Adriano/PythagorasGoal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(regime_lab): vrp annualizzato per timeframe + report() safe su 0 entries

Browse Source 
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
Adriano Dal Pastro
2026-06-02 00:08:33 +00:00
parent 7169614506
commit 5f229cd66e
1 changed files with 17 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/analysis/regime_lab.py
+17 -9
View File
@@ -74,15 +74,20 @@ def _rolling_pct(x: np.ndarray, win: int) -> np.ndarray:
lambda w: (w.iloc[-1] >= w).mean(), raw=False).values lambda w: (w.iloc[-1] >= w).mean(), raw=False).values
def regime_features(df: pd.DataFrame, pct_win: int = 252, rv_win: int = 24, fund_win: int = 168) -> dict: _BARS_PER_YEAR = {"1h": 24 * 365, "4h": 6 * 365, "1d": 365}
"""Tutte causali. dvol_pct/funding_z usano solo finestra passata. vrp = dvol - rv annualizz."""
def regime_features(df: pd.DataFrame, tf: str = "1h", pct_win: int = 252, rv_win: int = 24,
fund_win: int = 168) -> dict:
"""Tutte causali. dvol_pct/funding_z usano solo finestra passata. vrp = dvol - rv annualizz.
tf serve ad annualizzare correttamente la realized vol (sqrt barre/anno per timeframe)."""
c = df["close"].values.astype(float) c = df["close"].values.astype(float)
dvol = df["dvol"].values.astype(float) dvol = df["dvol"].values.astype(float)
fund = df["funding"].values.astype(float) fund = df["funding"].values.astype(float)
ret = np.zeros_like(c); ret[1:] = np.diff(np.log(c)) ret = np.zeros_like(c); ret[1:] = np.diff(np.log(c))
# realized vol annualizzata (in punti %, scala come DVOL): std rolling * sqrt(barre/anno) # realized vol annualizzata (punti %, scala come DVOL): std rolling * sqrt(barre/anno del tf)
bars_per_year = {1: 24 * 365}.get(1, 24 * 365) # default 1h; per tf diversi e' un proxy bpy = _BARS_PER_YEAR.get(tf, 24 * 365)
rv = pd.Series(ret).rolling(rv_win).std().values * np.sqrt(24 * 365) * 100 rv = pd.Series(ret).rolling(rv_win).std().values * np.sqrt(bpy) * 100
dvol_pct = _rolling_pct(dvol, pct_win) dvol_pct = _rolling_pct(dvol, pct_win)
fmean = pd.Series(fund).rolling(fund_win).mean().values fmean = pd.Series(fund).rolling(fund_win).mean().values
fstd = pd.Series(fund).rolling(fund_win).std().values fstd = pd.Series(fund).rolling(fund_win).std().values
@@ -142,7 +147,7 @@ def _cache_path(asset: str, tf: str) -> Path:
def build_cache(asset: str, tf: str, frac_step: int = 6) -> pd.DataFrame: def build_cache(asset: str, tf: str, frac_step: int = 6) -> pd.DataFrame:
"""Precompute OHLCV + regime + frattali -> parquet condiviso (per i 100 agenti).""" """Precompute OHLCV + regime + frattali -> parquet condiviso (per i 100 agenti)."""
df = load(asset, tf) df = load(asset, tf)
R = regime_features(df) R = regime_features(df, tf=tf)
F = frac_features(df, step=frac_step) F = frac_features(df, step=frac_step)
for k in _FEATCOLS_R: for k in _FEATCOLS_R:
df[k] = R[k] df[k] = R[k]
@@ -166,9 +171,12 @@ def report(name: str, entries: list[dict], df: pd.DataFrame, asset: str = "", tf
"""Netto-fee full + OOS (ultimo 30%) + sweep fee, via engine onesto di explore_lab. """Netto-fee full + OOS (ultimo 30%) + sweep fee, via engine onesto di explore_lab.
Ritorna dict compatto: trades, full/oos (ret%, sharpe, dd, acc), robust (OK su tutte le fee).""" Ritorna dict compatto: trades, full/oos (ret%, sharpe, dd, acc), robust (OK su tutte le fee)."""
if not entries: if not entries:
return {"name": name, "trades": 0, "verdict": False, "note": "no entries"} # struttura compatibile con robust() (tutti zero) -> robust()=False pulito, niente crash
ev = evaluate(name, entries, df) # full + oos + fee sweep z = {"ret": 0.0, "sharpe": 0.0, "dd": 0.0, "trades": 0, "win": 0.0, "exposure": 0.0, "yearly": {}}
return ev print(f" {name:<24s} NO ENTRIES")
return {"full": dict(z), "oos": dict(z), "sweep": {0.0: 0.0, 0.0005: 0.0, 0.001: 0.0, 0.002: 0.0},
"sweep_oos": {0.0: 0.0, 0.0005: 0.0, 0.001: 0.0, 0.002: 0.0}, "pos_yrs": 0, "n_yrs": 0}
return evaluate(name, entries, df) # full + oos + fee sweep
if __name__ == "__main__": if __name__ == "__main__":