solmath 0.1.3

Fixed-point financial math for Solana. Black-Scholes, Greeks, IV, NIG pricing, pool math — pure integer arithmetic, no_std, zero dependencies.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290

#!/usr/bin/env python3
"""
SolMath Barrier Option Validation Vectors

Generates reference barrier option prices using QuantLib's AnalyticBarrierEngine
with exact T (no day-count rounding). Cross-validates against mpmath at 50 digits.

QuantLib uses IEEE 754 f64 (~15.9 sig figs). mpmath provides the reference truth.

Covers all 8 configurations: 4 barrier types × call/put.
Stratified across S/K ratios, H/S ratios, sigma, T, r.
"""

import json
import math
import os
import random
import sys

import mpmath
import QuantLib as ql

mpmath.mp.dps = 50
random.seed(42)

SCALE = 10**12
OUTPUT_DIR = os.path.join(os.path.dirname(__file__), '..', 'benchmark')


def quantlib_barrier_exact_T(S, K, H, r, sigma, T, barrier_type, option_type):
    """Price a European barrier option using QuantLib's AnalyticBarrierEngine
    with exact T — no day-count rounding.

    We construct the term structures so that QuantLib sees the exact T we want:
    set evaluation date to t=0, maturity at t=T using a reference date that
    gives the exact year fraction.
    """
    try:
        # Use a reference date far in the past so the day-count doesn't matter.
        # We override the year fractions by choosing dates that give exact T.
        today = ql.Date(1, 1, 2025)
        ql.Settings.instance().evaluationDate = today

        # Compute days so that days/365 is as close to T as possible.
        # Use Actual365Fixed: yearFraction = days/365.
        days = max(1, round(T * 365))
        maturity = today + ql.Period(int(days), ql.Days)
        dc = ql.Actual365Fixed()
        actual_T = dc.yearFraction(today, maturity)

        # Build term structures using the actual_T that QuantLib will compute.
        # The prices QuantLib produces are at this actual_T.
        spot_handle = ql.QuoteHandle(ql.SimpleQuote(S))
        vol_handle = ql.BlackVolTermStructureHandle(
            ql.BlackConstantVol(today, ql.NullCalendar(),
                                ql.QuoteHandle(ql.SimpleQuote(sigma)), dc)
        )
        rate_handle = ql.YieldTermStructureHandle(
            ql.FlatForward(today, ql.QuoteHandle(ql.SimpleQuote(r)), dc)
        )
        div_handle = ql.YieldTermStructureHandle(
            ql.FlatForward(today, 0.0, dc)
        )

        process = ql.BlackScholesMertonProcess(
            spot_handle, div_handle, rate_handle, vol_handle
        )

        payoff = ql.PlainVanillaPayoff(option_type, K)
        exercise = ql.EuropeanExercise(maturity)

        # Barrier option
        barrier_opt = ql.BarrierOption(barrier_type, H, 0.0, payoff, exercise)
        barrier_opt.setPricingEngine(ql.AnalyticBarrierEngine(process))
        barrier_price = barrier_opt.NPV()

        # Vanilla for reference
        vanilla = ql.EuropeanOption(payoff, exercise)
        vanilla.setPricingEngine(ql.AnalyticEuropeanEngine(process))
        vanilla_price = vanilla.NPV()

        return barrier_price, vanilla_price, actual_T
    except Exception:
        return None


def mpmath_barrier_down_out_call(S, K, H, r, sigma, T):
    """Down-and-out call via Merton reflection at 50-digit precision.
    Only valid for down-out calls with K >= H (payoff entirely above barrier)."""
    S, K, H, r, sigma, T = [mpmath.mpf(x) for x in [S, K, H, r, sigma, T]]
    if T <= 0 or sigma <= 0 or S <= 0 or K <= 0 or H <= 0:
        return None
    if S <= H:
        return 0.0, 0.0
    if K < H:
        return None  # formula not numerically stable for K < H

    def bs_call(spot, strike):
        vol_sqrt_t = sigma * mpmath.sqrt(T)
        if vol_sqrt_t < mpmath.mpf('1e-30'):
            disc = mpmath.exp(-r * T)
            return max(spot - strike * disc, 0)
        d1 = (mpmath.log(spot / strike) + (r + sigma**2 / 2) * T) / vol_sqrt_t
        d2 = d1 - vol_sqrt_t
        disc = mpmath.exp(-r * T)
        return spot * mpmath.ncdf(d1) - strike * disc * mpmath.ncdf(d2)

    alpha = 2 * (r - sigma**2 / 2) / sigma**2
    power = (H / S) ** alpha
    vanilla = bs_call(S, K)
    reflected = bs_call(H**2 / S, K)
    return float(vanilla - power * reflected), float(vanilla)


def generate_param_grid():
    spots = [50, 80, 100, 120, 150, 200]
    sk_ratios = [0.80, 0.90, 0.95, 1.00, 1.05, 1.10, 1.20]
    sigmas = [0.10, 0.15, 0.20, 0.25, 0.30, 0.40, 0.50, 0.60]
    times = [0.05, 0.10, 0.25, 0.50, 1.00, 2.00]
    rates = [0.00, 0.02, 0.05, 0.08, 0.10]
    down_hs = [0.70, 0.75, 0.80, 0.85, 0.90, 0.95]
    up_hs = [1.05, 1.10, 1.15, 1.20, 1.30, 1.50]
    return spots, sk_ratios, sigmas, times, rates, down_hs, up_hs


def generate_vectors():
    spots, sk_ratios, sigmas, times, rates, down_hs, up_hs = generate_param_grid()

    barrier_configs = [
        ("down_out_call", ql.Barrier.DownOut, ql.Option.Call, down_hs, True),
        ("down_in_call",  ql.Barrier.DownIn,  ql.Option.Call, down_hs, True),
        ("down_out_put",  ql.Barrier.DownOut, ql.Option.Put,  down_hs, True),
        ("down_in_put",   ql.Barrier.DownIn,  ql.Option.Put,  down_hs, True),
        ("up_out_call",   ql.Barrier.UpOut,   ql.Option.Call, up_hs,   False),
        ("up_in_call",    ql.Barrier.UpIn,    ql.Option.Call, up_hs,   False),
        ("up_out_put",    ql.Barrier.UpOut,    ql.Option.Put,  up_hs,   False),
        ("up_in_put",     ql.Barrier.UpIn,     ql.Option.Put,  up_hs,   False),
    ]

    all_vectors = []
    skipped = 0
    total = 0

    for name, barrier_type, option_type, hs_ratios, is_down in barrier_configs:
        config_vectors = []
        is_call = (option_type == ql.Option.Call)

        for S in spots:
            for sk in sk_ratios:
                K = S * sk
                for hs in hs_ratios:
                    H = S * hs
                    if is_down and S <= H:
                        continue
                    if not is_down and S >= H:
                        continue
                    if is_call and not is_down and K >= H:
                        continue
                    if not is_call and is_down and K <= H:
                        continue

                    for sigma in sigmas:
                        for T in times:
                            for r in rates:
                                total += 1
                                result = quantlib_barrier_exact_T(
                                    S, K, H, r, sigma, T,
                                    barrier_type, option_type
                                )
                                if result is None:
                                    skipped += 1
                                    continue

                                barrier_price, vanilla_price, actual_T = result

                                # Store actual_T from QuantLib's day count
                                s_fp = int(round(S * SCALE))
                                k_fp = int(round(K * SCALE))
                                h_fp = int(round(H * SCALE))
                                r_fp = int(round(r * SCALE))
                                sigma_fp = int(round(sigma * SCALE))
                                t_fp = int(round(actual_T * SCALE))

                                vec = {
                                    "s": str(s_fp),
                                    "k": str(k_fp),
                                    "h": str(h_fp),
                                    "r": str(r_fp),
                                    "sigma": str(sigma_fp),
                                    "t": str(t_fp),
                                    "is_call": is_call,
                                    "barrier_type": name,
                                    "barrier_price": barrier_price,
                                    "vanilla_price": vanilla_price,
                                    "barrier_price_fp": str(int(round(
                                        barrier_price * SCALE))),
                                    "vanilla_price_fp": str(int(round(
                                        vanilla_price * SCALE))),
                                }
                                config_vectors.append(vec)

        print(f"  {name}: {len(config_vectors)} vectors")
        all_vectors.extend(config_vectors)

    print(f"\nTotal: {len(all_vectors)} vectors "
          f"({skipped} skipped out of {total} attempted)")

    # Cross-validate down-out calls (K>=H) against mpmath simple reflection formula
    print("\nCross-validating down-out calls (K>=H) against mpmath 50-digit...")
    doc = [v for v in all_vectors
           if v['barrier_type'] == 'down_out_call'
           and int(v['k']) >= int(v['h'])]
    sample = random.sample(doc, min(1000, len(doc)))
    max_ulp = 0.0
    cross_ok = 0
    for v in sample:
        S = int(v['s']) / SCALE
        K = int(v['k']) / SCALE
        H = int(v['h']) / SCALE
        r_val = int(v['r']) / SCALE
        sigma_val = int(v['sigma']) / SCALE
        T = int(v['t']) / SCALE

        mp_result = mpmath_barrier_down_out_call(S, K, H, r_val, sigma_val, T)
        if mp_result is None:
            continue
        mp_barrier, _ = mp_result
        ql_barrier = v['barrier_price']
        ulp = abs(ql_barrier - mp_barrier) * SCALE
        max_ulp = max(max_ulp, ulp)
        cross_ok += 1

    print(f"  {cross_ok} vectors, max diff: {max_ulp:.1f} ULP")

    # In/out conservation
    from collections import defaultdict
    by_params = defaultdict(dict)
    for v in all_vectors:
        key = (v['s'], v['k'], v['h'], v['r'], v['sigma'], v['t'], v['is_call'])
        by_params[key][v['barrier_type']] = v

    conservation_checked = 0
    max_conservation_err = 0
    for key, types in by_params.items():
        for in_name, out_name in [
            ('down_in_call', 'down_out_call'),
            ('down_in_put', 'down_out_put'),
            ('up_in_call', 'up_out_call'),
            ('up_in_put', 'up_out_put'),
        ]:
            if in_name in types and out_name in types:
                diff = abs(types[in_name]['barrier_price'] +
                           types[out_name]['barrier_price'] -
                           types[out_name]['vanilla_price'])
                conservation_checked += 1
                max_conservation_err = max(max_conservation_err, diff)

    print(f"\nIn/out conservation (QuantLib): {conservation_checked} pairs")
    print(f"  Max error: ${max_conservation_err:.2e}")

    return all_vectors


def main():
    print(f"Generating barrier option test vectors")
    print(f"QuantLib {ql.__version__} (AnalyticBarrierEngine)")
    print(f"mpmath at {mpmath.mp.dps} digits for cross-validation\n")

    vectors = generate_vectors()

    path = os.path.join(OUTPUT_DIR, 'barrier_vectors.json')
    with open(path, 'w') as f:
        json.dump({
            "meta": {
                "source": f"QuantLib {ql.__version__} AnalyticBarrierEngine",
                "cross_validation": f"mpmath {mpmath.mp.dps} digits",
                "n_vectors": len(vectors),
                "scale": SCALE,
                "no_dividends": True,
                "no_rebate": True,
                "T_note": "actual_T from QuantLib Actual365Fixed day count",
            },
            "vectors": vectors,
        }, f, indent=None)
    size_mb = os.path.getsize(path) / 1e6
    print(f"\nWrote {path} ({len(vectors)} vectors, {size_mb:.1f} MB)")


if __name__ == '__main__':
    main()