feat(letsplot): implement bar-3d (#2901)

github-actions[bot] · claude · web-flow · commit b6fb9860a966 · 2025-12-31T00:22:25.000Z
## Implementation: `bar-3d` - letsplot Implements the **letsplot** version of `bar-3d`. **File:** `plots/bar-3d/implementations/letsplot.py` **Parent Issue:** #2857 --- :robot: *[impl-generate workflow](https://github.com/MarkusNeusinger/pyplots/actions/runs/20608481317)* --------- Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
diff --git a/plots/bar-3d/implementations/letsplot.py b/plots/bar-3d/implementations/letsplot.py
@@ -0,0 +1,306 @@
+""" pyplots.ai
+bar-3d: 3D Bar Chart
+Library: letsplot 4.8.2 | Python 3.13.11
+Quality: 91/100 | Created: 2025-12-31
+"""
+
+import numpy as np
+import pandas as pd
+from lets_plot import (
+    LetsPlot,
+    aes,
+    element_blank,
+    element_text,
+    geom_polygon,
+    geom_segment,
+    geom_text,
+    ggplot,
+    ggsize,
+    labs,
+    layer_tooltips,
+    scale_fill_viridis,
+    theme,
+    theme_minimal,
+)
+from lets_plot.export import ggsave
+
+
+LetsPlot.setup_html()
+
+# Data - Quarterly sales performance across product categories
+np.random.seed(42)
+
+products = ["Electronics", "Clothing", "Home", "Sports", "Books"]
+quarters = ["Q1", "Q2", "Q3", "Q4"]
+
+# Revenue data in millions
+revenue = np.array(
+    [
+        [45, 52, 48, 65],  # Electronics
+        [32, 38, 42, 55],  # Clothing
+        [28, 30, 35, 40],  # Home
+        [20, 35, 45, 30],  # Sports
+        [15, 18, 22, 28],  # Books
+    ]
+)
+
+# 3D to 2D isometric projection settings
+elev = np.radians(25)  # elevation angle
+azim = np.radians(-50)  # azimuth angle
+cos_azim, sin_azim = np.cos(azim), np.sin(azim)
+cos_elev, sin_elev = np.cos(elev), np.sin(elev)
+
+# Bar dimensions - good spacing for clarity
+bar_width = 0.5
+bar_depth = 0.5
+spacing_x = 1.4
+spacing_y = 1.6
+
+# Create bar polygons (each bar is a 3D box with visible faces)
+bar_polygons = []
+bar_id = 0
+
+for i, product in enumerate(products):
+    for j, quarter in enumerate(quarters):
+        x_base = j * spacing_x
+        y_base = i * spacing_y
+        height = revenue[i, j]
+
+        # Define the 8 corners of the bar
+        corners_3d = {
+            "front_bottom_left": (x_base, y_base, 0),
+            "front_bottom_right": (x_base + bar_width, y_base, 0),
+            "front_top_left": (x_base, y_base, height),
+            "front_top_right": (x_base + bar_width, y_base, height),
+            "back_bottom_left": (x_base, y_base + bar_depth, 0),
+            "back_bottom_right": (x_base + bar_width, y_base + bar_depth, 0),
+            "back_top_left": (x_base, y_base + bar_depth, height),
+            "back_top_right": (x_base + bar_width, y_base + bar_depth, height),
+        }
+
+        # Project all corners to 2D (inline projection math)
+        corners_2d = {}
+        for name, (cx, cy, cz) in corners_3d.items():
+            x_rot = cx * cos_azim - cy * sin_azim
+            y_rot = cx * sin_azim + cy * cos_azim
+            px = x_rot
+            py = y_rot * sin_elev + cz * cos_elev
+            corners_2d[name] = (px, py)
+
+        # Calculate bar center depth for ordering (inline projection)
+        center_x = x_base + bar_width / 2
+        center_y = y_base + bar_depth / 2
+        center_z = height / 2
+        cx_rot = center_x * cos_azim - center_y * sin_azim
+        cy_rot = center_x * sin_azim + center_y * cos_azim
+        bar_depth_value = cy_rot * cos_elev - center_z * sin_elev
+
+        # Define the 3 visible faces (front, top, right side)
+        front_face = [
+            corners_2d["front_bottom_left"],
+            corners_2d["front_bottom_right"],
+            corners_2d["front_top_right"],
+            corners_2d["front_top_left"],
+        ]
+
+        top_face = [
+            corners_2d["front_top_left"],
+            corners_2d["front_top_right"],
+            corners_2d["back_top_right"],
+            corners_2d["back_top_left"],
+        ]
+
+        right_face = [
+            corners_2d["front_bottom_right"],
+            corners_2d["back_bottom_right"],
+            corners_2d["back_top_right"],
+            corners_2d["front_top_right"],
+        ]
+
+        # Add each face as a polygon
+        for face_name, face_coords in [("front", front_face), ("top", top_face), ("right", right_face)]:
+            for k, (fx, fy) in enumerate(face_coords):
+                bar_polygons.append(
+                    {
+                        "x": fx,
+                        "y": fy,
+                        "bar_id": bar_id,
+                        "face": face_name,
+                        "face_id": f"{bar_id}_{face_name}",
+                        "height": height,
+                        "depth": bar_depth_value,
+                        "product": product,
+                        "quarter": quarter,
+                        "order": k,
+                    }
+                )
+        bar_id += 1
+
+df_bars = pd.DataFrame(bar_polygons)
+
+# Sort by depth (back to front) for proper rendering order
+df_bars = df_bars.sort_values(["depth", "face_id", "order"])
+
+# Create floor grid for depth perception
+grid_lines = []
+x_min, x_max = -0.3, (len(quarters) - 1) * spacing_x + bar_width + 0.3
+y_min, y_max = -0.3, (len(products) - 1) * spacing_y + bar_depth + 0.3
+z_floor = -3
+
+# Grid lines parallel to X axis
+for i in range(len(products) + 1):
+    y_line = i * spacing_y - 0.3 + bar_depth / 2
+    # Inline projection for start point
+    x_rot_s = x_min * cos_azim - y_line * sin_azim
+    y_rot_s = x_min * sin_azim + y_line * cos_azim
+    start_x = x_rot_s
+    start_y = y_rot_s * sin_elev + z_floor * cos_elev
+    # Inline projection for end point
+    x_rot_e = x_max * cos_azim - y_line * sin_azim
+    y_rot_e = x_max * sin_azim + y_line * cos_azim
+    end_x = x_rot_e
+    end_y = y_rot_e * sin_elev + z_floor * cos_elev
+    grid_lines.append({"x": start_x, "y": start_y, "xend": end_x, "yend": end_y})
+
+# Grid lines parallel to Y axis
+for j in range(len(quarters) + 1):
+    x_line = j * spacing_x - 0.3 + bar_width / 2
+    # Inline projection for start point
+    x_rot_s = x_line * cos_azim - y_min * sin_azim
+    y_rot_s = x_line * sin_azim + y_min * cos_azim
+    start_x = x_rot_s
+    start_y = y_rot_s * sin_elev + z_floor * cos_elev
+    # Inline projection for end point
+    x_rot_e = x_line * cos_azim - y_max * sin_azim
+    y_rot_e = x_line * sin_azim + y_max * cos_azim
+    end_x = x_rot_e
+    end_y = y_rot_e * sin_elev + z_floor * cos_elev
+    grid_lines.append({"x": start_x, "y": start_y, "xend": end_x, "yend": end_y})
+
+df_grid = pd.DataFrame(grid_lines)
+
+# Create axis lines from corner (inline projection)
+origin_x, origin_y, origin_z = x_min - 1.5, y_min - 1.5, z_floor
+x_rot_o = origin_x * cos_azim - origin_y * sin_azim
+y_rot_o = origin_x * sin_azim + origin_y * cos_azim
+axis_origin_x = x_rot_o
+axis_origin_y = y_rot_o * sin_elev + origin_z * cos_elev
+
+axis_len_x = (x_max - x_min) * 0.4
+axis_len_y = (y_max - y_min) * 0.4
+axis_len_z = revenue.max() * 0.4
+
+# X axis end (inline projection)
+x_end_3d = origin_x + axis_len_x
+x_rot_xe = x_end_3d * cos_azim - origin_y * sin_azim
+y_rot_xe = x_end_3d * sin_azim + origin_y * cos_azim
+x_axis_end_x = x_rot_xe
+x_axis_end_y = y_rot_xe * sin_elev + origin_z * cos_elev
+
+# Y axis end (inline projection)
+y_end_3d = origin_y + axis_len_y
+x_rot_ye = origin_x * cos_azim - y_end_3d * sin_azim
+y_rot_ye = origin_x * sin_azim + y_end_3d * cos_azim
+y_axis_end_x = x_rot_ye
+y_axis_end_y = y_rot_ye * sin_elev + origin_z * cos_elev
+
+# Z axis end (inline projection)
+z_end_3d = origin_z + axis_len_z
+z_axis_end_x = x_rot_o
+z_axis_end_y = y_rot_o * sin_elev + z_end_3d * cos_elev
+
+axes_df = pd.DataFrame(
+    {
+        "x": [axis_origin_x, axis_origin_x, axis_origin_x],
+        "y": [axis_origin_y, axis_origin_y, axis_origin_y],
+        "xend": [x_axis_end_x, y_axis_end_x, z_axis_end_x],
+        "yend": [x_axis_end_y, y_axis_end_y, z_axis_end_y],
+    }
+)
+
+# Revenue label on Z axis (positioned at top of z-axis, slightly right to avoid cutoff)
+z_label_df = pd.DataFrame({"x": [z_axis_end_x + 0.3], "y": [z_axis_end_y + 1.0], "label": ["Revenue ($M)"]})
+
+# Product labels positioned at front-left, well away from the bars
+product_labels = []
+for i, product in enumerate(products):
+    y_pos = i * spacing_y + bar_depth / 2
+    # Position labels at front-left corner, outside the grid area
+    label_x_3d = -2.5
+    label_y_3d = -3.5
+    x_rot_p = label_x_3d * cos_azim - label_y_3d * sin_azim
+    y_rot_p = label_x_3d * sin_azim + label_y_3d * cos_azim
+    # Stack vertically with consistent spacing
+    px = x_rot_p
+    py = y_rot_p * sin_elev + z_floor * cos_elev + i * 2.5
+    product_labels.append({"x": px, "y": py, "label": product})
+df_product_labels = pd.DataFrame(product_labels)
+
+# Quarter labels positioned further forward to avoid overlap with bars
+quarter_labels = []
+for j, quarter in enumerate(quarters):
+    x_pos = j * spacing_x + bar_width / 2
+    # Inline projection - move labels much further forward (-4.0 instead of -2.0)
+    label_y_3d = -4.0
+    x_rot_q = x_pos * cos_azim - label_y_3d * sin_azim
+    y_rot_q = x_pos * sin_azim + label_y_3d * cos_azim
+    px = x_rot_q
+    py = y_rot_q * sin_elev + z_floor * cos_elev
+    quarter_labels.append({"x": px, "y": py, "label": quarter})
+df_quarter_labels = pd.DataFrame(quarter_labels)
+
+# Create the plot
+plot = (
+    ggplot()
+    # Floor grid
+    + geom_segment(
+        data=df_grid, mapping=aes(x="x", y="y", xend="xend", yend="yend"), color="#CCCCCC", size=0.5, alpha=0.5
+    )
+    # Axis lines
+    + geom_segment(
+        data=axes_df, mapping=aes(x="x", y="y", xend="xend", yend="yend"), color="#666666", size=1.5, alpha=0.8
+    )
+    # Bar faces (rendered back to front due to sorting)
+    + geom_polygon(
+        data=df_bars,
+        mapping=aes(x="x", y="y", group="face_id", fill="height"),
+        color="#333333",
+        size=0.6,
+        alpha=0.85,
+        tooltips=layer_tooltips().line("@product").line("@quarter").line("Revenue|$@{height}M"),
+    )
+    # Product labels (positioned to left of bars, left-aligned to avoid cutoff)
+    + geom_text(
+        data=df_product_labels,
+        mapping=aes(x="x", y="y", label="label"),
+        color="#1a1a1a",
+        size=11,
+        hjust=0,
+        fontface="bold",
+    )
+    # Quarter labels (further forward)
+    + geom_text(
+        data=df_quarter_labels, mapping=aes(x="x", y="y", label="label"), color="#1a1a1a", size=11, fontface="bold"
+    )
+    # Revenue axis label
+    + geom_text(data=z_label_df, mapping=aes(x="x", y="y", label="label"), color="#1a1a1a", size=12, fontface="bold")
+    + scale_fill_viridis(name="Revenue ($M)", option="viridis")
+    + labs(x="", y="", title="bar-3d \u00b7 letsplot \u00b7 pyplots.ai")
+    + theme_minimal()
+    + theme(
+        axis_title=element_blank(),
+        axis_text=element_blank(),
+        axis_ticks=element_blank(),
+        panel_grid=element_blank(),
+        plot_title=element_text(size=32, face="bold"),
+        legend_text=element_text(size=16),
+        legend_title=element_text(size=18),
+    )
+    + ggsize(1600, 900)
+)
+
+# Save PNG (scale=3 gives 4800x2700)
+ggsave(plot, "plot.png", path=".", scale=3)
+
+# Save HTML for interactivity (interactive rotation available via HTML export)
+ggsave(plot, "plot.html", path=".")
diff --git a/plots/bar-3d/metadata/letsplot.yaml b/plots/bar-3d/metadata/letsplot.yaml
@@ -0,0 +1,24 @@
+library: letsplot
+specification_id: bar-3d
+created: '2025-12-31T00:00:17Z'
+updated: '2025-12-31T00:21:37Z'
+generated_by: claude-opus-4-5-20251101
+workflow_run: 20608481317
+issue: 2857
+python_version: 3.13.11
+library_version: 4.8.2
+preview_url: https://storage.googleapis.com/pyplots-images/plots/bar-3d/letsplot/plot.png
+preview_thumb: https://storage.googleapis.com/pyplots-images/plots/bar-3d/letsplot/plot_thumb.png
+preview_html: https://storage.googleapis.com/pyplots-images/plots/bar-3d/letsplot/plot.html
+quality_score: 91
+review:
+  strengths:
+  - Excellent 3D isometric projection with clear depth perception using geom_polygon
+    for bar faces
+  - Good use of viridis color scale reinforcing z-values for colorblind accessibility
+  - Clean separation of product categories and quarters with readable labels
+  - Interactive tooltips provide rich data exploration when viewing HTML export
+  - Well-structured code with proper depth sorting for correct face rendering
+  weaknesses:
+  - Product labels on left side appear stacked vertically rather than aligned with
+    their corresponding bar rows in the 3D space
