Functional Code with CPU benchmarks

Author: adish29

.gitignore

@@ -3,4 +3,5 @@
 *.out
 *.o
 *.exp
-*.lib
+*.lib
+*.tmp

CPU/vector_initialize.h CPU/initialization.hCPU/vector_initialize.h renamed to CPU/initialization.h

@@ -2,8 +2,14 @@
 #define ITEMS_H
 
 #include <vector>
+#include <fstream>
+#include <iostream>
 #include <cmath>
+#include <memory>
+#include <string>
+#include <sstream>
 #include "omp.h"
+using namespace std;
 
 struct Item {
     Item(std::vector<double> _values) : values(_values) {}

Project Document.docx

@@ -1,10 +1,14 @@
 # Approximate-Nearest-Neighbor-Framework
 
+# HOW TO RUN
+
+bash compile.sh <br />
+bash run.sh [input_file] [output_file] [Integer value of k] <br />
 
 # DEMO RUN
 
 bash compile.sh <br />
-bash run.sh sample_inputs/sample4.in sample_outputs/sample4.out <br />
+bash run.sh sample_inputs/sample4.in sample_outputs/sample4.out 5 <br />
 
 # DEMO CONSOLE OUTPUT
 

README.md

@@ -8,9 +8,15 @@
 #include <unordered_set>
 #include <vector>
 #include <omp.h>
+#include <cmath> 
+#include <limits> 
 using namespace std;
 
 vector<int> HNSWGraph::searchLayer(Item& q, int entry_point, int ef, int lc) {
+	if (lc < 0 || lc >= layerEdgeLists.size()) {
+        cerr << "Error: Invalid layer index." << endl;
+        exit(EXIT_FAILURE); 
+    }
 	set<pair<double, int>> candidates;
 	set<pair<double, int>> nearestNeighbors;
 	unordered_set<int> isVisited;
@@ -49,7 +55,18 @@ vector<int> HNSWGraph::searchLayer(Item& q, int entry_point, int ef, int lc) {
 	return results;
 }
 
-vector<int> HNSWGraph::KNNSearch(Item& q, int K) {
+vector<int> HNSWGraph::KNNSearch(Item& q, int K, int N) {
+
+	if (K <= 0 || std::ceil(K) != K || K > std::numeric_limits<int>::max()) {
+		cerr << "Error: Invalid value of K for KNNSearch." << endl;
+		exit(EXIT_FAILURE); 
+	}
+
+	if (K > N) {
+        std::cerr << "Error: Value of K (" << K << ") is greater than the number of data points (N = " << N << ")." << std::endl;
+        exit(EXIT_FAILURE);
+    }
+	
 	int maxLyer = layerEdgeLists.size() - 1;
 	int entry_point = enterNode;
 	for (int l = maxLyer; l >= 1; l--) {
@@ -59,6 +76,10 @@ vector<int> HNSWGraph::KNNSearch(Item& q, int K) {
 }
 
 void HNSWGraph::addEdge(int st, int ed, int lc) {
+	if (lc < 0 || lc >= layerEdgeLists.size()) {
+        cerr << "Error: Invalid layer index." << endl;
+        return;
+    }
 	if (st == ed) {
 		return;
 	}
@@ -91,10 +112,12 @@ void HNSWGraph::Insert(Item& q) {
 		int MM = l == 0 ? MMax0 : MMax;
 		vector<int> neighbors = searchLayer(q, entry_point, efConstruction, i);
 		vector<int> selectedNeighbors = vector<int>(neighbors.begin(), neighbors.begin()+min(int(neighbors.size()), M));
-		for (int n: selectedNeighbors) {
-			addEdge(n, nid, i);
-		}
-		for (int n: selectedNeighbors) {
+		for (size_t j = 0; j < selectedNeighbors.size(); j++) {
+            int n = selectedNeighbors[j];
+            addEdge(n, nid, i);
+        }
+		for (size_t j = 0; j < selectedNeighbors.size(); j++)  {
+			int n = selectedNeighbors[j];
 			if (layerEdgeLists[i][n].size() > MM) {
 				vector<pair<double, int>> distPairs;
 				for (int nn: layerEdgeLists[i][n]) {

hnsw_implementation/hnsw.cpp

@@ -6,7 +6,7 @@
 #include <unordered_map>
 #include <iostream>
 
-#include "../CPU/vector_initialize.h"
+#include "../CPU/initialization.h"
 using namespace std;
 
 
@@ -28,7 +28,7 @@ struct HNSWGraph {
 	void addEdge(int st, int ed, int lc);
 	vector<int> searchLayer(Item& q, int ep, int ef, int lc);
 	void Insert(Item& q);
-	vector<int> KNNSearch(Item& q, int K);
+	vector<int> KNNSearch(Item& q, int K, int N);
 
 	void printGraph() {
 		for (int l = 0; l < layerEdgeLists.size(); l++) {

hnsw_implementation/hnsw.h

@@ -1,4 +1,5 @@
 #include "hnsw_implementation/hnsw.h"
+#include "CPU/initialization.h"
 #include <algorithm>
 #include <ctime>
 #include <iostream>
@@ -54,17 +55,26 @@ void readInputFromFile(const string& filename, int& D, int& N, int& M, vector<It
     infile.close();
 }
 
-
 int main(int argc, char* argv[]) {
 
-    if (argc != 3) {
-        cerr << "Usage: " << argv[0] << " <input_filename> <output_filename>" << endl;
+    if (argc != 4) {
+        cerr << "Usage: " << argv[0] << " <input_filename> <output_filename> <Integer value of k>" << endl;
         return EXIT_FAILURE;
     }
 
     string filename = argv[1];
     string outputFilename = argv[2];
-    int K = 5;
+    
+    int K;
+    try {
+        K = std::stoi(argv[3]); 
+    } catch (std::invalid_argument const& e) {
+        std::cerr << "Invalid argument: " << e.what() << std::endl;
+        return EXIT_FAILURE;
+    } catch (std::out_of_range const& e) {
+        std::cerr << "Out of range argument: " << e.what() << std::endl;
+        return EXIT_FAILURE;
+    }
 
     ifstream infile(filename);
     if (!infile.is_open()) {
@@ -78,6 +88,11 @@ int main(int argc, char* argv[]) {
 
     readInputFromFile(filename, D, N, M, base, queries);
 
+    if (K > N) {
+        std::cerr << "Error: Value of K (" << K << ") is greater than the number of data points (N = " << N << ")." << std::endl;
+        return EXIT_FAILURE;
+    }
+
     HNSWGraph myHNSWGraph(10, 30, 30, 10, 2);
 
     for (int i = 0; i < N; ++i) {
@@ -118,7 +133,7 @@ int main(int argc, char* argv[]) {
 
         begin_time = clock();
 
-        vector<int> knns = myHNSWGraph.KNNSearch(query, K);
+        vector<int> knns = myHNSWGraph.KNNSearch(query, K, N);
         for (size_t idx = 0; idx < knns.size(); ++idx) {
             outfile << knns[idx];
             if (idx != knns.size() - 1) {

main.cpp

@@ -1 +1 @@
-./my_program $1 $2
+./my_program $1 $2 $3