sqlacodegen 4.0.4

Features

• Supports SQLAlchemy 2.x

• Produces declarative code that almost looks like it was hand written

• Produces PEP 8 compliant code

• Accurately determines relationships, including many-to-many, one-to-one

• Automatically detects joined table inheritance

• Excellent test coverage

Installation

To install, do:

pip install sqlacodegen

To include support for the PostgreSQL CITEXT extension type (which should be considered as tested only under a few environments) specify the citext extra:

pip install sqlacodegen[citext]

To include support for the PostgreSQL GEOMETRY, GEOGRAPHY, and RASTER types (which should be considered as tested only under a few environments) specify the geoalchemy2 extra:

To include support for the PostgreSQL PGVECTOR extension type, specify the pgvector extra:

pip install sqlacodegen[pgvector]

pip install sqlacodegen[geoalchemy2]

Quickstart

At the minimum, you have to give sqlacodegen a database URL. The URL is passed directly to SQLAlchemy’s create_engine() method so please refer to SQLAlchemy’s documentation for instructions on how to construct a proper URL.

Examples:

sqlacodegen postgresql:///some_local_db
sqlacodegen --generator tables mysql+pymysql://user:password@localhost/dbname
sqlacodegen --generator dataclasses sqlite:///database.db
# --engine-arg values are parsed with ast.literal_eval
sqlacodegen oracle+oracledb://user:pass@127.0.0.1:1521/XE --engine-arg thick_mode=True
sqlacodegen oracle+oracledb://user:pass@127.0.0.1:1521/XE --engine-arg thick_mode=True --engine-arg connect_args='{"user": "user", "dsn": "..."}'

To see the list of generic options:

sqlacodegen --help

Available generators

The selection of a generator determines the

The following built-in generators are available:

• tables (only generates Table objects, for those who don’t want to use the ORM)

• declarative (the default; generates classes inheriting from declarative_base()

• dataclasses (generates dataclass-based models; v1.4+ only)

• sqlmodels (generates model classes for SQLModel)

Generator-specific options

The following options can be turned on by passing them using --options (multiple values must be delimited by commas, e.g. --options noconstraints,nobidi):

• tables

  • noconstraints: ignore constraints (foreign key, unique etc.)

  • nocomments: ignore table/column comments

  • noindexes: ignore indexes

  • nonativeenums: don’t generate Python enum classes for native database ENUM types (e.g., PostgreSQL ENUM); use plain string mapping instead

  • nosyntheticenums: don’t generate Python enum classes from CHECK constraints with IN clauses (e.g., column IN ('value1', 'value2', ...)); preserves CHECK constraints as-is

  • noidsuffix: prevent the special naming logic for single column many-to-one and one-to-one relationships (see Relationship naming logic for details)

  • include_dialect_options: render a table’ dialect options, such as starrocks_partition for StarRocks’ specific options.

  • keep_dialect_types: preserve dialect-specific column types instead of adapting to generic SQLAlchemy types.

• declarative

  • all the options from tables

  • use_inflect: use the inflect library when naming classes and relationships (turning plural names into singular; see below for details)

  • nojoined: don’t try to detect joined-class inheritance (see below for details)

  • nobidi: generate relationships in a unidirectional fashion, so only the many-to-one or first side of many-to-many relationships gets a relationship attribute, as on v2.X

  • nofknames: disable improved relationship naming when multiple FKs or junction tables connect to the same target. By default, uses FK column names for one-to-many (e.g., simple_items_parent_container) and junction table names for many-to-many (e.g., students_enrollments). Reverts to underscore suffixes (simple_items_, student_).

• dataclasses

  • all the options from declarative

• sqlmodels

  • all the options from declarative

Customizing code generation logic

If the built-in generators with all their options don’t quite do what you want, you can customize the logic by subclassing one of the existing code generator classes. Override whichever methods you need, and then add an entry point in the sqlacodegen.generators namespace that points to your new class. Once the entry point is in place (you typically have to install the project with pip install), you can use --generator <yourentrypoint> to invoke your custom code generator.

For examples, you can look at sqlacodegen’s own entry points in its pyproject.toml.

Getting help

If you have problems or other questions, you should start a discussion on the sqlacodegen discussion forum. As an alternative, you could also try your luck on the sqlalchemy room on Gitter.

Security contact information

To report a security vulnerability, please use the Tidelift security contact. Tidelift will coordinate the fix and disclosure.