be/src/exprs/function/variant_inverted_index_search.h
Line | Count | Source |
1 | | // Licensed to the Apache Software Foundation (ASF) under one |
2 | | // or more contributor license agreements. See the NOTICE file |
3 | | // distributed with this work for additional information |
4 | | // regarding copyright ownership. The ASF licenses this file |
5 | | // to you under the Apache License, Version 2.0 (the |
6 | | // "License"); you may not use this file except in compliance |
7 | | // with the License. You may obtain a copy of the License at |
8 | | // |
9 | | // http://www.apache.org/licenses/LICENSE-2.0 |
10 | | // |
11 | | // Unless required by applicable law or agreed to in writing, |
12 | | // software distributed under the License is distributed on an |
13 | | // "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
14 | | // KIND, either express or implied. See the License for the |
15 | | // specific language governing permissions and limitations |
16 | | // under the License. |
17 | | |
18 | | #pragma once |
19 | | |
20 | | #include <CLucene.h> |
21 | | #include <gen_cpp/Exprs_types.h> |
22 | | |
23 | | #include <cstdint> |
24 | | #include <functional> |
25 | | #include <map> |
26 | | #include <memory> |
27 | | #include <roaring/roaring.hh> |
28 | | #include <string> |
29 | | #include <unordered_map> |
30 | | #include <unordered_set> |
31 | | #include <utility> |
32 | | #include <vector> |
33 | | |
34 | | #include "common/status.h" |
35 | | #include "core/block/columns_with_type_and_name.h" |
36 | | #include "core/data_type/data_type.h" |
37 | | #include "storage/index/index_query_context.h" |
38 | | #include "storage/index/inverted/inverted_index_cache.h" |
39 | | #include "storage/index/inverted/inverted_index_iterator.h" |
40 | | #include "storage/index/inverted/inverted_index_reader.h" |
41 | | #include "storage/index/inverted/query_v2/query.h" |
42 | | #include "storage/index/inverted/query_v2/weight.h" |
43 | | #include "storage/olap_common.h" |
44 | | #include "storage/segment/column_reader.h" |
45 | | |
46 | | namespace doris::segment_v2::inverted_index::query_v2 { |
47 | | class Query; |
48 | | } |
49 | | |
50 | | namespace doris::segment_v2 { |
51 | | class NestedGroupReadProvider; |
52 | | struct NestedGroupReader; |
53 | | class VariantColumnReader; |
54 | | } // namespace doris::segment_v2 |
55 | | |
56 | | namespace doris { |
57 | | class TabletColumn; |
58 | | |
59 | | using namespace doris::segment_v2; |
60 | | |
61 | | class FunctionSearch; |
62 | | class IndexExecContext; |
63 | | |
64 | | using SearchLeafQueryMapper = std::function<Status( |
65 | | const std::string&, std::shared_ptr<segment_v2::inverted_index::query_v2::Query>*)>; |
66 | | |
67 | | enum class SearchFieldBindingState { |
68 | | BOUND, |
69 | | MISSING_IN_SEGMENT, |
70 | | }; |
71 | | |
72 | | enum class SearchFieldExecutionMode { |
73 | | UNBOUND, |
74 | | CLUCENE, |
75 | | DIRECT_INDEX, |
76 | | SNII_NATIVE, |
77 | | }; |
78 | | |
79 | | struct FieldReaderBinding { |
80 | | std::string logical_field_name; |
81 | | std::string stored_field_name; |
82 | | std::wstring stored_field_wstr; |
83 | | DataTypePtr column_type; |
84 | | InvertedIndexQueryType query_type; |
85 | | InvertedIndexReaderPtr inverted_reader; |
86 | | std::shared_ptr<lucene::index::IndexReader> lucene_reader; |
87 | | std::map<std::string, std::string> index_properties; |
88 | | std::string binding_key; |
89 | | std::string analyzer_key; |
90 | | InvertedIndexAnalyzerCtxSPtr analyzer_context; |
91 | | SearchFieldBindingState state = SearchFieldBindingState::MISSING_IN_SEGMENT; |
92 | | SearchFieldExecutionMode execution_mode = SearchFieldExecutionMode::UNBOUND; |
93 | | |
94 | 80 | bool is_bound() const { |
95 | 80 | return state == SearchFieldBindingState::BOUND || inverted_reader != nullptr || |
96 | 80 | lucene_reader != nullptr; |
97 | 80 | } |
98 | 8 | bool use_direct_index_reader() const { |
99 | 8 | return is_bound() && execution_mode == SearchFieldExecutionMode::DIRECT_INDEX; |
100 | 8 | } |
101 | 45 | bool use_snii_native_reader() const { |
102 | 45 | return is_bound() && execution_mode == SearchFieldExecutionMode::SNII_NATIVE; |
103 | 45 | } |
104 | | }; |
105 | | |
106 | | // Build the analyzer context a SEARCH binding executes with, converting a failure to build the |
107 | | // analyzer provider into a Status instead of letting the exception escape a Status-returning caller. |
108 | | Result<InvertedIndexAnalyzerCtxSPtr> build_search_analyzer_context( |
109 | | const std::map<std::string, std::string>& properties, const std::string& analyzer_key); |
110 | | |
111 | | class FieldReaderResolver { |
112 | | public: |
113 | | FieldReaderResolver( |
114 | | const std::unordered_map<std::string, IndexFieldNameAndTypePair>& data_type_with_names, |
115 | | const std::unordered_map<std::string, IndexIterator*>& iterators, |
116 | | std::shared_ptr<IndexQueryContext> context, |
117 | | const std::vector<TSearchFieldBinding>& field_bindings = {}); |
118 | | |
119 | | Status resolve(const std::string& field_name, InvertedIndexQueryType query_type, |
120 | | FieldReaderBinding* binding); |
121 | | |
122 | | Status resolve_with_analyzer_context(const std::string& field_name, |
123 | | InvertedIndexQueryType query_type, |
124 | | FieldReaderBinding* binding); |
125 | | |
126 | 65 | bool is_variant_subcolumn(const std::string& field_name) const { |
127 | 65 | return _variant_subcolumn_fields.count(field_name) > 0; |
128 | 65 | } |
129 | | |
130 | 11 | const std::vector<std::shared_ptr<lucene::index::IndexReader>>& readers() const { |
131 | 11 | return _readers; |
132 | 11 | } |
133 | | |
134 | | const std::unordered_map<std::string, std::shared_ptr<lucene::index::IndexReader>>& |
135 | 10 | reader_bindings() const { |
136 | 10 | return _binding_readers; |
137 | 10 | } |
138 | | |
139 | | const std::unordered_map<std::wstring, std::shared_ptr<lucene::index::IndexReader>>& |
140 | 10 | field_readers() const { |
141 | 10 | return _field_readers; |
142 | 10 | } |
143 | | |
144 | 14 | const std::unordered_map<std::string, FieldReaderBinding>& binding_cache() const { |
145 | 14 | return _cache; |
146 | 14 | } |
147 | | |
148 | 2 | IndexIterator* get_iterator(const std::string& field_name) const { |
149 | 2 | auto it = _iterators.find(field_name); |
150 | 2 | return (it != _iterators.end()) ? it->second : nullptr; |
151 | 2 | } |
152 | | |
153 | 2 | void set_leaf_query_mapper(SearchLeafQueryMapper mapper) { |
154 | 2 | _leaf_query_mapper = std::move(mapper); |
155 | 2 | } |
156 | | |
157 | | Status map_leaf_query( |
158 | | const std::string& field_name, |
159 | 24 | std::shared_ptr<segment_v2::inverted_index::query_v2::Query>* query) const { |
160 | 24 | if (!_leaf_query_mapper || query == nullptr || *query == nullptr) { |
161 | 22 | return Status::OK(); |
162 | 22 | } |
163 | 2 | return _leaf_query_mapper(field_name, query); |
164 | 24 | } |
165 | | |
166 | | private: |
167 | | std::string binding_key_for(const std::string& stored_field_name, |
168 | 64 | InvertedIndexQueryType query_type) const { |
169 | 64 | return stored_field_name + "#" + std::to_string(static_cast<int>(query_type)); |
170 | 64 | } |
171 | | |
172 | | const std::unordered_map<std::string, IndexFieldNameAndTypePair>& _data_type_with_names; |
173 | | const std::unordered_map<std::string, IndexIterator*>& _iterators; |
174 | | std::shared_ptr<IndexQueryContext> _context; |
175 | | std::vector<TSearchFieldBinding> _field_bindings; |
176 | | std::unordered_map<std::string, const TSearchFieldBinding*> _field_binding_map; |
177 | | std::unordered_set<std::string> _variant_subcolumn_fields; |
178 | | std::unordered_map<std::string, FieldReaderBinding> _cache; |
179 | | std::vector<std::shared_ptr<lucene::index::IndexReader>> _readers; |
180 | | std::unordered_map<std::string, std::shared_ptr<lucene::index::IndexReader>> _binding_readers; |
181 | | std::unordered_map<std::wstring, std::shared_ptr<lucene::index::IndexReader>> _field_readers; |
182 | | std::vector<segment_v2::InvertedIndexCacheHandle> _searcher_cache_handles; |
183 | | SearchLeafQueryMapper _leaf_query_mapper; |
184 | | }; |
185 | | |
186 | | class VariantSearchNullBitmapAdapter final : public inverted_index::query_v2::NullBitmapResolver { |
187 | | public: |
188 | | explicit VariantSearchNullBitmapAdapter(const FieldReaderResolver& resolver) |
189 | 2 | : _resolver(resolver) {} |
190 | | |
191 | | segment_v2::IndexIterator* iterator_for(const inverted_index::query_v2::Scorer& scorer, |
192 | | const std::string& logical_field) const override; |
193 | | |
194 | | private: |
195 | | const FieldReaderResolver& _resolver; |
196 | | }; |
197 | | |
198 | | void populate_variant_search_binding_context( |
199 | | const FieldReaderResolver& resolver, |
200 | | inverted_index::query_v2::QueryExecutionContext* exec_ctx); |
201 | | |
202 | | inverted_index::query_v2::QueryExecutionContext build_variant_search_query_execution_context( |
203 | | uint32_t segment_num_rows, const FieldReaderResolver& resolver, |
204 | | inverted_index::query_v2::NullBitmapResolver* null_resolver); |
205 | | |
206 | | struct VariantNestedDocMapperContext { |
207 | | std::string root_field; |
208 | | std::vector<const segment_v2::NestedGroupReader*> active_group_chain; |
209 | | const segment_v2::VariantColumnReader* variant_reader = nullptr; |
210 | | const segment_v2::NestedGroupReadProvider* read_provider = nullptr; |
211 | | segment_v2::ColumnIteratorOptions column_iter_opts; |
212 | | }; |
213 | | |
214 | | Status map_variant_nested_leaf_query_to_active_group(const VariantNestedDocMapperContext& context, |
215 | | const std::string& logical_field_name, |
216 | | inverted_index::query_v2::QueryPtr* query); |
217 | | |
218 | | inverted_index::query_v2::QueryPtr make_variant_nested_doc_mapping_query( |
219 | | inverted_index::query_v2::QueryPtr child_query, |
220 | | std::vector<const segment_v2::NestedGroupReader*> child_to_parent_chain, |
221 | | const segment_v2::NestedGroupReadProvider* read_provider, |
222 | | segment_v2::ColumnIteratorOptions column_iter_opts); |
223 | | |
224 | | class VariantNestedSearchEvaluator { |
225 | | public: |
226 | | explicit VariantNestedSearchEvaluator(const FunctionSearch& function_search) |
227 | 8 | : _function_search(function_search) {} |
228 | | |
229 | | Status evaluate(const TSearchParam& search_param, const TSearchClause& nested_clause, |
230 | | const std::shared_ptr<segment_v2::IndexQueryContext>& context, |
231 | | FieldReaderResolver& resolver, uint32_t num_rows, |
232 | | const IndexExecContext* index_exec_ctx, const TabletColumn* nested_column, |
233 | | std::shared_ptr<roaring::Roaring>& result_bitmap) const; |
234 | | |
235 | | private: |
236 | | const FunctionSearch& _function_search; |
237 | | }; |
238 | | |
239 | | } // namespace doris |