feat(plotnine): implement radar-basic

[github-actions]

Implementation: radar-basic - python/plotnine

Implements the python/plotnine version of radar-basic.

File: plots/radar-basic/implementations/python/plotnine.py

Parent Issue: #744

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

Light render (plot-light.png): Radar chart on a warm off-white #FAF8F1 background with 6 axes (Technical, Communication, Leadership, Creativity, Teamwork, Problem Solving). Two filled polygon series: Alice in #009E73 (bluish green) and Bob in #D55E00 (orange), each with alpha=0.25 transparency and 1.5-width line borders. Circular dashed gridlines at 20/40/60/80/100 intervals with spokes from center. Category labels at outer edge. Legend on the right titled "Employee" showing both series. Title "radar-basic · plotnine · anyplot.ai" is clearly visible in dark text at 24pt. All text readable against the light background.

Dark render (plot-dark.png): Same layout on a near-black #1A1A17 background. Data colors are identical to the light render — Alice still #009E73, Bob still #D55E00. Title and category labels are rendered in light text (#F0EFE8). Legend box uses elevated dark fill. No dark-on-dark failures — all chrome elements are correctly theme-adapted. The circular gridlines appear in a darker gray. Legibility verdict: PASS for both renders.

Both paragraphs are required. A review that only describes one render is invalid.

Score: 83/100

Category	Score	Max
Visual Quality	28	30
Design Excellence	10	20
Spec Compliance	15	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	5	10
Total	83	100

Visual Quality (28/30)

• VQ-01: Text Legibility (7/8) — Title 24pt, legend 16pt, category labels (geom_text) 16pt all explicitly set; category labels are slightly below the ≥20pt axis-label guideline
• VQ-02: No Overlap (6/6) — Category labels well-positioned at radius 122, no collisions
• VQ-03: Element Visibility (6/6) — Polygon fills, line borders, and vertex points all clearly visible
• VQ-04: Color Accessibility (2/2) — Okabe-Ito green/orange combo is CVD-safe and has adequate contrast
• VQ-05: Layout & Canvas (3/4) — Square 12×12 canvas (3600×3600 at 300 DPI) used well; minor empty space at bottom-left
• VQ-06: Axis Labels & Title (2/2) — Title in correct format; axis/tick text correctly blanked for radar type
• VQ-07: Palette Compliance (2/2) — Alice=#009E73 (position 1), Bob=#D55E00 (position 2); backgrounds #FAF8F1/#1A1A17; both renders theme-correct

Design Excellence (10/20)

• DE-01: Aesthetic Sophistication (4/8) — Correct colors and clean layout, but the dashed circular gridlines create a slightly busy visual texture at render resolution; typography is functional but not refined (no weight variation, plain title)
• DE-02: Visual Refinement (4/6) — Good refinement: all standard spines/ticks/axis grid removed, replaced with custom circular grid; dashed linetype on circular geom_line creates an optical stripe artifact — solid low-opacity lines would be cleaner
• DE-03: Data Storytelling (2/6) — Data is displayed but viewer must discover on their own that Alice excels in Creativity/Technical while Bob leads in Communication/Teamwork; no visual hierarchy, emphasis, or focal point

Spec Compliance (15/15)

• SC-01: Plot Type (5/5) — Correct radar/spider chart with all required structural elements
• SC-02: Required Features (4/4) — Filled polygons (alpha=0.25) ✓, gridlines at 20/40/60/80/100 ✓, outer edge labels ✓, polygon closure ✓, multi-series legend ✓
• SC-03: Data Mapping (3/3) — Values correctly mapped to axes, both series fully visible
• SC-04: Title & Legend (3/3) — "radar-basic · plotnine · anyplot.ai" ✓; legend titled "Employee" with correct Alice/Bob labels

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — 6 axes (within 4-8 range), 2 series (within 1-3 range), all polygon and grid spec elements present
• DQ-02: Realistic Context (5/5) — Employee performance evaluation is real-world, neutral, business scenario
• DQ-03: Appropriate Scale (4/4) — Values 60-90 range on a 0-100 scale; realistic competency scores; series have meaningfully different profiles

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Linear: imports → tokens → data → grid/spokes/labels → plot → save; no functions or classes
• CQ-02: Reproducibility (2/2) — All data is hardcoded; fully deterministic
• CQ-03: Clean Imports (2/2) — All imported symbols are used
• CQ-04: Code Elegance (2/2) — Clean coordinate computation, sys.path guard is necessary and documented
• CQ-05: Output & API (1/1) — Saves as plot-{THEME}.png ✓

Library Mastery (5/10)

• LM-01: Idiomatic Usage (3/5) — Correct ggplot grammar with layered geoms and scale_*_manual; but implementing radar via manual Cartesian math means the library's grammar is used more as a canvas than idiomatically
• LM-02: Distinctive Features (2/5) — Layer composition (geom_polygon + geom_line + geom_point + geom_text in one grammar expression) is distinctively plotnine, but the radar implementation doesn't showcase a standout plotnine-specific capability

Score Caps Applied

• None — DE-01=4 exceeds the ≤2 threshold; no other caps triggered

Strengths

• Perfect spec compliance: all radar features present (filled polygons with alpha, concentric gridlines, closed polygon, spokes, outer labels, legend)
• Correct Okabe-Ito palette with fully theme-adaptive chrome in both renders
• Clean KISS code with deterministic data and proper sys.path guard
• Realistic, neutral employee performance context with meaningfully differentiated series

Weaknesses

• Dashed linetype on circular geom_line creates an optical stripe/hatch artifact at render resolution; replace with solid lines at low opacity (e.g., alpha=0.2)
• No visual storytelling: the complementary skill profiles (Alice: Creativity/Technical, Bob: Communication/Teamwork) are not highlighted — consider adding a subtitle or using slightly different fill intensities to create a focal point
• Category label size at 16pt (geom_text) is below the ≥20pt axis-label guideline — increase to 18-20

Issues Found

1. DE-02 MEDIUM: Dashed circular gridlines create visual noise at high DPI
   • Fix: Change linetype from "dashed" to "solid" and lower opacity (use color with alpha or a lighter GRID_COLOR value)
2. DE-03 LOW: No visual hierarchy or emphasis
   • Fix: Add a descriptive subtitle (e.g., "Skill profile comparison: Alice vs. Bob"), or use slightly bolder polygon borders for emphasis
3. VQ-01 MINOR: Category labels at size=16 are slightly below ≥20pt label guideline
   • Fix: Increase geom_text size to 18 or 20
4. LM-02 LOW: Generic layer composition without leveraging distinctive plotnine features
   • Fix: Consider using annotate or theme features that are unique to plotnine vs. generic matplotlib usage

AI Feedback for Next Attempt

Replace dashed gridlines with solid lines at 10-15% opacity to eliminate the stripe artifact. Increase category label size to 18-20. Add a descriptive subtitle to create data storytelling. These three changes — cleaner grid, larger labels, and a brief narrative subtitle — will push DE-02 from 4→5, VQ-01 from 7→8, and DE-03 from 2→4, gaining ~6 points to reach 89+.

Verdict: REJECTED

[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 radar chart is rendered on a warm off-white background (#FAF8F1). The hexagonal polygon for Alice (brand green #009E73) and Bob (vermillion #D55E00) are filled with 25% alpha transparency and clearly overlapping, allowing both polygons to be seen simultaneously. Six category labels — Creativity, Leadership, Communication, Technical, Problem Solving, Teamwork — are positioned radially at the outer edge of the chart in dark ink text. Dashed circular gridlines at radii 20/40/60/80/100 are subtle but visible in light gray. A legend box in the upper-right shows "Employee" title with Alice and Bob labels. Title "radar-basic · plotnine · anyplot.ai" and a subtitle ("Alice leads in Technical & Creativity; Bob excels in Communication & Teamwork") appear at top-left in dark ink. All text is clearly readable against the light background. Legibility verdict: PASS.

Dark render (plot-dark.png): The same chart is rendered on a warm near-black background (#1A1A17). The two polygons use the identical data colors — green for Alice, orange-brown for Bob — confirming Okabe-Ito data colors are theme-invariant. All category labels, the title, subtitle, and legend text are rendered in light cream/off-white tones adapting correctly to the dark surface. The circular gridlines are very faint (dark gray on near-black: #3A3A35 on #1A1A17) but still discernible. No dark-on-dark failures detected — all text reads clearly against the dark background. The legend box has a slightly elevated dark fill (#242420). Data colors are identical to the light render. Legibility verdict: PASS.

Both paragraphs are required. A review that only describes one render is invalid.

Score: 86/100

Category	Score	Max
Visual Quality	26	30
Design Excellence	13	20
Spec Compliance	15	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	7	10
Total	86	100

Visual Quality (26/30)

• VQ-01: Text Legibility (7/8) — Title 24pt, legend 16pt, category labels size=16 all explicitly set; all readable in both themes. Minor: if subtitle is present, its size uses a default.
• VQ-02: No Overlap (5/6) — Category labels positioned radially with good spacing; no collisions observed. Minor imprecision at edges.
• VQ-03: Element Visibility (5/6) — Polygons, outlines, and data points clearly visible. Dark-mode gridlines (#3A3A35 on #1A1A17) are very faint — functional but barely visible.
• VQ-04: Color Accessibility (2/2) — Okabe-Ito positions 1 and 2; CVD-safe, distinguishable without hue alone.
• VQ-05: Layout & Canvas (3/4) — Square 12×12 at 300 DPI = 3600×3600 ✓; circular chart fills canvas well with balanced margins. Slight whitespace in corners is inherent to circular charts.
• VQ-06: Axis Labels & Title (2/2) — Title format correct; category labels serve as axis labels for radar; no redundant x/y axes needed.
• VQ-07: Palette Compliance (2/2) — First series (Alice) = #009E73 ✓; second series (Bob) = #D55E00 ✓; light background #FAF8F1 ✓; dark background #1A1A17 ✓; theme chrome adapts correctly in both renders.

Design Excellence (13/20)

• DE-01: Aesthetic Sophistication (5/8) — Custom Okabe-Ito palette, theme-adaptive chrome, transparent polygon fills for overlap visibility. Clearly above library defaults, professionally executed.
• DE-02: Visual Refinement (4/6) — All default ggplot axes/grid removed (element_blank). Custom circular gridlines and spoke lines drawn. Generous whitespace. Clean radar-specific design.
• DE-03: Data Storytelling (4/6) — Subtitle explicitly states the comparative insight ("Alice leads in Technical & Creativity; Bob excels in Communication & Teamwork"). The contrasting polygon shapes visually reinforce this story.

Spec Compliance (15/15)

• SC-01: Plot Type (5/5) — Correct radar/spider/web chart implemented.
• SC-02: Required Features (4/4) — Filled polygons with alpha ✓; gridlines at 20/40/60/80/100 ✓; axis labels at outer edge ✓; distinct colors with legend ✓; polygon closed by repeating first point ✓.
• SC-03: Data Mapping (3/3) — Values mapped to radial distances correctly; all 6 axes shown; both series visible.
• SC-04: Title & Legend (3/3) — Title: "radar-basic · plotnine · anyplot.ai" ✓; legend labeled "Employee" with Alice/Bob ✓.

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — Two contrasting employee profiles across 6 competency dimensions; different strengths in each series demonstrates the chart's comparison capability fully.
• DQ-02: Realistic Context (5/5) — Employee performance review scenario (Technical, Communication, Leadership, Creativity, Teamwork, Problem Solving) is real-world, neutral, and immediately recognizable.
• DQ-03: Appropriate Scale (4/4) — 0–100 scale as recommended in spec; values (60–90 range) are plausible competency scores with meaningful variation.

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Linear: imports → data → plot → save. No functions or classes.
• CQ-02: Reproducibility (2/2) — Fully deterministic data; no random elements.
• CQ-03: Clean Imports (2/2) — All imported symbols used; sys.path manipulation prevents shadowing.
• CQ-04: Code Elegance (2/2) — Appropriate complexity for manual radar coordinate construction; clean loop-based data assembly.
• CQ-05: Output & API (1/1) — Saves as plot-{THEME}.png ✓.

Library Mastery (7/10)

• LM-01: Idiomatic Usage (4/5) — Correct ggplot grammar with scale_color_manual, scale_fill_manual, geom_polygon; idiomatic layer composition. Manual Cartesian conversion is necessary because plotnine lacks native polar projection.
• LM-02: Distinctive Features (3/5) — Creative use of plotnine's layered grammar (geom_polygon + geom_line + geom_point + geom_text) to construct a chart type not natively supported by the library. Somewhat distinctive; not trivially portable to a non-grammar library.

Score Caps Applied

• None — no caps apply (DE-01=5 > 2, DE-02=4 > 2, no zero scores in VQ/SC/DQ).

Strengths

• Full spec compliance: all required radar features implemented (filled polygons, gridlines, axis labels, legend, closed polygons)
• Perfect data and code quality: realistic employee performance scenario, clean deterministic code
• Correct theme-adaptive chrome: backgrounds, ink colors, legend styling all flip correctly between light and dark
• Correct Okabe-Ito usage: #009E73 first, #D55E00 second, consistent across both renders
• Subtitle adds immediate data storytelling value

Weaknesses

• Dark-mode circular gridlines (#3A3A35 on #1A1A17) are barely visible — consider increasing contrast (e.g., rgba(240,239,232,0.15) for dark gridlines)
• DE-01 could reach 6+ with subtle typography improvements: slightly larger title, bold weight, or font family differentiation
• LM-02 could score higher with a more distinctive plotnine-specific technique (e.g., using facet_wrap for small multiples comparison or coord_polar if applicable)

Issues Found

1. DE-01 MODERATE: Good but not exceptional aesthetic sophistication — below FiveThirtyEight-level polish
   • Fix: Consider adding bold title weight, slightly refined legend styling, or more intentional whitespace tuning
2. VQ-03 MINOR: Dark-mode gridlines barely visible at #3A3A35 on #1A1A17
   • Fix: Use a slightly brighter gridline color for dark mode, e.g., rgba(240,239,232,0.12) equivalent hex

AI Feedback for Next Attempt

The implementation is strong overall and meets all spec requirements cleanly. To reach 90+: (1) boost gridline visibility in dark mode by using a slightly lighter GRID_COLOR (e.g., #4A4A45 → #5A5A55), (2) enhance DE-01 with explicit bold title weight via element_text(weight='bold') and potentially a cleaner font, (3) consider if any plotnine-distinctive feature (theme customization, stat functions, or faceting pattern) could be applied to differentiate from a generic implementation.

Verdict: APPROVED