kakoune/src/ranked_match.cc
Maxime Coste 08df409a53 RankedMatch: Do not compare word boundary match count on single word matches
As the computation of word boundary matches is separate from the
actual subsequence matching, we sometimes have candidate that match
as a single word while still having multiple word boundary matches.

For example, with query "expresins", candidate "expressionism's"
will match as single word ("expressins" is a subsequence of
"expressionism"), and will have two word boundaries match (it does
match the last "s", which is considered as a separate word).

This should not be taken into account when compared against
candidate "expresions", which should be considered a better
match.

Fixes #1925
2018-03-15 23:20:03 +11:00

255 lines
8.0 KiB
C++

#include "ranked_match.hh"
#include "flags.hh"
#include "unit_tests.hh"
#include "utf8_iterator.hh"
#include "optional.hh"
#include <algorithm>
namespace Kakoune
{
UsedLetters used_letters(StringView str)
{
UsedLetters res = 0;
for (auto c : str)
{
if (c >= 'a' and c <= 'z')
res |= 1uLL << (c - 'a');
else if (c >= 'A' and c <= 'Z')
res |= 1uLL << (c - 'A' + 26);
else if (c == '_')
res |= 1uLL << 53;
else if (c == '-')
res |= 1uLL << 54;
else
res |= 1uLL << 63;
}
return res;
}
bool matches(UsedLetters query, UsedLetters letters)
{
return (query & letters) == query;
}
using Utf8It = utf8::iterator<const char*>;
static int count_word_boundaries_match(StringView candidate, StringView query)
{
int count = 0;
Utf8It query_it{query.begin(), query};
Codepoint prev = 0;
for (Utf8It it{candidate.begin(), candidate}; it != candidate.end(); ++it)
{
const Codepoint c = *it;
const bool is_word_boundary = prev == 0 or
(!iswalnum((wchar_t)prev) and iswalnum((wchar_t)c)) or
(iswlower((wchar_t)prev) and iswupper((wchar_t)c));
prev = c;
if (not is_word_boundary)
continue;
const Codepoint lc = to_lower(c);
for (auto qit = query_it; qit != query.end(); ++qit)
{
const Codepoint qc = *qit;
if (qc == (iswlower((wchar_t)qc) ? lc : c))
{
++count;
query_it = qit+1;
break;
}
}
if (query_it == query.end())
break;
}
return count;
}
static bool smartcase_eq(Codepoint candidate, Codepoint query)
{
return query == (iswlower((wchar_t)query) ? to_lower(candidate) : candidate);
}
struct SubseqRes
{
int max_index;
bool single_word;
};
static Optional<SubseqRes> subsequence_match_smart_case(StringView str, StringView subseq)
{
bool single_word = true;
int max_index = -1;
auto it = str.begin();
int index = 0;
for (auto subseq_it = subseq.begin(); subseq_it != subseq.end();)
{
if (it == str.end())
return {};
const Codepoint c = utf8::read_codepoint(subseq_it, subseq.end());
while (true)
{
auto str_c = utf8::read_codepoint(it, str.end());
if (smartcase_eq(str_c, c))
break;
if (max_index != -1 and single_word and not is_word(str_c))
single_word = false;
++index;
if (it == str.end())
return {};
}
max_index = index++;
}
return SubseqRes{max_index, single_word};
}
template<typename TestFunc>
RankedMatch::RankedMatch(StringView candidate, StringView query, TestFunc func)
{
if (candidate.empty() or query.length() > candidate.length())
return;
if (query.empty())
{
m_candidate = candidate;
return;
}
if (not func())
return;
auto res = subsequence_match_smart_case(candidate, query);
if (not res)
return;
m_candidate = candidate;
m_max_index = res->max_index;
if (res->single_word)
m_flags |= Flags::SingleWord;
if (smartcase_eq(candidate[0], query[0]))
m_flags |= Flags::FirstCharMatch;
auto it = std::search(candidate.begin(), candidate.end(),
query.begin(), query.end(), smartcase_eq);
if (it != candidate.end())
{
m_flags |= Flags::Contiguous;
if (it == candidate.begin())
{
m_flags |= Flags::Prefix;
if (query.length() == candidate.length())
{
m_flags |= Flags::SmartFullMatch;
if (candidate == query)
m_flags |= Flags::FullMatch;
}
}
}
m_word_boundary_match_count = count_word_boundaries_match(candidate, query);
if (m_word_boundary_match_count == query.length())
m_flags |= Flags::OnlyWordBoundary;
}
RankedMatch::RankedMatch(StringView candidate, UsedLetters candidate_letters,
StringView query, UsedLetters query_letters)
: RankedMatch{candidate, query, [&] {
return matches(to_lower(query_letters), to_lower(candidate_letters)) and
matches(query_letters & upper_mask, candidate_letters & upper_mask);
}} {}
RankedMatch::RankedMatch(StringView candidate, StringView query)
: RankedMatch{candidate, query, [] { return true; }}
{
}
bool RankedMatch::operator<(const RankedMatch& other) const
{
kak_assert((bool)*this and (bool)other);
const auto diff = m_flags ^ other.m_flags;
// flags are different, use their ordering to return the first match
if (diff != Flags::None)
return (int)(m_flags & diff) > (int)(other.m_flags & diff);
// If we are SingleWord, FirstCharMatch will do the job, and we dont want to take
// other words boundaries into account.
if (not (m_flags & (Flags::Prefix | Flags::SingleWord)) and
m_word_boundary_match_count != other.m_word_boundary_match_count)
return m_word_boundary_match_count > other.m_word_boundary_match_count;
if (m_max_index != other.m_max_index)
return m_max_index < other.m_max_index;
// Reorder codepoints to improve matching behaviour
auto order = [](Codepoint cp) { return cp == '/' ? 0 : cp; };
auto it1 = m_candidate.begin(), it2 = other.m_candidate.begin();
const auto end1 = m_candidate.end(), end2 = other.m_candidate.end();
auto last1 = it1, last2 = it2;
while (true)
{
// find next mismatch
while (it1 != end1 and it2 != end2 and *it1 == *it2)
++it1, ++it2;
if (it1 == end1 or it2 == end2)
return it1 == end1 and it2 != end2;
// compare codepoints
it1 = utf8::character_start(it1, last1);
it2 = utf8::character_start(it2, last2);
const auto cp1 = utf8::read_codepoint(it1, end1);
const auto cp2 = utf8::read_codepoint(it2, end2);
if (cp1 != cp2)
{
const bool punct1 = iswpunct((wchar_t)cp1);
const bool punct2 = iswpunct((wchar_t)cp2);
if (punct1 != punct2)
return punct1;
const bool low1 = iswlower((wchar_t)cp1);
const bool low2 = iswlower((wchar_t)cp2);
if (low1 != low2)
return low1;
return order(cp1) < order(cp2);
}
last1 = it1; last2 = it2;
}
}
UnitTest test_ranked_match{[] {
kak_assert(count_word_boundaries_match("run_all_tests", "rat") == 3);
kak_assert(count_word_boundaries_match("run_all_tests", "at") == 2);
kak_assert(count_word_boundaries_match("countWordBoundariesMatch", "wm") == 2);
kak_assert(count_word_boundaries_match("countWordBoundariesMatch", "cobm") == 3);
kak_assert(count_word_boundaries_match("countWordBoundariesMatch", "cWBM") == 4);
kak_assert(RankedMatch{"source", "so"} < RankedMatch{"source_data", "so"});
kak_assert(not (RankedMatch{"source_data", "so"} < RankedMatch{"source", "so"}));
kak_assert(not (RankedMatch{"source", "so"} < RankedMatch{"source", "so"}));
kak_assert(RankedMatch{"single/word", "wo"} < RankedMatch{"multiw/ord", "wo"});
kak_assert(RankedMatch{"foo/bar/foobar", "foobar"} < RankedMatch{"foo/bar/baz", "foobar"});
kak_assert(RankedMatch{"delete-buffer", "db"} < RankedMatch{"debug", "db"});
kak_assert(RankedMatch{"create_task", "ct"} < RankedMatch{"constructor", "ct"});
kak_assert(RankedMatch{"class", "cla"} < RankedMatch{"class::attr", "cla"});
kak_assert(RankedMatch{"meta/", "meta"} < RankedMatch{"meta-a/", "meta"});
kak_assert(RankedMatch{"find(1p)", "find"} < RankedMatch{"findfs(8)", "find"});
kak_assert(RankedMatch{"expresions", "expresins"} < RankedMatch{"expressionism's", "expresins"});
}};
UnitTest test_used_letters{[]()
{
kak_assert(used_letters("abcd") == to_lower(used_letters("abcdABCD")));
}};
}