresh() raw_results = EmptyResults() if self.index.doc_count(): query = "%s:%s" % (ID, get_identifier(model_instance)) searcher = self.index.searcher() parsed_query = self.parser.parse(query) results = searcher.search(parsed_query) if len(results): raw_results = results[0].more_like_this(field_name, top=end_offset) # Handle the case where the results have been narrowed. if narrowed_results is not None and hasattr(raw_results, 'filter'): raw_results.filter(narrowed_results) try: raw_page = ResultsPage(raw_results, page_num, page_length) except ValueError: if not self.silently_fail: raise return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } results = self._process_results(raw_page, result_class=result_class) searcher.close() if hasattr(narrow_searcher, 'close'): narrow_searcher.close() return results
def more_like_this(self, model_instance, additional_query_string=None, start_offset=0, end_offset=None, models=None, limit_to_registered_models=None, result_class=None, **kwargs): if not self.setup_complete: self.setup() # Deferred models will have a different class ("RealClass_Deferred_fieldname") # which won't be in our registry: model_klass = model_instance._meta.concrete_model field_name = self.content_field_name narrow_queries = set() narrowed_results = None self.index = self.index.refresh() if limit_to_registered_models is None: limit_to_registered_models = getattr(settings, 'HAYSTACK_LIMIT_TO_REGISTERED_MODELS', True) if models and len(models): model_choices = sorted(['%s.%s' % (model._meta.app_label, model._meta.module_name) for model in models]) elif limit_to_registered_models: # Using narrow queries, limit the results to only models handled # with the current routers. model_choices = self.build_models_list() else: model_choices = [] if len(model_choices) > 0: if narrow_queries is None: narrow_queries = set() narrow_queries.add(' OR '.join(['%s:%s' % (DJANGO_CT, rm) for rm in model_choices])) if additional_query_string and additional_query_string != '*': narrow_queries.add(additional_query_string) narrow_searcher = None #added by gan narrow_queries = None if narrow_queries is not None: # Potentially expensive? I don't see another way to do it in Whoosh... narrow_searcher = self.index.searcher(weighting=self.weight_score) for nq in narrow_queries: recent_narrowed_results = narrow_searcher.search(self.parser.parse(force_unicode(nq))) if len(recent_narrowed_results) <= 0: return { 'results': [], 'hits': 0, } if narrowed_results: narrowed_results.filter(recent_narrowed_results) else: narrowed_results = recent_narrowed_results # Prevent against Whoosh throwing an error. Requires an end_offset # greater than 0. if not end_offset is None and end_offset <= 0: end_offset = 1 # Determine the page. page_num = 0 if end_offset is None: end_offset = 1000000 if start_offset is None: start_offset = 0 page_length = end_offset - start_offset if page_length and page_length > 0: page_num = start_offset / page_length # Increment because Whoosh uses 1-based page numbers. page_num += 1 self.index = self.index.refresh() raw_results = EmptyResults() if self.index.doc_count(): query = "%s:%s" % (ID, get_identifier(model_instance)) searcher = self.index.searcher(weighting=self.weight_score) parsed_query = self.parser.parse(query) results = searcher.search(parsed_query) #print 'begin.......' #for keyword, score in results.key_terms("text", docs=20, numterms=10): # print keyword, score if len(results): raw_results = results[0].more_like_this(field_name, top=end_offset) # Handle the case where the results have been narrowed. if narrowed_results is not None and hasattr(raw_results, 'filter'): raw_results.filter(narrowed_results) try: raw_page = ResultsPage(raw_results, page_num, page_length) except ValueError: if not self.silently_fail: raise return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } results = self._process_results(raw_page, result_class=result_class, s=searcher) searcher.close() if hasattr(narrow_searcher, 'close'): narrow_searcher.close() return results
def more_like_this(self, model_instance, additional_query_string=None, start_offset=0, end_offset=None, models=None, limit_to_registered_models=None, result_class=None, **kwargs): if not self.setup_complete: self.setup() model_klass = model_instance._meta.concrete_model field_name = self.content_field_name narrow_queries = set() narrowed_results = None self.index = self.index.refresh() if limit_to_registered_models is None: limit_to_registered_models = getattr( settings, "HAYSTACK_LIMIT_TO_REGISTERED_MODELS", True) if models and len(models): model_choices = sorted(get_model_ct(model) for model in models) elif limit_to_registered_models: model_choices = self.build_models_list() else: model_choices = [] if len(model_choices) > 0: if narrow_queries is None: narrow_queries = set() narrow_queries.add('OR'.join( ['%s: %s' % (DJANGO_CT, rm) for rm in model_choices])) if additional_query_string and additional_query_string != "*": narrow_queries.add(additional_query_string) narrow_searcher = True if narrow_queries is not None: narrow_searcher = self.index.searcher() for nq in narrow_queries: recent_narrowed_results = narrow_searcher(self.parse.parse( force_text(nq)), limit=None) if len(recent_narrowed_results) <= 0: return { 'results': [], 'hits': 0, } if narrowed_results: narrowed_results.filter(recent_narrowed_results) else: narrowed_results = recent_narrowed_results page_num, page_length = self.calculate_page( start_offset, end_offset) self.index = self.index.refresh() raw_results = EmptyResults() if self.index.doc_count(): query = "%s: %s" % (ID, get_identifier(model_instance)) searcher = self.index.searcher() parsed_query = self.parser.parse(query) results = searcher.search(parsed_query) if len(results): raw_results = results[0].more_like_this(field_name, top=end_offset) if narrowed_results is not None and hasattr( raw_results, 'filter'): raw_results.filter(narrowed_results) try: raw_page = ResultsPage(raw_results, page_num, page_length) except ValueError: if not self.silently_fail: raise return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } if raw_page.pagenum < page_num: return { 'results': [], 'hits': 0, 'spelling_suggestion': None, } results = self._process_results(raw_page, result_class=result_class) searcher.close() if hasattr(narrow_searcher, 'close'): narrow_searcher.close() return results