Upload folder using huggingface_hub

app.py

@@ -205,13 +205,15 @@ def load_governance():
            .replace("impl_alert", "implementation-drift alert").replace("SPY/MTUM", "the S&P 500 and the momentum factor"))
     if len(why) > 420:
         why = why[:420].rsplit(" ", 1)[0] + "…"
+    CONMAP = {"abs10_sec5": "10% cap · ≤5/sector", "rel12b2_sec5": "index-aware caps (≤12%, index wt +2%) · ≤5/sector",
+              "abs10_sec6": "10% cap · ≤6/sector"}
     sig = " + ".join(SIGMAP.get(x, x) for x in (inc.get("signal") or []))
     return (f"### 🤝 Strategy governance — the crew's latest decision\n"
             f"**{act.strip()}** — {why}\n"
             f"- **Tracking:** 8-week gap vs S&P {tr.get('gap_8w_pp', '—')}pp — the {tr.get('pctile_8w', '—')}th percentile "
             f"of the strategy's 20-year history · implementation drift {tr.get('impl_drift_pp', '—')}pp vs its own simulation\n"
             f"- **Live strategy (v{inc.get('version', '?')}):** {sig} selection · {inc.get('regime', '?')} trend rule · "
-            f"defense = {DEFMAP.get(inc.get('defense'), inc.get('defense'))} · rebalanced every {inc.get('cadence_weeks', 1)} week(s)\n"
+            f"defense = {DEFMAP.get(inc.get('defense'), inc.get('defense'))} · {CONMAP.get(inc.get('construction'), '10% cap · ≤5/sector')} · rebalanced every {inc.get('cadence_weeks', 1)} week(s)\n"
             f"- **Change policy:** decided weekly by an agentic crew — Analyst → adversarial Critic → PM (Claude Opus 5) — "
             f"gated by a code validator. A variant is adopted only if it beats the incumbent on the full 20-year "
             f"walk-forward, net of costs — never a re-fit to recent weeks; at most one change per 90 days. Any AI failure "

rebalance_engine.py

@@ -43,7 +43,29 @@ def cap_weights(w, mw=MAX_W):
 
 
 DEFAULT_CFG = {"signal": ["n_mom_12_1", "n_roe"], "regime": "cross10/40",
-               "defense": "mom60_agg20_gld20", "cadence_weeks": 1, "version": 1}
+               "defense": "mom60_agg20_gld20", "cadence_weeks": 1, "construction": "abs10_sec5", "version": 1}
+
+
+CONSTRUCTIONS = {
+    "abs10_sec5": {"rule": "abs", "hard": 0.10, "buffer": 0.0,  "sector_cap": 5},
+    "rel12b2_sec5": {"rule": "rel", "hard": 0.12, "buffer": 0.02, "sector_cap": 5},
+    "abs10_sec6": {"rule": "abs", "hard": 0.10, "buffer": 0.0,  "sector_cap": 6},
+}
+
+
+def cap_vec(w, caps):
+    w = np.asarray(w, float); caps = np.asarray(caps, float)
+    for _ in range(50):
+        over = w > caps + 1e-9
+        if not over.any():
+            break
+        ex = (w[over] - caps[over]).sum(); w[over] = caps[over]
+        under = ~over
+        room = w[under].sum()
+        if room <= 0:
+            break
+        w[under] += ex * w[under] / room
+    return w
 
 
 def load_config(token):
@@ -83,21 +105,27 @@ def detect_regime(rule="cross10/40"):
                 "warn": f"SPY fetch failed ({type(e).__name__}); defaulting to OFFENSE"}
 
 
-def _rank_sleeve(df, cols, sign=1):
+def _rank_sleeve(df, cols, sign=1, con=None):
+    con = con or CONSTRUCTIONS["abs10_sec5"]
     df = df[df["marketcap"].notna() & (df["marketcap"] > 0)].copy()
     df = df.sort_values("marketcap", ascending=False).head(SEG_K)
+    top500 = float(df["marketcap"].head(500).sum()) or 1.0
     df["score"] = sum(sign * df[c].fillna(0) for c in cols)
     df = df.sort_values("score", ascending=False)
     picked, sec = [], {}
     for _, r in df.iterrows():
-        if sec.get(r["sector"], 0) >= SECTOR_CAP:
+        if sec.get(r["sector"], 0) >= con["sector_cap"]:
             continue
         sec[r["sector"]] = sec.get(r["sector"], 0) + 1
         picked.append(r)
         if len(picked) >= SIZE:
             break
     mc = np.array([r["marketcap"] for r in picked], float)
-    w = cap_weights(mc / mc.sum())
+    if con["rule"] == "rel":
+        caps = np.maximum(np.minimum(con["hard"], mc / top500 + con["buffer"]), 0.02)
+    else:
+        caps = np.full(len(picked), con["hard"])
+    w = cap_vec(mc / mc.sum(), caps)
     return [{"ticker": r["ticker"], "name": r.get("name"), "sector": r["sector"], "weight": float(wi)}
             for r, wi in zip(picked, w)]
 
@@ -112,7 +140,8 @@ def get_target(token):
     cfg = load_config(token)
     reg = detect_regime(cfg["regime"])
     snap = pd.read_parquet(hf_hub_download(PANEL_REPO, "marketfm_snapshot.parquet", repo_type="dataset", token=token))
-    sleeve = _rank_sleeve(snap, cfg["signal"])
+    con = CONSTRUCTIONS.get(cfg.get("construction", "abs10_sec5"), CONSTRUCTIONS["abs10_sec5"])
+    sleeve = _rank_sleeve(snap, cfg["signal"], con=con)
     if reg["offense"]:
         rows = [dict(x) for x in sleeve]
     else:
@@ -121,7 +150,7 @@ def get_target(token):
             rows = []
             eq_frac, agg_w, gld_w = 0.0, 0.5, 0.5
         elif d == "lowvol":
-            rows = _rank_sleeve(snap, ["n_vol_13w"], sign=-1)
+            rows = _rank_sleeve(snap, ["n_vol_13w"], sign=-1, con=con)
             eq_frac, agg_w, gld_w = 1.0, 0.0, 0.0
         elif d == "mom40_agg30_gld30":
             rows = [dict(x, weight=0.4 * x["weight"]) for x in sleeve]; eq_frac, agg_w, gld_w = 0.4, 0.3, 0.3