malkhut: DuckDB file-backed asset store — schema, sync, queries, persistence

DuckDB store (asset_store.py): - 4 tables: assets (28 cols), exchanges (10 cols), asset_exchanges (junction), behavior_profiles (28 cols) - Schema with indexes on blockchain, coingecko_id, cmc_id, exchange_id - sync_from_profiles(): populate from in-memory dicts in one call - Query API: query_assets(**filters), get_asset(), symbols_for_exchange(), assets_on_blockchain(), asset_count(), exchange_count() - Case-insensitive exchange lookup - File-backed persistence: data survives connection close/reopen - Performance: <1s sync, 100 queries in <1s, 1000 gets in <1s 30 tests covering: - Schema creation and table structure (3 tests) - Sync from profiles (6 tests, idempotent) - Asset CRUD (7 tests, query by blockchain/coingecko/sector) - Exchange mapping (5 tests, case-insensitive) - Behavior profiles (4 tests) - Persistence across connections (2 tests) - Performance baseline (3 tests) Total: 1276 tests across 50 files, all green, zero regressions.
2026-07-12 12:21:48 +02:00
parent 7544955634
commit 7f27ed22c8
4 changed files with 533 additions and 3 deletions
--- a/MALKHUT/malkhut/tests/test_asset_store.py
+++ b/MALKHUT/malkhut/tests/test_asset_store.py
@@ -0,0 +1,253 @@
+"""
+Tests for DuckDB asset store — schema, sync, queries, persistence.
+
+Covers:
+- Schema creation and table structure
+- Sync from in-memory profiles → DuckDB
+- Asset CRUD (upsert, query, get)
+- Exchange-asset mapping
+- Behavior profiles
+- Cross-exchange queries
+- Blockchain filtering
+- Case-insensitive exchange lookup
+- Persistence across connections
+- Performance baseline
+"""
+import pytest
+import os
+from pathlib import Path
+
+from malkhut.storage.asset_store import AssetStore
+
+
+@pytest.fixture
+def store(tmp_path):
+    """Create a fresh DuckDB store for each test."""
+    db_path = str(tmp_path / "test_assets.duckdb")
+    s = AssetStore(db_path=db_path)
+    yield s
+    s.close()
+
+
+class TestSchema:
+    def test_tables_created(self, store):
+        tables = store.conn.execute(
+            "SELECT table_name FROM information_schema.tables WHERE table_schema='main'"
+        ).fetchall()
+        names = {t[0] for t in tables}
+        assert "assets" in names
+        assert "exchanges" in names
+        assert "asset_exchanges" in names
+        assert "behavior_profiles" in names
+
+    def test_asset_table_columns(self, store):
+        cols = store.conn.execute(
+            "SELECT column_name FROM information_schema.columns WHERE table_name='assets'"
+        ).fetchall()
+        col_names = [c[0] for c in cols]
+        assert "symbol" in col_names
+        assert "base_asset" in col_names
+        assert "coingecko_id" in col_names
+        assert "sectors" in col_names
+        assert "tick_size" in col_names
+
+    def test_indexes_created(self, store):
+        indexes = store.conn.execute(
+            "SELECT name FROM sqlite_master WHERE type='index' AND tbl_name='assets'"
+        ).fetchall()
+        idx_names = {i[0] for i in indexes}
+        assert any("blockchain" in n for n in idx_names)
+        assert any("coingecko" in n for n in idx_names)
+
+
+class TestSyncFromProfiles:
+    def test_sync_returns_count(self, store):
+        count = store.sync_from_profiles()
+        assert count >= 13
+
+    def test_sync_populates_assets(self, store):
+        store.sync_from_profiles()
+        assert store.asset_count() >= 13
+
+    def test_sync_populates_exchanges(self, store):
+        store.sync_from_profiles()
+        assert store.exchange_count() >= 3
+
+    def test_sync_populates_asset_exchanges(self, store):
+        store.sync_from_profiles()
+        exs = store.get_asset_exchanges("BTCUSDT")
+        assert len(exs) >= 1
+
+    def test_sync_populates_behaviors(self, store):
+        store.sync_from_profiles()
+        row = store.conn.execute(
+            "SELECT * FROM behavior_profiles WHERE symbol = 'BTCUSDT'"
+        ).fetchone()
+        assert row is not None
+
+    def test_sync_is_idempotent(self, store):
+        store.sync_from_profiles()
+        count1 = store.asset_count()
+        store.sync_from_profiles()
+        count2 = store.asset_count()
+        assert count1 == count2
+
+
+class TestAssetCRUD:
+    def test_get_asset_hit(self, store):
+        store.sync_from_profiles()
+        btc = store.get_asset("BTCUSDT")
+        assert btc is not None
+        assert btc["base_asset"] == "BTC"
+        assert btc["name"] == "BTC"
+
+    def test_get_asset_miss(self, store):
+        store.sync_from_profiles()
+        assert store.get_asset("NONEXISTENT") is None
+
+    def test_get_asset_all_fields(self, store):
+        store.sync_from_profiles()
+        btc = store.get_asset("BTCUSDT")
+        assert btc["coingecko_id"] == "bitcoin"
+        assert btc["cmc_id"] == 1
+        assert btc["blockchain"] == "bitcoin"
+        assert btc["unified_symbol"] == "BTC/USDT"
+        assert btc["quote_currency"] == "USDT"
+        assert btc["tick_size"] == 0.1
+        assert btc["lot_size"] == 0.001
+
+    def test_query_assets_by_blockchain(self, store):
+        store.sync_from_profiles()
+        eth_assets = store.query_assets(blockchain="ethereum")
+        assert len(eth_assets) >= 5
+        symbols = {a["symbol"] for a in eth_assets}
+        assert "ETHUSDT" in symbols
+        assert "UNIUSDT" in symbols
+
+    def test_query_assets_by_coingecko(self, store):
+        store.sync_from_profiles()
+        results = store.query_assets(coingecko_id="bitcoin")
+        assert len(results) == 1
+        assert results[0]["symbol"] == "BTCUSDT"
+
+    def test_query_assets_by_sector(self, store):
+        store.sync_from_profiles()
+        defi = store.query_assets(sectors=["defi"])
+        assert len(defi) >= 2
+        symbols = {a["symbol"] for a in defi}
+        assert "UNIUSDT" in symbols
+
+    def test_asset_count(self, store):
+        store.sync_from_profiles()
+        assert store.asset_count() == 13
+
+
+class TestExchangeMapping:
+    def test_symbols_for_exchange(self, store):
+        store.sync_from_profiles()
+        binance_assets = store.symbols_for_exchange("binance")
+        assert len(binance_assets) >= 13
+
+    def test_symbols_for_exchange_case_insensitive(self, store):
+        store.sync_from_profiles()
+        upper = store.symbols_for_exchange("BINANCE")
+        lower = store.symbols_for_exchange("binance")
+        assert upper == lower
+
+    def test_symbols_for_unknown_exchange(self, store):
+        store.sync_from_profiles()
+        assert store.symbols_for_exchange("KRAKEN") == []
+
+    def test_assets_on_blockchain(self, store):
+        store.sync_from_profiles()
+        eth = store.assets_on_blockchain("ethereum")
+        assert len(eth) >= 5
+
+    def test_assets_on_unknown_blockchain(self, store):
+        store.sync_from_profiles()
+        assert store.assets_on_blockchain("nonexistent") == []
+
+
+class TestBehaviorProfiles:
+    def test_behavior_stored(self, store):
+        store.sync_from_profiles()
+        row = store.conn.execute(
+            "SELECT depth_amplitude_usd, vol_annualized_normal FROM behavior_profiles WHERE symbol='BTCUSDT'"
+        ).fetchone()
+        assert row[0] == 750_000.0
+        assert row[1] == 35.0
+
+    def test_behavior_template(self, store):
+        store.sync_from_profiles()
+        row = store.conn.execute(
+            "SELECT template_name FROM behavior_profiles WHERE symbol='BTCUSDT'"
+        ).fetchone()
+        assert row[0] == "institutional_blue_chip"
+
+    def test_behavior_reference_price(self, store):
+        store.sync_from_profiles()
+        row = store.conn.execute(
+            "SELECT reference_price FROM behavior_profiles WHERE symbol='BTCUSDT'"
+        ).fetchone()
+        assert row[0] == 64000.0
+
+    def test_behavior_liquidation(self, store):
+        store.sync_from_profiles()
+        row = store.conn.execute(
+            "SELECT liq_speed, liq_recovery FROM behavior_profiles WHERE symbol='BTCUSDT'"
+        ).fetchone()
+        assert row[0] == "slow"
+        assert row[1] == "fast"
+
+
+class TestPersistence:
+    def test_data_survives_reopen(self, tmp_path):
+        db_path = str(tmp_path / "persist.duckdb")
+        s1 = AssetStore(db_path=db_path)
+        s1.sync_from_profiles()
+        s1.close()
+
+        s2 = AssetStore(db_path=db_path)
+        assert s2.asset_count() >= 13
+        btc = s2.get_asset("BTCUSDT")
+        assert btc["base_asset"] == "BTC"
+        s2.close()
+
+    def test_exchange_mapping_survives(self, tmp_path):
+        db_path = str(tmp_path / "persist.duckdb")
+        s1 = AssetStore(db_path=db_path)
+        s1.sync_from_profiles()
+        exs = s1.get_asset_exchanges("BTCUSDT")
+        s1.close()
+
+        s2 = AssetStore(db_path=db_path)
+        exs2 = s2.get_asset_exchanges("BTCUSDT")
+        assert exs == exs2
+        s2.close()
+
+
+class TestPerformanceBaseline:
+    def test_sync_speed(self, store):
+        import time
+        t0 = time.time()
+        store.sync_from_profiles()
+        elapsed = time.time() - t0
+        assert elapsed < 1.0, f"Sync took {elapsed:.2f}s"
+
+    def test_query_speed(self, store):
+        import time
+        store.sync_from_profiles()
+        t0 = time.time()
+        for _ in range(100):
+            store.query_assets(blockchain="ethereum")
+        elapsed = time.time() - t0
+        assert elapsed < 1.0, f"100 queries took {elapsed:.2f}s"
+
+    def test_get_speed(self, store):
+        import time
+        store.sync_from_profiles()
+        t0 = time.time()
+        for _ in range(1000):
+            store.get_asset("BTCUSDT")
+        elapsed = time.time() - t0
+        assert elapsed < 1.0, f"1000 gets took {elapsed:.2f}s"