lumibot 4.0.23__py3-none-any.whl → 4.1.0__py3-none-any.whl

tests/backtest/test_daily_data_timestamp_comparison.py

@@ -0,0 +1,801 @@
+"""
+CRITICAL TEST: Daily Data Timestamp & Price Accuracy Comparison
+
+This test verifies that daily (day timeframe) data from ThetaData and Polygon:
+1. Has IDENTICAL timestamps (no day shifts, no hour shifts, no timezone bugs)
+2. Has matching OHLC prices (within penny-level tolerance)
+3. Covers FULL MONTH of data (minimum 20 trading days)
+4. Tests MULTIPLE symbols (different exchanges, characteristics)
+5. Handles edge cases (holidays, month boundaries, extended hours)
+
+ANY failure in this test indicates a CRITICAL bug that could cause:
+- Incorrect backtests
+- Wrong trading signals
+- Financial losses
+- Lawsuits
+
+ZERO TOLERANCE for failures.
+"""
+
+import os
+import pytest
+import datetime
+import pandas as pd
+from dotenv import load_dotenv
+from lumibot.backtesting import ThetaDataBacktesting, PolygonDataBacktesting
+from lumibot.entities import Asset
+from lumibot.tools import thetadata_helper
+from lumibot.tools.polygon_helper import get_price_data_from_polygon as polygon_get_price_data
+
+load_dotenv()
+
+
+@pytest.mark.apitest
+class TestDailyDataTimestampComparison:
+    """
+    Comprehensive daily data comparison between ThetaData and Polygon.
+    Tests full month, multiple symbols, penny-level accuracy.
+    """
+
+    def test_daily_data_full_month_pltr(self):
+        """Test PLTR daily data for full September 2025 - ZERO tolerance."""
+        self._test_symbol_daily_data(
+            symbol="PLTR",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=19
+        )
+
+    def test_daily_data_full_month_spy(self):
+        """Test SPY daily data for full September 2025 - ZERO tolerance."""
+        self._test_symbol_daily_data(
+            symbol="SPY",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=19
+        )
+
+    def test_daily_data_full_month_aapl(self):
+        """Test AAPL daily data for full September 2025 - ZERO tolerance."""
+        self._test_symbol_daily_data(
+            symbol="AAPL",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=19
+        )
+
+    def test_daily_data_full_month_amzn(self):
+        """Test AMZN daily data for full September 2025 - ZERO tolerance."""
+        self._test_symbol_daily_data(
+            symbol="AMZN",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=19
+        )
+
+    # ========== INDEX TESTS ==========
+    def test_daily_data_full_month_spx_index(self):
+        """Test SPX index daily data for full September 2025 - ZERO tolerance."""
+        self._test_symbol_daily_data(
+            symbol="SPX",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=19,
+            asset_type="index"
+        )
+
+    def test_daily_data_full_month_vix_index(self):
+        """Test VIX index daily data for full September 2025 - ZERO tolerance."""
+        self._test_symbol_daily_data(
+            symbol="VIX",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=19,
+            asset_type="index"
+        )
+
+    def test_daily_data_full_month_ndx_index(self):
+        """Test SPX index daily data for full August 2024 - ZERO tolerance."""
+        self._test_symbol_daily_data(
+            symbol="SPX",
+            start_date=datetime.datetime(2024, 8, 1),
+            end_date=datetime.datetime(2024, 8, 31),
+            min_trading_days=21,
+            asset_type="index"
+        )
+
+    # ========== OPTION TESTS ==========
+    def test_daily_data_spy_call_option(self):
+        """Test SPY call option daily data for September 2025 - ZERO tolerance."""
+        self._test_option_daily_data(
+            symbol="SPY",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=15,  # Options may have less liquidity
+            expiration=datetime.datetime(2025, 12, 19),  # Dec 2025 expiry
+            strike=580.0,  # ATM/slightly OTM for SPY ~$570
+            right="CALL"
+        )
+
+    def test_daily_data_spy_put_option(self):
+        """Test SPY put option daily data for September 2025 - ZERO tolerance."""
+        self._test_option_daily_data(
+            symbol="SPY",
+            start_date=datetime.datetime(2025, 9, 1),
+            end_date=datetime.datetime(2025, 9, 30),
+            min_trading_days=15,
+            expiration=datetime.datetime(2025, 12, 19),
+            strike=560.0,  # ATM/slightly ITM for SPY ~$570
+            right="PUT"
+        )
+
+    def _test_option_daily_data(self, symbol, start_date, end_date, min_trading_days, expiration, strike, right):
+        """
+        Test function for option daily data validation.
+
+        NOTE: Options comparison is challenging because:
+        - Yahoo Finance historical option data is limited (often <1 week)
+        - Different providers use different quote/trade data
+        - Options have wider bid-ask spreads than stocks
+        - No universal "official" option price like stocks have
+
+        This test validates:
+        1. ThetaData returns option data successfully
+        2. Minimum number of trading days
+        3. Price values are reasonable (not zero, not negative)
+        4. Volume data exists
+        """
+        username = os.environ.get("THETADATA_USERNAME")
+        password = os.environ.get("THETADATA_PASSWORD")
+
+        asset = Asset(symbol, asset_type="option", expiration=expiration, strike=strike, right=right)
+
+        print(f"\n{'='*80}")
+        print(f"TESTING {symbol} {right} ${strike} (exp {expiration.date()}) OPTION DAILY DATA")
+        print(f"Period: {start_date.date()} to {end_date.date()}")
+        print(f"{'='*80}")
+
+        # ==== GET THETADATA OPTION DATA ====
+        print(f"\n1. Fetching ThetaData option daily data...")
+        try:
+            theta_df = thetadata_helper.get_price_data(
+                username=username,
+                password=password,
+                asset=asset,
+                start=start_date,
+                end=end_date,
+                timespan="day"
+            )
+        except Exception as e:
+            pytest.fail(f"CRITICAL: ThetaData option daily data FAILED: {e}")
+
+        if theta_df is None or len(theta_df) == 0:
+            pytest.fail(f"CRITICAL: ThetaData returned NO option daily data")
+
+        print(f"   ✓ ThetaData: {len(theta_df)} daily bars")
+        print(f"   Date range: {theta_df.index[0]} to {theta_df.index[-1]}")
+
+        # ==== GET POLYGON OPTION DATA FOR COMPARISON ====
+        print(f"\n2. Fetching Polygon option data for validation...")
+        polygon_api_key = os.environ.get("POLYGON_API_KEY")
+
+        try:
+            polygon_df = polygon_get_price_data(
+                api_key=polygon_api_key,
+                asset=asset,
+                start=start_date,
+                end=end_date,
+                timespan="day",
+                quote_asset=Asset("USD", asset_type="forex")
+            )
+
+            if polygon_df is None or len(polygon_df) == 0:
+                print(f"   ⚠ WARNING: Polygon returned NO option data - skipping price comparison")
+                polygon_df = None
+            else:
+                print(f"   ✓ Polygon: {len(polygon_df)} daily bars")
+                print(f"   Date range: {polygon_df.index[0]} to {polygon_df.index[-1]}")
+        except Exception as e:
+            print(f"   ⚠ WARNING: Polygon failed ({e}) - skipping price comparison")
+            polygon_df = None
+
+        # ==== CHECK: Minimum Trading Days ====
+        print(f"\n3. Verifying minimum trading days...")
+        assert len(theta_df) >= min_trading_days, \
+            f"CRITICAL: Expected at least {min_trading_days} days, got {len(theta_df)}"
+        print(f"   ✓ Sufficient trading days: {len(theta_df)} >= {min_trading_days}")
+
+        # ==== CHECK: Price Comparison (if Polygon data available) ====
+        if polygon_df is not None and len(polygon_df) > 0:
+            print(f"\n4. Verifying OHLC prices vs Polygon (half-penny tolerance: $0.005)...")
+
+            # Check same number of days
+            if len(theta_df) != len(polygon_df):
+                print(f"\n   ✗ MISMATCH: ThetaData={len(theta_df)} days, Polygon={len(polygon_df)} days")
+                pytest.fail(f"CRITICAL: Different number of trading days")
+
+            # Align data
+            max_diff = {'open': 0.0, 'high': 0.0, 'low': 0.0, 'close': 0.0}
+            comparison_data = []
+
+            for theta_idx, polygon_idx in zip(theta_df.index, polygon_df.index):
+                theta_row = theta_df.loc[theta_idx]
+                polygon_row = polygon_df.loc[polygon_idx]
+
+                diffs = {
+                    'open': abs(theta_row['open'] - polygon_row['open']),
+                    'high': abs(theta_row['high'] - polygon_row['high']),
+                    'low': abs(theta_row['low'] - polygon_row['low']),
+                    'close': abs(theta_row['close'] - polygon_row['close'])
+                }
+
+                for field in ['open', 'high', 'low', 'close']:
+                    max_diff[field] = max(max_diff[field], diffs[field])
+
+                comparison_data.append({
+                    'date': theta_idx.date(),
+                    'theta_close': theta_row['close'],
+                    'polygon_close': polygon_row['close'],
+                    'diff_close': diffs['close'],
+                })
+
+            # HALF-PENNY tolerance ($0.005) - anything more is unacceptable
+            tolerance = 0.005
+            failures = []
+
+            for field in ['open', 'high', 'low', 'close']:
+                if max_diff[field] > tolerance:
+                    failures.append(f"{field}: max diff ${max_diff[field]:.4f}")
+
+            if failures:
+                print(f"\n   ✗ PRICE TOLERANCE EXCEEDED:")
+                for failure in failures:
+                    print(f"      {failure}")
+
+                print(f"\n   PRICE COMPARISON (first 10 days):")
+                print(f"   {'Date':<12} {'Theta':<10} {'Polygon':<10} {'Diff':<10}")
+                print(f"   {'-'*50}")
+                for row in comparison_data[:10]:
+                    t_close = row['theta_close']
+                    p_close = row['polygon_close']
+                    diff = row['diff_close']
+                    match_str = "✅" if diff <= tolerance else "❌"
+                    print(f"   {row['date']} ${t_close:<9.2f} ${p_close:<9.2f} ${diff:<9.4f} {match_str}")
+
+                pytest.fail(f"CRITICAL: Option price tolerance exceeded: {', '.join(failures)}")
+
+            print(f"   ✓ All prices within ${tolerance:.3f} tolerance")
+            print(f"      Max differences: open=${max_diff['open']:.4f}, high=${max_diff['high']:.4f}, "
+                  f"low=${max_diff['low']:.4f}, close=${max_diff['close']:.4f}")
+
+        # ==== CHECK: Price Data Sanity ====
+        print(f"\n5. Verifying price data sanity...")
+
+        # Check for zero or negative prices (invalid)
+        zero_prices = (theta_df['close'] <= 0).sum()
+        if zero_prices > 0:
+            pytest.fail(f"CRITICAL: {zero_prices} bars have zero/negative close prices")
+
+        # Check for reasonable price ranges
+        min_price = theta_df['close'].min()
+        max_price = theta_df['close'].max()
+        avg_price = theta_df['close'].mean()
+
+        print(f"   ✓ All prices positive")
+        print(f"      Price range: ${min_price:.2f} - ${max_price:.2f} (avg: ${avg_price:.2f})")
+
+        # ==== CHECK: OHLC Consistency ====
+        print(f"\n4. Verifying OHLC consistency...")
+
+        # High should be >= Low for every bar
+        invalid_hl = (theta_df['high'] < theta_df['low']).sum()
+        if invalid_hl > 0:
+            pytest.fail(f"CRITICAL: {invalid_hl} bars have high < low")
+
+        # High should be >= Open and Close
+        invalid_h = ((theta_df['high'] < theta_df['open']) | (theta_df['high'] < theta_df['close'])).sum()
+        if invalid_h > 0:
+            pytest.fail(f"CRITICAL: {invalid_h} bars have high < open/close")
+
+        # Low should be <= Open and Close
+        invalid_l = ((theta_df['low'] > theta_df['open']) | (theta_df['low'] > theta_df['close'])).sum()
+        if invalid_l > 0:
+            pytest.fail(f"CRITICAL: {invalid_l} bars have low > open/close")
+
+        print(f"   ✓ OHLC relationships valid (high >= low, high >= open/close, low <= open/close)")
+
+        # ==== CHECK: Volume Data ====
+        print(f"\n5. Verifying volume data...")
+        zero_volume = (theta_df['volume'] == 0).sum()
+        pct_zero_vol = (zero_volume / len(theta_df)) * 100
+
+        print(f"   ✓ Volume data present ({zero_volume}/{len(theta_df)} bars with zero volume = {pct_zero_vol:.1f}%)")
+        if pct_zero_vol > 50:
+            print(f"   ⚠ WARNING: >50% of bars have zero volume (may indicate low liquidity)")
+
+        print(f"\n{'='*80}")
+        print(f"✓✓✓ {symbol} OPTION DATA VALIDATION PASSED ✓✓✓")
+        print(f"    Trading days: {len(theta_df)}")
+        print(f"    Price range: ${min_price:.2f} - ${max_price:.2f}")
+        print(f"    OHLC relationships: VALID")
+        print(f"    Period: {theta_df.index[0].date()} to {theta_df.index[-1].date()}")
+        print(f"{'='*80}\n")
+
+    def _test_symbol_daily_data(self, symbol, start_date, end_date, min_trading_days, asset_type="stock"):
+        """
+        Core test function that validates daily data for a symbol.
+
+        CRITICAL CHECKS:
+        1. Both sources return data
+        2. Same number of trading days
+        3. IDENTICAL timestamps (no shifts)
+        4. OHLC within 0.01 (penny) tolerance
+        5. Volume reasonable
+        6. No duplicate dates
+        7. No missing dates (within market calendar)
+        """
+        username = os.environ.get("THETADATA_USERNAME")
+        password = os.environ.get("THETADATA_PASSWORD")
+        polygon_api_key = os.environ.get("POLYGON_API_KEY")
+
+        asset = Asset(symbol, asset_type=asset_type)
+
+        print(f"\n{'='*80}")
+        print(f"TESTING {symbol} DAILY DATA: {start_date.date()} to {end_date.date()}")
+        print(f"{'='*80}")
+
+        # ==== GET THETADATA DAILY DATA ====
+        print(f"\n1. Fetching ThetaData daily data...")
+        try:
+            theta_df = thetadata_helper.get_price_data(
+                username=username,
+                password=password,
+                asset=asset,
+                start=start_date,
+                end=end_date,
+                timespan="day"
+            )
+        except Exception as e:
+            pytest.fail(f"CRITICAL: ThetaData daily data FAILED for {symbol}: {e}")
+
+        if theta_df is None or len(theta_df) == 0:
+            pytest.fail(f"CRITICAL: ThetaData returned NO daily data for {symbol}")
+
+        print(f"   ✓ ThetaData: {len(theta_df)} daily bars")
+        print(f"   Date range: {theta_df.index[0]} to {theta_df.index[-1]}")
+
+        # ==== GET POLYGON OR YAHOO DAILY DATA ====
+        # NOTE: Polygon requires paid plan for indexes, so we use Yahoo Finance for indexes
+        if asset_type == "index":
+            print(f"\n2. Fetching Yahoo Finance daily data (indexes not available in free Polygon)...")
+            import yfinance as yf
+
+            # Yahoo Finance uses ^SPX for SPX, ^VIX for VIX, ^NDX for NDX
+            yahoo_symbol = f"^{symbol}" if symbol in ["SPX", "VIX", "NDX", "RUT", "DJI"] else symbol
+            ticker = yf.Ticker(yahoo_symbol)
+
+            try:
+                from datetime import timedelta as td
+                # Yahoo requires end_date to be exclusive (next day)
+                yahoo_end = (end_date + td(days=1)).strftime('%Y-%m-%d')
+                yahoo_start = start_date.strftime('%Y-%m-%d')
+                yahoo_hist = ticker.history(start=yahoo_start, end=yahoo_end, interval='1d')
+
+                if yahoo_hist is None or len(yahoo_hist) == 0:
+                    pytest.fail(f"CRITICAL: Yahoo Finance returned NO daily data for {symbol}")
+
+                # Convert Yahoo data to match our format
+                polygon_df = pd.DataFrame({
+                    'open': yahoo_hist['Open'],
+                    'high': yahoo_hist['High'],
+                    'low': yahoo_hist['Low'],
+                    'close': yahoo_hist['Close'],
+                    'volume': yahoo_hist['Volume']
+                })
+                polygon_df.index = pd.to_datetime(polygon_df.index).tz_convert('UTC')
+
+            except Exception as e:
+                pytest.fail(f"CRITICAL: Yahoo Finance daily data FAILED for {symbol}: {e}")
+        else:
+            print(f"\n2. Fetching Polygon daily data...")
+            try:
+                polygon_df = polygon_get_price_data(
+                    api_key=polygon_api_key,
+                    asset=asset,
+                    start=start_date,
+                    end=end_date,
+                    timespan="day",
+                    quote_asset=Asset("USD", asset_type="forex")
+                )
+            except Exception as e:
+                pytest.fail(f"CRITICAL: Polygon daily data FAILED for {symbol}: {e}")
+
+            if polygon_df is None or len(polygon_df) == 0:
+                pytest.fail(f"CRITICAL: Polygon returned NO daily data for {symbol}")
+
+        comparison_source = "Yahoo Finance" if asset_type == "index" else "Polygon"
+        print(f"   ✓ {comparison_source}: {len(polygon_df)} daily bars")
+        print(f"   Date range: {polygon_df.index[0]} to {polygon_df.index[-1]}")
+
+        # ==== CHECK 1: Minimum Trading Days ====
+        print(f"\n3. Verifying minimum trading days...")
+        assert len(theta_df) >= min_trading_days, \
+            f"CRITICAL: ThetaData has only {len(theta_df)} days, expected >={min_trading_days}"
+        assert len(polygon_df) >= min_trading_days, \
+            f"CRITICAL: Polygon has only {len(polygon_df)} days, expected >={min_trading_days}"
+        print(f"   ✓ Both sources have >={min_trading_days} trading days")
+
+        # ==== CHECK 2: Same Number of Days ====
+        print(f"\n4. Verifying same number of trading days...")
+        if len(theta_df) != len(polygon_df):
+            print(f"\n   ✗ MISMATCH: ThetaData={len(theta_df)} days, Polygon={len(polygon_df)} days")
+            print(f"\n   ThetaData dates:")
+            for dt in theta_df.index:
+                print(f"      {dt.date()}")
+            print(f"\n   Polygon dates:")
+            for dt in polygon_df.index:
+                print(f"      {dt.date()}")
+            pytest.fail(f"CRITICAL: Different number of trading days: Theta={len(theta_df)}, Polygon={len(polygon_df)}")
+        print(f"   ✓ Same number of trading days: {len(theta_df)}")
+
+        # ==== CHECK 3: IDENTICAL TIMESTAMPS ====
+        print(f"\n5. Verifying IDENTICAL timestamps (ZERO tolerance for shifts)...")
+
+        # Convert to date for comparison (ignore time component)
+        theta_dates = [dt.date() for dt in theta_df.index]
+        polygon_dates = [dt.date() for dt in polygon_df.index]
+
+        mismatched_dates = []
+        for i, (theta_date, polygon_date) in enumerate(zip(theta_dates, polygon_dates)):
+            if theta_date != polygon_date:
+                mismatched_dates.append((i, theta_date, polygon_date))
+
+        if mismatched_dates:
+            print(f"\n   ✗ CRITICAL: TIMESTAMP MISMATCH DETECTED!")
+            print(f"\n   {'Index':<10} {'ThetaData':<15} {'Polygon':<15} {'Shift (days)'}")
+            print(f"   {'-'*60}")
+            for idx, theta_date, polygon_date in mismatched_dates:
+                shift = (theta_date - polygon_date).days
+                print(f"   {idx:<10} {theta_date} {polygon_date} {shift:+d}")
+            pytest.fail(f"CRITICAL: {len(mismatched_dates)} timestamp mismatches found!")
+
+        print(f"   ✓ ALL timestamps match perfectly (0 shifts)")
+
+        # ==== CHECK 4: OHLC PRICE ACCURACY ====
+        print(f"\n6. Verifying OHLC prices (penny-level tolerance: $0.01)...")
+
+        # Create aligned DataFrame for comparison
+        comparison_data = []
+        max_diff = {'open': 0.0, 'high': 0.0, 'low': 0.0, 'close': 0.0}
+
+        for theta_idx, polygon_idx in zip(theta_df.index, polygon_df.index):
+            theta_row = theta_df.loc[theta_idx]
+            polygon_row = polygon_df.loc[polygon_idx]
+
+            diffs = {
+                'open': abs(theta_row['open'] - polygon_row['open']),
+                'high': abs(theta_row['high'] - polygon_row['high']),
+                'low': abs(theta_row['low'] - polygon_row['low']),
+                'close': abs(theta_row['close'] - polygon_row['close'])
+            }
+
+            for field in ['open', 'high', 'low', 'close']:
+                max_diff[field] = max(max_diff[field], diffs[field])
+
+            comparison_data.append({
+                'date': theta_idx.date(),
+                'theta_open': theta_row['open'],
+                'poly_open': polygon_row['open'],
+                'diff_open': diffs['open'],
+                'theta_close': theta_row['close'],
+                'poly_close': polygon_row['close'],
+                'diff_close': diffs['close'],
+            })
+
+        # TOLERANCE: Stocks require ZERO tolerance, indexes allow fractional cent (rounding)
+        # Stocks: ZERO tolerance - regulated data must match EXACTLY
+        # Indexes: $0.001 tolerance - calculated values may have fractional cent rounding differences
+        tolerance = 0.001 if asset_type == "index" else 0.00
+        failures = []
+
+        for field in ['open', 'high', 'low', 'close']:
+            if max_diff[field] > tolerance:
+                failures.append(f"{field}: max diff ${max_diff[field]:.4f}")
+
+        if failures:
+            # Add Yahoo Finance 3-way comparison for failed days
+            import yfinance as yf
+
+            print(f"\n   ✗ PRICE TOLERANCE EXCEEDED:")
+            for failure in failures:
+                print(f"      {failure}")
+
+            print(f"\n   3-WAY COMPARISON (ThetaData vs Polygon vs Yahoo):")
+            print(f"   {'Date':<12} {'Theta':<10} {'Polygon':<10} {'Yahoo':<10} {'Which Match?':<20}")
+            print(f"   {'-'*70}")
+
+            ticker = yf.Ticker(symbol)
+            for row in comparison_data[:10]:
+                try:
+                    from datetime import timedelta as td
+                    date_obj = row['date']
+                    date_str = date_obj.strftime('%Y-%m-%d')
+                    next_date = (date_obj + td(days=1)).strftime('%Y-%m-%d')
+                    yahoo_hist = ticker.history(start=date_str, end=next_date, interval='1d')
+                    yahoo_close = yahoo_hist.iloc[0]['Close'] if len(yahoo_hist) > 0 else None
+
+                    t_close = row['theta_close']
+                    p_close = row['poly_close']
+                    y_close = yahoo_close
+
+                    # Check which ones match
+                    tp_match = abs(t_close - p_close) < 0.01
+                    ty_match = abs(t_close - y_close) < 0.01 if y_close else False
+                    py_match = abs(p_close - y_close) < 0.01 if y_close else False
+
+                    if tp_match and ty_match and py_match:
+                        match_str = "✅ All match"
+                    elif py_match:
+                        match_str = "❌ Polygon+Yahoo (Theta wrong)"
+                    elif ty_match:
+                        match_str = "❌ Theta+Yahoo (Polygon wrong)"
+                    elif tp_match:
+                        match_str = "❌ Theta+Polygon (Yahoo wrong)"
+                    else:
+                        match_str = "❌ None match!"
+
+                    print(f"   {date_str:<12} ${t_close:<9.2f} ${p_close:<9.2f} ${y_close:<9.2f} {match_str}")
+                except:
+                    print(f"   {date_str:<12} ${row['theta_close']:<9.2f} ${row['poly_close']:<9.2f} {'N/A':<9} Yahoo error")
+
+            pytest.fail(f"CRITICAL: Price tolerance exceeded: {', '.join(failures)}")
+
+        print(f"   ✓ All prices within ${tolerance:.2f} tolerance")
+        print(f"      Max differences: open=${max_diff['open']:.4f}, high=${max_diff['high']:.4f}, "
+              f"low=${max_diff['low']:.4f}, close=${max_diff['close']:.4f}")
+
+        # ==== CHECK 5: Exact Timestamp Alignment ====
+        print(f"\n6. Verifying EXACT timestamp alignment (no shifts allowed)...")
+        timestamp_mismatches = []
+        for i, (theta_ts, polygon_ts) in enumerate(zip(theta_df.index, polygon_df.index)):
+            if theta_ts.date() != polygon_ts.date():
+                timestamp_mismatches.append((i, theta_ts, polygon_ts))
+
+        if timestamp_mismatches:
+            print(f"\n   ✗ TIMESTAMP MISMATCH DETECTED:")
+            for idx, theta_ts, polygon_ts in timestamp_mismatches[:10]:
+                print(f"      Index {idx}: Theta={theta_ts.date()}, Polygon={polygon_ts.date()}")
+            pytest.fail(f"CRITICAL: {len(timestamp_mismatches)} timestamp mismatches!")
+
+        print(f"   ✓ ALL timestamps match EXACTLY (0 day shifts)")
+
+        # ==== CHECK 6: No Duplicates ====
+        print(f"\n7. Verifying no duplicate dates...")
+        theta_duplicates = theta_df.index[theta_df.index.duplicated()].tolist()
+        polygon_duplicates = polygon_df.index[polygon_df.index.duplicated()].tolist()
+
+        if theta_duplicates:
+            pytest.fail(f"CRITICAL: ThetaData has duplicate dates: {theta_duplicates}")
+        if polygon_duplicates:
+            pytest.fail(f"CRITICAL: Polygon has duplicate dates: {polygon_duplicates}")
+
+        print(f"   ✓ No duplicate dates in either source")
+
+        # ==== CHECK 6: Volume Sanity ====
+        print(f"\n8. Verifying volume data...")
+        if 'volume' in theta_df.columns and 'volume' in polygon_df.columns:
+            theta_zero_vol = (theta_df['volume'] == 0).sum()
+            polygon_zero_vol = (polygon_df['volume'] == 0).sum()
+
+            if theta_zero_vol > len(theta_df) * 0.1:  # More than 10% zero volume
+                print(f"   ⚠ WARNING: ThetaData has {theta_zero_vol}/{len(theta_df)} days with zero volume")
+            if polygon_zero_vol > len(polygon_df) * 0.1:
+                print(f"   ⚠ WARNING: Polygon has {polygon_zero_vol}/{len(polygon_df)} days with zero volume")
+
+            print(f"   ✓ Volume data present (Theta: {theta_zero_vol} zero days, Polygon: {polygon_zero_vol} zero days)")
+
+        # ==== FINAL SUMMARY ====
+        print(f"\n{'='*80}")
+        print(f"✓✓✓ {symbol} DAILY DATA VALIDATION PASSED ✓✓✓")
+        print(f"    Trading days: {len(theta_df)}")
+        print(f"    Timestamps: PERFECT MATCH (0 shifts)")
+        print(f"    Prices: ALL within $0.01")
+        print(f"    Period: {theta_df.index[0].date()} to {theta_df.index[-1].date()}")
+        print(f"{'='*80}\n")
+
+
+@pytest.mark.apitest
+class TestIntradayDataComparison:
+    """
+    Comprehensive intraday interval comparison (5min, 10min, 15min, 30min, hour).
+    Tests ThetaData server-side intervals vs Polygon client-side aggregation.
+    ZERO TOLERANCE: Exact bar counts, exact timestamps, half-penny price accuracy.
+    """
+
+    @pytest.mark.parametrize("interval,resample_rule,expected_bars", [
+        ("5minute", "5min", 78),
+        ("10minute", "10min", 39),
+        ("15minute", "15min", 26),
+        ("30minute", "30min", 13),
+        ("hour", "1h", 7),
+    ])
+    def test_theta_vs_polygon_intervals(self, interval, resample_rule, expected_bars):
+        """Test ThetaData intervals match Polygon aggregated data EXACTLY."""
+        import pytz
+        from lumibot import LUMIBOT_DEFAULT_PYTZ
+
+        username = os.environ.get("THETADATA_USERNAME")
+        password = os.environ.get("THETADATA_PASSWORD")
+        polygon_api_key = os.environ.get("POLYGON_API_KEY")
+
+        asset = Asset("SPY", asset_type="stock")
+
+        # Use timezone-aware datetimes (ET) to properly filter RTH
+        et_tz = pytz.timezone("America/New_York")
+        start_et = et_tz.localize(datetime.datetime(2025, 9, 15, 9, 30))  # 9:30 AM ET
+        end_et = et_tz.localize(datetime.datetime(2025, 9, 15, 16, 0))    # 4:00 PM ET
+        start = start_et.astimezone(pytz.UTC)
+        end = end_et.astimezone(pytz.UTC)
+
+        print(f"\n{'='*80}")
+        print(f"TESTING {asset.symbol} {interval.upper()} INTERVAL: {start_et.date()}")
+        print(f"{'='*80}")
+
+        # ==== GET THETADATA SERVER-SIDE AGGREGATED DATA ====
+        print(f"\n1. Fetching ThetaData {interval} data...")
+        try:
+            theta_df = thetadata_helper.get_price_data(
+                username=username,
+                password=password,
+                asset=asset,
+                start=start,
+                end=end,
+                timespan=interval,
+                include_after_hours=False  # RTH only for fair comparison with Polygon
+            )
+        except Exception as e:
+            pytest.fail(f"CRITICAL: ThetaData {interval} FAILED: {e}")
+
+        if theta_df is None or len(theta_df) == 0:
+            pytest.fail(f"CRITICAL: ThetaData returned NO {interval} data")
+
+        print(f"   ✓ ThetaData: {len(theta_df)} {interval} bars")
+        print(f"   First bar: {theta_df.index[0]}")
+        print(f"   Last bar:  {theta_df.index[-1]}")
+
+        # ==== GET POLYGON MINUTE DATA AND AGGREGATE CLIENT-SIDE ====
+        print(f"\n2. Fetching Polygon minute data and aggregating to {interval}...")
+        try:
+            polygon_minute_df = polygon_get_price_data(
+                api_key=polygon_api_key,
+                asset=asset,
+                start=start,
+                end=end,
+                timespan="minute",
+                quote_asset=Asset("USD", asset_type="forex")
+            )
+        except Exception as e:
+            pytest.fail(f"CRITICAL: Polygon minute data FAILED: {e}")
+
+        if polygon_minute_df is None or len(polygon_minute_df) == 0:
+            pytest.fail(f"CRITICAL: Polygon returned NO minute data")
+
+        # Filter to RTH only (9:30 AM - 4:00 PM ET) before aggregating
+        # Polygon may return extended hours data - we need to filter it manually
+        polygon_minute_rth = polygon_minute_df[(polygon_minute_df.index >= start) & (polygon_minute_df.index <= end)]
+
+        if polygon_minute_rth is None or len(polygon_minute_rth) == 0:
+            pytest.fail(f"CRITICAL: Polygon returned NO RTH minute data")
+
+        # Aggregate Polygon minute data
+        # For hourly, offset to align with market open (9:30 AM = 13:30 UTC)
+        if interval == "hour":
+            polygon_agg_df = polygon_minute_rth.resample(resample_rule, offset='30min').agg({
+                'open': 'first',
+                'high': 'max',
+                'low': 'min',
+                'close': 'last',
+                'volume': 'sum'
+            }).dropna()
+        else:
+            polygon_agg_df = polygon_minute_rth.resample(resample_rule).agg({
+                'open': 'first',
+                'high': 'max',
+                'low': 'min',
+                'close': 'last',
+                'volume': 'sum'
+            }).dropna()
+
+        print(f"   ✓ Polygon: {len(polygon_agg_df)} {interval} bars (aggregated from {len(polygon_minute_rth)} RTH minute bars)")
+        print(f"   First bar: {polygon_agg_df.index[0]}")
+        print(f"   Last bar:  {polygon_agg_df.index[-1]}")
+
+        # ==== CHECK 1: Bar Count - Allow ±1 for 16:00 bar edge case ====
+        print(f"\n3. Verifying bar count match...")
+
+        # ThetaData RTH ends at 15:55 for intraday (no 16:00 bar), Polygon may include 16:00
+        # This is acceptable behavior - both are correct interpretations of "4 PM close"
+        bar_diff = abs(len(theta_df) - len(polygon_agg_df))
+
+        if bar_diff > 1:
+            print(f"\n   ✗ CRITICAL: Bar count MISMATCH!")
+            print(f"      ThetaData: {len(theta_df)} bars")
+            print(f"      Polygon:   {len(polygon_agg_df)} bars")
+            print(f"      Difference: {bar_diff} bars")
+            pytest.fail(f"CRITICAL: Bar count diff {bar_diff} > 1. Theta={len(theta_df)}, Polygon={len(polygon_agg_df)}")
+
+        if bar_diff == 1:
+            print(f"   ⚠ Bar count off by 1 (acceptable for 16:00 bar edge case)")
+            print(f"      ThetaData: {len(theta_df)} bars (ends {theta_df.index[-1]})")
+            print(f"      Polygon:   {len(polygon_agg_df)} bars (ends {polygon_agg_df.index[-1]})")
+            # Use shorter dataset for comparison
+            min_len = min(len(theta_df), len(polygon_agg_df))
+            theta_df = theta_df.iloc[:min_len]
+            polygon_agg_df = polygon_agg_df.iloc[:min_len]
+        else:
+            print(f"   ✓ EXACT match: {len(theta_df)} bars")
+
+        # ==== CHECK 2: EXACT Timestamp Match ====
+        print(f"\n4. Verifying EXACT timestamp alignment...")
+        timestamp_mismatches = []
+        for i, (theta_ts, polygon_ts) in enumerate(zip(theta_df.index, polygon_agg_df.index)):
+            if theta_ts != polygon_ts:
+                timestamp_mismatches.append((i, theta_ts, polygon_ts))
+
+        if timestamp_mismatches:
+            print(f"\n   ✗ TIMESTAMP MISMATCH DETECTED!")
+            print(f"\n   {'Index':<8} {'ThetaData':<25} {'Polygon':<25} {'Shift (seconds)'}")
+            print(f"   {'-'*75}")
+            for idx, theta_ts, polygon_ts in timestamp_mismatches[:10]:
+                shift = (theta_ts - polygon_ts).total_seconds()
+                print(f"   {idx:<8} {theta_ts} {polygon_ts} {shift:+.0f}s")
+            pytest.fail(f"CRITICAL: {len(timestamp_mismatches)} timestamp mismatches!")
+
+        print(f"   ✓ ALL timestamps match EXACTLY (0 shifts)")
+
+        # ==== CHECK 3: Price Accuracy (half-penny tolerance) ====
+        print(f"\n5. Verifying OHLC prices (half-penny tolerance: $0.005)...")
+
+        max_diff = {'open': 0.0, 'high': 0.0, 'low': 0.0, 'close': 0.0}
+        price_failures = []
+
+        for theta_ts, polygon_ts in zip(theta_df.index, polygon_agg_df.index):
+            theta_row = theta_df.loc[theta_ts]
+            polygon_row = polygon_agg_df.loc[polygon_ts]
+
+            for field in ['open', 'high', 'low', 'close']:
+                diff = abs(theta_row[field] - polygon_row[field])
+                max_diff[field] = max(max_diff[field], diff)
+
+                if diff > 0.005:  # Half-penny tolerance
+                    price_failures.append({
+                        'timestamp': theta_ts,
+                        'field': field,
+                        'theta': theta_row[field],
+                        'polygon': polygon_row[field],
+                        'diff': diff
+                    })
+
+        if price_failures:
+            print(f"\n   ✗ PRICE TOLERANCE EXCEEDED ({len(price_failures)} failures):")
+            for failure in price_failures[:10]:
+                print(f"      {failure['timestamp']} {failure['field']}: Theta=${failure['theta']:.4f}, "
+                      f"Polygon=${failure['polygon']:.4f}, Diff=${failure['diff']:.4f}")
+            pytest.fail(f"CRITICAL: {len(price_failures)} price differences exceed $0.005")
+
+        print(f"   ✓ All prices within $0.005 tolerance")
+        print(f"      Max differences: open=${max_diff['open']:.4f}, high=${max_diff['high']:.4f}, "
+              f"low=${max_diff['low']:.4f}, close=${max_diff['close']:.4f}")
+
+        # ==== FINAL SUMMARY ====
+        print(f"\n{'='*80}")
+        print(f"✓✓✓ {asset.symbol} {interval.upper()} VALIDATION PASSED ✓✓✓")
+        print(f"    Bars: {len(theta_df)} (EXACT match)")
+        print(f"    Timestamps: PERFECT MATCH (0 shifts)")
+        print(f"    Prices: ALL within $0.005 (half-penny)")
+        print(f"    Period: {theta_df.index[0]} to {theta_df.index[-1]}")
+        print(f"{'='*80}\n")
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v", "-s"])