mirror of
https://github.com/xtr-dev/rondevu-server.git
synced 2025-12-10 10:53:24 +00:00
Bas van den Aakster
8111cb9cec
- Changed service FQN format: service:version@username (colon instead of @) - Added service discovery: direct lookup, random selection, paginated queries - Updated parseServiceFqn to handle optional username for discovery - Removed UUID privacy layer (service_index table) - Updated storage interface with discovery methods (discoverServices, getRandomService, getServiceByFqn) - Removed deprecated methods (getServiceByUuid, queryService, listServicesForUsername, findServicesByName, touchUsername, batchCreateServices) - Updated API routes: /services/:fqn with three modes (direct, random, paginated) - Changed offer/answer/ICE routes to offer-specific: /services/:fqn/offers/:offerId/* - Added extracted fields to services table (service_name, version, username) for efficient discovery - Created migration 0007 to update schema and migrate existing data - Added discovery indexes for performance 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
55 lines
2.0 KiB
SQL
55 lines
2.0 KiB
SQL
-- V0.4.1 Migration: Simplify schema and add service discovery
|
|
-- Remove privacy layer (service_index) and add extracted fields for discovery
|
|
|
|
-- Step 1: Drop service_index table (privacy layer removal)
|
|
DROP TABLE IF EXISTS service_index;
|
|
|
|
-- Step 2: Create new services table with extracted fields for discovery
|
|
CREATE TABLE services_new (
|
|
id TEXT PRIMARY KEY,
|
|
service_fqn TEXT NOT NULL,
|
|
service_name TEXT NOT NULL,
|
|
version TEXT NOT NULL,
|
|
username TEXT NOT NULL,
|
|
created_at INTEGER NOT NULL,
|
|
expires_at INTEGER NOT NULL,
|
|
FOREIGN KEY (username) REFERENCES usernames(username) ON DELETE CASCADE,
|
|
UNIQUE(service_fqn)
|
|
);
|
|
|
|
-- Step 3: Migrate existing data (if any) - parse FQN to extract components
|
|
-- Note: This migration assumes FQN format is already "service:version@username"
|
|
-- If there's old data with different format, manual intervention may be needed
|
|
INSERT INTO services_new (id, service_fqn, service_name, version, username, created_at, expires_at)
|
|
SELECT
|
|
id,
|
|
service_fqn,
|
|
-- Extract service_name: everything before first ':'
|
|
substr(service_fqn, 1, instr(service_fqn, ':') - 1) as service_name,
|
|
-- Extract version: between ':' and '@'
|
|
substr(
|
|
service_fqn,
|
|
instr(service_fqn, ':') + 1,
|
|
instr(service_fqn, '@') - instr(service_fqn, ':') - 1
|
|
) as version,
|
|
username,
|
|
created_at,
|
|
expires_at
|
|
FROM services
|
|
WHERE service_fqn LIKE '%:%@%'; -- Only migrate properly formatted FQNs
|
|
|
|
-- Step 4: Drop old services table
|
|
DROP TABLE services;
|
|
|
|
-- Step 5: Rename new table to services
|
|
ALTER TABLE services_new RENAME TO services;
|
|
|
|
-- Step 6: Create indexes for efficient querying
|
|
CREATE INDEX idx_services_fqn ON services(service_fqn);
|
|
CREATE INDEX idx_services_discovery ON services(service_name, version);
|
|
CREATE INDEX idx_services_username ON services(username);
|
|
CREATE INDEX idx_services_expires ON services(expires_at);
|
|
|
|
-- Step 7: Create index on offers for available offer filtering
|
|
CREATE INDEX IF NOT EXISTS idx_offers_available ON offers(answerer_peer_id) WHERE answerer_peer_id IS NULL;
|