feat(letsplot): implement errorbar-basic

[github-actions]

Implementation: errorbar-basic - python/letsplot

Implements the python/letsplot version of errorbar-basic.

File: plots/errorbar-basic/implementations/python/letsplot.py

Parent Issue: #973

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

Light render (plot-light.png): The plot shows five experimental groups (Control, Treatment A, Treatment B, Treatment C, Treatment D) on the x-axis against Measured Value in mg/dL on the y-axis, rendered on a warm off-white background (#FAF8F1). Four groups (Control, Treatment A, Treatment B, Treatment C) are drawn in brand green (#009E73), while Treatment D — the group with the largest error spread — is highlighted in lavender (#C475FD, Imprint position 2). Each data point has a visible center marker (size 6) and capped error bars. Horizontal-only grid lines are subtle and unobtrusive. Title "errorbar-basic · python · letsplot · anyplot.ai" appears in dark ink at the top left. All text (title, axis labels "Experimental Group" / "Measured Value (mg/dL)", and tick labels) is clearly readable against the light background. No dark-on-light legibility issues.

Dark render (plot-dark.png): Identical data layout on the warm near-black background (#1A1A17). Data colors are unchanged — green groups and lavender focal (Treatment D) are identical to the light render; only chrome flips. Title and all axis/tick labels render in light ink and are clearly legible against the dark background. No dark-on-dark failures observed; no black text against near-black background. Grid lines remain subtle but perceptible. Brand green (#009E73) is comfortably visible on the dark surface.

Both renders pass the theme-readability check. No edge clipping detected on either render.

Score: 87/100

Category	Score	Max
Visual Quality	29	30
Design Excellence	12	20
Spec Compliance	15	15
Data Quality	14	15
Code Quality	10	10
Library Mastery	7	10
Total	87	100

Visual Quality (29/30)

• VQ-01: Text Legibility (7/8) — All text readable at full size in both themes. Title (16pt), axis labels (12pt), tick labels (10pt) are proportional and legible. Minor: title is left-aligned at top edge with no breathing room above; not a legibility failure but slightly cramped.
• VQ-02: No Overlap (6/6) — Five well-spaced groups, no overlapping text or data elements.
• VQ-03: Element Visibility (6/6) — Error bars with visible caps (width=0.3), center points (size=6) prominent for 5 sparse data points. Good density-to-size match.
• VQ-04: Color Accessibility (2/2) — Green and lavender from Imprint palette are perceptually distinct; chart encodes group identity via x-axis labels as redundant encoding.
• VQ-05: Layout & Canvas (4/4) — Canvas gate passed (3200×1800). Good proportions, nothing cut off, adequate whitespace.
• VQ-06: Axis Labels & Title (2/2) — X: "Experimental Group", Y: "Measured Value (mg/dL)" with units. Title in canonical {spec-id} · python · letsplot · anyplot.ai format.
• VQ-07: Palette Compliance (2/2) — First (base) series is #009E73 ✓. Focal series uses #C475FD (Imprint position 2) ✓. Backgrounds #FAF8F1 (light) / #1A1A17 (dark) ✓. Data colors identical across themes ✓.

Design Excellence (12/20)

• DE-01: Aesthetic Sophistication (5/8) — Thoughtful focal-highlight pattern: Treatment D (largest error spread) is visually distinguished in lavender, creating intentional hierarchy. Legend omitted cleanly since x-axis labels are self-sufficient. Above default, but no additional refinements (data labels, annotations, reference lines).
• DE-02: Visual Refinement (4/6) — X-grid suppressed; Y-only horizontal grid at 15% opacity via 8-digit hex RGBA. Panel border removed. Axis line and ticks styled with INK_SOFT. Legend suppressed. Clean composition; does not reach publication-ready polish.
• DE-03: Data Storytelling (3/6) — The focal color on the highest-error group tells a partial story (variability as the narrative), but there is no title annotation, no data label on the focal point, and no textual callout to reinforce what the viewer should take away.

Spec Compliance (15/15)

• SC-01: Plot Type (5/5) — Correct error bar plot combining geom_errorbar + geom_point.
• SC-02: Required Features (4/4) — Visible end-caps (width=0.3) ✓. Consistent cap widths across all groups ✓. Color distinction for grouping ✓ (focal vs base).
• SC-03: Data Mapping (3/3) — X: categorical groups; Y: mean values; error bars via ymin/ymax derived from ±error.
• SC-04: Title & Legend (3/3) — Title matches required format exactly. No legend needed (groups labeled on x-axis; single-purpose focal color needs no legend in this context).

Data Quality (14/15)

• DQ-01: Feature Coverage (5/6) — Symmetric error bars with capped ends, center points, and color-distinguished focal group all demonstrated. Asymmetric error bars (mentioned as a "may be needed" consideration in the spec) are not shown, leaving one aspect of the full error-bar feature set uncovered.
• DQ-02: Realistic Context (5/5) — Clinical trial scenario (control + four treatment groups, mg/dL units) is realistic, domain-appropriate, and neutral.
• DQ-03: Appropriate Scale (4/4) — Y-axis range ~40–67 mg/dL, error magnitudes ±3.9–7.1, all plausible for clinical measurements.

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Flat script, no functions or classes.
• CQ-02: Reproducibility (2/2) — All data hardcoded; deterministic output, no random seed needed.
• CQ-03: Clean Imports (2/2) — Every import is used; no dead imports.
• CQ-04: Code Elegance (2/2) — Focal-highlight pattern (idxmax → boolean list → color_map dict) is clean and idiomatic. No fake UI or interactivity.
• CQ-05: Output & API (1/1) — Saves plot-{THEME}.png (scale=4 → 3200×1800) and plot-{THEME}.html. Current API throughout.

Library Mastery (7/10)

• LM-01: Idiomatic Usage (4/5) — Correct grammar-of-graphics layering (ggplot + geom + scale + theme), proper theme element types, ggsize for canvas control. Well-structured lets-plot code.
• LM-02: Distinctive Features (3/5) — HTML interactive export is a key lets-plot differentiator (automatic tooltips in the .html output). scale=4 ggsave for high-DPI raster export is also lets-plot-specific. Could have demonstrated additional letsplot-specific capabilities (e.g., tooltips configuration in the HTML output, geom_crossbar, or theme_grey).

Score Caps Applied

• None

Strengths

• Focal highlight design (Treatment D in Imprint lavender) adds intentional visual hierarchy and communicates which group has the most variability
• Full adaptive chrome: all theme tokens correctly applied for both light and dark renders with no dark-on-dark failures
• Clean, well-layered lets-plot code using geom_errorbar + geom_point with correct canvas sizing (3200×1800 via ggsize(800,450) + scale=4)
• Correct Imprint palette usage: #009E73 as first (base) series, #C475FD as focal; data colors are theme-invariant as required
• HTML export included — leverages lets-plot's interactive output capability
• Clinical trial data context is realistic and neutral with appropriate mg/dL units and sensible error magnitudes

Weaknesses

• Design storytelling is incomplete: the focal highlight on Treatment D draws the eye but there is no annotation, subtitle, or data label to explain WHY Treatment D is highlighted or what the viewer should conclude — the narrative stops halfway
• Asymmetric error bars (a feature type explicitly noted in the spec as relevant to skewed distributions / log-transformed data) are not demonstrated; only symmetric errors are shown
• No top/right spine removal via explicit theme override — theme_minimal() handles this, but it would be stronger to explicitly suppress panel_grid_minor in addition to the x-grid (already done) and add a comment or verify spine removal is complete

Issues Found

1. DE-03 MODERATE: Focal highlight creates visual attention but no narrative anchor to close the loop
   • Fix: Add a short annotation near Treatment D (e.g., geom_text labeling the error value) or a subtitle line in labs() explaining the focal point
2. DQ-01 MINOR: Asymmetric error bars not demonstrated despite being a key feature of the errorbar-basic spec
   • Fix: Add a secondary dataset or note with ymin/ymax that differ asymmetrically from the center value to illustrate non-symmetric errors

AI Feedback for Next Attempt

Strong implementation — palette, canvas, and theme adaptation are all correct. To improve Design Excellence: (1) add a brief annotation or subtitle that contextualizes the focal highlight on Treatment D (the storytelling gap); (2) consider demonstrating asymmetric error bars on one or two data points to fully cover the spec's feature set; (3) optionally configure tooltip content in the HTML export to leverage lets-plot's interactive strengths.

Verdict: APPROVED

[github-actions]

🔧 Repair Attempt 1/4

Applied fixes based on AI review feedback.

Status: Repair completed, re-triggering review...

---

🤖 impl-repair

[claude]

AI Review - Attempt 2/3

Image Description

Light render (plot-light.png): The plot renders on a warm off-white background consistent with #FAF8F1. The title "errorbar-basic · python · letsplot · anyplot.ai" appears in dark ink (#1A1A17) at 16pt, spanning approximately 60% of the canvas width. The subtitle "Treatment D (highlighted) has the widest error margin (±7.1 mg/dL), indicating higher measurement variability" is in medium gray (INK_SOFT) at 11pt — clearly readable. X-axis label "Experimental Group" and Y-axis label "Measured Value (mg/dL)" are in dark ink at 12pt; tick labels are in INK_SOFT at 10pt. Four groups (Control, Treatment A, Treatment B, Treatment C) are rendered in brand green #009E73 with T-capped error bars and prominent circular center points (size=6). Treatment D is rendered in Imprint palette position 2 lavender #C475FD, visually standing out as the focal group. Subtle horizontal Y-axis grid lines are present (no X-axis grid). L-shaped axis frame (left + bottom only). No legend. All text is clearly readable against the light background — no legibility issues.

Dark render (plot-dark.png): The plot renders on a warm near-black background consistent with #1A1A17. Title and subtitle switch to near-white (#F0EFE8 / #B8B7B0) and are fully legible against the dark surface. Axis labels, tick labels, and axis lines all correctly adopt light-theme tokens. Data colors are identical to the light render — brand green #009E73 for the four base groups and lavender #C475FD for Treatment D — both reading clearly on the dark background. Grid lines remain subtle. No dark-on-dark failures observed. All text is readable against the dark background.

Both paragraphs confirmed. Both renders pass theme-readability check.

Score: 89/100

Category	Score	Max
Visual Quality	30	30
Design Excellence	13	20
Spec Compliance	15	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	6	10
Total	89	100

Visual Quality (30/30)

• VQ-01: Text Legibility (8/8) — All font sizes explicitly set (title 16pt, axis 12pt, ticks 10pt, subtitle 11pt). Readable in both themes at full resolution.
• VQ-02: No Overlap (6/6) — No overlapping text or data elements. 5 well-spaced groups.
• VQ-03: Element Visibility (6/6) — geom_point(size=6) appropriately prominent for 5 sparse points. T-caps clearly visible on all groups.
• VQ-04: Color Accessibility (2/2) — Green + lavender are CVD-safe (distinct by both hue and lightness).
• VQ-05: Layout & Canvas (4/4) — Correct 3200×1800 canvas. Balanced margins, no overflow or clipping.
• VQ-06: Axis Labels & Title (2/2) — "Measured Value (mg/dL)" has units; "Experimental Group" is descriptive.
• VQ-07: Palette Compliance (2/2) — Base series #009E73 (Imprint pos 1), focal #C475FD (Imprint pos 2). Backgrounds #FAF8F1 light / #1A1A17 dark. Chrome theme-adaptive in both renders.

Design Excellence (13/20)

• DE-01: Aesthetic Sophistication (5/8) — Above well-configured default (4). Intentional focal highlighting using Imprint palette positions 1+2, informative subtitle replacing legend, clean professional appearance. Not yet publication-ready (8).
• DE-02: Visual Refinement (4/6) — Y-axis only grid with subtle opacity, panel border removed, axis lines styled with INK_SOFT, legend removed. Solid refinement above library defaults.
• DE-03: Data Storytelling (4/6) — Focal highlight (Treatment D in lavender) creates clear visual hierarchy. Subtitle explicitly communicates the key insight. Viewer is guided to the finding.

Spec Compliance (15/15)

• SC-01: Plot Type (5/5) — Correct error bar chart using geom_errorbar + geom_point layers.
• SC-02: Required Features (4/4) — Visible T-caps (width=0.3), consistent widths, symmetric errors, color distinguishes focal group.
• SC-03: Data Mapping (3/3) — Categorical X, numeric Y, ymin/ymax from symmetric error. 5 groups in spec range.
• SC-04: Title & Legend (3/3) — Title matches required format. Subtitle replaces legend effectively.

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — Center points, T-capped error bars, varying error magnitudes (3.9–7.1), varying means (45.2–61.3). All key error bar aspects shown.
• DQ-02: Realistic Context (5/5) — Clinical trial / treatment group comparison in mg/dL is a real, neutral, scientifically relevant scenario.
• DQ-03: Appropriate Scale (4/4) — mg/dL values 45–61 and error magnitudes 4–7 are plausible for biological/clinical measurements.

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Flat code: constants → data → theme → plot → save.
• CQ-02: Reproducibility (2/2) — Static deterministic data. No randomness.
• CQ-03: Clean Imports (2/2) — All imported symbols used.
• CQ-04: Code Elegance (2/2) — Clean lets-plot grammar. No over-engineering, no fake UI.
• CQ-05: Output & API (1/1) — Saves plot-{THEME}.png and plot-{THEME}.html.

Library Mastery (6/10)

• LM-01: Idiomatic Usage (4/5) — Comprehensive ggplot grammar: geom layering, scale_color_manual with named values, full theme customization with element_text/line/rect/blank, correct ggsize+ggsave with scale. Above default.
• LM-02: Distinctive Features (2/5) — HTML export shows lets-plot interactive capability. element_blank/line usage demonstrates library theme API. No distinctively lets-plot-exclusive features (e.g. geom_livemap, sampling) leveraged.

Score Caps Applied

• None

Strengths

• Focal highlighting using Imprint palette positions 1+2 (brand green base, lavender for widest-error group) creates intentional visual hierarchy with a clear focal point
• Subtitle explicitly communicates the key insight, removing the need for a legend while telling the data story
• Y-axis-only grid with subtle opacity, panel border removed, axis line styling — solid visual refinement above library defaults
• Theme-adaptive chrome fully correct in both renders; warm off-white / near-black backgrounds, all text readable in both themes
• Correct 3200×1800 canvas, all font sizes explicitly set, no overflow or clipping in either render
• Clean KISS code structure with both PNG and HTML outputs

Weaknesses

• LM-02: lets-plot's interactive HTML output is exported but no distinctively idiomatic lets-plot features (e.g. geom_livemap, sampling, animation) are leveraged — implementation is solid but generic across ggplot2-style libraries
• DE-01: Design is intentional and clean but does not yet reach publication-ready level — the focal highlight idea is effective but a reference baseline annotation or a subtle horizontal reference line at the control group mean would further anchor the story

Issues Found

1. LM-02 MODERATE: Generic ggplot2-style usage — lets-plot's HTML interactivity is exported but not enhanced
   • Fix (optional for future): use geom_text() with nudge for data labels, or leverage Kotlin-interop sampling for larger datasets to show a distinctively lets-plot pattern

AI Feedback for Next Attempt

Implementation is solid and approved. For any future refinement: consider adding a text annotation via geom_text() labeling each group's error value directly on the plot (avoids the need to read the subtitle), or use geom_hline() to show the control group baseline for comparison. This would push DE-01 and DE-03 closer to their maximums.

Verdict: APPROVED