Multi-Database Quick Start (Supabase Edition)

pip install supabase python-dotenv
 
# Optional dependencies
pip install neo4j           # For graph relationships
pip install pinecone-client # For vector search
pip install openai          # For embeddings

# Required
SUPABASE_URL=postgresql://postgres:password@db.project.supabase.co:5432/postgres
SUPABASE_API_KEY=your_anon_key_here
 
# Optional
NEO4J_URI=bolt://localhost:7687
NEO4J_USERNAME=neo4j
NEO4J_PASSWORD=your_password
 
PINECONE_API_KEY=your_pinecone_key
OPENAI_API_KEY=your_openai_key

-- Sport Archetypes
CREATE TABLE IF NOT EXISTS sport_archetypes (
  id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
  archetype_name TEXT UNIQUE NOT NULL,
  display_name TEXT NOT NULL,
  code_pattern TEXT NOT NULL,
  characteristics JSONB DEFAULT '{}',
  embedding_vector JSONB,
  neo4j_node_id TEXT,
  created_at TIMESTAMPTZ DEFAULT NOW(),
  updated_at TIMESTAMPTZ DEFAULT NOW()
);
 
-- Prospective Leagues (with multi-DB sync tracking)
CREATE TABLE IF NOT EXISTS prospective_leagues (
  id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
  league_tag TEXT UNIQUE NOT NULL,
  league_name TEXT,
  sport_name TEXT,
  sport_tier TEXT DEFAULT 'TIER2',
  sport_archetype TEXT,
  status TEXT DEFAULT 'unverified',
  verification_status TEXT DEFAULT 'unverified',
  source_type TEXT DEFAULT 'web_scrape',
 
  -- Embedding fields
  embedding_vector JSONB,
  embedding_model TEXT,
  embedding_updated_at TIMESTAMPTZ,
 
  -- Neo4j sync tracking
  neo4j_node_id TEXT,
  neo4j_synced_at TIMESTAMPTZ,
  neo4j_sync_status TEXT DEFAULT 'pending',
 
  -- Pinecone sync tracking
  pinecone_id TEXT,
  pinecone_namespace TEXT,
  pinecone_synced_at TIMESTAMPTZ,
  pinecone_sync_status TEXT DEFAULT 'pending',
 
  created_at TIMESTAMPTZ DEFAULT NOW(),
  updated_at TIMESTAMPTZ DEFAULT NOW()
);
 
-- Few-Shot Examples
CREATE TABLE IF NOT EXISTS few_shot_examples (
  id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
  example_id TEXT UNIQUE NOT NULL,
  category TEXT NOT NULL,
  scenario TEXT,
  sport TEXT DEFAULT 'general',
  tier TEXT DEFAULT 'mixed',
  complexity TEXT DEFAULT 'moderate',
  version TEXT DEFAULT '2.0',
  quality_score DECIMAL DEFAULT 0.80,
  usage_count INTEGER DEFAULT 0,
  input_data JSONB DEFAULT '{}',
  output_data JSONB DEFAULT '{}',
  tags JSONB DEFAULT '[]',
  embedding_vector JSONB,
  embedding_model TEXT,
  pinecone_sync_status TEXT DEFAULT 'pending',
  created_at TIMESTAMPTZ DEFAULT NOW(),
  updated_at TIMESTAMPTZ DEFAULT NOW()
);
 
-- Indexes for performance
CREATE INDEX IF NOT EXISTS idx_leagues_archetype ON prospective_leagues(sport_archetype);
CREATE INDEX IF NOT EXISTS idx_leagues_tier ON prospective_leagues(sport_tier);
CREATE INDEX IF NOT EXISTS idx_leagues_status ON prospective_leagues(status);
CREATE INDEX IF NOT EXISTS idx_leagues_verification ON prospective_leagues(verification_status);
CREATE INDEX IF NOT EXISTS idx_examples_category ON few_shot_examples(category);
CREATE INDEX IF NOT EXISTS idx_examples_sport ON few_shot_examples(sport);

cd database
python scripts/validate_supabase_multi_db.py

✅ SUPABASE_URL: Supabase project URL
✅ SUPABASE_API_KEY: Supabase API key
✅ Supabase connection successful
✅ OpenAI connection successful

python scripts/seed_supabase_multi_db.py --archetypes

{
  "league_tag": "ufc",
  "league_name": "Ultimate Fighting Championship",
  "sport_name": "Mixed Martial Arts",
  "sport_tier": "TIER1",
  "sport_archetype": "combat",
  "status": "verified",
  "verification_status": "human_verified"
}

python scripts/seed_supabase_multi_db.py --leagues

python scripts/seed_supabase_multi_db.py --all

from supabase import create_client
import os
 
# Connect
url = os.getenv('SUPABASE_URL')
key = os.getenv('SUPABASE_API_KEY')
supabase = create_client(url, key)
 
# Query combat sports leagues
result = supabase.table('prospective_leagues')\
    .select('*')\
    .eq('sport_archetype', 'combat')\
    .execute()
 
for league in result.data:
    print(f"{league['league_name']} - {league['status']}")

-- Find all combat sports leagues
SELECT league_name, sport_tier, verification_status
FROM prospective_leagues
WHERE sport_archetype = 'combat'
ORDER BY league_name;
 
-- Get sync status summary
SELECT
  neo4j_sync_status,
  pinecone_sync_status,
  COUNT(*) as count
FROM prospective_leagues
GROUP BY neo4j_sync_status, pinecone_sync_status;
 
-- Find high-quality examples by category
SELECT example_id, scenario, quality_score
FROM few_shot_examples
WHERE category = 'triage'
  AND quality_score > 0.85
ORDER BY quality_score DESC
LIMIT 10;

JSONL Seed Files (source of truth)
         ↓
    Supabase PostgreSQL
         ↓
    ┌────┴────┐
    ↓         ↓
  Neo4j    Pinecone
  (graph)   (vectors)

result = supabase.table('prospective_leagues')\
    .select('*')\
    .eq('sport_archetype', 'team_sport')\
    .execute()

result = supabase.table('prospective_leagues')\
    .select('*')\
    .eq('verification_status', 'unverified')\
    .eq('sport_tier', 'TIER1')\
    .execute()

result = supabase.table('few_shot_examples')\
    .select('*')\
    .eq('category', 'contract_generation')\
    .gte('quality_score', 0.85)\
    .limit(5)\
    .execute()