# Advanced

# CLI vs API

pg-generator may be used as CLI or as API. To use it as a CLI tool, execute the pgen command. To use via API within Node.js, import and execute async generate function. Options for CLI and API are the same most of the time.

$ pgen example --outDir models

is equal to

await generate("example" { outDir: "models" });

# Context

Generators may add and use extra context data which may be used as a configuration for the templates they provide. It should be described in the generator's documents. You may provide a context file or a context module to override those values and modify the behavior of the template.

For example pg-generator-examples "sequelize6" sub-generator uses global.addSchemaName boolean value to determine whether to include PostgreSQL schema names to the created object's names.

With the contextFile option, you can provide a path to a JSON file or a JavaScript file that exports a context object as default.

$ pgen example:sequelize6 --outDir models --contextFile path/to/my-context.js


module.exports = {
  global: {
    addSchemaName: false,

# Relation Names

pg-generator uses pg-structure to reverse engineer a PostgreSQL Database. Since databases do not have actual relations and relation names, naming a relation is an opinionated process. You can use built-in functions or provide your naming functions using relationNameFunctions option.

Below is the summary of built-in methods:

Built-in Functions Type Features
short Deterministic Produces short names, but may result in naming collisions, if the database contains complex relations.
descriptive Deterministic Produces long names especially if foreign keys are named well. Less chance of naming collisions.
optimal Non-Deterministic Produces short names where possible, otherwise produces descriptive names.

See latest documentation (opens new window) about relation naming methods.

If you would like to use your relation name functions create a file exporting m2m, m2o, and m2m functions, and provide the path of the file using relationNameFunctions option.

export function o2m(relation: O2MRelation): string {},
export function m2o(relation: M2ORelation): string {},
export function m2m(relation: M2MRelation): string {},

Please see documentation for relation name functions (opens new window).

Input arguments are corresponding relations. Relation objects provide a rich set of data via attributes. You can generate the relation name using that information according to your needs. See documentation for O2MRelation Class (opens new window), M2ORelation Class (opens new window), M2MRelation Class (opens new window).

# Including & Excluding Schemas

If you want to include or exclude specific schemas, you can use includeSchemas and excludeSchemas options, or their alias -i and -e. Their syntax is the same.

For example, to include all schemas starting with "public_", use the following:

$ pgen example -i: "public_%"

or to include only "public" and "extra" schemas, use the "-i" alias multiple times as the following:

$ pgen example -i public -i extra