pub_soft/rdkit
3
0
Fork 0
mirror of https://github.com/rdkit/rdkit.git synced 2026-06-03 21:44:30 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fixes #9033 (#9036)

Browse Source 
Also fixes the other similarity metrics.
A very small amount of refactoring

Co-authored-by: = <=>
This commit is contained in:
Greg Landrum
2026-01-03 06:42:29 +01:00
committed by GitHub
parent f8b6776af1
commit e0f3ada0d7
2 changed files with 36 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
Code/DataStructs/BitOps.cpp
View File

@@ -290,7 +290,7 @@ double TanimotoSimilarity(const T1 &bv1, const T2 &bv2) {
}
unsigned int total = bv1.getNumOnBits() + bv2.getNumOnBits();
if (total == 0) {
return 1.0;
return 0.0;
}
unsigned int common = NumOnBitsInCommon(bv1, bv2);
return (double)common / (double)(total - common);
@@ -304,8 +304,11 @@ double TverskySimilarity(const T1 &bv1, const T2 &bv2, double a, double b) {
throw ValueErrorException("BitVects must be same length");
}
double x = NumOnBitsInCommon(bv1, bv2);
double y = bv1.getNumOnBits();
double z = bv2.getNumOnBits();
auto y = bv1.getNumOnBits();
auto z = bv2.getNumOnBits();
if (y == 0 || z == 0) {
return 0.0;
}
double denom = a * y + b * z + (1 - a - b) * x;
if (denom == 0.0) {
return 1.0;
@@ -368,10 +371,13 @@ double SokalSimilarity(const T1 &bv1, const T2 &bv2) {
throw ValueErrorException("BitVects must be same length");
}
double x = NumOnBitsInCommon(bv1, bv2);
double y = bv1.getNumOnBits();
double z = bv2.getNumOnBits();
auto y = bv1.getNumOnBits();
auto z = bv2.getNumOnBits();
if (y == 0 || z == 0) {
return 0.0;
}
return x / (2 * y + 2 * z - 3 * x);
return x / (2. * y + 2. * z - 3. * x);
}
template <typename T1, typename T2>
@@ -390,16 +396,6 @@ double McConnaugheySimilarity(const T1 &bv1, const T2 &bv2) {
}
}
template <typename T>
inline T tmin(T v1, T v2) {
return std::min(v2, v1);
}
template <typename T>
inline T tmax(T v1, T v2) {
return std::max(v2, v1);
}
template <typename T1, typename T2>
double AsymmetricSimilarity(const T1 &bv1, const T2 &bv2) {
if (bv1.getNumBits() != bv2.getNumBits()) {
@@ -409,7 +405,7 @@ double AsymmetricSimilarity(const T1 &bv1, const T2 &bv2) {
double y = bv1.getNumOnBits();
double z = bv2.getNumOnBits();
double min = tmin(y, z);
double min = std::min(y, z);
if (min > 0.0) {
return x / min;
} else {
@@ -426,7 +422,7 @@ double BraunBlanquetSimilarity(const T1 &bv1, const T2 &bv2) {
double y = bv1.getNumOnBits();
double z = bv2.getNumOnBits();
double max = tmax(y, z);
double max = std::max(y, z);
if (max > 0.0) {
return x / max;
} else {
@@ -439,6 +435,7 @@ double RusselSimilarity(const T1 &bv1, const T2 &bv2) {
if (bv1.getNumBits() != bv2.getNumBits()) {
throw ValueErrorException("BitVects must be same length");
}
double x = NumOnBitsInCommon(bv1, bv2);
return x / bv1.getNumBits();
}
@@ -449,8 +446,12 @@ double RogotGoldbergSimilarity(const T1 &bv1, const T2 &bv2) {
throw ValueErrorException("BitVects must be same length");
}
double x = NumOnBitsInCommon(bv1, bv2);
double y = bv1.getNumOnBits();
double z = bv2.getNumOnBits();
auto y = bv1.getNumOnBits();
auto z = bv2.getNumOnBits();
if (y == 0 || z == 0) {
return 0.0;
}
double l = bv1.getNumBits();
double d = l - y - z + x;

15
Code/DataStructs/catch_tests.cpp
View File

@@ -14,6 +14,7 @@
#include "BitVects.h"
#include "BitOps.h"
#include "BitVectUtils.h"
#include "ExplicitBitVect.h"
#include "SparseIntVect.h"
#include <limits>
@@ -26,4 +27,18 @@ TEST_CASE("special cases for the limits of sparse vectors") {
CHECK(!sbv.setBit(std::numeric_limits<unsigned int>::max()));
CHECK(sbv.getBit(std::numeric_limits<unsigned int>::max()) == 1);
}
}
TEST_CASE("github #9033: tversky is 1 when no bits are set") {
ExplicitBitVect bv1(8);
ExplicitBitVect bv2(8);
CHECK(TverskySimilarity(bv1, bv2, 0.5, 0.5) == 0.0);
CHECK(TanimotoSimilarity(bv1, bv2) == 0.0);
CHECK(CosineSimilarity(bv1, bv2) == 0.0);
CHECK(KulczynskiSimilarity(bv1, bv2) == 0.0);
CHECK(SokalSimilarity(bv1, bv2) == 0.0);
CHECK(McConnaugheySimilarity(bv1, bv2) == 0.0);
CHECK(BraunBlanquetSimilarity(bv1, bv2) == 0.0);
CHECK(RusselSimilarity(bv1, bv2) == 0.0);
CHECK(RogotGoldbergSimilarity(bv1, bv2) == 0.0);
}