ClickHouse/src/Functions/FunctionsStringSimilarity.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

547 lines
21 KiB
C++
Raw Normal View History

#include <Functions/FunctionsStringSimilarity.h>
#include <Functions/FunctionFactory.h>
#include <Functions/FunctionsHashing.h>
#include <Common/HashTable/ClearableHashMap.h>
#include <Common/HashTable/Hash.h>
#include <Common/UTF8Helpers.h>
#include <Core/Defines.h>
2021-10-02 07:13:14 +00:00
#include <base/unaligned.h>
#include <algorithm>
#include <climits>
#include <cstring>
#include <limits>
#include <memory>
#include <utility>
#ifdef __SSE4_2__
# include <nmmintrin.h>
#endif
#if defined(__aarch64__) && defined(__ARM_FEATURE_CRC32)
# include <arm_acle.h>
#endif
namespace DB
{
/** Distance function implementation.
* We calculate all the n-grams from left string and count by the index of
* 16 bits hash of them in the map.
* Then calculate all the n-grams from the right string and calculate
* the n-gram distance on the flight by adding and subtracting from the hashmap.
* Then return the map into the condition of which it was after the left string
* calculation. If the right string size is big (more than 2**15 bytes),
* the strings are not similar at all and we return 1.
*/
2019-10-25 03:25:02 +00:00
template <size_t N, class CodePoint, bool UTF8, bool case_insensitive, bool symmetric>
struct NgramDistanceImpl
{
using ResultType = Float32;
/// map_size for ngram difference.
2019-02-22 03:50:06 +00:00
static constexpr size_t map_size = 1u << 16;
/// If the haystack size is bigger than this, behaviour is unspecified for this function.
2019-02-22 03:50:06 +00:00
static constexpr size_t max_string_size = 1u << 15;
/// Default padding to read safely.
static constexpr size_t default_padding = 16;
/// Max codepoints to store at once. 16 is for batching usage and PODArray has this padding.
static constexpr size_t simultaneously_codepoints_num = default_padding + N - 1;
2020-06-08 17:35:45 +00:00
/** map_size of this fits mostly in L2 cache all the time.
* Actually use UInt16 as addings and subtractions do not UB overflow. But think of it as a signed
* integer array.
*/
2020-06-08 17:35:45 +00:00
using NgramCount = UInt16;
static ALWAYS_INLINE UInt16 calculateASCIIHash(const CodePoint * code_points)
{
return intHashCRC32(unalignedLoad<UInt32>(code_points)) & 0xFFFFu;
}
static ALWAYS_INLINE UInt16 calculateUTF8Hash(const CodePoint * code_points)
{
UInt64 combined = (static_cast<UInt64>(code_points[0]) << 32) | code_points[1];
#ifdef __SSE4_2__
return _mm_crc32_u64(code_points[2], combined) & 0xFFFFu;
#elif defined(__aarch64__) && defined(__ARM_FEATURE_CRC32)
return __crc32cd(code_points[2], combined) & 0xFFFFu;
#else
return (intHashCRC32(combined) ^ intHashCRC32(code_points[2])) & 0xFFFFu;
#endif
}
template <size_t Offset, class Container, size_t... I>
static ALWAYS_INLINE inline void unrollLowering(Container & cont, const std::index_sequence<I...> &)
{
((cont[Offset + I] = std::tolower(cont[Offset + I])), ...);
}
static ALWAYS_INLINE size_t readASCIICodePoints(CodePoint * code_points, const char *& pos, const char * end)
{
/// Offset before which we copy some data.
constexpr size_t padding_offset = default_padding - N + 1;
/// We have an array like this for ASCII (N == 4, other cases are similar)
/// |a0|a1|a2|a3|a4|a5|a6|a7|a8|a9|a10|a11|a12|a13|a14|a15|a16|a17|a18|
/// And we copy ^^^^^^^^^^^^^^^ these bytes to the start
/// Actually it is enough to copy 3 bytes, but memcpy for 4 bytes translates into 1 instruction
memcpy(code_points, code_points + padding_offset, roundUpToPowerOfTwoOrZero(N - 1) * sizeof(CodePoint));
/// Now we have an array
/// |a13|a14|a15|a16|a4|a5|a6|a7|a8|a9|a10|a11|a12|a13|a14|a15|a16|a17|a18|
/// ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
/// Doing unaligned read of 16 bytes and copy them like above
/// 16 is also chosen to do two `movups`.
2019-03-05 23:08:08 +00:00
/// Such copying allow us to have 3 codepoints from the previous read to produce the 4-grams with them.
memcpy(code_points + (N - 1), pos, default_padding * sizeof(CodePoint));
2019-10-25 03:25:02 +00:00
if constexpr (case_insensitive)
2019-02-22 03:07:35 +00:00
{
/// We really need template lambdas with C++20 to do it inline
unrollLowering<N - 1>(code_points, std::make_index_sequence<padding_offset>());
}
pos += padding_offset;
if (pos > end)
return default_padding - (pos - end);
return default_padding;
}
static ALWAYS_INLINE size_t readUTF8CodePoints(CodePoint * code_points, const char *& pos, const char * end)
{
/// The same copying as described in the function above.
memcpy(code_points, code_points + default_padding - N + 1, roundUpToPowerOfTwoOrZero(N - 1) * sizeof(CodePoint));
size_t num = N - 1;
while (num < default_padding && pos < end)
{
size_t length = UTF8::seqLength(*pos);
if (pos + length > end)
length = end - pos;
CodePoint res;
/// This is faster than just memcpy because of compiler optimizations with moving bytes.
switch (length)
{
case 1:
res = 0;
memcpy(&res, pos, 1);
break;
case 2:
res = 0;
memcpy(&res, pos, 2);
break;
case 3:
res = 0;
memcpy(&res, pos, 3);
break;
default:
memcpy(&res, pos, 4);
}
/// This is not a really true case insensitive utf8. We zero the 5-th bit of every byte.
2019-05-25 18:47:26 +00:00
/// And first bit of first byte if there are two bytes.
2020-06-27 19:05:00 +00:00
/// For ASCII it works https://catonmat.net/ascii-case-conversion-trick. For most cyrillic letters also does.
/// For others, we don't care now. Lowering UTF is not a cheap operation.
2019-10-25 03:25:02 +00:00
if constexpr (case_insensitive)
{
switch (length)
{
case 4:
res &= ~(1u << (5 + 3 * CHAR_BIT));
[[fallthrough]];
case 3:
res &= ~(1u << (5 + 2 * CHAR_BIT));
[[fallthrough]];
case 2:
2019-05-25 18:47:26 +00:00
res &= ~(1u);
res &= ~(1u << (5 + CHAR_BIT));
[[fallthrough]];
default:
res &= ~(1u << 5);
}
}
pos += length;
code_points[num++] = res;
}
return num;
}
2019-10-25 03:25:02 +00:00
template <bool save_ngrams>
static ALWAYS_INLINE inline size_t calculateNeedleStats(
const char * data,
const size_t size,
2020-06-08 17:35:45 +00:00
NgramCount * ngram_stats,
[[maybe_unused]] NgramCount * ngram_storage,
size_t (*read_code_points)(CodePoint *, const char *&, const char *),
UInt16 (*hash_functor)(const CodePoint *))
{
const char * start = data;
const char * end = data + size;
CodePoint cp[simultaneously_codepoints_num] = {};
/// read_code_points returns the position of cp where it stopped reading codepoints.
size_t found = read_code_points(cp, start, end);
/// We need to start for the first time here, because first N - 1 codepoints mean nothing.
size_t i = N - 1;
2019-05-05 20:48:46 +00:00
size_t len = 0;
do
{
for (; i + N <= found; ++i)
2019-05-05 20:48:46 +00:00
{
++len;
UInt16 hash = hash_functor(cp + i);
2019-10-25 03:25:02 +00:00
if constexpr (save_ngrams)
2019-05-05 20:48:46 +00:00
*ngram_storage++ = hash;
++ngram_stats[hash];
}
i = 0;
} while (start < end && (found = read_code_points(cp, start, end)));
return len;
}
2019-10-25 03:25:02 +00:00
template <bool reuse_stats>
static ALWAYS_INLINE inline UInt64 calculateHaystackStatsAndMetric(
const char * data,
const size_t size,
2020-06-08 17:35:45 +00:00
NgramCount * ngram_stats,
size_t & distance,
2019-05-05 20:48:46 +00:00
[[maybe_unused]] UInt16 * ngram_storage,
size_t (*read_code_points)(CodePoint *, const char *&, const char *),
UInt16 (*hash_functor)(const CodePoint *))
{
size_t ngram_cnt = 0;
const char * start = data;
const char * end = data + size;
CodePoint cp[simultaneously_codepoints_num] = {};
/// read_code_points returns the position of cp where it stopped reading codepoints.
size_t found = read_code_points(cp, start, end);
/// We need to start for the first time here, because first N - 1 codepoints mean nothing.
size_t iter = N - 1;
do
{
for (; iter + N <= found; ++iter)
{
UInt16 hash = hash_functor(cp + iter);
2019-05-25 18:47:26 +00:00
/// For symmetric version we should add when we can't subtract to get symmetric difference.
if (static_cast<Int16>(ngram_stats[hash]) > 0)
--distance;
2019-10-25 03:25:02 +00:00
else if constexpr (symmetric)
++distance;
2019-10-25 03:25:02 +00:00
if constexpr (reuse_stats)
2019-05-05 20:48:46 +00:00
ngram_storage[ngram_cnt] = hash;
++ngram_cnt;
--ngram_stats[hash];
}
iter = 0;
} while (start < end && (found = read_code_points(cp, start, end)));
/// Return the state of hash map to its initial.
2019-10-25 03:25:02 +00:00
if constexpr (reuse_stats)
2019-05-05 20:48:46 +00:00
{
for (size_t i = 0; i < ngram_cnt; ++i)
++ngram_stats[ngram_storage[i]];
}
return ngram_cnt;
}
template <class Callback, class... Args>
2019-05-05 20:48:46 +00:00
static inline auto dispatchSearcher(Callback callback, Args &&... args)
{
if constexpr (!UTF8)
return callback(std::forward<Args>(args)..., readASCIICodePoints, calculateASCIIHash);
else
return callback(std::forward<Args>(args)..., readUTF8CodePoints, calculateUTF8Hash);
}
static void constantConstant(std::string data, std::string needle, Float32 & res)
{
2020-06-08 17:35:45 +00:00
std::unique_ptr<NgramCount[]> common_stats{new NgramCount[map_size]{}};
/// We use unsafe versions of getting ngrams, so I decided to use padded strings.
const size_t needle_size = needle.size();
const size_t data_size = data.size();
needle.resize(needle_size + default_padding);
data.resize(data_size + default_padding);
2020-06-08 17:35:45 +00:00
size_t second_size = dispatchSearcher(calculateNeedleStats<false>, needle.data(), needle_size, common_stats.get(), nullptr);
size_t distance = second_size;
if (data_size <= max_string_size)
2019-02-22 03:07:35 +00:00
{
2020-06-08 17:35:45 +00:00
size_t first_size = dispatchSearcher(calculateHaystackStatsAndMetric<false>, data.data(), data_size, common_stats.get(), distance, nullptr);
2019-10-25 03:25:02 +00:00
/// For !symmetric version we should not use first_size.
if constexpr (symmetric)
res = distance * 1.f / std::max(first_size + second_size, static_cast<size_t>(1));
else
res = 1.f - distance * 1.f / std::max(second_size, static_cast<size_t>(1));
2019-02-22 03:07:35 +00:00
}
else
{
2019-10-25 03:25:02 +00:00
if constexpr (symmetric)
res = 1.f;
else
res = 0.f;
}
}
static void vectorVector(
2019-05-05 20:48:46 +00:00
const ColumnString::Chars & haystack_data,
const ColumnString::Offsets & haystack_offsets,
const ColumnString::Chars & needle_data,
const ColumnString::Offsets & needle_offsets,
PaddedPODArray<Float32> & res)
{
const size_t haystack_offsets_size = haystack_offsets.size();
size_t prev_haystack_offset = 0;
size_t prev_needle_offset = 0;
2020-06-08 17:35:45 +00:00
std::unique_ptr<NgramCount[]> common_stats{new NgramCount[map_size]{}};
2019-05-05 20:48:46 +00:00
/// The main motivation is to not allocate more on stack because we have already allocated a lot (128Kb).
/// And we can reuse these storages in one thread because we care only about what was written to first places.
std::unique_ptr<UInt16[]> needle_ngram_storage(new UInt16[max_string_size]);
std::unique_ptr<UInt16[]> haystack_ngram_storage(new UInt16[max_string_size]);
2019-05-05 20:48:46 +00:00
for (size_t i = 0; i < haystack_offsets_size; ++i)
{
const char * haystack = reinterpret_cast<const char *>(&haystack_data[prev_haystack_offset]);
const size_t haystack_size = haystack_offsets[i] - prev_haystack_offset - 1;
const char * needle = reinterpret_cast<const char *>(&needle_data[prev_needle_offset]);
const size_t needle_size = needle_offsets[i] - prev_needle_offset - 1;
if (needle_size <= max_string_size && haystack_size <= max_string_size)
{
/// Get needle stats.
const size_t needle_stats_size = dispatchSearcher(
calculateNeedleStats<true>,
needle,
needle_size,
2020-06-08 17:35:45 +00:00
common_stats.get(),
2019-05-05 20:48:46 +00:00
needle_ngram_storage.get());
size_t distance = needle_stats_size;
/// Combine with haystack stats, return to initial needle stats.
const size_t haystack_stats_size = dispatchSearcher(
calculateHaystackStatsAndMetric<true>,
haystack,
haystack_size,
2020-06-08 17:35:45 +00:00
common_stats.get(),
2019-05-05 20:48:46 +00:00
distance,
haystack_ngram_storage.get());
/// Return to zero array stats.
for (size_t j = 0; j < needle_stats_size; ++j)
--common_stats[needle_ngram_storage[j]];
/// For now, common stats is a zero array.
2019-05-25 18:47:26 +00:00
2019-10-25 03:25:02 +00:00
/// For !symmetric version we should not use haystack_stats_size.
if constexpr (symmetric)
res[i] = distance * 1.f / std::max(haystack_stats_size + needle_stats_size, static_cast<size_t>(1));
else
res[i] = 1.f - distance * 1.f / std::max(needle_stats_size, static_cast<size_t>(1));
2019-05-05 20:48:46 +00:00
}
else
{
2019-05-05 21:37:58 +00:00
/// Strings are too big, we are assuming they are not the same. This is done because of limiting number
2019-05-05 20:48:46 +00:00
/// of bigrams added and not allocating too much memory.
2019-10-25 03:25:02 +00:00
if constexpr (symmetric)
res[i] = 1.f;
else
res[i] = 0.f;
2019-05-05 20:48:46 +00:00
}
prev_needle_offset = needle_offsets[i];
prev_haystack_offset = haystack_offsets[i];
}
}
static void constantVector(
2019-05-25 18:47:26 +00:00
std::string haystack,
const ColumnString::Chars & needle_data,
const ColumnString::Offsets & needle_offsets,
PaddedPODArray<Float32> & res)
{
/// For symmetric version it is better to use vector_constant
2019-10-25 03:25:02 +00:00
if constexpr (symmetric)
2019-05-25 18:47:26 +00:00
{
vectorConstant(needle_data, needle_offsets, std::move(haystack), res);
2019-05-25 18:47:26 +00:00
}
else
{
const size_t haystack_size = haystack.size();
haystack.resize(haystack_size + default_padding);
/// For logic explanation see vector_vector function.
const size_t needle_offsets_size = needle_offsets.size();
size_t prev_offset = 0;
2020-06-08 17:35:45 +00:00
std::unique_ptr<NgramCount[]> common_stats{new NgramCount[map_size]{}};
2019-05-25 18:47:26 +00:00
std::unique_ptr<UInt16[]> needle_ngram_storage(new UInt16[max_string_size]);
std::unique_ptr<UInt16[]> haystack_ngram_storage(new UInt16[max_string_size]);
for (size_t i = 0; i < needle_offsets_size; ++i)
{
const char * needle = reinterpret_cast<const char *>(&needle_data[prev_offset]);
const size_t needle_size = needle_offsets[i] - prev_offset - 1;
if (needle_size <= max_string_size && haystack_size <= max_string_size)
{
const size_t needle_stats_size = dispatchSearcher(
calculateNeedleStats<true>,
needle,
needle_size,
2020-06-08 17:35:45 +00:00
common_stats.get(),
2019-05-25 18:47:26 +00:00
needle_ngram_storage.get());
size_t distance = needle_stats_size;
dispatchSearcher(
calculateHaystackStatsAndMetric<true>,
haystack.data(),
haystack_size,
2020-06-08 17:35:45 +00:00
common_stats.get(),
2019-05-25 18:47:26 +00:00
distance,
haystack_ngram_storage.get());
for (size_t j = 0; j < needle_stats_size; ++j)
--common_stats[needle_ngram_storage[j]];
res[i] = 1.f - distance * 1.f / std::max(needle_stats_size, static_cast<size_t>(1));
2019-05-25 18:47:26 +00:00
}
else
{
res[i] = 0.f;
2019-05-25 18:47:26 +00:00
}
prev_offset = needle_offsets[i];
}
}
}
static void vectorConstant(
2019-05-05 20:48:46 +00:00
const ColumnString::Chars & data,
const ColumnString::Offsets & offsets,
std::string needle,
PaddedPODArray<Float32> & res)
{
/// zeroing our map
2020-06-08 17:35:45 +00:00
std::unique_ptr<NgramCount[]> common_stats{new NgramCount[map_size]{}};
2020-06-08 17:35:45 +00:00
/// We can reuse these storages in one thread because we care only about what was written to first places.
std::unique_ptr<UInt16[]> ngram_storage(new NgramCount[max_string_size]);
/// We use unsafe versions of getting ngrams, so I decided to use padded_data even in needle case.
const size_t needle_size = needle.size();
needle.resize(needle_size + default_padding);
2020-06-08 17:35:45 +00:00
const size_t needle_stats_size = dispatchSearcher(calculateNeedleStats<false>, needle.data(), needle_size, common_stats.get(), nullptr);
size_t distance = needle_stats_size;
size_t prev_offset = 0;
for (size_t i = 0; i < offsets.size(); ++i)
{
const UInt8 * haystack = &data[prev_offset];
const size_t haystack_size = offsets[i] - prev_offset - 1;
2019-02-22 03:50:06 +00:00
if (haystack_size <= max_string_size)
2019-02-22 03:07:35 +00:00
{
size_t haystack_stats_size = dispatchSearcher(
2019-05-05 20:48:46 +00:00
calculateHaystackStatsAndMetric<true>,
reinterpret_cast<const char *>(haystack),
2020-06-08 17:35:45 +00:00
haystack_size, common_stats.get(),
2019-05-05 20:48:46 +00:00
distance,
ngram_storage.get());
2019-10-25 03:25:02 +00:00
/// For !symmetric version we should not use haystack_stats_size.
if constexpr (symmetric)
res[i] = distance * 1.f / std::max(haystack_stats_size + needle_stats_size, static_cast<size_t>(1));
else
res[i] = 1.f - distance * 1.f / std::max(needle_stats_size, static_cast<size_t>(1));
2019-02-22 03:07:35 +00:00
}
else
{
/// if the strings are too big, we say they are completely not the same
2019-10-25 03:25:02 +00:00
if constexpr (symmetric)
res[i] = 1.f;
else
res[i] = 0.f;
}
distance = needle_stats_size;
prev_offset = offsets[i];
}
}
};
2019-03-29 01:02:05 +00:00
struct NameNgramDistance
{
static constexpr auto name = "ngramDistance";
};
2019-03-29 01:02:05 +00:00
struct NameNgramDistanceCaseInsensitive
{
static constexpr auto name = "ngramDistanceCaseInsensitive";
};
2019-03-29 01:02:05 +00:00
struct NameNgramDistanceUTF8
{
static constexpr auto name = "ngramDistanceUTF8";
};
2019-03-29 01:02:05 +00:00
struct NameNgramDistanceUTF8CaseInsensitive
{
static constexpr auto name = "ngramDistanceCaseInsensitiveUTF8";
};
2019-05-27 09:05:02 +00:00
struct NameNgramSearch
2019-05-25 18:47:26 +00:00
{
2019-05-27 09:05:02 +00:00
static constexpr auto name = "ngramSearch";
2019-05-25 18:47:26 +00:00
};
2019-05-27 09:05:02 +00:00
struct NameNgramSearchCaseInsensitive
2019-05-25 18:47:26 +00:00
{
2019-05-27 09:05:02 +00:00
static constexpr auto name = "ngramSearchCaseInsensitive";
2019-05-25 18:47:26 +00:00
};
2019-05-27 09:05:02 +00:00
struct NameNgramSearchUTF8
2019-05-25 18:47:26 +00:00
{
2019-05-27 09:05:02 +00:00
static constexpr auto name = "ngramSearchUTF8";
2019-05-25 18:47:26 +00:00
};
2019-05-27 09:05:02 +00:00
struct NameNgramSearchUTF8CaseInsensitive
2019-05-25 18:47:26 +00:00
{
2019-05-27 09:05:02 +00:00
static constexpr auto name = "ngramSearchCaseInsensitiveUTF8";
2019-05-25 18:47:26 +00:00
};
using FunctionNgramDistance = FunctionsStringSimilarity<NgramDistanceImpl<4, UInt8, false, false, true>, NameNgramDistance>;
using FunctionNgramDistanceCaseInsensitive = FunctionsStringSimilarity<NgramDistanceImpl<4, UInt8, false, true, true>, NameNgramDistanceCaseInsensitive>;
using FunctionNgramDistanceUTF8 = FunctionsStringSimilarity<NgramDistanceImpl<3, UInt32, true, false, true>, NameNgramDistanceUTF8>;
using FunctionNgramDistanceCaseInsensitiveUTF8 = FunctionsStringSimilarity<NgramDistanceImpl<3, UInt32, true, true, true>, NameNgramDistanceUTF8CaseInsensitive>;
2019-05-27 09:05:02 +00:00
using FunctionNgramSearch = FunctionsStringSimilarity<NgramDistanceImpl<4, UInt8, false, false, false>, NameNgramSearch>;
using FunctionNgramSearchCaseInsensitive = FunctionsStringSimilarity<NgramDistanceImpl<4, UInt8, false, true, false>, NameNgramSearchCaseInsensitive>;
using FunctionNgramSearchUTF8 = FunctionsStringSimilarity<NgramDistanceImpl<3, UInt32, true, false, false>, NameNgramSearchUTF8>;
using FunctionNgramSearchCaseInsensitiveUTF8 = FunctionsStringSimilarity<NgramDistanceImpl<3, UInt32, true, true, false>, NameNgramSearchUTF8CaseInsensitive>;
2019-05-25 18:47:26 +00:00
REGISTER_FUNCTION(StringSimilarity)
{
factory.registerFunction<FunctionNgramDistance>();
factory.registerFunction<FunctionNgramDistanceCaseInsensitive>();
factory.registerFunction<FunctionNgramDistanceUTF8>();
factory.registerFunction<FunctionNgramDistanceCaseInsensitiveUTF8>();
2019-05-25 18:47:26 +00:00
2019-05-27 09:05:02 +00:00
factory.registerFunction<FunctionNgramSearch>();
factory.registerFunction<FunctionNgramSearchCaseInsensitive>();
factory.registerFunction<FunctionNgramSearchUTF8>();
factory.registerFunction<FunctionNgramSearchCaseInsensitiveUTF8>();
}
}