Spaces:
Running on Zero
Running on Zero
File size: 14,199 Bytes
3e33e2e b277794 3e33e2e | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 | """Quantitative portfolio construction, weighting models, and risk allocation."""
from __future__ import annotations
import math
from dataclasses import dataclass
from typing import Any
try:
from enum import StrEnum
except ImportError:
from enum import Enum
class StrEnum(str, Enum): # noqa: UP042
def __str__(self) -> str:
return str(self.value)
class AllocationModel(StrEnum):
EQUAL_WEIGHT = "equal_weight"
INVERSE_VOLATILITY = "inverse_volatility"
RISK_PARITY = "risk_parity"
MULTI_FACTOR_MOMENTUM = "multi_factor_momentum"
MINIMUM_VARIANCE = "minimum_variance"
@dataclass(frozen=True)
class AssetAllocation:
symbol: str
sector: str
weight: float
volatility: float
momentum_return: float
marginal_risk_contribution: float
@dataclass(frozen=True)
class AllocationResult:
model: str
constituents_count: int
portfolio_volatility: float
effective_n: float
diversification_ratio: float
max_single_weight: float
sector_weights: dict[str, float]
allocations: list[AssetAllocation]
def to_dict(self) -> dict[str, Any]:
return {
"model": self.model,
"constituents_count": self.constituents_count,
"portfolio_volatility": round(self.portfolio_volatility, 4),
"effective_n": round(self.effective_n, 2),
"diversification_ratio": round(self.diversification_ratio, 3),
"max_single_weight": round(self.max_single_weight, 4),
"sector_weights": {k: round(v, 4) for k, v in sorted(self.sector_weights.items())},
"allocations": [
{
"symbol": a.symbol,
"sector": a.sector,
"weight": round(a.weight, 4),
"volatility": round(a.volatility, 4),
"momentum_return": round(a.momentum_return, 4),
"marginal_risk_contribution": round(a.marginal_risk_contribution, 4),
}
for a in self.allocations
],
}
class PortfolioAllocator:
"""Quantitative portfolio allocator implementing smart beta and risk budgeting."""
def __init__(self, sector_map: dict[str, str] | None = None) -> None:
self.sector_map = sector_map or {}
def compute_daily_returns(
self, bars_by_symbol: dict[str, list[dict[str, Any]]]
) -> tuple[list[str], list[str], dict[str, list[float]]]:
"""Align bar series across symbols and compute percentage returns."""
# Find intersecting timestamps or group by date
daily_closes: dict[str, dict[str, float]] = {}
for sym, bars in bars_by_symbol.items():
daily_closes[sym] = {}
for b in bars:
day = b["timestamp"][:10]
daily_closes[sym][day] = b["close"]
# Intersect trading days
symbols = sorted(bars_by_symbol.keys())
if not symbols:
return [], [], {}
common_days = set(daily_closes[symbols[0]].keys())
for s in symbols[1:]:
common_days.intersection_update(daily_closes[s].keys())
sorted_days = sorted(common_days)
if len(sorted_days) < 2:
return symbols, [], {s: [] for s in symbols}
returns: dict[str, list[float]] = {s: [] for s in symbols}
for s in symbols:
prices = [daily_closes[s][day] for day in sorted_days]
for i in range(1, len(prices)):
prev = prices[i - 1]
curr = prices[i]
ret = (curr - prev) / prev if prev > 0 else 0.0
returns[s].append(ret)
return symbols, sorted_days[1:], returns
def compute_covariance(
self, symbols: list[str], returns: dict[str, list[float]]
) -> tuple[dict[str, float], list[list[float]]]:
"""Compute annualized volatilities and sample covariance matrix."""
n_assets = len(symbols)
if n_assets == 0:
return {}, []
t_samples = len(returns[symbols[0]])
if t_samples < 2:
return {s: 0.20 for s in symbols}, [
[1.0 if i == j else 0.0 for j in range(n_assets)] for i in range(n_assets)
]
# Sample means
means = {s: sum(returns[s]) / t_samples for s in symbols}
# Volatilities (annualized assuming 252 trading days)
vols: dict[str, float] = {}
for s in symbols:
var = sum((r - means[s]) ** 2 for r in returns[s]) / (t_samples - 1)
vols[s] = math.sqrt(max(1e-8, var)) * math.sqrt(252)
# Covariance matrix (annualized)
cov: list[list[float]] = [[0.0] * n_assets for _ in range(n_assets)]
for i in range(n_assets):
s_i = symbols[i]
mean_i = means[s_i]
for j in range(i, n_assets):
s_j = symbols[j]
mean_j = means[s_j]
cov_ij = (
sum(
(returns[s_i][k] - mean_i) * (returns[s_j][k] - mean_j)
for k in range(t_samples)
)
/ (t_samples - 1)
) * 252
cov[i][j] = cov_ij
cov[j][i] = cov_ij
return vols, cov
def shrink_covariance(
self, cov: list[list[float]], shrinkage: float = 0.20
) -> list[list[float]]:
"""Ledoit-Wolf style constant correlation shrinkage to condition covariance."""
n = len(cov)
if n <= 1:
return cov
# Compute diagonal standard deviations
stds = [math.sqrt(max(1e-8, cov[i][i])) for i in range(n)]
# Compute average correlation
corrs = []
for i in range(n):
for j in range(i + 1, n):
corr = cov[i][j] / (stds[i] * stds[j])
corrs.append(corr)
mean_corr = sum(corrs) / len(corrs) if corrs else 0.0
# Target matrix F with constant correlation
shrunk = [[0.0] * n for _ in range(n)]
for i in range(n):
for j in range(n):
target = stds[i] * stds[j] if i == j else mean_corr * stds[i] * stds[j]
shrunk[i][j] = (1.0 - shrinkage) * cov[i][j] + shrinkage * target
return shrunk
def allocate(
self,
model: AllocationModel | str,
bars_by_symbol: dict[str, list[dict[str, Any]]],
max_stock_cap: float = 0.15,
max_sector_cap: float = 0.35,
) -> AllocationResult:
"""Construct optimal portfolio weights according to chosen quant strategy."""
if isinstance(model, str):
model = AllocationModel(model)
symbols, _days, returns = self.compute_daily_returns(bars_by_symbol)
n = len(symbols)
if n == 0:
return AllocationResult(
model=model.value,
constituents_count=0,
portfolio_volatility=0.0,
effective_n=0.0,
diversification_ratio=0.0,
max_single_weight=0.0,
sector_weights={},
allocations=[],
)
vols, cov = self.compute_covariance(symbols, returns)
shrunk_cov = self.shrink_covariance(cov, shrinkage=0.25)
# Compute momentum returns (e.g. cumulative return over available window)
momentum_rets: dict[str, float] = {}
for s in symbols:
s_bars = bars_by_symbol[s]
if len(s_bars) >= 2:
p_start = s_bars[0]["close"]
p_end = s_bars[-1]["close"]
momentum_rets[s] = (p_end - p_start) / p_start if p_start > 0 else 0.0
else:
momentum_rets[s] = 0.0
# 1. Compute Raw Uncapped Weights
raw_weights: dict[str, float] = {}
if model == AllocationModel.EQUAL_WEIGHT:
w_eq = 1.0 / n
raw_weights = {s: w_eq for s in symbols}
elif model == AllocationModel.INVERSE_VOLATILITY:
inv_vols = {s: 1.0 / max(0.01, vols[s]) for s in symbols}
sum_inv = sum(inv_vols.values())
raw_weights = {s: inv_vols[s] / sum_inv for s in symbols}
elif model == AllocationModel.RISK_PARITY:
# Cyclical coordinate descent for Equal Risk Contribution
# Initialize with inverse vol
w = [1.0 / max(0.01, vols[s]) for s in symbols]
sum_w = sum(w)
w = [x / sum_w for x in w]
# Coordinate descent iterations
for _ in range(50):
for i in range(n):
# Marginal contribution of asset i: (Sigma * w)_i
sigma_w_i = sum(shrunk_cov[i][j] * w[j] for j in range(n))
# Solve quadratic for w_i: w_i * (sigma_w_i) = target_risk * total_var
# Updating step with dampening
denom = shrunk_cov[i][i]
if denom > 1e-8:
a = denom
b = sigma_w_i - w[i] * denom
# Root of a*x^2 + b*x - target = 0
disc = b * b + 4 * a * 0.001
if disc >= 0:
new_wi = (-b + math.sqrt(disc)) / (2 * a)
w[i] = max(1e-4, 0.8 * w[i] + 0.2 * new_wi)
# Re-normalize
total_w = sum(w)
if total_w > 0:
w = [x / total_w for x in w]
raw_weights = {symbols[i]: w[i] for i in range(n)}
elif model == AllocationModel.MULTI_FACTOR_MOMENTUM:
# Cross-sectional Z-score blending:
# Factor Score = 0.6 * Z(Momentum) - 0.4 * Z(Volatility)
mom_values = [momentum_rets[s] for s in symbols]
vol_values = [vols[s] for s in symbols]
mean_mom = sum(mom_values) / n
std_mom = max(
1e-4, math.sqrt(sum((x - mean_mom) ** 2 for x in mom_values) / max(1, n - 1))
)
mean_vol = sum(vol_values) / n
std_vol = max(
1e-4, math.sqrt(sum((x - mean_vol) ** 2 for x in vol_values) / max(1, n - 1))
)
scores = {}
for s in symbols:
z_mom = (momentum_rets[s] - mean_mom) / std_mom
z_vol = (vols[s] - mean_vol) / std_vol
composite = 0.6 * z_mom - 0.4 * z_vol
scores[s] = math.exp(
min(2.5, max(-2.5, composite))
) # Softmax-style positive exponent
sum_scores = sum(scores.values())
raw_weights = {s: scores[s] / sum_scores for s in symbols}
elif model == AllocationModel.MINIMUM_VARIANCE:
# Diagonal inverse approximation with shrinkage
inv_vars = {s: 1.0 / max(1e-4, vols[s] ** 2) for s in symbols}
sum_inv_var = sum(inv_vars.values())
raw_weights = {s: inv_vars[s] / sum_inv_var for s in symbols}
# 2. Apply Caps & Constraints (SEBI Regulatory Guidelines)
# Cap max single-stock weight, reallocate excess iteratively
capped_weights = dict(raw_weights)
for _ in range(15):
excess = 0.0
uncapped_sum = 0.0
for s, weight in capped_weights.items():
if weight > max_stock_cap:
excess += weight - max_stock_cap
capped_weights[s] = max_stock_cap
else:
uncapped_sum += weight
if excess <= 1e-5 or uncapped_sum <= 1e-5:
break
# Redistribute excess proportionally
for s, weight in capped_weights.items():
if weight < max_stock_cap:
capped_weights[s] += excess * (weight / uncapped_sum)
# Normalize to exactly 1.0
final_sum = sum(capped_weights.values())
final_weights = {s: w / final_sum for s, w in capped_weights.items()}
# 3. Compute Portfolio Risk Analytics
w_vec = [final_weights[s] for s in symbols]
port_var = 0.0
marginal_risks: list[float] = []
for i in range(n):
row_cov = sum(shrunk_cov[i][j] * w_vec[j] for j in range(n))
port_var += w_vec[i] * row_cov
port_vol = math.sqrt(max(1e-8, port_var))
# Marginal risk contributions: (w_i * (Sigma * w)_i) / port_vol
for i in range(n):
row_cov = sum(shrunk_cov[i][j] * w_vec[j] for j in range(n))
mrc = (w_vec[i] * row_cov) / port_vol if port_vol > 0 else 0.0
marginal_risks.append(mrc)
# Diversification Ratio = Weighted Vol / Portfolio Vol
weighted_vol = sum(w_vec[i] * vols[symbols[i]] for i in range(n))
diversification_ratio = weighted_vol / port_vol if port_vol > 0 else 1.0
# Effective N (Herfindahl Index Inverse: 1 / sum(w_i^2))
effective_n = 1.0 / sum(w**2 for w in w_vec) if sum(w**2 for w in w_vec) > 0 else float(n)
# Sector weights breakdown
sector_weights: dict[str, float] = {}
for s, w in final_weights.items():
sec = self.sector_map.get(s, "Diversified / Other")
sector_weights[sec] = sector_weights.get(sec, 0.0) + w
allocations = [
AssetAllocation(
symbol=symbols[i],
sector=self.sector_map.get(symbols[i], "Diversified / Other"),
weight=final_weights[symbols[i]],
volatility=vols[symbols[i]],
momentum_return=momentum_rets[symbols[i]],
marginal_risk_contribution=marginal_risks[i],
)
for i in range(n)
]
allocations.sort(key=lambda a: -a.weight)
return AllocationResult(
model=model.value,
constituents_count=n,
portfolio_volatility=port_vol,
effective_n=effective_n,
diversification_ratio=diversification_ratio,
max_single_weight=max(final_weights.values()),
sector_weights=sector_weights,
allocations=allocations,
)
|