README

Very Simple Vector Database in C

This repository contains a simple implementation of a non-permanent vector database in C. The primary purpose of this code is to demonstrate basic vector operations and provide a foundation for understanding cosine distance calculations. The database allows for the insertion of vectors with associated metadata and provides a function to find the closest vector to a given query vector.

Table of Contents

• Overview
• Usage
• Functions
• Sample Usage

---

Overview

The code defines a structure called Embedding to represent vectors with associated metadata. The database (db) is an array of Embedding structures, and it can hold up to MAX_EMBEDDINGS entries. The vectors are three-dimensional, as specified by VECTORS_DIMENSIONS.

Usage

To use this code, follow these steps:

1. Clone the repository.

   git clone https://github.com/tommaso-merlini/HedgeHog.git

2. Navigate to the directory containing the code.

   cd HedgeHog

3. Compile the code using your preferred C compiler.

   gcc main.c -o main -lm

4. Run the compiled executable.

   ./main

Functions

1. main function:

   • Initializes the vector database (db).
   • Creates sample vectors (vec1, vec2, vec3) with associated metadata.
   • Inserts vec1 and vec2 into the database.
   • Searches for the closest vector to vec3 and prints the associated metadata.

2. topEmbedding function:

   • Takes a query vector, the database, and the number of dimensions as input.
   • Returns the closest vector in the database to the query vector based on cosine distance.

3. Cosine Distance-related functions:

   • cosineDistance: Calculates the cosine distance between two vectors.
   • cosineNumerator: Calculates the numerator part of the cosine distance formula.
   • cosineDenominator: Calculates the denominator part of the cosine distance formula.

Sample Usage

// Create sample vectors
Embedding vec1 = {.vector = {-0.021426061, -0.00665704, 0.020520154}, .metadata = "vec1"};
Embedding vec2 = {.vector = {-0.0204378, -0.017829886, 0.008949257}, .metadata = "vec2"};
Embedding vec3 = {.vector = {-0.01877697, 0.010980684, -0.01781616}, .metadata = "vec3"};

// Insert vectors into the database
db[0] = vec1;
db[1] = vec2;

// Search for the closest vector to vec3
Embedding result = topEmbedding(vec3.vector, db, VECTORS_DIMENSIONS);

// Print the metadata of the closest vector
printf("Closest Vector: %s\n", result.metadata);

Feel free to explore and modify the code for your specific use case or integrate it into a larger project.