feat(altair): implement line-confidence

plots/line-confidence/implementations/altair.py

@@ -0,0 +1,147 @@
+""" pyplots.ai
+line-confidence: Line Plot with Confidence Interval
+Library: altair 6.0.0 | Python 3.13.11
+Quality: 78/100 | Created: 2025-12-26
+"""
+
+import altair as alt
+import numpy as np
+import pandas as pd
+
+
+# Data: Simulating a time series forecast with 95% confidence interval
+np.random.seed(42)
+
+# Generate time points (50 days of daily data)
+n_points = 50
+days = np.arange(n_points)
+
+# Create a trend with some curvature (e.g., a model prediction)
+trend = 100 + 0.5 * days + 0.02 * days**2 + np.sin(days / 5) * 5
+
+# Add noise for the central line
+noise = np.random.normal(0, 3, n_points)
+y_mean = trend + noise
+
+# Confidence interval that widens over time (typical for forecasts)
+uncertainty = 5 + 0.15 * days  # Uncertainty grows with time
+y_lower = y_mean - 1.96 * uncertainty / 2
+y_upper = y_mean + 1.96 * uncertainty / 2
+
+# Create DataFrame
+df = pd.DataFrame({"Day": days, "Value": y_mean, "Lower": y_lower, "Upper": y_upper})
+
+# Calculate appropriate y-axis domain with padding (avoid unnecessary whitespace from 0)
+y_min = df["Lower"].min()
+y_max = df["Upper"].max()
+y_padding = (y_max - y_min) * 0.1
+y_domain = [y_min - y_padding, y_max + y_padding]
+
+# Create shared y-scale for consistent rendering across all layers
+y_scale = alt.Scale(domain=y_domain)
+
+# Create the confidence band (area) with fill encoding for legend
+band = (
+    alt.Chart(df)
+    .transform_calculate(legend_label='"95% Confidence Interval"')
+    .mark_area(opacity=0.3)
+    .encode(
+        x=alt.X("Day:Q", title="Day", axis=alt.Axis(values=list(range(0, 51, 5)))),
+        y=alt.Y("Lower:Q", title="Predicted Value", scale=y_scale),
+        y2="Upper:Q",
+        fill=alt.Fill(
+            "legend_label:N",
+            scale=alt.Scale(domain=["95% Confidence Interval"], range=["#306998"]),
+            legend=alt.Legend(
+                title="Legend",
+                orient="right",
+                labelFontSize=16,
+                titleFontSize=18,
+                symbolType="square",
+                symbolSize=300,
+                symbolOpacity=0.3,
+            ),
+        ),
+    )
+)
+
+# Create the central line with stroke encoding for legend
+line = (
+    alt.Chart(df)
+    .transform_calculate(legend_label='"Predicted Mean"')
+    .mark_line(strokeWidth=4)
+    .encode(
+        x="Day:Q",
+        y=alt.Y("Value:Q", scale=y_scale),
+        stroke=alt.Stroke(
+            "legend_label:N",
+            scale=alt.Scale(domain=["Predicted Mean"], range=["#306998"]),
+            legend=alt.Legend(
+                title="Legend", orient="right", labelFontSize=16, titleFontSize=18, symbolStrokeWidth=4, symbolSize=300
+            ),
+        ),
+    )
+)
+
+# Add point markers on the line for clarity (increased size for better visibility)
+points = (
+    alt.Chart(df)
+    .mark_point(size=100, filled=True, color="#306998")
+    .encode(x="Day:Q", y=alt.Y("Value:Q", scale=y_scale))
+)
+
+# Combine band, line, and points with resolved legends
+chart = (
+    alt.layer(band, line, points)
+    .resolve_legend(fill="independent", stroke="independent")
+    .properties(
+        width=1600, height=900, title=alt.Title("line-confidence · altair · pyplots.ai", fontSize=28, anchor="middle")
+    )
+    .configure_axis(labelFontSize=18, titleFontSize=22, gridOpacity=0.3)
+    .configure_view(strokeWidth=0)
+)
+
+# Save as PNG (scale_factor=3 gives 4800x2700)
+chart.save("plot.png", scale_factor=3.0)
+
+# Save as HTML for interactivity
+interactive_band = (
+    alt.Chart(df)
+    .mark_area(opacity=0.3, color="#306998")
+    .encode(
+        x=alt.X("Day:Q", title="Day"),
+        y=alt.Y("Lower:Q", title="Predicted Value", scale=y_scale),
+        y2="Upper:Q",
+        tooltip=[
+            alt.Tooltip("Day:Q", title="Day"),
+            alt.Tooltip("Lower:Q", title="Lower Bound", format=".1f"),
+            alt.Tooltip("Upper:Q", title="Upper Bound", format=".1f"),
+        ],
+    )
+)
+
+interactive_line = (
+    alt.Chart(df).mark_line(strokeWidth=3, color="#306998").encode(x="Day:Q", y=alt.Y("Value:Q", scale=y_scale))
+)
+
+interactive_points = (
+    alt.Chart(df)
+    .mark_point(size=60, filled=True, color="#306998")
+    .encode(
+        x="Day:Q",
+        y=alt.Y("Value:Q", scale=y_scale),
+        tooltip=[alt.Tooltip("Day:Q", title="Day"), alt.Tooltip("Value:Q", title="Predicted Value", format=".1f")],
+    )
+)
+
+interactive_chart = (
+    alt.layer(interactive_band, interactive_line, interactive_points)
+    .properties(
+        width=800, height=450, title=alt.Title("line-confidence · altair · pyplots.ai", fontSize=20, anchor="middle")
+    )
+    .configure_axis(labelFontSize=14, titleFontSize=16, gridOpacity=0.3)
+    .configure_view(strokeWidth=0)
+    .interactive()
+)
+
+interactive_chart.save("plot.html")

plots/line-confidence/metadata/altair.yaml

@@ -0,0 +1,28 @@
+library: altair
+specification_id: line-confidence
+created: '2025-12-26T16:14:06Z'
+updated: '2025-12-26T16:33:33Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20525434309
+issue: 0
+python_version: 3.13.11
+library_version: 6.0.0
+preview_url: https://storage.googleapis.com/pyplots-images/plots/line-confidence/altair/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/line-confidence/altair/plot_thumb.png
+preview_html: https://storage.googleapis.com/pyplots-images/plots/line-confidence/altair/plot.html
+quality_score: 78
+review:
+  strengths:
+  - Clean code structure following KISS principles with clear data generation logic
+  - Confidence band appropriately widens over time showing realistic forecast uncertainty
+    behavior
+  - Uses Altair layered chart composition effectively (band + line + points)
+  - Interactive HTML export includes useful tooltips for exploration
+  - Consistent color scheme with good contrast between line and band
+  weaknesses:
+  - Legend is not visible in the rendered PNG despite code attempting to create one
+    using resolve_legend - this violates spec requirement for legend that clearly
+    identifies both the central line and confidence band
+  - Y-axis starts at 0 but data ranges ~90-180, wasting nearly half the vertical canvas
+    space (the code calculates y_domain but Altair appears to ignore it or it is not
+    applied correctly to all layers)