@forge-clients/generator

CLI and programmatic library for generating @forge-clients/jira and @forge-clients/confluence from the cleaned Atlassian OpenAPI specifications in @forge-clients/specs.

CLI Usage

# Install globally
npm install -g @forge-clients/generator

# Or use via npx
npx @forge-clients/generator generate

# Update specs from Atlassian and regenerate
npx @forge-clients/generator update-specs
npx @forge-clients/generator generate

CLI Commands

generate

Generates TypeScript client code from the cleaned OpenAPI specs.

forge-clients-gen generate [options]

Options:
  --spec <spec>   Which spec to generate: jira-v3, jira-v2, confluence-v2,
                  confluence-v1, jira-software, jira-sm, all (default: all)
  --out <dir>     Output directory (overrides defaults)

update-specs

Downloads the latest Atlassian OpenAPI specs, applies the post-processing pipeline, and writes cleaned specs to @forge-clients/specs.

forge-clients-gen update-specs [options]

Options:
  --dry-run       Show what would change without writing files

Programmatic API

import type { GeneratorOptions } from '@forge-clients/generator';

const options: GeneratorOptions = {
  specs: ['jira-v3', 'confluence-v2'],
  format: true,
  lint: true,
};

// Implementation pending

Architecture

The generator consists of:

1. Spec Pipeline (src/pipeline/) — Downloads, validates, transforms, and patches the raw Atlassian specs
2. Intermediate Representation (src/ir/) — Converts the cleaned OpenAPI object into a normalised IR
3. Emitters (src/emitters/) — Converts the IR into TypeScript source files using ts-morph

See ai-planning/ for the full design proposal.