Definitions of technical terms used throughout OSBridge documentation.
4 min readA set of protocols and tools for building software applications. In OSBridge, we use REST APIs to communicate between the frontend and backend, and external APIs (GitHub, LeetCode, Google Gemini) for data and AI capabilities.
A mathematical measure of similarity between two vectors. Values range from -1 (opposite) to 1 (identical), with 0 meaning no similarity. OSBridge uses cosine similarity via pgvector's <=> operator to compare skill profile embeddings with issue embeddings.
A file found in many open-source repositories that explains how to contribute to the project. It typically includes setup instructions, coding standards, and PR guidelines. Repositories with this file tend to be more welcoming to first-time contributors.
A lightweight, type-safe SQL ORM (Object-Relational Mapper) for TypeScript. OSBridge uses Drizzle to define database schemas and execute queries against PostgreSQL with full type safety.
A high-dimensional vector representation of text that captures its semantic meaning. OSBridge uses 768-dimensional embeddings generated by Google Gemini's text-embedding-004 model to represent both user skills and issue descriptions.
A personal copy of someone else's GitHub repository. Forking creates a linked copy under your account where you can make changes without affecting the original project. You submit changes back via pull requests.
A GitHub label used by maintainers to mark issues that are suitable for new contributors. OSBridge goes beyond this label by using AI to match issues to your specific skills rather than relying on self-reported labels.
An open standard for access delegation. OSBridge uses GitHub OAuth to authenticate users — you sign in with your GitHub account, and we receive a limited access token to read your public profile information.
A PostgreSQL extension that adds support for vector data types and similarity search operations. OSBridge uses pgvector to store embeddings and run cosine similarity queries efficiently directly in the database.
A proposal to merge your changes from one branch into another. In open source, you typically create a PR from your fork's feature branch to the upstream repository's main branch. Maintainers review the changes before merging.
A mathematical representation of your algorithmic abilities based on your LeetCode solved problem tags. It consists of a JSON object of tag counts and a 768-dimensional vector embedding stored in the skill_profiles table.
In the context of AI and embeddings, a vector is an ordered array of floating-point numbers that represents text in a high-dimensional mathematical space. Similar texts produce vectors that are closer together, enabling similarity search.