Chronon AI

0.0.111 verified Fri May 01 auth: no python

Chronon is a platform for defining, managing and executing feature engineering pipelines in production. The Python client library provides APIs to interact with the Chronon backend, define feature tables, and run jobs. Current version: 0.0.111. Release cadence: active development.

pip install chronon-ai

Common errors

error ModuleNotFoundError: No module named 'chronon' ↓

cause Installed package name is 'chronon-ai', but code tries to import 'chronon'.

fix

Ensure you installed 'chronon-ai' (pip install chronon-ai) and import using 'import chronon'.

error pyspark.sql.utils.IllegalArgumentException: 'Unsupported data type' ↓

cause Chronon may expect specific Spark SQL types that are not automatically converted.

fix

Explicitly cast columns to supported types (e.g., using .cast() in your feature expressions).

Warnings

breaking The package name on PyPI is 'chronon-ai', but the import statement uses 'chronon'. This mismatch can cause confusion. ↓

fix Install with 'pip install chronon-ai' but import using 'import chronon'.

gotcha Chronon heavily relies on PySpark and Thrift. If these are not installed, imports may fail silently or at runtime. ↓

fix Install with extras: 'pip install chronon-ai[spark,thrift]' or ensure compatible versions of PySpark and Thrift are available.

deprecated The 'chronon.api' module is being deprecated in favor of 'chronon.feature_table' and 'chronon.utils'. ↓

fix Replace 'from chronon.api import ...' with 'from chronon.feature_table import ...'.

Imports

FeatureTable
```
from chronon import feature_table as ft
ft.FeatureTable(...)
```
The correct import path uses the 'chronon' package; direct imports like 'from chronon.feature_table import ...' should be avoided due to internal restructuring.

Quickstart

Define a basic feature table using Chronon Python API.

from chronon import feature_table as ft
from chronon import utils

# Define a simple feature table
table = ft.FeatureTable(
    name="user_features",
    source="events",
    keys=["user_id"],
    features=[
        ft.Feature(name="total_purchase", expression="SUM(purchase_amount)"),
    ]
)
print(table.name)