LBRYCommunity/lbrycrd
1
0
Fork 0
Code Issues 29 Pull requests 1 Projects Releases 32 Packages Wiki Activity

blockfilter: Refactor GCS params into struct.

Browse source
This commit is contained in:
Jim Posen 2018-08-27 15:04:43 -07:00
parent 6af27b8157
commit 20b812993a
4 changed files with 66 additions and 41 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
src/bench/gcs_filter.cpp
View file

@ -17,7 +17,7 @@ static void ConstructGCSFilter(benchmark::State& state)
uint64_t siphash_k0 = 0;
while (state.KeepRunning()) {
GCSFilter filter(siphash_k0, 0, 20, 1 << 20, elements);
GCSFilter filter({siphash_k0, 0, 20, 1 << 20}, elements);
siphash_k0++;
}
@ -32,7 +32,7 @@ static void MatchGCSFilter(benchmark::State& state)
element[1] = static_cast<unsigned char>(i >> 8);
elements.insert(std::move(element));
}
GCSFilter filter(0, 0, 20, 1 << 20, elements);
GCSFilter filter({0, 0, 20, 1 << 20}, elements);
while (state.KeepRunning()) {
filter.Match(GCSFilter::Element());

48
src/blockfilter.cpp
View file

@ -79,7 +79,7 @@ static uint64_t MapIntoRange(uint64_t x, uint64_t n)
uint64_t GCSFilter::HashToRange(const Element& element) const
{
uint64_t hash = CSipHasher(m_siphash_k0, m_siphash_k1)
uint64_t hash = CSipHasher(m_params.m_siphash_k0, m_params.m_siphash_k1)
.Write(element.data(), element.size())
.Finalize();
return MapIntoRange(hash, m_F);
@ -96,16 +96,13 @@ std::vector<uint64_t> GCSFilter::BuildHashedSet(const ElementSet& elements) cons
return hashed_elements;
}
GCSFilter::GCSFilter(uint64_t siphash_k0, uint64_t siphash_k1, uint8_t P, uint32_t M)
: m_siphash_k0(siphash_k0), m_siphash_k1(siphash_k1), m_P(P), m_M(M), m_N(0), m_F(0)
GCSFilter::GCSFilter(const Params& params)
: m_params(params), m_N(0), m_F(0), m_encoded{0}
{}
GCSFilter::GCSFilter(uint64_t siphash_k0, uint64_t siphash_k1, uint8_t P, uint32_t M,
std::vector<unsigned char> encoded_filter)
: GCSFilter(siphash_k0, siphash_k1, P, M)
GCSFilter::GCSFilter(const Params& params, std::vector<unsigned char> encoded_filter)
: m_params(params), m_encoded(std::move(encoded_filter))
{
m_encoded = std::move(encoded_filter);
VectorReader stream(GCS_SER_TYPE, GCS_SER_VERSION, m_encoded, 0);
uint64_t N = ReadCompactSize(stream);
@ -113,29 +110,28 @@ GCSFilter::GCSFilter(uint64_t siphash_k0, uint64_t siphash_k1, uint8_t P, uint32
if (m_N != N) {
throw std::ios_base::failure("N must be <2^32");
}
m_F = static_cast<uint64_t>(m_N) * static_cast<uint64_t>(m_M);
m_F = static_cast<uint64_t>(m_N) * static_cast<uint64_t>(m_params.m_M);
// Verify that the encoded filter contains exactly N elements. If it has too much or too little
// data, a std::ios_base::failure exception will be raised.
BitStreamReader<VectorReader> bitreader(stream);
for (uint64_t i = 0; i < m_N; ++i) {
GolombRiceDecode(bitreader, m_P);
GolombRiceDecode(bitreader, m_params.m_P);
}
if (!stream.empty()) {
throw std::ios_base::failure("encoded_filter contains excess data");
}
}
GCSFilter::GCSFilter(uint64_t siphash_k0, uint64_t siphash_k1, uint8_t P, uint32_t M,
const ElementSet& elements)
: GCSFilter(siphash_k0, siphash_k1, P, M)
GCSFilter::GCSFilter(const Params& params, const ElementSet& elements)
: m_params(params)
{
size_t N = elements.size();
m_N = static_cast<uint32_t>(N);
if (m_N != N) {
throw std::invalid_argument("N must be <2^32");
}
m_F = static_cast<uint64_t>(m_N) * static_cast<uint64_t>(m_M);
m_F = static_cast<uint64_t>(m_N) * static_cast<uint64_t>(m_params.m_M);
CVectorWriter stream(GCS_SER_TYPE, GCS_SER_VERSION, m_encoded, 0);
@ -150,7 +146,7 @@ GCSFilter::GCSFilter(uint64_t siphash_k0, uint64_t siphash_k1, uint8_t P, uint32
uint64_t last_value = 0;
for (uint64_t value : BuildHashedSet(elements)) {
uint64_t delta = value - last_value;
GolombRiceEncode(bitwriter, m_P, delta);
GolombRiceEncode(bitwriter, m_params.m_P, delta);
last_value = value;
}
@ -170,7 +166,7 @@ bool GCSFilter::MatchInternal(const uint64_t* element_hashes, size_t size) const
uint64_t value = 0;
size_t hashes_index = 0;
for (uint32_t i = 0; i < m_N; ++i) {
uint64_t delta = GolombRiceDecode(bitreader, m_P);
uint64_t delta = GolombRiceDecode(bitreader, m_params.m_P);
value += delta;
while (true) {
@ -227,17 +223,29 @@ static GCSFilter::ElementSet BasicFilterElements(const CBlock& block,
BlockFilter::BlockFilter(BlockFilterType filter_type, const CBlock& block, const CBlockUndo& block_undo)
: m_filter_type(filter_type), m_block_hash(block.GetHash())
{
GCSFilter::Params params;
if (!BuildParams(params)) {
throw std::invalid_argument("unknown filter_type");
}
m_filter = GCSFilter(params, BasicFilterElements(block, block_undo));
}
bool BlockFilter::BuildParams(GCSFilter::Params& params) const
{
switch (m_filter_type) {
case BlockFilterType::BASIC:
m_filter = GCSFilter(m_block_hash.GetUint64(0), m_block_hash.GetUint64(1),
BASIC_FILTER_P, BASIC_FILTER_M,
BasicFilterElements(block, block_undo));
params.m_siphash_k0 = m_block_hash.GetUint64(0);
params.m_siphash_k1 = m_block_hash.GetUint64(1);
params.m_P = BASIC_FILTER_P;
params.m_M = BASIC_FILTER_M;
break;
default:
throw std::invalid_argument("unknown filter_type");
return false;
}
return true;
}
uint256 BlockFilter::GetHash() const

34
src/blockfilter.h
View file

@ -25,11 +25,20 @@ public:
typedef std::vector<unsigned char> Element;
typedef std::unordered_set<Element, ByteVectorHash> ElementSet;
private:
struct Params
{
uint64_t m_siphash_k0;
uint64_t m_siphash_k1;
uint8_t m_P; //!< Golomb-Rice coding parameter
uint32_t m_M; //!< Inverse false positive rate
Params(uint64_t siphash_k0 = 0, uint64_t siphash_k1 = 0, uint8_t P = 0, uint32_t M = 1)
: m_siphash_k0(siphash_k0), m_siphash_k1(siphash_k1), m_P(P), m_M(M)
{}
};
private:
Params m_params;
uint32_t m_N; //!< Number of elements in the filter
uint64_t m_F; //!< Range of element hashes, F = N * M
std::vector<unsigned char> m_encoded;
@ -45,19 +54,16 @@ private:
public:
/** Constructs an empty filter. */
GCSFilter(uint64_t siphash_k0 = 0, uint64_t siphash_k1 = 0, uint8_t P = 0, uint32_t M = 0);
explicit GCSFilter(const Params& params = Params());
/** Reconstructs an already-created filter from an encoding. */
GCSFilter(uint64_t siphash_k0, uint64_t siphash_k1, uint8_t P, uint32_t M,
std::vector<unsigned char> encoded_filter);
GCSFilter(const Params& params, std::vector<unsigned char> encoded_filter);
/** Builds a new filter from the params and set of elements. */
GCSFilter(uint64_t siphash_k0, uint64_t siphash_k1, uint8_t P, uint32_t M,
const ElementSet& elements);
GCSFilter(const Params& params, const ElementSet& elements);
uint8_t GetP() const { return m_P; }
uint32_t GetN() const { return m_N; }
uint32_t GetM() const { return m_M; }
const Params& GetParams() const { return m_params; }
const std::vector<unsigned char>& GetEncoded() const { return m_encoded; }
/**
@ -93,6 +99,8 @@ private:
uint256 m_block_hash;
GCSFilter m_filter;
bool BuildParams(GCSFilter::Params& params) const;
public:
// Construct a new BlockFilter of the specified type from a block.
@ -131,15 +139,11 @@ public:
m_filter_type = static_cast<BlockFilterType>(filter_type);
switch (m_filter_type) {
case BlockFilterType::BASIC:
m_filter = GCSFilter(m_block_hash.GetUint64(0), m_block_hash.GetUint64(1),
BASIC_FILTER_P, BASIC_FILTER_M, std::move(encoded_filter));
break;
default:
GCSFilter::Params params;
if (!BuildParams(params)) {
throw std::ios_base::failure("unknown filter_type");
}
m_filter = GCSFilter(params, std::move(encoded_filter));
}
};

15
src/test/blockfilter_tests.cpp
View file

@ -29,7 +29,7 @@ BOOST_AUTO_TEST_CASE(gcsfilter_test)
excluded_elements.insert(std::move(element2));
}
GCSFilter filter(0, 0, 10, 1 << 10, included_elements);
GCSFilter filter({0, 0, 10, 1 << 10}, included_elements);
for (const auto& element : included_elements) {
BOOST_CHECK(filter.Match(element));
@ -39,6 +39,19 @@ BOOST_AUTO_TEST_CASE(gcsfilter_test)
}
}
BOOST_AUTO_TEST_CASE(gcsfilter_default_constructor)
{
GCSFilter filter;
BOOST_CHECK_EQUAL(filter.GetN(), 0);
BOOST_CHECK_EQUAL(filter.GetEncoded().size(), 1);
const GCSFilter::Params& params = filter.GetParams();
BOOST_CHECK_EQUAL(params.m_siphash_k0, 0);
BOOST_CHECK_EQUAL(params.m_siphash_k1, 0);
BOOST_CHECK_EQUAL(params.m_P, 0);
BOOST_CHECK_EQUAL(params.m_M, 1);
}
BOOST_AUTO_TEST_CASE(blockfilter_basic_test)
{
CScript included_scripts[5], excluded_scripts[3];