Replace and expand StarForge database schema

[CorentynDevPro]

Replaces the previous schema with a comprehensive, idempotent Supabase/Postgres schema for StarForge. Adds canonical game data tables (troops, kingdoms, classes, weapons, pets, spells, aspects, medals, events), player data tables (heroes, hero_troops, pets, artifacts, teams, progress, PvP, guilds, roles, permissions, raw dumps, job logs), utility functions, triggers, materialized views, and indexes. Designed for Supabase Cloud and includes detailed instructions and notes for setup and usage.

[Copilot]

Pull Request Overview

This PR replaces the entire StarForge database schema with a comprehensive, idempotent Supabase/Postgres schema. The new schema significantly expands functionality by adding canonical game data tables for a game (troops, kingdoms, classes, weapons, pets, spells, aspects, medals, events) and player data tables (heroes, hero_troops, pets, artifacts, teams, progress, PvP data, and more).

Key changes:

• Complete schema replacement with 40+ new tables for game and player data
• Idempotent design using IF NOT EXISTS and DO blocks for triggers
• Added materialized view for optimized troop data exports
• Comprehensive trigger system for auto-updating timestamps across all tables

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

[nitpick] Missing index on the foreign key team_saves.hero_id. Although there is an index idx_team_saves_hero defined at line 1149, it should be created earlier in the index section (lines 2046-2055) for consistency with other indexes.

[Copilot]

[nitpick] Missing index on the foreign key hero_pvp_stats.hero_id. Although there is a UNIQUE constraint on hero_id (line 1277-1280), a UNIQUE constraint automatically creates an index, so this is fine. However, for clarity and consistency with other tables, consider explicitly documenting this in comments.

[Copilot]

Missing index on the foreign key hero_progress_weapons.hero_id. Add an index to improve query performance:

CREATE INDEX IF NOT EXISTS idx_hero_progress_weapons_hero ON hero_progress_weapons (hero_id);

Suggested change

	-- Add index to improve query performance for foreign key hero_id
	CREATE INDEX IF NOT EXISTS idx_hero_progress_weapons_hero ON hero_progress_weapons (hero_id);

[Copilot]

The DROP MATERIALIZED VIEW IF EXISTS statement should be followed by CASCADE to avoid errors if there are dependent objects:

DROP MATERIALIZED VIEW IF EXISTS troops_master_light CASCADE;

Suggested change

	MATERIALIZED VIEW IF EXISTS troops_master_light;
	MATERIALIZED VIEW IF EXISTS troops_master_light CASCADE;

[Copilot]

Consider adding an index on raw_profiles.captured_at to improve performance when querying for recent profiles or filtering by time range:

CREATE INDEX IF NOT EXISTS idx_raw_profiles_captured_at ON raw_profiles (captured_at);

Suggested change

	-- Index to improve performance for time-based queries on raw_profiles
	CREATE INDEX IF NOT EXISTS idx_raw_profiles_captured_at ON raw_profiles (captured_at);

[Copilot]

The table app_users is referenced in foreign keys, but the old schema used users. This is a breaking change from the previous schema. Ensure that any existing application code that references users table is updated to use app_users, or consider maintaining backward compatibility with a view.

Suggested change

	-- Backward compatibility: provide a view named "users" for legacy code
	CREATE OR REPLACE VIEW users AS
	SELECT
	id,
	email,
	username,
	display_name,
	avatar_url,
	created_at,
	updated_at
	FROM app_users;

[Copilot]

The old schema had analytics_events, scheduled_jobs, and battle_simulations tables that are not present in the new schema. If these features are still needed, ensure they are implemented elsewhere or document their removal.

Suggested change

	-- NOTE: The old schema included the following tables, which have been removed in this version:
	-- - analytics_events
	-- - scheduled_jobs
	-- - battle_simulations
	-- If these features are still needed, please implement them elsewhere or restore the relevant tables.
	-- Their removal is intentional; see project documentation for details.

[Copilot]

The audit_logs.ip_address column is defined as text but was previously defined as INET in the old schema. The INET data type provides better validation and storage efficiency for IP addresses. Consider using inet instead of text:

ip_address inet,

Suggested change

	ip_address text,
	ip_address inet,

[Copilot]

The hero_meta_json table does not have an updated_at column, but a trigger hero_meta_json_updated_at_tr is created for it (lines 1993-2000). This will cause a runtime error when the trigger attempts to set NEW.updated_at. Add an updated_at timestamptz DEFAULT now() column to the table.

[Copilot]

The cache_invalidation table does not have an updated_at column, but a trigger cache_invalidation_updated_at_tr is created for it (lines 2036-2043). This will cause a runtime error when the trigger attempts to set NEW.updated_at. Add an updated_at timestamptz DEFAULT now() column to the table.

Suggested change

	)
	);
	),
	updated_at timestamptz DEFAULT now()
	);