remove migrations table, use schema files instead

Author: ScriptRaccoon

CONTRIBUTING.md

@@ -69,7 +69,7 @@ to continuously run this update when a file in the subfolder [/databases/catdat/
 
 ### Troubleshooting
 
-- If the local database is corrupted, delete the `catdat.db` file and recreate it using `pnpm db:update`.
+- If the local database is corrupted, recreate it using `pnpm db:update`.
 - If the `pnpm db:update` command fails, examine the error message to determine the cause. It could be due to malformed SQL, a contradictory property, or a failing test in the `pnpm db:test` script (which also runs as part of the update command), as explained below.
 
 ### Tests for Data Quality

DATABASE.md

@@ -42,11 +42,11 @@ For functors there are similar tables, such as:
 - `functor_implications`
 - `functor_property_assignments`
 
-## Migrations vs. Data
+## Schema vs. Data
 
-Migrations update the database structure: tables, views, indexes, and triggers. They are defined in SQL files located in the subfolder [/databases/catdat/migrations](/databases/catdat/migrations/). The command `pnpm db:migrate` applies any new migrations.
+The schema defines the structure of the database: tables, views, indexes, and triggers. It is specified in several SQL files located in the subfolder [/databases/catdat/schema](/databases/catdat/schema/). The command `pnpm db:setup` deletes the old database file (if it exists) and creates a new one using this schema.
 
-Database entries (categories, properties, implications, etc.) are defined via SQL files in the subfolder [/databases/catdat/data](/databases/catdat/data). The command `pnpm db:seed` replaces the current database by clearing all existing data and inserting the entries defined in these SQL files.
+Database entries (categories, properties, implications, etc.) are defined in SQL files located in the subfolder [/databases/catdat/data](/databases/catdat/data/). The command `pnpm db:seed` replaces the current contents of the database by clearing all existing data and inserting the entries defined in these SQL files.
 
 ## Derived Data
 
@@ -60,7 +60,7 @@ The command `pnpm db:test` executes some tests and verifies that the data behave
 
 ## One command for everything
 
-Use `pnpm db:update` to run all the commands in sequence: `pnpm db:migrate`, `pnpm db:seed`,`pnpm db:deduce`, and `pnpm db:test`.
+Use `pnpm db:update` to run all the commands in sequence: `pnpm db:setup`, `pnpm db:seed`,`pnpm db:deduce`, and `pnpm db:test`.
 
 Use `pnpm db:watch` to run this command automatically every time a file in the subfolder [/databases/catdat/data](/databases/catdat/data) changes. This is useful in particular during development.
 

databases/app/scripts/migrate.ts databases/app/scripts/setup.tsdatabases/app/scripts/migrate.ts renamed to databases/app/scripts/setup.ts

@@ -13,13 +13,13 @@ const db = createClient({
 	authToken: APP_DB_AUTH_TOKEN,
 })
 
-migrate()
+setup()
 
 /**
  * Creates the tables in the app database.
  */
-async function migrate() {
-	console.info('\n--- Migrate App database ---')
+async function setup() {
+	console.info('\n--- Setup App database ---')
 	await create_visits_table()
 	await create_submissions_table()
 }

databases/catdat/migrations/014_order_assignments.sql

@@ -20,15 +20,15 @@ CREATE TABLE related_category_properties (
 );
 
 CREATE TABLE category_property_assignments (
+    id INTEGER PRIMARY KEY,
     category_id TEXT NOT NULL,
     property_id TEXT NOT NULL,
     is_satisfied INTEGER NOT NULL
         CHECK (is_satisfied in (TRUE, FALSE)),
     reason TEXT NOT NULL CHECK (length(reason) > 0),
     is_deduced INTEGER NOT NULL DEFAULT FALSE
         CHECK (is_deduced in (TRUE, FALSE)),
-    position INTEGER NOT NULL DEFAULT 0,
-    PRIMARY KEY (category_id, property_id),
+    UNIQUE (category_id, property_id),
     FOREIGN KEY (category_id) REFERENCES categories (id) ON DELETE CASCADE,
     FOREIGN KEY (property_id) REFERENCES category_properties (id) ON DELETE CASCADE
 );