Kibo

Kibo is the code generator. It reads a DSM model plus a Kibo template and emits source code — typed C++, Python and TypeScript surfaces wired for the Viper runtime when used with kibo-template-viper.

Kibo is the second step of the code-generation pipeline: DSM → Kibo → kibo-template-viper. Kibo itself is template-agnostic — change the template and you change the target language and runtime. This section covers Kibo’s command line, the Template Model a template reads, and kibo-project, which drives the generation of a whole project. The catalogue of templated features it produces in the Viper world lives in kibo-template-viper.

This is kibo 2, which exposes Template Model 2 and generates over the same 1.2 runtime as kibo 1.2. A template pack written for kibo 1.2 moves with Migrating a template pack from Template Model 1 to Template Model 2.

Place in the ecosystem

• Depends on — DSM models (.dsm.json files), a Kibo template pack.

• Consumed by — developers, through kibo-project and a project’s kibo.toml, or directly by running the JAR.

• Source repositories — digital-substrate/kibo, a Java tool bridging DSM and StringTemplate, and digital-substrate/kibo-project.

• Distribution — a single JAR (kibo-2.0.2.jar) and the kibo_project.py script.

Quickstart

A project states its generation once, in a kibo.toml:

[project]
definitions = "definitions"
infrastructure = "myapp"

[generator]
templates = "2"

[target.python]
features = ["Base", "Wheel"]
output = "python/generated"

[target.cpp]
features = ["Base", "Attachments"]
output = "cpp/generated"

python3 kibo_project.py generate kibo.toml

The Python result is a typed package, import myapp, ready to be used through dsviper; the C++ result is headers and .cpp files that link against the Viper C++ runtime. Calling the JAR directly is covered in Command-line usage.

Topics

• Templates
  • What templates know about DSM
  • Templates form an ecosystem
  • Two ways to encounter templates
• Command-line usage
  • Synopsis
  • Options
  • What a run renders
  • Names
  • Example
  • Coming from kibo 1.2
• kibo-project
  • Usage
  • The project file
  • Where the generator comes from
  • Coming from a generate.py
• Template Model Reference
  • Code Generation Pipeline
  • What a Template Is Rendered For
  • Type Suffix
  • The Three Type Spaces
  • Template Naming Convention
  • String Formats
  • Template Model
  • DSM Model
• Migrating a template pack from Template Model 1 to Template Model 2
  • The method: a diff you can account for
  • Names you build yourself
  • Read the diagnostics, but do not stop there
  • The renames
  • What reads differently
  • Does this affect your pack?
  • Which space to name where
  • Dropping your leaf table
  • What Model 2 adds, without breaking anything
  • Checklist

Status

Kibo 2 — Template Model 2, over the 1.2 runtime. Kibo 1.2 and Template Model 1 are documented in the Kibo 1 section.