feat(libs): add incremental version of murmurhash2 calculation

This does two passes for a given file, which is kinda slow, but I don't
know how else to get the size excluding the filtered ones :<

Signed-off-by: flow <flowlnlnln@gmail.com>
This commit is contained in:
flow 2022-07-23 23:14:49 -03:00
parent 15ec1abb6a
commit f95bcf45ad
No known key found for this signature in database
GPG Key ID: 8D0F221F0A59F469
2 changed files with 106 additions and 82 deletions

View File

@ -1,86 +1,110 @@
//----------------------------------------------------------------------------- //-----------------------------------------------------------------------------
// MurmurHash2 was written by Austin Appleby, and is placed in the public // MurmurHash2 was written by Austin Appleby, and is placed in the public
// domain. The author hereby disclaims copyright to this source code. // domain. The author hereby disclaims copyright to this source code.
//
// Note - This code makes a few assumptions about how your machine behaves - // This was modified as to possibilitate it's usage incrementally.
// Those modifications are also placed in the public domain, and the author of
// 1. We can read a 4-byte value from any address without crashing // such modifications hereby disclaims copyright to this source code.
// 2. sizeof(int) == 4
// And it has a few limitations -
// 1. It will not work incrementally.
// 2. It will not produce the same results on little-endian and big-endian
// machines.
#include "MurmurHash2.h" #include "MurmurHash2.h"
//----------------------------------------------------------------------------- //-----------------------------------------------------------------------------
// Platform-specific functions and macros
// Microsoft Visual Studio // 'm' and 'r' are mixing constants generated offline.
// They're not really 'magic', they just happen to work well.
const uint32_t m = 0x5bd1e995;
const int r = 24;
#if defined(_MSC_VER) uint32_t MurmurHash2(std::ifstream&& file_stream, std::size_t buffer_size, std::function<bool(char)> filter_out)
#define BIG_CONSTANT(x) (x)
// Other compilers
#else // defined(_MSC_VER)
#define BIG_CONSTANT(x) (x##LLU)
#endif // !defined(_MSC_VER)
//-----------------------------------------------------------------------------
uint64_t MurmurHash2 ( const void* key, int len, uint32_t seed )
{ {
// 'm' and 'r' are mixing constants generated offline. auto* buffer = new char[buffer_size];
// They're not really 'magic', they just happen to work well. char data[4];
const uint32_t m = 0x5bd1e995; int read = 0;
const int r = 24; uint32_t size = 0;
// Initialize the hash to a 'random' value // We need the size without the filtered out characters before actually calculating the hash,
// to setup the initial value for the hash.
do {
file_stream.read(buffer, buffer_size);
read = file_stream.gcount();
for (int i = 0; i < read; i++) {
if (!filter_out(buffer[i]))
size += 1;
}
} while (!file_stream.eof());
uint32_t h = seed ^ len; file_stream.clear();
file_stream.seekg(0, file_stream.beg);
// Mix 4 bytes at a time into the hash int index = 0;
const auto* data = (const unsigned char*) key;
while(len >= 4)
{
uint32_t k = *(uint32_t*)data;
k *= m; // This forces a seed of 1.
k ^= k >> r; IncrementalHashInfo info{ (uint32_t)1 ^ size, (uint32_t)size };
k *= m; do {
file_stream.read(buffer, buffer_size);
read = file_stream.gcount();
for (int i = 0; i < read; i++) {
char c = buffer[i];
h *= m; if (filter_out(c))
h ^= k; continue;
data += 4*sizeof(char); data[index] = c;
len -= 4; index = (index + 1) % 4;
}
// Handle the last few bytes of the input array // Mix 4 bytes at a time into the hash
if (index == 0)
FourBytes_MurmurHash2((unsigned char*)&data, info);
}
} while (!file_stream.eof());
switch(len) // Do one last bit shuffle in the hash
{ FourBytes_MurmurHash2((unsigned char*)&data, info);
case 3: h ^= data[2] << 16;
case 2: h ^= data[1] << 8;
case 1: h ^= data[0];
h *= m;
};
// Do a few final mixes of the hash to ensure the last few delete[] buffer;
// bytes are well-incorporated.
h ^= h >> 13; file_stream.close();
h *= m; return info.h;
h ^= h >> 15; }
return h; void FourBytes_MurmurHash2(const unsigned char* data, IncrementalHashInfo& prev)
} {
if (prev.len >= 4) {
// Not the final mix
uint32_t k = *(uint32_t*)data;
k *= m;
k ^= k >> r;
k *= m;
prev.h *= m;
prev.h ^= k;
prev.len -= 4;
} else {
// The final mix
// Handle the last few bytes of the input array
switch (prev.len) {
case 3:
prev.h ^= data[2] << 16;
case 2:
prev.h ^= data[1] << 8;
case 1:
prev.h ^= data[0];
prev.h *= m;
};
// Do a few final mixes of the hash to ensure the last few
// bytes are well-incorporated.
prev.h ^= prev.h >> 13;
prev.h *= m;
prev.h ^= prev.h >> 15;
prev.len = 0;
}
}
//----------------------------------------------------------------------------- //-----------------------------------------------------------------------------

View File

@ -1,30 +1,30 @@
//----------------------------------------------------------------------------- //-----------------------------------------------------------------------------
// MurmurHash2 was written by Austin Appleby, and is placed in the public // The original MurmurHash2 was written by Austin Appleby, and is placed in the
// domain. The author hereby disclaims copyright to this source code. // public domain. The author hereby disclaims copyright to this source code.
//
// This was modified as to possibilitate it's usage incrementally.
// Those modifications are also placed in the public domain, and the author of
// such modifications hereby disclaims copyright to this source code.
#pragma once #pragma once
//----------------------------------------------------------------------------- #include <cstdint>
// Platform-specific functions and macros #include <fstream>
// Microsoft Visual Studio #include <functional>
#if defined(_MSC_VER) && (_MSC_VER < 1600)
typedef unsigned char uint8_t;
typedef unsigned int uint32_t;
typedef unsigned __int64 uint64_t;
// Other compilers
#else // defined(_MSC_VER)
#include <stdint.h>
#endif // !defined(_MSC_VER)
//----------------------------------------------------------------------------- //-----------------------------------------------------------------------------
uint64_t MurmurHash2 ( const void* key, int len, uint32_t seed = 1 ); uint32_t MurmurHash2(
std::ifstream&& file_stream,
std::size_t buffer_size = 4096,
std::function<bool(char)> filter_out = [](char) { return true; });
struct IncrementalHashInfo {
uint32_t h;
uint32_t len;
};
void FourBytes_MurmurHash2(const unsigned char* data, IncrementalHashInfo& prev);
//----------------------------------------------------------------------------- //-----------------------------------------------------------------------------