ParallelLLC/algorithmic_trading
2732
1"""Plotly figures for the Lab.2 3Colour system (dark surface, validated for CVD separation):4 5* **blue** is always *your strategy's honest result* — the realised equity6 curve, the out-of-sample fold, the observed Sharpe.7* **orange** is always *the thing it is measured against* — buy & hold, the8 in-sample fold, the null distribution.9 10Holding that mapping across every figure means a reader learns it once.11"""12 13from __future__ import annotations14 15from typing import Dict, Optional16 17import numpy as np18import pandas as pd19import plotly.graph_objects as go20 21SURFACE = "#1a1a19"22PAGE = "#0d0d0d"23INK = "#ffffff"24INK_SECONDARY = "#c3c2b7"25INK_MUTED = "#898781"26GRID = "#2c2c2a"27AXIS = "#383835"28 29SUBJECT = "#3987e5" # categorical slot 130REFERENCE = "#d95926" # categorical slot 231NEGATIVE = "#e66767" # negative arm of the diverging pair (drawdowns)32 33FONT = 'system-ui, -apple-system, "Segoe UI", sans-serif'34 35_EMPTY_NOTE = "Run an analysis to populate this chart."36 37 38def _base_layout(title: str, height: int = 340, **kwargs) -> dict:39 return dict(40 title=dict(text=title, font=dict(size=15, color=INK), x=0, xanchor="left", pad=dict(b=8)),41 paper_bgcolor=PAGE,42 plot_bgcolor=SURFACE,43 font=dict(family=FONT, size=12, color=INK_SECONDARY),44 height=height,45 margin=dict(l=56, r=24, t=48, b=40),46 hovermode="x unified",47 hoverlabel=dict(bgcolor=SURFACE, bordercolor=AXIS, font=dict(color=INK, family=FONT)),48 xaxis=dict(gridcolor=GRID, linecolor=AXIS, zeroline=False, tickfont=dict(color=INK_MUTED)),49 yaxis=dict(gridcolor=GRID, linecolor=AXIS, zeroline=False, tickfont=dict(color=INK_MUTED)),50 legend=dict(51 orientation="h", yanchor="bottom", y=1.02, xanchor="left", x=0,52 font=dict(color=INK_SECONDARY, size=11), bgcolor="rgba(0,0,0,0)",53 ),54 **kwargs,55 )56 57 58def empty_figure(message: str = _EMPTY_NOTE, height: int = 340) -> go.Figure:59 fig = go.Figure()60 fig.update_layout(**_base_layout("", height=height))61 fig.update_xaxes(visible=False)62 fig.update_yaxes(visible=False)63 fig.add_annotation(64 text=message, showarrow=False, xref="paper", yref="paper", x=0.5, y=0.5,65 font=dict(color=INK_MUTED, size=13),66 )67 return fig68 69 70def equity_chart(report, benchmark_label: str = "Buy & hold") -> go.Figure:71 """Strategy equity against its benchmark, both indexed to the same start."""72 bt = report.backtest73 strat = bt.equity / bt.equity.iloc[0] * 100.074 bench = bt.benchmark_equity / bt.benchmark_equity.iloc[0] * 100.075 76 fig = go.Figure()77 fig.add_trace(78 go.Scatter(79 x=bench.index, y=bench.to_numpy(), name=benchmark_label, mode="lines",80 line=dict(color=REFERENCE, width=2, dash="dash"),81 hovertemplate=benchmark_label + " %{y:.1f}<extra></extra>",82 )83 )84 fig.add_trace(85 go.Scatter(86 x=strat.index, y=strat.to_numpy(), name=report.strategy.name, mode="lines",87 line=dict(color=SUBJECT, width=2),88 hovertemplate=report.strategy.name + " %{y:.1f}<extra></extra>",89 )90 )91 92 # Direct-label the two endpoints; the axis and tooltip carry everything else.93 for series, color, label in ((strat, SUBJECT, report.strategy.name), (bench, REFERENCE, benchmark_label)):94 fig.add_annotation(95 x=series.index[-1], y=float(series.iloc[-1]),96 text=f" {label}: {series.iloc[-1]:.0f}", showarrow=False,97 xanchor="left", font=dict(color=color, size=11),98 )99 100 fig.update_layout(**_base_layout("Growth of 100 (net of costs)", height=360))101 fig.update_layout(margin=dict(l=56, r=140, t=48, b=40))102 return fig103 104 105def drawdown_chart(report) -> go.Figure:106 """Underwater plot — how deep, and for how long."""107 from .metrics import drawdown_series108 109 dd = drawdown_series(report.backtest.equity) * 100.0110 fig = go.Figure(111 go.Scatter(112 x=dd.index, y=dd.to_numpy(), mode="lines", name="Drawdown",113 line=dict(color=NEGATIVE, width=2), fill="tozeroy",114 fillcolor="rgba(230,103,103,0.18)",115 hovertemplate="Drawdown %{y:.1f}%<extra></extra>",116 )117 )118 trough = float(dd.min())119 fig.add_annotation(120 x=dd.idxmin(), y=trough, text=f"worst {trough:.1f}%", showarrow=True,121 arrowhead=0, arrowcolor=AXIS, ay=24, font=dict(color=INK_SECONDARY, size=11),122 )123 fig.update_layout(**_base_layout("Drawdown", height=240, showlegend=False))124 fig.update_yaxes(ticksuffix="%")125 return fig126 127 128def permutation_chart(report) -> go.Figure:129 """The headline chart: your Sharpe against Sharpes from shuffled markets."""130 perm = report.permutation131 if perm is None or perm.null.size == 0:132 return empty_figure("Permutation test was skipped.", height=320)133 134 null = perm.null135 fig = go.Figure()136 fig.add_trace(137 go.Histogram(138 x=null, name="Shuffled markets (no real edge)", nbinsx=44,139 marker=dict(color="rgba(217,89,38,0.55)", line=dict(color=REFERENCE, width=1)),140 hovertemplate="Sharpe %{x:.2f}<br>%{y} shuffles<extra></extra>",141 )142 )143 144 top = np.histogram(null, bins=44)[0].max() if null.size else 1145 fig.add_trace(146 go.Scatter(147 x=[perm.observed, perm.observed], y=[0, top * 1.08], mode="lines",148 name="Your strategy", line=dict(color=SUBJECT, width=2),149 hovertemplate="Your Sharpe %{x:.2f}<extra></extra>",150 )151 )152 fig.add_annotation(153 x=perm.observed, y=top * 1.08, text=f" your Sharpe {perm.observed:.2f}",154 showarrow=False, xanchor="left", font=dict(color=SUBJECT, size=11),155 )156 157 beats = (null >= perm.observed).mean() * 100.0158 fig.update_layout(159 **_base_layout(160 f"Permutation test — {beats:.0f}% of structure-free markets did this well or better "161 f"(p = {perm.p_value:.3f})",162 height=320,163 )164 )165 fig.update_layout(hovermode="closest", bargap=0.02)166 fig.update_xaxes(title=dict(text="Annualised Sharpe ratio", font=dict(color=INK_MUTED, size=11)))167 # Headroom so the "your Sharpe" label never collides with the plot edge.168 fig.update_yaxes(169 title=dict(text="Shuffled markets", font=dict(color=INK_MUTED, size=11)),170 range=[0, top * 1.28],171 )172 return fig173 174 175def walkforward_chart(report) -> go.Figure:176 """In-sample vs out-of-sample Sharpe, fold by fold."""177 folds = report.walkforward.get("folds") or []178 if not folds:179 return empty_figure(report.walkforward.get("note") or _EMPTY_NOTE, height=300)180 181 labels = [f"Fold {f['fold']}<br><span style='font-size:10px'>{f['test_start'][:7]}</span>" for f in folds]182 fig = go.Figure()183 fig.add_trace(184 go.Bar(185 x=labels, y=[f["is_sharpe"] for f in folds], name="In-sample (tuned)",186 marker=dict(color=REFERENCE, line=dict(color=SURFACE, width=2)),187 hovertemplate="In-sample Sharpe %{y:.2f}<extra></extra>",188 )189 )190 fig.add_trace(191 go.Bar(192 x=labels, y=[f["oos_sharpe"] for f in folds], name="Out-of-sample (blind)",193 marker=dict(color=SUBJECT, line=dict(color=SURFACE, width=2)),194 hovertemplate="Out-of-sample Sharpe %{y:.2f}<extra></extra>",195 )196 )197 eff = report.walkforward.get("efficiency", 0.0)198 fig.update_layout(199 **_base_layout(f"Walk-forward — {eff:.0%} of the tuned Sharpe survived out of sample", height=300)200 )201 fig.update_layout(barmode="group", bargap=0.35, bargroupgap=0.08, hovermode="x unified")202 fig.add_hline(y=0, line=dict(color=AXIS, width=1))203 return fig204 205 206def score_chart(verdict: Dict[str, object], significance_label: str = "Beats shuffled markets") -> go.Figure:207 """The five components behind the Reality Score."""208 components = verdict.get("components") or {}209 if not components:210 return empty_figure(height=260)211 212 pretty = {213 "significance": significance_label,214 "selection": "Survives selection bias",215 "walk_forward": "Holds up walking forward",216 "overfitting": "Not overfit (PBO)",217 "robustness": "Survives 3x costs",218 }219 keys = list(pretty)220 values = [float(components.get(k, 0.0)) for k in keys]221 222 fig = go.Figure(223 go.Bar(224 x=values, y=[pretty[k] for k in keys], orientation="h",225 marker=dict(color=SUBJECT, line=dict(color=SURFACE, width=2)),226 text=[f"{v:.0f}" for v in values], textposition="outside",227 textfont=dict(color=INK_SECONDARY, size=11),228 hovertemplate="%{y}: %{x:.0f}/100<extra></extra>",229 )230 )231 fig.update_layout(**_base_layout("Where the score comes from", height=260, showlegend=False))232 fig.update_layout(margin=dict(l=190, r=48, t=48, b=32), hovermode="closest")233 fig.update_xaxes(range=[0, 108], tickvals=[0, 25, 50, 75, 100])234 fig.update_yaxes(autorange="reversed")235 return fig236 237 238def arena_chart(table: pd.DataFrame) -> go.Figure:239 """Leaderboard bars. One measure, one colour — the table carries the rest."""240 if table is None or table.empty:241 return empty_figure(height=380)242 243 ordered = table.iloc[::-1]244 fig = go.Figure(245 go.Bar(246 x=ordered["Sharpe"].to_numpy(), y=ordered["Strategy"].tolist(), orientation="h",247 marker=dict(color=SUBJECT, line=dict(color=SURFACE, width=2)),248 customdata=np.column_stack([ordered["p-value"].to_numpy(), ordered["DSR"].to_numpy()]),249 hovertemplate="%{y}<br>Sharpe %{x:.2f}<br>p = %{customdata[0]:.3f}"250 "<br>Deflated Sharpe %{customdata[1]:.2f}<extra></extra>",251 )252 )253 # Direct-label only what matters: the ones that actually cleared significance.254 for _, row in ordered.iterrows():255 if np.isfinite(row["p-value"]) and row["p-value"] < 0.05:256 fig.add_annotation(257 x=row["Sharpe"], y=row["Strategy"], text=" p < 0.05", showarrow=False,258 xanchor="left" if row["Sharpe"] >= 0 else "right",259 font=dict(color=INK_SECONDARY, size=10),260 )261 fig.update_layout(262 **_base_layout(263 "Strategy arena — Sharpe ratio, ordered by strength of evidence",264 height=max(300, 42 * len(table)),265 showlegend=False,266 )267 )268 fig.update_layout(margin=dict(l=180, r=96, t=48, b=32), hovermode="closest")269 fig.add_vline(x=0, line=dict(color=AXIS, width=1))270 return fig271 272 273def cross_permutation_chart(report) -> go.Figure:274 """Sharpe against books of identical shape holding randomly chosen names."""275 perm = report.permutation276 if perm is None or perm.null.size == 0:277 return empty_figure("Name-shuffle test was skipped.", height=320)278 279 fig = go.Figure()280 fig.add_trace(281 go.Histogram(282 x=perm.null, name="Same book, random names", nbinsx=40,283 marker=dict(color="rgba(217,89,38,0.55)", line=dict(color=REFERENCE, width=1)),284 hovertemplate="Sharpe %{x:.2f}<br>%{y} shuffles<extra></extra>",285 )286 )287 top = np.histogram(perm.null, bins=40)[0].max() if perm.null.size else 1288 fig.add_trace(289 go.Scatter(290 x=[perm.observed, perm.observed], y=[0, top * 1.08], mode="lines",291 name="Your book", line=dict(color=SUBJECT, width=2),292 hovertemplate="Your Sharpe %{x:.2f}<extra></extra>",293 )294 )295 fig.add_annotation(296 x=perm.observed, y=top * 1.08, text=f" your Sharpe {perm.observed:.2f}",297 showarrow=False, xanchor="left", font=dict(color=SUBJECT, size=11),298 )299 beats = (perm.null >= perm.observed).mean() * 100.0300 fig.update_layout(301 **_base_layout(302 f"Name-shuffle test — {beats:.0f}% of books with the same shape but random names "303 f"did this well or better (p = {perm.p_value:.3f})",304 height=320,305 )306 )307 fig.update_layout(hovermode="closest", bargap=0.02)308 fig.update_xaxes(title=dict(text="Annualised Sharpe ratio", font=dict(color=INK_MUTED, size=11)))309 fig.update_yaxes(310 title=dict(text="Shuffled books", font=dict(color=INK_MUTED, size=11)),311 range=[0, top * 1.28],312 )313 return fig314 315 316def attribution_chart(attribution: Dict[str, object]) -> go.Figure:317 """Factor betas. One measure across categories, so one colour."""318 if not attribution or not attribution.get("available"):319 return empty_figure((attribution or {}).get("note", _EMPTY_NOTE), height=280)320 321 betas = attribution.get("betas") or {}322 if not betas:323 return empty_figure("No factor exposures to show.", height=280)324 325 names = list(betas)326 values = [betas[n] for n in names]327 fig = go.Figure(328 go.Bar(329 x=values, y=[n.replace("_", " ") for n in names], orientation="h",330 marker=dict(color=SUBJECT, line=dict(color=SURFACE, width=2)),331 text=[f"{v:+.2f}" for v in values], textposition="outside",332 textfont=dict(color=INK_SECONDARY, size=11),333 hovertemplate="%{y} beta %{x:.2f}<extra></extra>",334 )335 )336 alpha = attribution.get("alpha_annual", 0.0)337 t_stat = attribution.get("alpha_t_stat", 0.0)338 fig.update_layout(339 **_base_layout(340 f"Style exposure — alpha {alpha:+.1%}/yr (t = {t_stat:.1f}), "341 f"R² {attribution.get('r_squared', 0):.0%}",342 height=280,343 showlegend=False,344 )345 )346 fig.update_layout(margin=dict(l=120, r=88, t=48, b=32), hovermode="closest")347 # Outside labels need room or the widest beta reads as "+0".348 span = max(abs(min(values)), abs(max(values)), 0.1)349 fig.update_xaxes(range=[min(0, min(values)) - 0.25 * span, max(0, max(values)) + 0.35 * span])350 fig.add_vline(x=0, line=dict(color=AXIS, width=1))351 return fig352 353 354def weights_chart(report) -> go.Figure:355 """Gross and net exposure over time — is the book actually neutral?"""356 held = report.backtest.held357 gross = held.abs().sum(axis=1)358 net = held.sum(axis=1)359 360 fig = go.Figure()361 fig.add_trace(362 go.Scatter(363 x=gross.index, y=gross.to_numpy(), name="Gross", mode="lines",364 line=dict(color=REFERENCE, width=2, dash="dash"),365 hovertemplate="Gross %{y:.2f}x<extra></extra>",366 )367 )368 fig.add_trace(369 go.Scatter(370 x=net.index, y=net.to_numpy(), name="Net", mode="lines",371 line=dict(color=SUBJECT, width=2),372 hovertemplate="Net %{y:.2f}x<extra></extra>",373 )374 )375 fig.update_layout(**_base_layout("Book exposure", height=240))376 fig.add_hline(y=0, line=dict(color=AXIS, width=1))377 return fig378 379 380def exposure_chart(report) -> go.Figure:381 """What the strategy was actually holding, over time."""382 pos = report.backtest.position383 fig = go.Figure(384 go.Scatter(385 x=pos.index, y=pos.to_numpy(), mode="lines", name="Exposure",386 line=dict(color=SUBJECT, width=2, shape="hv"), fill="tozeroy",387 fillcolor="rgba(57,135,229,0.16)",388 hovertemplate="Exposure %{y:.2f}x<extra></extra>",389 )390 )391 fig.update_layout(**_base_layout("Position held", height=200, showlegend=False))392 fig.add_hline(y=0, line=dict(color=AXIS, width=1))393 return fig394