Collections

Collections are the foundational data model in Minerva. Every collection represents a logical table in your database, but with more capability than a raw SQL table.

Definition

A collection is stored in the Blueprint and encapsulates table_metadata, an include_map for relational querying, a record_label expression, and metadata for the five standard routes.

1. Column Types

Minerva supports several rich data types tailored for business apps:

Text: Strings, enums, long-form content
Numeric: Integers, decimals, currency
Boolean: True/false flags
Date/Timestamp: Time-sensitive workflows
Record Link: Relations to rows in other collections

Minerva handles the underlying foreign key constraints and join tables automatically.

2. The `include_map`

Record Links let you build complex relationship graphs. The include_map determines which relations are eagerly fetched when you query a collection.

When you configure an include_map, Minerva handles the corresponding SQL joins automatically while still enforcing field-level access.

3. Record Labels (CEL)

The record_label is a CEL expression that computes a string label for a record.

This is useful for:

Combining name fields
Appending status text
Formatting a more useful display label than a raw ID

4. DDL Generation Pipeline

When you save changes to a collection:

Minerva validates the schema update
Generates the necessary PostgreSQL DDL statements
Computes the reverse delta for history
Applies the migration to the tenant database