1 /*!
2 @file
3 Forward declares `boost::hana::Searchable`.
4 
5 @copyright Louis Dionne 2013-2017
6 Distributed under the Boost Software License, Version 1.0.
7 (See accompanying file LICENSE.md or copy at http://boost.org/LICENSE_1_0.txt)
8  */
9 
10 #ifndef BOOST_HANA_FWD_CONCEPT_SEARCHABLE_HPP
11 #define BOOST_HANA_FWD_CONCEPT_SEARCHABLE_HPP
12 
13 #include <boost/hana/config.hpp>
14 
15 
16 BOOST_HANA_NAMESPACE_BEGIN
17     //! @ingroup group-concepts
18     //! @defgroup group-Searchable Searchable
19     //! The `Searchable` concept represents structures that can be searched.
20     //!
21     //! Intuitively, a `Searchable` is any structure, finite or infinite,
22     //! containing elements that can be searched using a predicate. Sometimes,
23     //! `Searchable`s will associate keys to values; one can search for a key
24     //! with a predicate, and the value associated to it is returned. This
25     //! gives rise to map-like data structures. Other times, the elements of
26     //! the structure that are searched (i.e. those to which the predicate is
27     //! applied) are the same that are returned, which gives rise to set-like
28     //! data structures. In general, we will refer to the _keys_ of a
29     //! `Searchable` structure as those elements that are used for searching,
30     //! and to the _values_ of a `Searchable` as those elements that are
31     //! returned when a search is successful. As was explained, there is no
32     //! requirement that both notions differ, and it is often useful to have
33     //! keys and values coincide (think about `std::set`).
34     //!
35     //! Some methods like `any_of`, `all_of` and `none_of` allow simple queries
36     //! to be performed on the keys of the structure, while other methods like
37     //! `find` and `find_if` make it possible to find the value associated
38     //! to a key. The most specific method should always be used if one
39     //! cares about performance, because it is usually the case that heavy
40     //! optimizations can be performed in more specific methods. For example,
41     //! an associative data structure implemented as a hash table will be much
42     //! faster to access using `find` than `find_if`, because in the second
43     //! case it will have to do a linear search through all the entries.
44     //! Similarly, using `contains` will likely be much faster than `any_of`
45     //! with an equivalent predicate.
46     //!
47     //! > __Insight__\n
48     //! > In a lazy evaluation context, any `Foldable` can also become a model
49     //! > of `Searchable` because we can search lazily through the structure
50     //! > with `fold_right`. However, in the context of C++, some `Searchable`s
51     //! > can not be folded; think for example of an infinite set.
52     //!
53     //!
54     //! Minimal complete definition
55     //! ---------------------------
56     //! `find_if` and `any_of`
57     //!
58     //! When `find_if` and `any_of` are provided, the other functions are
59     //! implemented according to the laws explained below.
60     //!
61     //! @note
62     //! We could implement `any_of(xs, pred)` by checking whether
63     //! `find_if(xs, pred)` is an empty `optional` or not, and then reduce
64     //! the minimal complete definition to `find_if`. However, this is not
65     //! done because that implementation requires the predicate of `any_of`
66     //! to return a compile-time `Logical`, which is more restrictive than
67     //! what we have right now.
68     //!
69     //!
70     //! Laws
71     //! ----
72     //! In order for the semantics of the methods to be consistent, some
73     //! properties must be satisfied by any model of the `Searchable` concept.
74     //! Rigorously, for any `Searchable`s  `xs` and `ys` and any predicate `p`,
75     //! the following laws should be satisfied:
76     //! @code
77     //!     any_of(xs, p) <=> !all_of(xs, negated p)
78     //!                   <=> !none_of(xs, p)
79     //!
80     //!     contains(xs, x) <=> any_of(xs, equal.to(x))
81     //!
82     //!     find(xs, x) == find_if(xs, equal.to(x))
83     //!     find_if(xs, always(false_)) == nothing
84     //!
85     //!     is_subset(xs, ys) <=> all_of(xs, [](auto x) { return contains(ys, x); })
86     //!     is_disjoint(xs, ys) <=> none_of(xs, [](auto x) { return contains(ys, x); })
87     //! @endcode
88     //!
89     //! Additionally, if all the keys of the `Searchable` are `Logical`s,
90     //! the following laws should be satisfied:
91     //! @code
92     //!     any(xs)  <=> any_of(xs, id)
93     //!     all(xs)  <=> all_of(xs, id)
94     //!     none(xs) <=> none_of(xs, id)
95     //! @endcode
96     //!
97     //!
98     //! Concrete models
99     //! ---------------
100     //! `hana::map`, `hana::optional`, `hana::range`, `hana::set`,
101     //! `hana::string`, `hana::tuple`
102     //!
103     //!
104     //! Free model for builtin arrays
105     //! -----------------------------
106     //! Builtin arrays whose size is known can be searched as-if they were
107     //! homogeneous tuples. However, since arrays can only hold objects of
108     //! a single type and the predicate to `find_if` must return a compile-time
109     //! `Logical`, the `find_if` method is fairly useless. For similar reasons,
110     //! the `find` method is also fairly useless. This model is provided mainly
111     //! because of the `any_of` method & friends, which are both useful and
112     //! compile-time efficient.
113     //!
114     //!
115     //! Structure preserving functions
116     //! ------------------------------
117     //! Given two `Searchables` `S1` and `S2`, a function
118     //! @f$ f : S_1(X) \to S_2(X) @f$ is said to preserve the `Searchable`
119     //! structure if for all `xs` of data type `S1(X)` and predicates
120     //! @f$ \mathtt{pred} : X \to Bool @f$ (for a `Logical` `Bool`),
121     //! @code
122     //!     any_of(xs, pred)  if and only if  any_of(f(xs), pred)
123     //!     find_if(xs, pred) == find_if(f(xs), pred)
124     //! @endcode
125     //!
126     //! This is really just a generalization of the following, more intuitive
127     //! requirements. For all `xs` of data type `S1(X)` and `x` of data type
128     //! `X`,
129     //! @code
130     //!     x ^in^ xs  if and only if  x ^in^ f(xs)
131     //!     find(xs, x) == find(f(xs), x)
132     //! @endcode
133     //!
134     //! These requirements can be understood as saying that `f` does not
135     //! change the content of `xs`, although it may reorder elements.
136     //! As usual, such a structure-preserving transformation is said to
137     //! be an embedding if it is also injective, i.e. if it is a lossless
138     //! transformation.
139     template <typename S>
140     struct Searchable;
141 BOOST_HANA_NAMESPACE_END
142 
143 #endif // !BOOST_HANA_FWD_CONCEPT_SEARCHABLE_HPP
144