Terminology

In this section, we’ll cover the key terms and concepts used throughout the Databraid documentation. Understanding these terms will help you navigate and utilize the platform more effectively.

Core Concepts

• Braid: A braid is a visual representation of a workflow or pipeline in Databraid. It consists of nodes connected by wires, forming a directed graph that defines the flow of data and the sequence of operations. Braids are executed by the Databraid runtime engine.

• Node: Nodes are the building blocks of braids. They represent individual units of functionality, such as data processing, manipulation, or visualization. Nodes have input and output ports that allow data to flow between them. Databraid provides a wide range of built-in nodes, and users can also create custom nodes using code or Jupyter notebooks.

• Bead: Beads are specialized nodes that provide additional functionality and integration with external services. They extend the capabilities of Databraid making use of third-party APIs, libraries, and tools. Beads can be used for tasks such as data retrieval, machine learning, or API integration.

• Event: Events in Databraid represent triggers or actions that initiate the execution of a braid. They can be based on various factors, such as time, data changes, user interactions, or external triggers. Events allow you to automate and schedule the execution of your workflows.

• Data Type: Databraid supports a variety of data types, including primitives (e.g., numbers, strings, booleans), arrays, objects, and custom data structures. Understanding the data types used in your braids is crucial for ensuring compatibility and proper data flow between nodes.

• Execution: Execution refers to the process of running a braid and processing data through the connected nodes. Databraid’s execution engine handles the flow of data, ensures the proper sequence of operations, and manages the lifecycle of nodes and beads.

Language Model Concepts

• Prompt: Prompts are text-based instructions or queries that are provided to language models to guide their output. They serve as the input to the model and can be used to generate text, answer questions, or perform specific tasks. Crafting effective prompts is essential for obtaining desired results from language models.

• Language Model: Language models are machine learning models trained on large amounts of text data. They are capable of understanding and generating human-like text based on the patterns and relationships learned from the training data. Databraid integrates with various language models, such as GPT-3, to enable natural language processing capabilities.

• Token: Tokens are the basic units of input and output for language models. They represent individual words, subwords, or characters that the model processes. Language models have a maximum token limit, which determines the length of the input and output sequences they can handle.

• Embedding: Embeddings are dense vector representations of text or other data types. They capture the semantic meaning and relationships between different pieces of data. Embeddings are commonly used for tasks such as similarity search, clustering, or as input features for downstream machine learning models.

API and Integration Concepts

• API: API stands for Application Programming Interface. It is a set of rules and protocols that define how different software components should interact with each other. APIs allow Databraid to integrate with external services, databases, or libraries, enabling seamless data exchange and functionality extension.

• Third-Party Services: Databraid integrates with various third-party services to enhance its capabilities. These services can include cloud platforms, databases, machine learning frameworks, or external APIs. Databraid provides beads and nodes specifically designed to interact with popular third-party services.

• Secrets: Secrets are sensitive pieces of information, such as API keys, access tokens, or database credentials, that are required to authenticate and authorize access to external services or resources. Databraid provides a secure way to manage and store secrets, ensuring they are not exposed or compromised.

Databraid Interface Concepts

• Braid Editor: The braid editor is the visual interface where you create, edit, and manage your braids. It provides a drag-and-drop interface for adding nodes, connecting them with wires, and configuring their properties. The braid editor also allows you to preview and test your braids before deployment.

• Node Library: The node library is a collection of pre-built nodes and beads available in Databraid. It provides a wide range of functionality, including data processing, machine learning, visualization, and integration with external services. The node library is continuously expanded and updated to offer new capabilities.

• Braid Execution: Braid execution refers to the process of running a braid and processing data through the connected nodes. Databraid provides a runtime environment that handles the execution of braids, manages the flow of data, and ensures the proper sequence of operations.

• Monitoring and Logging: Databraid offers monitoring and logging capabilities to track the execution of your braids, monitor performance, and diagnose issues. You can access logs, metrics, and visualizations to gain insights into the behavior and health of your workflows.

By familiarizing yourself with these key terms and concepts, you’ll be well-equipped to navigate the Databraid documentation and leverage the platform’s capabilities effectively. If you come across any additional terms that are not covered here, please refer to the specific documentation sections or reach out to our support channels for further clarification.