Introducing Daggr: Chain Apps Programmatically, Inspect Visually

Daggr is a novel, open-source Python library designed to streamline the process of building AI workflows by connecting Gradio apps, machine learning models, and custom functions. It allows developers to define workflows in Python and automatically generates a visual canvas for inspecting intermediate outputs, rerunning individual steps, and managing state for complex pipelines.

Why This Matters

The development of AI applications often involves combining multiple models or processing steps, which can be cumbersome and prone to errors. Traditional approaches either rely on fragile scripts that are hard to debug or utilize heavy orchestration platforms intended for production pipelines, not rapid experimentation. Daggr addresses these challenges by providing a code-first approach with automatic visual canvas generation, enabling efficient inspection and rerunning of individual steps within a workflow.

Key Insights

Daggr supports three types of nodes: GradioNode for calling Gradio Space API endpoints, FnNode for running custom Python functions, and InferenceNode for calling models via Hugging Face Inference Providers.
The library allows for state persistence, saving workflow state, input values, cached results, and canvas position, enabling developers to pick up where they left off.
Daggr integrates seamlessly with Gradio Spaces, enabling the use of public or private Spaces as nodes in workflows without requiring adapters or wrappers.

Working Example

import gradio as gr
from daggr import GradioNode, Graph

# Generate an image using a Gradio Space
image_gen = GradioNode(
    "hf-applications/Z-Image-Turbo",
    api_name="/generate_image",
    inputs={
        "prompt": gr.Textbox(
            label="Prompt",
            value="A cheetah sprints across the grassy savanna.",
            lines=3,
        ),
        "height": 1024,
        "width": 1024,
        "seed": 123,
    },
    outputs={
        "image": gr.Image(label="Generated Image"),
    },
)

# Remove background using another Gradio Space
bg_remover = GradioNode(
    "hf-applications/background-removal",
    api_name="/image",
    inputs={
        "image": image_gen.image,  # Connect to previous node's output
    },
    outputs={
        "original_image": None,  # Hide this output
        "final_image": gr.Image(label="Final Image"),
    },
)

graph = Graph(
    name="Transparent Background Generator",
    nodes=[image_gen, bg_remover]
)
graph.launch()

Practical Applications

Use Case: Developers can utilize Daggr to create complex AI workflows for image processing, such as generating images, removing backgrounds, and enhancing images, all while visually inspecting and managing the workflow.
Pitfall: A common anti-pattern is to overlook the importance of state persistence and workflow management, leading to inefficiencies in debugging and rerunning workflows. Daggr addresses this by automatically saving workflow state and allowing for easy inspection and modification of individual steps.

References:

On This Page

Introducing Daggr: Chain Apps Programmatically, Inspect Visually