REPACK

repack is a tool to define data models and generate corresponding code in various languages and formats. You define your models in .repack files using a simple syntax.

• Note: Repack is not an ORM. Repack is a codegen tool. *

• Another note: documentation below is AI generated and has only been lightly reviewed. While this will change in the near future, to learn more now, check out the example files in test/test.repack *

Key Concepts

• Snippets: Reusable blocks of fields. Include them in records using !snippet_name. Example: base snippet for id and created_date.
• Records: Define your data models (e.g., User, Organization). Similar to classes or tables.
• Structs: Define data structures not necessarily database models but useful for your application (e.g., OrganizationDirectory).
• Outputs: Specify desired output formats and locations. Currently supported outputs (more to come):
  • description: Text description of models.
  • postgres: SQL for PostgreSQL.
  • typescript_class: TypeScript classes.
  • typescript_interface: TypeScript interfaces.
  • typescript_drizzle: TypeScript code for Drizzle ORM.
  • rust: Rust structs.
• Tags: Categorize models using #tag_name (e.g., #models, #data, #private). Used by output generators.
• Fields: Define fields with types (e.g., name string, id int32).
• Functions Add functions like db:primary_key, db:default("NOW()"), db:index("org_id").
• References: Define relationships using ref(OtherRecord.field).
• Joins Define views and join data using from(local_ref_field.field).
• Inheritance/Composition: Create record variations by including all fields (*) and adding/removing specific fields (- field_name).
• Imports: Import definitions from other .repack files using import "filename.repack".

How to Create Your Own Models

1. Create a .repack file: This is where you'll define your models.
2. Define Snippets (Optional but Recommended):

   snippet base {
       id int32 db:primary_key db:identity db:unique
       created_date datetime db:default("NOW()")
   }
   

3. Define Records:

   record User @users #private #models {
       !base // Includes id and created_date
       name string
       email string
       // ... other fields
   }
   
   record Post @posts #models {
       !base
       title string
       content string
       author_id ref(User.id) // Foreign key to User table
       db:index("author_id")
   }
   

4. Define Structs (Optional):

   struct UserProfile #data {
       user User
       posts Post[]
   }
   

5. Specify Outputs:

   // Output all records tagged with #models to a PostgreSQL schema file
   output postgres @./output/postgres_schema #models;
   
   // Output all records and structs tagged with #data as TypeScript interfaces
   output typescript_interface @./output/ts_interfaces #data {
       make_index true // Option to create an index.ts file
   }
   
   // Output Rust structs for items tagged #data
   output rust @./output/rust_structs #data;
   

6. Run the repack CLI:

   repack your_definitions.repack
   

   This will process your definitions and generate files in the specified output directories.

By combining these features, you can create a rich and well-structured definition of your data model that can then be translated into various code artifacts, saving time and ensuring consistency.