Skip to content

SBT Changelog

🐞 Patch Release v0.3.1

Bug Fix

  • Fix bug where python model source references were not recognized if they were lowercase

⚙️ New Release: v0.3.0

Dynamic Config Powered by TOML

SBT has replaced it's .env based configuration with a more robust and dynamic configuration system powered by TOML.

  • TOML provides more structure than .env files
  • Multiple different configurations can be stored in a single file, and sbt commands can swap between configs based on CLI options
  • Use sbt init-config to create a new sbtconf.toml file in your project, it will walk you through setting up one block of configs, then you can customize the file as needed to support multiple configurations
  • See SBT Configuration for more information

Runtime Source + Param Overrides via TOML

  • Add [overrides.name.sources] and [overrides.name.models] to your sbtconf.toml to override sources and models at runtime
  • As many different named blocks of overrides as you need can be saved in the config file
  • In sbt run commands, pass the --overrides name option to pull the source and param value mappings from the TOML

See SBT Configuration for more information

Token Based Auth with Insight Cloud

Multi-App Deploys using tags

  • New YAML config key on the app object called tags
  • Use tags to group apps together for deployment
  • upload-app and publish-app now take an optional --tag flag which will run it's process for all apps that have that tag

🐛 Patch Release: v0.2.1

Bugs

  • Model casing bug

🚀 New Release: v0.2.0

This released focused on enhancing the user experience and improving the performance of the Seek platform. Here are the key highlights of this release:

Schema Adherence

  • Define schemas for your data with column names, data types, value constraints, and metadata
  • Attach those schemas to sources and models to ensure inbound and outbound data is formatted as expected
  • Use the new validate cli commands and flags to check your data against your schemas

Quality of Life Enhancements

  • Halt app processing should a model fail
  • Stricter validation of model and app configurations
  • Ensure required fields are present, ensure valid values are given, provide errors if not
  • Add --debug to run-app to get printed out call proc statements
  • Ensure all models and apps have unique names
  • Output model name before running it...easier to understand what model run is currently executing
  • Install model deps during local run
  • SBT will attempt to install any missing dependencies before running a model in the current python environment

New Helper Commands

This version of SBT introducers new helper commands to guide new SBT developers through setting up a new model or app.

Run sbt new model to walk through prompts that help you create a new model in your project.

Run sbt new app to walk through prompts that help you create a new app in your project.