Next Gen UI Core Functionality

This module is part of the Next Gen UI Agent project.

This module contains UI Agent Core functionality and frameworks.

Provides

• NextGenUIAgent providing agent configuration and methods for individual processing steps
• NextGenUIAgent produced UI Data Blocks
  • LLM selected and configured dynamic components
    • Supported components: one-card, image, video-player, set-of-cards, table, chart-bar, chart-line, chart-pie, chart-donut, chart-mirrored-bar
  • Hand Build Components
  • Extensible framework for "data transformation" step per dynamic component
• Plugable "UI renderer" framework for UI components rendering
  • Default json renderer used to send definitions to client-side renderers
• Pluggable and configurable "Input Data Transformation" framework
  • Provided transformers: json, yaml, csv-comma, csv-semicolon, csv-tab, fwctable, noop
• Abstraction of the LLM inference
  • InferenceBase inference interface used by UI Agent
  • LangChainModelInference inference implementation using LangChain chat_models
  • ProxiedAnthropicVertexAIInference inference implementation to call Anthropic/Claude models from proxied Google Vertex AI API endpoint
  • reusable inference provider builder from commandline arguments/env variables, used by all AI protocol servers

Installation

pip install -U next_gen_ui_agent

Interface usage

from next_gen_ui_agent import NextGenUIAgent

inference = # any AI framework inference

agent = NextGenUIAgent(
    inference=LlamaStackInference(model="ollama:llama3.2"),
)

# API is not very friendly ATM, as you have to call methods for individual processing steps. We plan to improve it iit the near future.

Links

• Documentation
• Source Codes
• Contributing