openconstruct-jupyter 0.1.0

Jupyter notebook integration for OpenConstruct — connect, sense, and build with agents

openconstruct-jupyter — Jupyter Integration for OpenConstruct

Connect OpenConstruct to your Jupyter notebooks in one line. Cell magic, widgets, fleet panels, and DataFrame-to-room conversion.

Part of SuperInstance OpenConstruct.

What This Gives You

• Cell magic — %%openconstruct --sense vision captures output as a sense shadow
• DataFrame → Plato room — turn any pandas DataFrame into a knowledge graph
• Python function → agent tool — oc.from_function(greet, name="my_tool")
• Fleet panel — inline widget showing discovered fleet nodes
• Room viewer — visualize Plato room tiles and dependencies

Quick Start

# Load the extension
%load_ext openconstruct_jupyter

# Connect
import openconstruct_jupyter as oc
oc.connect(notebook=True)

Cell Magic

%%openconstruct --sense vision
# Cell output captured as a vision shadow

%%openconstruct --tick "analysis complete"
# Post a tick to the fleet

%%openconstruct --room my-analysis
# Create or load a Plato room

Python API

import pandas as pd

# DataFrame → room
df = pd.DataFrame({"x": [1, 2, 3], "y": [4, 5, 6]})
oc.from_dataframe(df, name="my_data")

# Function → agent tool
def greet(name):
    return f"Hello, {name}!"
oc.from_function(greet, name="greet_tool")

# REST API → agent resource
oc.from_api("https://api.example.com/data", name="external_data")

Installation

pip install openconstruct-jupyter

Testing

pytest tests/ -v

License

MIT