Introduction
remodel-core is a Rust library for database modeling — build ER diagrams, convert them to relational schemas, and generate SQL DDL.
remodel-core is the engine that powers Remodel — a desktop application for visual database modeling. It is also published as an independent Rust crate so you can embed it in your own tools and pipelines.
What it does
The library works in three layers, each building on the previous one:
| Layer | Module | Purpose |
|---|---|---|
| Conceptual | models::conceptual | Entity-Relationship structures (entities, attributes, relationships, specializations) |
| Logical | models::logical | Relational schema (tables, columns, constraints, foreign keys) |
| Physical | sql | SQL DDL generation per dialect |
The transformation from conceptual → logical is handled by transform::conceptual_to_logical, and structural correctness is checked by the validation module.
Origin
remodel-core is a Rust reimplementation of the modeling engine of brModelo by Carlos Henrique Cândido. The type names and algorithm choices follow the original Java implementation closely, so brModelo .brM3 files can be migrated to the .remodel format with the companion migration functions.
Quick example
use remodel_core::prelude::*;
// 1. Build a conceptual model
let mut m = ConceptualModel::new("blog");
let post = m.add_entity("Post");
let user = m.add_entity("User");
m.add_primary_attribute(post, "id", DataType::Integer)?;
m.add_attribute(AttributeOwner::Entity(post), "title", DataType::Varchar(255))?;
m.add_primary_attribute(user, "id", DataType::Integer)?;
m.add_attribute(AttributeOwner::Entity(user), "email", DataType::Varchar(120))?;
m.relate("wrote", post, Cardinality::ZeroToMany)
.with(user, Cardinality::OneToMany);
// 2. Validate
let diags = validate_conceptual(&m);
assert!(diags.iter().all(|d| d.severity != Severity::Error));
// 3. Convert to logical
let logical = m.to_logical()?;
// 4. Generate SQL
let sql = logical.to_sql(SqlDialect::Postgres);
println!("{sql}");