Clarify now-snapshot, align Toto vs NWS forecast in text and chart

- src/weather_ui.py: hero now shows two explicit rows — '📡 Ecowitt now'
(measured) and '🌎 NWS this hour' (predicted) — with the gap between
them surfaced inline so the model error is visible at a glance.
- src/weather_ui.py: new aligned_comparison_markdown produces a +6h /
+12h / +18h / +24h table with Toto and NWS predictions for the same
wall-clock hour, side by side, plus a Δ column. Replaces the previous
high/low layout where the two models' peak times rarely matched.
- src/weather_ui.py: combined_figure traces now set mode='lines'
explicitly and a fully-transparent line on the band polygon so Plotly
doesn't auto-add markers at the band vertices. Toto median switches
from dashed to solid; NWS forecast from dotted to dashed.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

app.py

@@ -20,6 +20,7 @@ import pandas as pd
 from src import ecowitt, forecast_log, nws, persist
 from src.forecast import forecast_series
 from src.weather_ui import (
+    aligned_comparison_markdown,
     combined_figure,
     emoji_strip_markdown,
     headline_forecast_blocks,
@@ -149,15 +150,20 @@ def refresh(cycle_label: str = "Hourly", horizon_label: str = "24 h"):
             nws_temp=nws_aligned.get("temp_f"),
             tz=DISPLAY_TZ,
         )
+        comparison_md = "### 🆚 Same hour, side-by-side (temperature)\n\n" + aligned_comparison_markdown(
+            toto=totos["temp_f"],
+            nws_temp=nws_aligned.get("temp_f"),
+            tz=DISPLAY_TZ,
+        )
     else:
-        toto_md, nws_md = "", ""
+        toto_md, nws_md, comparison_md = "", "", ""
     strip = emoji_strip_markdown(nws_df_raw, DISPLAY_TZ, n=12)
     scoreboard = render_scoreboard(log_conn)
 
     # Backup the SQLite log to the HF dataset (non-blocking).
     persist.push_db_async()
 
-    return hero, toto_md, nws_md, strip, fig, scoreboard
+    return hero, toto_md, nws_md, comparison_md, strip, fig, scoreboard
 
 
 # --- scoreboard ----------------------------------------------------------
@@ -230,6 +236,7 @@ with gr.Blocks(title="Toto Weather Forecast", theme=gr.themes.Soft()) as demo:
     with gr.Row():
         toto_headline_md = gr.Markdown()
         nws_headline_md = gr.Markdown()
+    comparison_md = gr.Markdown()
     strip_md = gr.Markdown()
 
     with gr.Row():
@@ -246,7 +253,7 @@ with gr.Blocks(title="Toto Weather Forecast", theme=gr.themes.Soft()) as demo:
     scoreboard_md = gr.Markdown()
     plot = gr.Plot(label="Forecast")
 
-    outputs = [hero_md, toto_headline_md, nws_headline_md, strip_md, plot, scoreboard_md]
+    outputs = [hero_md, toto_headline_md, nws_headline_md, comparison_md, strip_md, plot, scoreboard_md]
     inputs = [cycle_dd, horizon_dd]
     demo.load(refresh, inputs=inputs, outputs=outputs)
     refresh_btn.click(refresh, inputs=inputs, outputs=outputs)

src/weather_ui.py

@@ -65,7 +65,9 @@ def hero_markdown(
     nws_first: pd.Series | None,
     tz: str,
 ) -> str:
-    """A 'now' tile: current temp/RH/P with hour-over-hour delta and weather words."""
+    """A 'now' tile that explicitly distinguishes Ecowitt's measured value
+    from NWS's forecast for the same hour, so the viewer can see the model
+    error live."""
     if history.empty:
         return "_(no current readings yet)_"
     last = history.dropna(how="all").index.max()
@@ -78,47 +80,115 @@ def hero_markdown(
             return ""
         d = cur[col] - prev[col]
         arrow = "▲" if d > 0 else ("▼" if d < 0 else "·")
-        return f" <span style='opacity:0.6'>{arrow} {d:{fmt}} {unit}/h</span>"
+        return f" <span style='opacity:0.55'>({arrow} {d:{fmt}} {unit}/h)</span>"
 
-    short = ""
+    when = last.tz_convert(tz).strftime("%-I:%M %p %Z")
+
+    # Pull NWS's forecast for the same wall-clock hour (its first period).
+    nws_temp_str = "—"
+    nws_short = "—"
     glyph = "🌡"
     if nws_first is not None and not nws_first.empty:
-        first_row = nws_first.iloc[0] if isinstance(nws_first, pd.DataFrame) else None
-        if isinstance(first_row, pd.Series) and "short_forecast" in first_row:
-            short = str(first_row["short_forecast"])
-            glyph = emoji_for(short)
+        row = nws_first.iloc[0] if isinstance(nws_first, pd.DataFrame) else nws_first
+        if isinstance(row, pd.Series):
+            if "temp_f" in row and pd.notna(row["temp_f"]):
+                nws_temp_str = f"{row['temp_f']:.0f}°F"
+            if "short_forecast" in row:
+                nws_short = str(row["short_forecast"])
+                glyph = emoji_for(nws_short)
+
+    # Highlight the gap between actual and NWS prediction for this hour.
+    gap_str = ""
+    if nws_first is not None and not nws_first.empty:
+        row = nws_first.iloc[0] if isinstance(nws_first, pd.DataFrame) else nws_first
+        if isinstance(row, pd.Series) and "temp_f" in row and pd.notna(row["temp_f"]):
+            gap = float(cur["temp_f"]) - float(row["temp_f"])
+            sign = "+" if gap >= 0 else ""
+            gap_str = f"<span style='opacity:0.55'>(NWS off by {sign}{gap:.1f}°F)</span>"
 
-    when = last.tz_convert(tz).strftime("%-I:%M %p %Z")
     lines = [
-        f"### {glyph} {place} · {cur['temp_f']:.1f}°F",
-        f"<span style='font-size:1.1em'>{cur['humidity']:.0f}% RH · {cur['pressure_inhg']:.2f} inHg{delta('temp_f','°F')}{delta('humidity','%','+.0f')}{delta('pressure_inhg','inHg','+.3f')}</span>",
-        f"<span style='opacity:0.6'>Last reading {when} · NWS now: {short or '—'}</span>",
+        f"### {glyph} {place}",
+        (
+            "| | Temperature | Humidity | Pressure | Conditions |\n"
+            "|---|---|---|---|---|\n"
+            f"| **📡 Ecowitt now** | **{cur['temp_f']:.1f}°F**{delta('temp_f','°F')}"
+            f" | **{cur['humidity']:.0f}%**{delta('humidity','%','+.0f')}"
+            f" | **{cur['pressure_inhg']:.2f} inHg**{delta('pressure_inhg','inHg','+.3f')}"
+            f" | _(measured)_ |\n"
+            f"| **🌎 NWS this hour** | **{nws_temp_str}** {gap_str} | — | — | {nws_short} |"
+        ),
+        f"<span style='opacity:0.55'>Last Ecowitt reading: {when}</span>",
     ]
     return "\n\n".join(lines)
 
 
+def aligned_comparison_markdown(
+    toto: TotoForecast,
+    nws_temp: pd.Series | None,
+    tz: str,
+    offsets_hours: list[int] = (6, 12, 18, 24),
+) -> str:
+    """Apples-to-apples table: at the same future hour, show Toto and NWS.
+
+    For each requested offset h, find the forecast point in each series
+    closest to t0 + h hours and report both numbers in the same row.
+    """
+    if toto is None or toto.median.empty:
+        return ""
+    base = toto.median.index[0] - (toto.median.index[1] - toto.median.index[0]) if len(toto.median) > 1 else toto.median.index[0]
+
+    def _nearest(series: pd.Series, target: pd.Timestamp):
+        if series is None or series.empty:
+            return None, None
+        idx = series.index.get_indexer([target], method="nearest")[0]
+        if idx < 0 or idx >= len(series):
+            return None, None
+        return series.index[idx], float(series.iloc[idx])
+
+    rows = ["| When | 🤖 Toto | 🌎 NWS | Δ |", "|---|---|---|---|"]
+    for h in offsets_hours:
+        target = base + pd.Timedelta(hours=h)
+        t_idx, t_val = _nearest(toto.median, target)
+        n_idx, n_val = _nearest(nws_temp, target) if nws_temp is not None else (None, None)
+        if t_val is None and n_val is None:
+            continue
+        when_label = (t_idx or n_idx).tz_convert(tz).strftime("%-I %p %a")
+        toto_str = f"**{t_val:.0f}°F**" if t_val is not None else "—"
+        nws_str = f"**{n_val:.0f}°F**" if n_val is not None else "—"
+        if t_val is not None and n_val is not None:
+            d = t_val - n_val
+            sign = "+" if d >= 0 else ""
+            delta_str = f"{sign}{d:.1f}°F"
+        else:
+            delta_str = "—"
+        rows.append(f"| +{h}h · {when_label} | {toto_str} | {nws_str} | {delta_str} |")
+    return "\n".join(rows)
+
+
 def headline_forecast_blocks(
     toto: TotoForecast,
     nws_temp: pd.Series | None,
     tz: str,
 ) -> tuple[str, str]:
-    """Return (toto_md, nws_md) for placement in side-by-side columns."""
+    """Side-by-side high/low summaries — same 24h window for both, so
+    different peak/trough times become a real comparison."""
     t_hi, t_hi_t, t_lo, t_lo_t = hi_lo(toto.median, tz)
     width24 = float(toto.p90.iloc[-1] - toto.p10.iloc[-1])
     toto_md = (
-        "### 🤖 Toto says (next 24h)\n\n"
-        f"High **{t_hi:.0f}°F** at {t_hi_t}  ·  Low **{t_lo:.0f}°F** at {t_lo_t}\n\n"
-        f"<span style='opacity:0.6'>80% interval at +24h: ±{width24/2:.1f}°F</span>"
+        "### 🤖 Toto's 24h forecast\n\n"
+        f"High **{t_hi:.0f}°F** at {t_hi_t}\n\n"
+        f"Low **{t_lo:.0f}°F** at {t_lo_t}\n\n"
+        f"<span style='opacity:0.55'>80% interval at +24h: ±{width24/2:.1f}°F</span>"
     )
-
     if nws_temp is None or nws_temp.empty:
-        nws_md = "### 🌎 NWS\n\n_(unavailable for this metric)_"
+        nws_md = "### 🌎 NWS 24h forecast\n\n_(unavailable)_"
     else:
         n_hi, n_hi_t, n_lo, n_lo_t = hi_lo(nws_temp, tz)
         nws_md = (
-            "### 🌎 NWS says (next 24h)\n\n"
-            f"High **{n_hi:.0f}°F** at {n_hi_t}  ·  Low **{n_lo:.0f}°F** at {n_lo_t}\n\n"
-            "<span style='opacity:0.6'>Point forecast (no interval)</span>"
+            "### 🌎 NWS 24h forecast\n\n"
+            f"High **{n_hi:.0f}°F** at {n_hi_t}\n\n"
+            f"Low **{n_lo:.0f}°F** at {n_lo_t}\n\n"
+            "<span style='opacity:0.55'>Point forecast (no interval)</span>"
         )
     return toto_md, nws_md
 
@@ -160,7 +230,7 @@ def combined_figure(
         fig.add_trace(
             go.Scatter(
                 x=hist.index, y=hist.values,
-                name="Ecowitt (past)", mode="lines",
+                name="📡 Ecowitt (measured)", mode="lines",
                 line=dict(color="#222", width=2),
                 showlegend=showlegend, legendgroup="hist",
             ),
@@ -172,8 +242,9 @@ def combined_figure(
                     x=list(toto.p90.index) + list(toto.p10.index[::-1]),
                     y=list(toto.p90.values) + list(toto.p10.values[::-1]),
                     fill="toself", fillcolor="rgba(31,119,180,0.18)",
-                    line=dict(width=0), hoverinfo="skip",
-                    name="Toto 10–90% interval",
+                    mode="lines", line=dict(width=0, color="rgba(0,0,0,0)"),
+                    hoverinfo="skip",
+                    name="🤖 Toto 80% interval",
                     showlegend=showlegend, legendgroup="toto-band",
                 ),
                 row=i, col=1,
@@ -181,8 +252,8 @@ def combined_figure(
             fig.add_trace(
                 go.Scatter(
                     x=toto.median.index, y=toto.median.values,
-                    name="Toto median", mode="lines",
-                    line=dict(color="#1f77b4", width=2, dash="dash"),
+                    name="🤖 Toto median", mode="lines",
+                    line=dict(color="#1f77b4", width=2.5),
                     showlegend=showlegend, legendgroup="toto-med",
                 ),
                 row=i, col=1,
@@ -193,8 +264,8 @@ def combined_figure(
                 fig.add_trace(
                     go.Scatter(
                         x=ns.index, y=ns.values,
-                        name="NWS forecast", mode="lines",
-                        line=dict(color="#d62728", width=2, dash="dot"),
+                        name="🌎 NWS forecast", mode="lines",
+                        line=dict(color="#d62728", width=2.5, dash="dash"),
                         showlegend=showlegend, legendgroup="nws",
                     ),
                     row=i, col=1,