nikola/treender
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add similaritys

Browse Source
This commit is contained in:
Nikola Petrov
2025-10-22 17:56:26 +02:00
parent 5f6305a2f2
commit 593b813988
6 changed files with 138 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
shared/inc/values/Dna.hpp
View File

@@ -9,8 +9,6 @@
#define MAX_DEPTH 8
#define MAX_POSIBLE_DEPTH 11
static_assert(MAX_DEPTH <= MAX_POSIBLE_DEPTH);
static_assert(180 == sizeof(Dna));
constexpr int SIZE_OF_DNA = sizeof(Dna);
struct Branch
{

10
shared/inc/values/Similarity.hpp
View File

@@ -3,14 +3,14 @@
namespace Similarity
{
// float euclidean_distance(Dna *d1, Dna *d2); direct distance betwen vector. wont give 0 and 1
// float dot_product(Dna *d1, Dna *d2); doent return betwen 0 to 1
// float cosine_similarity(Dna *d1, Dna *d2);
// float cosine_similarity_int(Dna *d1, Dna *d2);
float euclidean_distance(Dna *d1, Dna *d2);// direct distance betwen vector. wont give 0 and 1
// float dot_product(Dna *d1, Dna *d2); // doent return betwen 0 to 1
float cosine_similarity(Dna *d1, Dna *d2);
float cosine_similarity_int(Dna *d1, Dna *d2);
float hamming_distance(Dna *d1, Dna *d2);
float hamming_distance_without_seeds(Dna *d1, Dna *d2);
// float jaccard_index(Dna *d1, Dna *d2); // primerja unio genov naprimer gleda ce je gen za nebo isti z genom za barvo za liste, to nerabimo
// float levenshtein_distance(Dna *d1, Dna *d2); // odstranjen ker mi vrne iste podatke kot hamming distance ki je bolj enostaven za izracun
float levenshtein_distance(Dna *d1, Dna *d2); // odstranjen ker mi vrne iste podatke kot hamming distance ki je bolj enostaven za izracun
// float needleman_wunsch(Dna *d1, Dna *d2); used for bioinformatics and aligment. Dont need its aligned alredy
typedef float(simil_func)(Dna *d1, Dna *d2);

134
shared/src/values/Similarity.cpp
View File

@@ -1,56 +1,72 @@
#include "values/Similarity.hpp"
#include <cmath>
#include <algorithm>
#include <numeric>
#include <raylib.h>
namespace Similarity
{
float euclidean_distance(Dna *d1, Dna *d2)
{
uint8_t *a = (uint8_t *)d1;
uint8_t *b = (uint8_t *)d2;
float sum = 0.0f;
for (size_t i = 0; i < sizeof(Dna); ++i) {
float diff = static_cast<float>(a[i]) - static_cast<float>(b[i]);
sum += diff * diff;
}
float distance = std::sqrt(sum);
float max_distance = 255.0f * std::sqrt(static_cast<float>(sizeof(Dna)));
return distance / max_distance;
}
// todo: use int8_t insted of uint8_t and map data
// 0 -> -128
// 255 -> 127
// int8_t = uint8_t - 128
// float cosine_similarity(Dna *d1, Dna *d2)
// {
// uint8_t *d1a = (uint8_t *)d1;
// uint8_t *d2a = (uint8_t *)d2;
float cosine_similarity(Dna *d1, Dna *d2)
{
uint8_t *d1a = (uint8_t *)d1;
uint8_t *d2a = (uint8_t *)d2;
// float mag1 = 0.0f;
// float mag2 = 0.0f;
// float dot_prod = 0.0f;
// for (size_t i = 0; i < sizeof(Dna); i++)
// {
// dot_prod += d1a[i] * d2a[i];
// mag1 += d1a[i] * d1a[i];
// mag2 += d2a[i] * d2a[i];
// }
// mag1 = sqrt(mag1);
// mag2 = sqrt(mag2);
float mag1 = 0.0f;
float mag2 = 0.0f;
float dot_prod = 0.0f;
for (size_t i = 0; i < sizeof(Dna); i++)
{
dot_prod += d1a[i] * d2a[i];
mag1 += d1a[i] * d1a[i];
mag2 += d2a[i] * d2a[i];
}
mag1 = sqrt(mag1);
mag2 = sqrt(mag2);
// return dot_prod / (mag1 * mag2);
// }
return dot_prod / (mag1 * mag2);
}
// float cosine_similarity_int(Dna *d1, Dna *d2)
// {
// auto map = [](uint8_t a) -> int8_t
// { return a - 128; };
// uint8_t *d1a = (uint8_t *)d1;
// uint8_t *d2a = (uint8_t *)d2;
// float mag1 = 0.0f;
// float mag2 = 0.0f;
// float dot_prod = 0.0f;
// for (size_t i = 0; i < sizeof(Dna); i++)
// {
// int8_t a = map(d1a[i]);
// int8_t b = map(d2a[i]);
// dot_prod += a * b;
// mag1 += a * a;
// mag2 += b * b;
// }
// mag1 = sqrt(mag1);
// mag2 = sqrt(mag2);
// return dot_prod / (mag1 * mag2);
// }
float cosine_similarity_int(Dna *d1, Dna *d2)
{
auto map = [](uint8_t a) -> int8_t
{ return a - 128; };
uint8_t *d1a = (uint8_t *)d1;
uint8_t *d2a = (uint8_t *)d2;
float mag1 = 0.0f;
float mag2 = 0.0f;
float dot_prod = 0.0f;
for (size_t i = 0; i < sizeof(Dna); i++)
{
int8_t a = map(d1a[i]);
int8_t b = map(d2a[i]);
dot_prod += a * b;
mag1 += a * a;
mag2 += b * b;
}
mag1 = sqrt(mag1);
mag2 = sqrt(mag2);
return dot_prod / (mag1 * mag2);
}
float hamming_distance(Dna *d1, Dna *d2)
{
@@ -99,4 +115,40 @@ namespace Similarity
float average_similarity = total_similarity / num_pairs;
return average_similarity * 100.0f;
}
float levenshtein_distance(Dna *d1, Dna *d2)
{
size_t len = sizeof(Dna);
uint8_t *a = (uint8_t *)d1;
uint8_t *b = (uint8_t *)d2;
// Create a distance matrix
std::vector<std::vector<uint32_t>> dp(len + 1, std::vector<uint32_t>(len + 1, 0));
// Initialize the first row and column
for (size_t i = 0; i <= len; ++i)
{
dp[i][0] = i;
}
for (size_t j = 0; j <= len; ++j)
{
dp[0][j] = j;
}
// Fill the distance matrix
for (size_t i = 1; i <= len; ++i)
{
for (size_t j = 1; j <= len; ++j)
{
uint32_t cost = (a[i - 1] == b[j - 1]) ? 0 : 1;
dp[i][j] = std::min({
dp[i - 1][j] + 1, // deletion
dp[i][j - 1] + 1, // insertion
dp[i - 1][j - 1] + cost // substitution
});
}
}
return 1 - (dp[len][len] / float (len + len));
}
}