1 //===--- Iterator.h - Query Symbol Retrieval --------------------*- C++ -*-===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 ///
9 /// \file
10 /// Symbol index queries consist of specific requirements for the requested
11 /// symbol, such as high fuzzy matching score, scope, type etc. The lists of all
12 /// symbols matching some criteria (e.g. belonging to "clang::clangd::" scope)
13 /// are expressed in a form of Search Tokens which are stored in the inverted
14 /// index. Inverted index maps these tokens to the posting lists - sorted (by
15 /// symbol quality) sequences of symbol IDs matching the token, e.g. scope token
16 /// "clangd::clangd::" is mapped to the list of IDs of all symbols which are
17 /// declared in this namespace. Search queries are build from a set of
18 /// requirements which can be combined with each other forming the query trees.
19 /// The leafs of such trees are posting lists, and the nodes are operations on
20 /// these posting lists, e.g. intersection or union. Efficient processing of
21 /// these multi-level queries is handled by Iterators. Iterators advance through
22 /// all leaf posting lists producing the result of search query, which preserves
23 /// the sorted order of IDs. Having the resulting IDs sorted is important,
24 /// because it allows receiving a certain number of the most valuable items
25 /// (e.g. symbols with highest quality which was the sorting key in the first
26 /// place) without processing all items with requested properties (this might
27 /// not be computationally effective if search request is not very restrictive).
28 //
29 //===----------------------------------------------------------------------===//
30
31 #ifndef LLVM_CLANG_TOOLS_EXTRA_CLANGD_INDEX_DEX_ITERATOR_H
32 #define LLVM_CLANG_TOOLS_EXTRA_CLANGD_INDEX_DEX_ITERATOR_H
33
34 #include "llvm/ADT/ArrayRef.h"
35 #include "llvm/Support/raw_ostream.h"
36 #include <algorithm>
37 #include <memory>
38 #include <vector>
39
40 namespace clang {
41 namespace clangd {
42 namespace dex {
43
44 /// Symbol position in the list of all index symbols sorted by a pre-computed
45 /// symbol quality.
46 using DocID = uint32_t;
47
48 /// Iterator is the interface for Query Tree node. The simplest type of Iterator
49 /// is DocumentIterator which is simply a wrapper around PostingList iterator
50 /// and serves as the Query Tree leaf. More sophisticated examples of iterators
51 /// can manage intersection, union of the elements produced by other iterators
52 /// (their children) to form a multi-level Query Tree. The interface is designed
53 /// to be extensible in order to support multiple types of iterators.
54 class Iterator {
55 public:
56 /// Returns true if all valid DocIDs were processed and hence the iterator is
57 /// exhausted.
58 virtual bool reachedEnd() const = 0;
59 /// Moves to next valid DocID. If it doesn't exist, the iterator is exhausted
60 /// and proceeds to the END.
61 ///
62 /// Note: reachedEnd() must be false.
63 virtual void advance() = 0;
64 /// Moves to the first valid DocID which is equal or higher than given ID. If
65 /// it doesn't exist, the iterator is exhausted and proceeds to the END.
66 ///
67 /// Note: reachedEnd() must be false.
68 virtual void advanceTo(DocID ID) = 0;
69 /// Returns the current element this iterator points to.
70 ///
71 /// Note: reachedEnd() must be false.
72 virtual DocID peek() const = 0;
73 /// Informs the iterator that the current document was consumed, and returns
74 /// its boost.
75 ///
76 /// Note: If this iterator has any child iterators that contain the document,
77 /// consume() should be called on those and their boosts incorporated.
78 /// consume() must *not* be called on children that don't contain the current
79 /// doc.
80 virtual float consume() = 0;
81 /// Returns an estimate of advance() calls before the iterator is exhausted.
82 virtual size_t estimateSize() const = 0;
83
~Iterator()84 virtual ~Iterator() {}
85
86 /// Prints a convenient human-readable iterator representation by recursively
87 /// dumping iterators in the following format:
88 ///
89 /// (Type Child1 Child2 ...)
90 ///
91 /// Where Type is the iterator type representation: "&" for And, "|" for Or,
92 /// ChildN is N-th iterator child. Raw iterators over PostingList are
93 /// represented as "[... CurID ...]" where CurID is the current PostingList
94 /// entry being inspected.
95 friend llvm::raw_ostream &operator<<(llvm::raw_ostream &OS,
96 const Iterator &Iterator) {
97 return Iterator.dump(OS);
98 }
99
100 /// Inspect iterator type, used internally for optimizing query trees.
101 enum class Kind { And, Or, True, False, Other };
kind()102 Kind kind() const { return MyKind; }
103
104 protected:
MyKind(MyKind)105 Iterator(Kind MyKind = Kind::Other) : MyKind(MyKind) {}
106
107 private:
108 virtual llvm::raw_ostream &dump(llvm::raw_ostream &OS) const = 0;
109 Kind MyKind;
110 };
111
112 /// Advances the iterator until it is exhausted. Returns pairs of document IDs
113 /// with the corresponding boosting score.
114 ///
115 /// Boosting can be seen as a compromise between retrieving too many items and
116 /// calculating finals score for each of them (which might be very expensive)
117 /// and not retrieving enough items so that items with very high final score
118 /// would not be processed. Boosting score is a computationally efficient way
119 /// to acquire preliminary scores of requested items.
120 std::vector<std::pair<DocID, float>> consume(Iterator &It);
121
122 namespace detail {
123 // Variadic template machinery.
populateChildren(std::vector<std::unique_ptr<Iterator>> &)124 inline void populateChildren(std::vector<std::unique_ptr<Iterator>> &) {}
125 template <typename... TailT>
populateChildren(std::vector<std::unique_ptr<Iterator>> & Children,std::unique_ptr<Iterator> Head,TailT...Tail)126 void populateChildren(std::vector<std::unique_ptr<Iterator>> &Children,
127 std::unique_ptr<Iterator> Head, TailT... Tail) {
128 Children.push_back(move(Head));
129 populateChildren(Children, move(Tail)...);
130 }
131 } // namespace detail
132
133 // A corpus is a set of documents, and a factory for iterators over them.
134 class Corpus {
135 DocID Size;
136
137 public:
Corpus(DocID Size)138 explicit Corpus(DocID Size) : Size(Size) {}
139
140 /// Returns AND Iterator which performs the intersection of the PostingLists
141 /// of its children.
142 ///
143 /// consume(): AND Iterator returns the product of Childrens' boosting
144 /// scores.
145 std::unique_ptr<Iterator>
146 intersect(std::vector<std::unique_ptr<Iterator>> Children) const;
147
148 /// Returns OR Iterator which performs the union of the PostingLists of its
149 /// children.
150 ///
151 /// consume(): OR Iterator returns the highest boost value among children
152 /// containing the requested item.
153 std::unique_ptr<Iterator>
154 unionOf(std::vector<std::unique_ptr<Iterator>> Children) const;
155
156 /// Returns TRUE Iterator which iterates over "virtual" PostingList
157 /// containing all items in range [0, Size) in an efficient manner.
158 std::unique_ptr<Iterator> all() const;
159
160 /// Returns FALSE Iterator which iterates over no documents.
161 std::unique_ptr<Iterator> none() const;
162
163 /// Returns BOOST iterator which multiplies the score of each item by given
164 /// factor. Boosting can be used as a computationally inexpensive filtering.
165 /// Users can return significantly more items using consumeAndBoost() and
166 /// then trim Top K using retrieval score.
167 std::unique_ptr<Iterator> boost(std::unique_ptr<Iterator> Child,
168 float Factor) const;
169
170 /// Returns LIMIT iterator, which yields up to N elements of its child
171 /// iterator. Elements only count towards the limit if they are part of the
172 /// final result set. Therefore the following iterator (AND (2) (LIMIT (1 2)
173 /// 1)) yields (2), not ().
174 std::unique_ptr<Iterator> limit(std::unique_ptr<Iterator> Child,
175 size_t Limit) const;
176
177 /// This allows intersect(create(...), create(...)) syntax.
178 template <typename... Args>
intersect(Args...args)179 std::unique_ptr<Iterator> intersect(Args... args) const {
180 std::vector<std::unique_ptr<Iterator>> Children;
181 detail::populateChildren(Children, std::forward<Args>(args)...);
182 return intersect(move(Children));
183 }
184
185 /// This allows unionOf(create(...), create(...)) syntax.
186 template <typename... Args>
unionOf(Args...args)187 std::unique_ptr<Iterator> unionOf(Args... args) const {
188 std::vector<std::unique_ptr<Iterator>> Children;
189 detail::populateChildren(Children, std::forward<Args>(args)...);
190 return unionOf(move(Children));
191 }
192 };
193
194 } // namespace dex
195 } // namespace clangd
196 } // namespace clang
197
198 #endif // LLVM_CLANG_TOOLS_EXTRA_CLANGD_INDEX_DEX_ITERATOR_H
199