ddl-emitter

A small, self-contained TypeSpec emitter that turns your models into SQL DDL, using the same "simple ORM" conventions Hibernate applies by default.

Install

npm install --save-dev ddl-emitter

Usage

Via the config (tspconfig.yaml):

emit:
  - "ddl-emitter"
options:
  "ddl-emitter":
    drop-tables: true # default; set false to omit the leading `drop table ...` statements

Or via the command line:

tsp compile . --emit ddl-emitter

Each TypeSpec namespace is treated as one "package": every model declared directly in it lands in its own <packagename>.ddl file (namespace name, snake_cased). Models declared with no wrapping namespace fall back to schema.ddl.

Mapping rules

• model → table (name converted to snake_case)
• @key property → primary key column
• model with no @key → a surrogate id identity column is synthesized
• scalar / enum property → plain column (enums become varchar + a check constraint, i.e. the @Enumerated(STRING) style)
• other: Other property → many-to-one: an other_id foreign key column on this table
• others: Other[] property → one-to-many: a <table>_id foreign key column on the referenced table (a unidirectional collection with a join column, not a join table)
• tags: string[] property → element collection: a separate <table>_<column> join table, like @ElementCollection
• extends → merged into one table (SINGLE_TABLE inheritance, Hibernate's default)

Anything else (unions, tuples, anonymous nested objects, ...) is skipped with a warning diagnostic rather than failing the whole emit.

@comment decorator

Attach a real, persisted column comment (emitted as comment on column ...), the DDL equivalent of Hibernate's @Comment annotation:

import "ddl-emitter";
using DdlEmitter;

model BusinessPartner {
  @key
  @comment("Primary identifier of the business partner.")
  id: string;
}

Emitter options

| Option | Type | Default | Description | | --- | --- | --- | --- | | drop-tables | boolean | true | Prefix the script with drop table if exists ... cascade; for every table, so it can be re-run against an existing database. |

Dialect: generic SQL, close to PostgreSQL. Adjust the generated output as needed for your target database.