Skip to content
Docs

Search documentation

Start here

↑↓ Navigate↵ OpenEsc Close

Docs

Cookbook

Choose a task and run a complete script. Each example preserves the stdout recorded against the live API; it is a record, rather than a promised result for your next call.

Choose a task

Make classification and regression requestsRead one result per query row and the metadata returned with it.

Input: German credit and diabetes datasets; an API key.

Read the recorded stdout
Compare models on one splitCompare Causilo with XGBoost and LightGBM at their defaults, including the API round trip.

Input: German credit; an API key and the booster libraries.

Read the recorded stdout
Check limits and count a callInspect live ceilings, count cells and see what happens when you stop waiting.

Input: Diabetes dataset; an API key.

Read the recorded stdout
Reuse a context for new queriesUpload the context once, send three query batches and compare the cells used.

Input: Adult census dataset; an API key.

Read the recorded stdout

The scripts read your key from CAUSILO_KEY. CAUSILO_ENDPOINT overrides the API address. Start with the Quickstart for one minimal call, or inspect the API reference for exact request and response fields.

Original first-call example

You need an existing API key. See API access for availability and support. Install the Python client and pass the key to it:

TerminalBASH
pip install causilo-client
Python
import pandas as pd
from causilo_client import Causilo

cx = Causilo("https://api.nums.world", token="csl_...")

context = pd.DataFrame({
    "region":   ["EU", "EU", "US", "US", "APAC", "APAC"],
    "plan":     ["pro", "free", "pro", "free", "pro", "free"],
    "spend":    [120.0, 0.0, 340.0, 12.0, 90.0, 4.0],
    "renewed":  ["yes", "no", "yes", "no", "yes", "no"],
})
query = pd.DataFrame({
    "region": ["EU", "APAC"],
    "plan":   ["pro", "free"],
    "spend":  [200.0, 1.0],
})

print(cx.predict(context, query, target="renewed", model="causilo-clf"))

The quickstart makes the same kind of call on a real dataset for both classification and regression and shows the output it produced.