lab-mini

Minimal reference for repeatable data-science labbing: load, profile, analyze, claim, report.

lab-mini is the small research loop behind the bigger agent-infrastructure work: make the dataset shape explicit, make claims reproducible, and render the result as a report that another session can inspect.

What It Proves

• A lab run should produce a claim artifact, not just notebook state.
• Profiling comes before analysis so downstream claims know their data shape.
• Deterministic seeds and fixture-backed tests keep the loop repeatable.
• The same pattern works as a lightweight evaluation harness for agent behavior.

Install

pip install git+https://github.com/JustinJLeopard/lab-mini.git

For local development:

pip install -e '.[dev]'

30-second example

from lab_mini import claim_mean, load_csv, profile, render_markdown

frame = load_csv("tests/_fixtures/sample_lab.csv")
dataset_profile = profile(frame)
claim = claim_mean(frame, "value", seed=42)

print(dataset_profile.row_count)
print(render_markdown([claim]))

See docs/USAGE.md for the full loop.

Related

• JustAi — orchestration control plane.
• safe-mini — safe local execution substrate.
• memory-mini — durable agent memory reference.