Skip to content

Tempora Documentation

From Raw Data to Model-Ready Batches

Tempora helps you build training and inference-ready batches from structured or sequential data residing either locally, in cloud storage, or in enterprise data systems.

What You Can Do

  • Connect to a Tempora server and manage sessions: Server & Authentication
  • Load and prepare datasets from file systems, object stores, databases or data warehouses: Datasets
  • Sample model-ready batches (random, sequential, series): Samplers
  • Specify targets to be generated with each ML batch: Target Specification
  • Work directly with Batch objects and batch IO: Batch Objects
  • Define arbitrary transform functions to be applied to batch or target data: Transforms
  • Configuration options for file system datasets: File System Options

Installation

pip install tempora-ai

Environment Variables

These environment variables are used by tempora.login() and tempora.server_connection():

  • TEMPORA_SERVER_HOST: Server address in hostname:port format. If port is omitted, 7070 is used.
  • TEMPORA_SERVER_USERNAME: Username used to authenticate with the Tempora server.
  • TEMPORA_SERVER_PASSWORD: Password for TEMPORA_SERVER_USERNAME. If not set, you are prompted interactively.