diff --git a/kennis_streamlit.py b/kennis_streamlit.py index 1830a58..528e18e 100644 --- a/kennis_streamlit.py +++ b/kennis_streamlit.py @@ -5,61 +5,57 @@ import pandas as pd from io import BytesIO import plotly.graph_objects as go -# ---------- Config ---------- -st.set_page_config( - page_title="Kennis - Institucional FX Markets", - layout="wide", - initial_sidebar_state="expanded", -) - -# Small styling for institutional look +# ----- Page config and minimal styling (institutional) ----- +st.set_page_config(page_title="Kennis FX Markets", layout="wide", initial_sidebar_state="expanded") st.markdown( """ """, unsafe_allow_html=True, ) -# ---------- Header ---------- +# ----- Header / Hero ----- col1, col2 = st.columns([4,1]) with col1: - st.markdown("
Kennis - Institucional FX Markets
", unsafe_allow_html=True) - st.markdown("
Bayesian probabilistic FX signal — USD/JPY
", unsafe_allow_html=True) + st.markdown("
Kennis FX Markets
", unsafe_allow_html=True) + st.markdown("
Institutional Bayesian Intelligence Engine
", unsafe_allow_html=True) + st.markdown("
Detection and strategic allocation guidance.
", unsafe_allow_html=True) with col2: - st.markdown("") # space for logo placeholder - st.markdown("
© Kennis - Institucional FX Markets
", unsafe_allow_html=True) + st.markdown("
Kennis FX — Tactical edition
", unsafe_allow_html=True) st.markdown("---") -# ---------- Sidebar: Inputs ---------- -st.sidebar.header("Inputs (última fila / live)") -uploaded = st.sidebar.file_uploader("Sube CSV histórico (opcional) — columnas oblig.: cot_long,cot_short,fed_prob,retail_pct,target_up", type=["csv"]) -st.sidebar.markdown("O usa los controles para ingresar la fila *live*:") +# ----- Sidebar: inputs & upload ----- +st.sidebar.header("Live inputs (tactical / swing)") +uploaded = st.sidebar.file_uploader("Upload historical CSV (optional). Required columns: cot_long,cot_short,fed_prob,retail_pct,target_up", type=["csv"]) +st.sidebar.markdown("Or use the live controls below:") -cot_long = st.sidebar.number_input("COT Long (No-Commercial)", value=13662, step=1) -cot_short = st.sidebar.number_input("COT Short (No-Commercial)", value=13546, step=1) -fed_prob = st.sidebar.number_input("FedWatch Prob (%) para bin actual", min_value=0.0, max_value=100.0, value=96.0, step=0.1) +cot_long = st.sidebar.number_input("COT No-Commercial — Long", value=13662, step=1) +cot_short = st.sidebar.number_input("COT No-Commercial — Short", value=13546, step=1) +fed_prob = st.sidebar.number_input("FedWatch Prob (%) for current bin", min_value=0.0, max_value=100.0, value=96.0, step=0.1) retail_pct = st.sidebar.number_input("Retail % Buy (0-100)", min_value=0.0, max_value=100.0, value=47.0, step=0.1) -# Model hyperparams st.sidebar.markdown("---") -st.sidebar.header("Model") -prior_var = st.sidebar.number_input("Prior variance (gauss.)", value=0.5, step=0.1) -s_cot = st.sidebar.number_input("S_cot (escala normalización)", value=50000, step=1000) -retrain = st.sidebar.button("Retrain / Fit model (si subes CSV)") +st.sidebar.header("Model & normalization") +s_cot = st.sidebar.number_input("S_cot (scale for COT)", value=50000, step=1000) +prior_var = st.sidebar.number_input("Prior variance (gauss)", value=0.5, step=0.1) +retrain = st.sidebar.button("Retrain model (if CSV uploaded)") -# ---------- Utilities ---------- +# ----- Utilities ----- def sigmoid(x): return 1.0 / (1.0 + np.exp(-x)) def fit_bayesian_logit(X, y, prior_var=0.5, maxiter=200, tol=1e-8): - # MAP via Newton-Raphson with Gaussian prior (0, prior_var) + # MAP (Newton-Raphson) with Gaussian prior N(0, prior_var) n, p = X.shape beta = np.zeros(p) prior_prec = np.eye(p) / prior_var @@ -68,8 +64,7 @@ def fit_bayesian_logit(X, y, prior_var=0.5, maxiter=200, tol=1e-8): eta = X.dot(beta) p_vec = sigmoid(eta) W = p_vec * (1.0 - p_vec) - # avoid singular by flooring W - W = np.maximum(W, 1e-8) + W = np.clip(W, 1e-8, None) H = X.T.dot(W[:, None] * X) + prior_prec grad = X.T.dot(y - p_vec) - prior_prec.dot(beta) try: @@ -84,23 +79,23 @@ def fit_bayesian_logit(X, y, prior_var=0.5, maxiter=200, tol=1e-8): return beta, cov def to_excel(df): - output = BytesIO() - with pd.ExcelWriter(output, engine="openpyxl") as writer: + out = BytesIO() + with pd.ExcelWriter(out, engine="openpyxl") as writer: df.to_excel(writer, index=False, sheet_name="signal") - return output.getvalue() + return out.getvalue() -# ---------- Data preparation ---------- +# ----- Data ingestion and feature building ----- if uploaded is not None: df = pd.read_csv(uploaded) - st.info(f"CSV cargado: {uploaded.name} — {len(df)} filas") + st.info(f"CSV loaded: {uploaded.name} — {len(df)} rows") else: - # demo synthetic historical data for training (keeps app usable without CSV) + # synthetic demo history so app is usable without CSV np.random.seed(42) - N = 400 + N = 420 cot_long_demo = np.random.normal(12000, 3000, size=N).astype(int) cot_short_demo = np.random.normal(11000, 3000, size=N).astype(int) - fed_demo = np.clip(np.random.normal(60, 10, size=N), 0, 100) - retail_demo = np.clip(np.random.normal(50, 10, size=N), 0, 100) + fed_demo = np.clip(np.random.normal(60, 12, size=N), 0, 100) + retail_demo = np.clip(np.random.normal(50, 12, size=N), 0, 100) net_demo = cot_long_demo - cot_short_demo z_net_demo = (net_demo - net_demo.mean()) / (net_demo.std() if net_demo.std()>0 else 1) true_score = 0.8*z_net_demo + 1.5*((fed_demo - 50)/50) - 0.5*((retail_demo - 50)/50) @@ -111,14 +106,13 @@ else: "fed_prob": fed_demo, "retail_pct": retail_demo, "target_up": target_demo }) - st.info("Usando demo histórico (sube tu CSV para resultados reales).") + st.info("Running on demo historical data. Upload CSV for real-data calibration.") -# Build feature engineering (robust, reproducible) -def build_features(df, s_cot_local=50000): - df = df.copy() +def build_features(df_in, s_cot_local=50000): + df = df_in.copy() df["net"] = df["cot_long"] - df["cot_short"] - # normalized net (z-style): scale by s_cot (user set) and also produce standardized z df["net_scaled"] = df["net"] / float(s_cot_local) + # rolling z using up to 252 rows (business-year style) df["z_net"] = (df["net"] - df["net"].rolling(252, min_periods=1).mean()) / df["net"].rolling(252, min_periods=1).std().replace(0,1) df["delta_1w"] = df["net"].diff(5).fillna(0) df["delta_4w"] = df["net"].diff(20).fillna(0) @@ -130,122 +124,186 @@ def build_features(df, s_cot_local=50000): df_feat = build_features(df, s_cot_local=s_cot) -# training set (if user uploaded data and target_up exists, use it) +# features chosen for the tactical model feature_cols = ["z_net", "z_delta_1w", "z_delta_4w", "signal_fed", "signal_retail"] + +# prepare training data (if uploaded and has target_up will be used) if "target_up" in df_feat.columns: train_df = df_feat.copy() X_train = train_df[feature_cols].fillna(0).values y_train = train_df["target_up"].astype(int).values else: - # synthetic fallback + # fallback (demo includes target_up) train_df = df_feat.copy() X_train = train_df[feature_cols].fillna(0).values y_train = train_df["target_up"].astype(int).values -# Fit model -with st.spinner("Modelo: calibrando (MAP + Laplace)..."): +# If user pressed retrain and uploaded provided CSV, we would re-fit (button here for UI clarity) +if retrain and uploaded is None: + st.warning("No CSV uploaded — retrain requires historical CSV with 'target_up' labels.") +# Fit model (MAP + Laplace) +with st.spinner("Calibrating Bayesian tactical model (MAP + Laplace)..."): beta_map, cov_post = fit_bayesian_logit(X_train, y_train, prior_var=prior_var) -# ---------- Live row features (from sidebar inputs) ---------- -live = { - "cot_long": cot_long, - "cot_short": cot_short, - "fed_prob": fed_prob, - "retail_pct": retail_pct -} -live_df = pd.DataFrame([live]) -live_feat = build_features(pd.concat([df.head(1), live_df], ignore_index=True), s_cot_local=s_cot).iloc[-1] +# ----- Live row features (from sidebar) ----- +live = {"cot_long": float(cot_long), "cot_short": float(cot_short), "fed_prob": float(fed_prob), "retail_pct": float(retail_pct)} +# to compute z etc we build by appending the live row to the history head +aug = pd.concat([df.head(1).copy(), pd.DataFrame([live])], ignore_index=True) +live_feat = build_features(aug, s_cot_local=s_cot).iloc[-1] x_live = live_feat[feature_cols].fillna(0).values +net = live_feat["net"] -# Posterior sampling and predictive +# ----- Posterior sampling and predictive distribution ----- +# defensive: ensure covariance is PSD-ish try: samples = np.random.multivariate_normal(beta_map, cov_post, size=4000) sample_probs = sigmoid(samples.dot(x_live)) p_mean = float(sample_probs.mean()) p_low = float(np.percentile(sample_probs, 2.5)) p_high = float(np.percentile(sample_probs, 97.5)) -except Exception as e: - # fallback to point estimate if covariance numerically invalid - eta = x_live.dot(beta_map) +except Exception: + # fallback to point probability + eta = float(x_live.dot(beta_map)) p_mean = float(sigmoid(eta)) p_low, p_high = p_mean, p_mean -# ---------- Decision logic ---------- +# Conviction index: transforms p_mean and dispersion into 0-100 +dispersion = max(1e-6, float(np.std(sample_probs))) if 'sample_probs' in locals() else 0.0 +conviction = min(100.0, max(0.0, 100.0 * (abs(p_mean - 0.5) * 2.0) * (1.0 / (1.0 + 5.0 * dispersion)))) +conviction = round(conviction, 1) + +# Decision thresholds (tactical) if p_mean >= 0.60: decision = "COMPRAR (LONG)" - banner_color = "✅" - banner_style = "success" + decision_flag = "buy" elif p_mean <= 0.40: decision = "VENDER (SHORT)" - banner_color = "🚫" - banner_style = "error" + decision_flag = "sell" else: decision = "ESPERAR (NO OPERAR)" - banner_color = "🟡" - banner_style = "warning" + decision_flag = "wait" -# ---------- UI: top result ---------- -colA, colB = st.columns([2,3]) -with colA: +# Compute contribution breakdown (feature * beta) and normalize for narrative +contrib_raw = np.array(beta_map) * np.array(x_live) +# normalize to percent contribution with sign +if np.sum(np.abs(contrib_raw)) > 0: + contrib_pct = 100.0 * contrib_raw / np.sum(np.abs(contrib_raw)) +else: + contrib_pct = np.zeros_like(contrib_raw) + +# ----- Expert narrative generator (deterministic, audit-friendly) ----- +def narrative_tactical(p_mean, p_low, p_high, conviction, decision_flag, x_live, contrib_pct, feature_cols): + # Build human expert style narrative using magnitudes and signed contributions + lines = [] + # regime sentence + fed_s = x_live[3] # signal_fed + cot_s = x_live[0] # z_net + retail_s = x_live[4] + lines.append("Executive summary (tactical):") + lines.append(f"- Bayesian posterior P(USD/JPY ↑) = {p_mean*100:.1f}% (95% CI: {p_low*100:.1f}%–{p_high*100:.1f}%). Conviction: {conviction}/100.") + # interpretation + if fed_s > 0.4: + lines.append("- Macro: FedWatch is strongly hawkish — this structurally supports USD strength versus JPY over short horizon.") + elif fed_s < -0.4: + lines.append("- Macro: FedWatch is dovish — downside pressure on USD is likely.") + else: + lines.append("- Macro: FedWatch neutral/moderately balanced — macro is not the dominant driver.") + # COT + if abs(cot_s) < 0.15: + lines.append("- Positioning: COT non-commercials are currently near neutral (low extremeness); tail risk from crowded positioning is limited.") + elif cot_s >= 0.15: + lines.append("- Positioning: COT shows noticeable long bias among non-commercials — structural support exists but be wary of profit-taking risk.") + else: + lines.append("- Positioning: COT shows notable short bias — risk of short-squeeze exists if macro tilts hawkish.") + # Retail + if retail_s > 0.15: + lines.append("- Retail: retail crowd leaning long — often a contrarian warning for short-term reversals; exercise caution.") + elif retail_s < -0.15: + lines.append("- Retail: retail lean short — can be supportive for medium momentum if institutional flow aligns.") + else: + lines.append("- Retail: retail positioning near balanced — not a dominant contrarian signal now.") + # Contributions + lines.append("- Signal decomposition (signed % contribution):") + for name, pct, val in zip(feature_cols, contrib_pct, x_live): + sign = "+" if pct >= 0 else "-" + lines.append(f" • {name}: {sign}{abs(pct):.1f}% (value {val:.3f})") + # Decision rationale + if decision_flag == "buy": + lines.append("- Tactical recommendation: Gradual LONG accumulation. Rationale: hawkish Fed probabilities dominate neutral institutional positioning; market structure allows tactical upside.") + lines.append("- Suggested execution (tactical): scale-in (3 tranches), initial exposure 0.5–1.5% notional, use volatility stop (1.0–1.5 × recent ATR) and target R:R ≥ 1:1.5.") + elif decision_flag == "sell": + lines.append("- Tactical recommendation: Consider SHORT exposure or reduce existing long exposure. Rationale: posterior favors downside with sufficient conviction.") + lines.append("- Suggested execution: tight initial sizing, stop at 1.0–1.5 × ATR, target adapt to event risk.") + else: + lines.append("- Tactical recommendation: No trade. Rationale: posterior is near-neutral and uncertainty is significant; wait for clearer regime signal or confirmatory flows.") + # Caveats + lines.append("- Caveats: model uses weekly-positioning (COT) which is lagged; complement with intraday order flow and option skew for execution decisions.") + return "\n".join(lines) + +narrative = narrative_tactical(p_mean, p_low, p_high, conviction, decision_flag, x_live, contrib_pct, feature_cols) + +# ----- UI: top result and narrative ----- +left, right = st.columns([2,3]) +with left: st.markdown("
", unsafe_allow_html=True) - st.markdown(f"**Señal (decisión):**
{banner_color} {decision}
", unsafe_allow_html=True) - st.markdown(f"**Prob. (media):** {p_mean*100:.2f}%    **95% CI:** [{p_low*100:.1f}%, {p_high*100:.1f}%]") - st.markdown("**Resumen breve de por qué:**") - # compute component contributions (normalized by assumed weights) - w = np.array([0.30, 0.20, 0.10, 0.30, 0.10]) # chosen weights (documentado) - contribs = w * np.array([x_live[0], x_live[1], x_live[2], x_live[3], x_live[4]]) - comp_names = ["COT (z_net)", "Δ1w (z)", "Δ4w (z)", "Fed (signal)", "Retail (signal)"] - explanation_lines = [] - for nm, val, c in zip(comp_names, x_live, contribs): - explanation_lines.append(f"- {nm}: {val:.3f} → contrib {c:.3f}") - st.write("\n".join(explanation_lines)) + st.markdown(f"
{decision}
", unsafe_allow_html=True) + st.markdown(f"
Prob. (mean): {p_mean*100:.1f}%    95% CI: {p_low*100:.1f}%–{p_high*100:.1f}%
", unsafe_allow_html=True) + st.markdown(f"
Conviction: {conviction}/100
", unsafe_allow_html=True) + st.markdown("
", unsafe_allow_html=True) + st.markdown("### Tactical trade plan (concise)", unsafe_allow_html=True) + if decision_flag == "buy": + st.markdown("- Scale-in (3 tranches). Initial tranche: 0.5–1% notional.", unsafe_allow_html=True) + st.markdown("- Stop: technical stop (1.0–1.5×ATR) or below nearest structure.", unsafe_allow_html=True) + st.markdown("- Target: prefer R:R ≥ 1:1.5; manage increments.", unsafe_allow_html=True) + elif decision_flag == "sell": + st.markdown("- Consider reducing longs or initiating small short (0.5–1%).", unsafe_allow_html=True) + st.markdown("- Tight stops; monitor macro headlines closely.", unsafe_allow_html=True) + else: + st.markdown("- No trade. Await clearer signal or follow confirmatory flow.", unsafe_allow_html=True) st.markdown("
", unsafe_allow_html=True) -with colB: +with right: + # gauge fig = go.Figure(go.Indicator( - mode="gauge+number", + mode="gauge+number+delta", value=p_mean*100, number={'suffix': '%'}, domain={'x': [0,1], 'y': [0,1]}, - title={'text': "Probabilidad USD/JPY al alza"}, - gauge={'axis': {'range': [0,100]}, - 'bar': {'color': "darkcyan"}, - 'steps': [ - {'range': [0,40], 'color': "red"}, - {'range': [40,60], 'color': "gold"}, - {'range': [60,100], 'color': "green"}, - ]} )) + title={'text': "P(USD/JPY ↑)"}, + gauge={ + 'axis': {'range': [0,100]}, + 'bar': {'color': "#1f7a8c"}, + 'steps': [ + {'range':[0,40], 'color':'#a62b2b'}, + {'range':[40,60], 'color':'#bfae59'}, + {'range':[60,100], 'color':'#2a8f6b'} + ], + } + )) st.plotly_chart(fig, use_container_width=True) -# ---------- Component table ---------- -st.markdown("### Componentes y valores normalizados") +st.markdown("### Contextual narrative — expert style", unsafe_allow_html=True) +st.markdown(f"
{narrative.replace(chr(10), '
')}
", unsafe_allow_html=True) + +# ----- Component table and coefficients ----- +st.markdown("### Components (normalized features) and MAP coefficients") comp_df = pd.DataFrame({ - "Componente": comp_names, - "Valor (normalizado)": [f"{v:.4f}" for v in x_live], - "Peso aplicado": list(w) + "Componente": feature_cols, + "Valor (normalized)": [f"{v:.4f}" for v in x_live], + "beta_MAP": [float(b) for b in beta_map], + "Signed % contrib": [f"{float(c):.1f}%" for c in contrib_pct] }) st.table(comp_df) -# ---------- Optional: show coefficients and metrics ---------- -st.markdown("### Coeficientes MAP (modelo bayesiano)") -coef_df = pd.DataFrame({ - "feature": feature_cols, - "beta_map": [float(b) for b in beta_map], - "post_var_diag": [float(v) for v in np.diag(cov_post)] -}) -st.table(coef_df) - -# ---------- Export / download ---------- +# ----- Export & footer ----- out_df = pd.DataFrame([{ "cot_long": cot_long, "cot_short": cot_short, "net": net, "fed_prob": fed_prob, "retail_pct": retail_pct, - "p_mean": p_mean, "p_2.5": p_low, "p_97.5": p_high, "decision": decision + "p_mean": p_mean, "p_2.5": p_low, "p_97.5": p_high, + "conviction": conviction, "decision": decision }]) -st.download_button("Exportar resultado (Excel)", data=to_excel(out_df), - file_name="kennis_usdjpy_signal.xlsx", - mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet") +st.download_button("Export result (Excel)", data=to_excel(out_df), file_name="kennis_usdjpy_signal.xlsx") -# ---------- Footer ---------- st.markdown("---") -st.markdown("
Implementación: MAP logistic (Laplace) — Posterior sampling para incertidumbre. Reemplaza demo subiendo CSV histórico con 'target_up' (0/1) para entrenamiento real.
", unsafe_allow_html=True) -st.markdown("© Kennis - Institucional FX Markets") +st.markdown("
Implementation: Tactical Bayesian logistic (MAP + Laplace). Replace demo with real historical CSV (with 'target_up' labels) for production-grade calibration. Use with execution policy and risk controls.
", unsafe_allow_html=True) +st.markdown("© Kennis FX Markets — Institutional Bayesian Intelligence Engine", unsafe_allow_html=True)