feat(bokeh): implement rose-basic

[github-actions]

Implementation: rose-basic - python/bokeh

Implements the python/bokeh version of rose-basic.

File: plots/rose-basic/implementations/python/bokeh.py

Parent Issue: #1003

---

🤖 impl-generate workflow

[claude]

AI Review - Attempt 1/3

Image Description

Light render (plot-light.png): Warm off-white (#FAF8F1) background, correct. Title "Monthly Rainfall · rose-basic · bokeh · anyplot.ai" at top-center in dark ink — clearly readable. Twelve equal-angle wedges all in brand green (#009E73) with alpha ranging from ~0.35 (summer dry months) to 1.0 (Dec peak), creating a strong visual gradient. Dec and Jan at top are deep opaque; Jun and Jul at bottom are pale and short. Dashed concentric circles serve as radial gridlines. Radial scale labels ("90 mm", "68 mm", "45 mm", "23 mm") on the right at 15pt — slightly small but readable. All month labels at 20pt positioned around the perimeter without collision. All text is readable against the light background.

Dark render (plot-dark.png): Warm near-black (#1A1A17) background, correct. Title, month labels, and scale labels all switch to light ink (#F0EFE8 / #B8B7B0) — clearly readable, no dark-on-dark failures. Data colors are identical to the light render: same brand green, same alpha gradient. Radial gridlines are subtle but visible. The chart structure and layout match the light render exactly with only the chrome flipping. All text is readable against the dark background.

Both paragraphs are provided. Review is valid.

Score: 84/100

Category	Score	Max
Visual Quality	27	30
Design Excellence	13	20
Spec Compliance	13	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	6	10
Total	84	100

Visual Quality (27/30)

• VQ-01: Text Legibility (6/8) — Title 28pt and month labels 20pt explicitly set and readable; scale labels at 15pt are below the 18pt guideline for pixel-based libraries at 4800px
• VQ-02: No Overlap (6/6) — All 12 month labels and scale labels are well-separated with no collisions
• VQ-03: Element Visibility (6/6) — All wedges visible including smallest summer months; alpha floor of 0.35 keeps dry-season wedges distinguishable
• VQ-04: Color Accessibility (2/2) — Single brand green with luminance variation; CVD-safe by construction
• VQ-05: Layout & Canvas (3/4) — Chart fills ~50% of canvas; lower half is naturally sparse due to dry-season data, creating some empty space
• VQ-06: Axis Labels & Title (2/2) — Radial scale labels include units (mm); title is descriptive
• VQ-07: Palette Compliance (2/2) — Brand green #009E73 used throughout; backgrounds #FAF8F1/#1A1A17 correct; text tokens theme-adaptive in both renders

Design Excellence (13/20)

• DE-01: Aesthetic Sophistication (5/8) — Alpha gradient creates genuine depth; brand-only palette is intentional and coherent; above library defaults but not publication-ready
• DE-02: Visual Refinement (4/6) — Axes hidden (appropriate for rose chart), custom dashed gridlines, white dividers between wedges; good refinement with some polish opportunities remaining
• DE-03: Data Storytelling (4/6) — Seasonal wet/dry pattern immediately visible from shape + alpha; placing December at 12 o'clock foregrounds the rainfall peak effectively

Spec Compliance (13/15)

• SC-01: Plot Type (5/5) — Correct Nightingale/coxcomb rose chart using Bokeh wedge primitive
• SC-02: Required Features (3/4) — Equal-angle wedges, radius proportional to value, radial gridlines, month labels all present; start position places December at 12 o'clock rather than January (spec notes time data typically starts January at top)
• SC-03: Data Mapping (3/3) — Angular positions map to months, radius maps to rainfall value correctly
• SC-04: Title & Legend (2/3) — Attribution elements present but title starts with descriptive prefix; required format is rose-basic · bokeh · anyplot.ai

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — Shows full seasonal cycle with clear variation from peak (90 mm Dec) to trough (15 mm Jul)
• DQ-02: Realistic Context (5/5) — Monthly rainfall in mm: realistic, neutral, scientifically credible domain
• DQ-03: Appropriate Scale (4/4) — 15–90 mm monthly range plausible for a temperate climate; 6:1 ratio creates compelling shape

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Clean linear flow; no functions or classes
• CQ-02: Reproducibility (2/2) — All data is hardcoded; deterministic
• CQ-03: Clean Imports (2/2) — All imported names are used
• CQ-04: Code Elegance (2/2) — Pythonic; ColumnDataSource used for vectorized alpha
• CQ-05: Output & API (1/1) — Saves plot-{THEME}.png and plot-{THEME}.html; current Bokeh API

Library Mastery (6/10)

• LM-01: Idiomatic Usage (4/5) — ColumnDataSource with data-driven fill_alpha is idiomatic Bokeh; wedge primitive correctly used; figure model configuration is complete
• LM-02: Distinctive Features (2/5) — Data-driven alpha via ColumnDataSource is Bokeh-specific but basic; HTML output is not enhanced with HoverTool showing exact mm values on hover, which is the flagship Bokeh capability most distinguishable from static libraries

Score Caps Applied

• None — all cap conditions clear

Strengths

• Alpha-variation encoding immediately communicates the wet/dry seasonal cycle — viewer needs no legend to read the pattern
• Perfect theme adaptation in both renders: backgrounds, ink colors, and data colors all correct
• Full data quality: realistic meteorological context, wide value range (15–90 mm), 6:1 ratio creates compelling visual shape
• Perfect code quality: KISS structure, deterministic data, all imports used, saves both PNG and HTML

Weaknesses

• Scale labels at 15pt are below the 18pt minimum for pixel-based libraries at 4800px — bump to 18pt
• Title should be rose-basic · bokeh · anyplot.ai (or include a human-readable prefix separated by a different delimiter, e.g. Monthly Rainfall | rose-basic · bokeh · anyplot.ai)
• January (i=0) should be at 12 o'clock, not December — rotate the angle offset so start_angles[0] centers January at top (π/2)
• HTML output lacks HoverTool — add a tooltip showing month name and exact mm value; this is the most distinctive Bokeh feature vs. static libraries

Issues Found

1. SC-04 LOW: Title format deviates from {spec-id} · {library} · anyplot.ai
   • Fix: Change title to "rose-basic · bokeh · anyplot.ai" or add a subtitle; ensure spec-id comes first in the attribution chain
2. VQ-01 MINOR: Scale labels at 15pt below 18pt guideline
   • Fix: Change text_font_size="15pt" to text_font_size="18pt" for scale labels
3. SC-02 MINOR: December at 12 o'clock rather than January
   • Fix: Adjust start_angles offset so i=0 (Jan) sits centered at π/2 (top)
4. LM-02 LOW: HoverTool absent from HTML output
   • Fix: Add HoverTool(tooltips=[("Month", "@month"), ("Rainfall", "@rainfall mm")]) and include month/rainfall columns in the ColumnDataSource

AI Feedback for Next Attempt

Fix the four issues above. Priority order: (1) add HoverTool to the ColumnDataSource and figure — this is the single biggest LM score booster and requires adding "month" and "rainfall" columns to the ColumnDataSource dict; (2) fix the title format to start with the spec-id; (3) rotate the angle calculation so January is at 12 o'clock (change the base offset from np.pi/2 - angle_width/2 to np.pi/2 + angle_width/2 - angle_width or similar so i=0/Jan is centered at top); (4) bump scale label font size to 18pt. The alpha-gradient approach and data choice are strong — preserve them.

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 plot renders on a warm off-white #FAF8F1 background. A 12-wedge rose chart represents monthly rainfall using brand green (#009E73) with alpha varying from ~0.35 (summer, low rainfall) to ~1.0 (winter, high rainfall). Jan and Dec wedges are the tallest (north/upper), summer months (Jun–Aug) are the smallest. Concentric dashed gridlines at 25/50/75/100% radius are subtle and effective. Month labels (20pt) are positioned at each wedge's midpoint at radius 1.15. Rainfall scale labels ("23 mm", "45 mm", "68 mm", "90 mm") are displayed on the right at 15pt. Title reads "Monthly Rainfall · rose-basic · bokeh · anyplot.ai". All text is clearly readable against the light background.

Dark render (plot-dark.png): The same rose chart renders on #1A1A17 near-black background. Data colors are identical to the light render — the same brand green with same alpha gradation. Chrome flips appropriately: month labels, scale labels, and title all appear in light tones (INK = #F0EFE8). Gridlines use INK with 22% alpha, subtle on the dark surface. There are no dark-on-dark failures; all text is legible. The title "Monthly Rainfall · rose-basic · bokeh · anyplot.ai" reads clearly at the top center.

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

Score: 84/100

Category	Score	Max
Visual Quality	26	30
Design Excellence	13	20
Spec Compliance	14	15
Data Quality	15	15
Code Quality	10	10
Library Mastery	6	10
Total	84	100

Visual Quality (26/30)

• VQ-01: Text Legibility (6/8) — Title 28pt ✓, month labels 20pt ✓, but scale labels only 15pt (below the 18pt minimum for pixel-based libs)
• VQ-02: No Overlap (5/6) — Jan label is tight against the large Jan wedge in the light render; minor positional tightness near the top
• VQ-03: Element Visibility (6/6) — Wedges clearly visible across the full alpha range; gridlines subtle and appropriate
• VQ-04: Color Accessibility (2/2) — Single-hue approach with luminance variation; no red-green conflict; CVD-safe
• VQ-05: Layout & Canvas (3/4) — Chart is centered but summer months (May–Sep) produce very small petals, leaving the bottom third of the landscape canvas largely empty; scale labels float far right
• VQ-06: Axis Labels & Title (2/2) — Scale labels include "mm" units; title is descriptive; no traditional axes needed for this chart type
• VQ-07: Palette Compliance (2/2) — Primary series is #009E73 ✓; backgrounds are #FAF8F1 / #1A1A17 ✓; text tokens flip correctly in both themes ✓

Design Excellence (13/20)

• DE-01: Aesthetic Sophistication (5/8) — The alpha-based data encoding is elegant and intentional, lifting this clearly above defaults; single-color depth creates professional polish, though it falls short of publication-ready sophistication
• DE-02: Visual Refinement (4/6) — Dashed concentric gridlines at 22% alpha, radial dividers at 18% alpha, and hidden axes show deliberate refinement; PAGE_BG line_color between wedges creates clean separation
• DE-03: Data Storytelling (4/6) — The alpha gradient directly encodes rainfall magnitude, creating clear visual hierarchy where winter months dominate visually; the seasonal story is immediately legible

Spec Compliance (14/15)

• SC-01: Plot Type (5/5) — Correct rose/Nightingale chart; equal-angle wedges with radius ∝ value; starting from north ✓
• SC-02: Required Features (4/4) — Radial gridlines ✓, 12 monthly categories ✓, consistent color scheme ✓, top start position ✓
• SC-03: Data Mapping (3/3) — Angular positions = months, radius = rainfall values; all 12 months visible ✓
• SC-04: Title & Legend (2/3) — Title includes required "rose-basic · bokeh · anyplot.ai" components but uses non-standard format "Monthly Rainfall · rose-basic · bokeh · anyplot.ai" (extra prefix); no legend appropriate for single series

Data Quality (15/15)

• DQ-01: Feature Coverage (6/6) — Shows full range from peak winter (90mm Jan) to summer trough (15mm Jul); variation and cyclical pattern fully demonstrated
• DQ-02: Realistic Context (5/5) — Monthly rainfall in mm is a genuine, neutral, comprehensible scenario; Mediterranean-type climate pattern is plausible
• DQ-03: Appropriate Scale (4/4) — 15–90mm monthly range is realistic; seasonal gradient is factually consistent with temperate climates

Code Quality (10/10)

• CQ-01: KISS Structure (3/3) — Linear flow: imports → data → plot → save; no functions or classes
• CQ-02: Reproducibility (2/2) — Fully deterministic hardcoded data; no seed needed
• CQ-03: Clean Imports (2/2) — All imports used; os, sys, numpy, bokeh.* all needed
• CQ-04: Code Elegance (2/2) — Clean, Pythonic; sys.path manipulation is a legitimate fix for the bokeh.py shadowing issue; no fake UI
• CQ-05: Output & API (1/1) — Saves plot-{THEME}.png and plot-{THEME}.html; current Bokeh API

Library Mastery (6/10)

• LM-01: Idiomatic Usage (4/5) — Correct use of figure(), ColumnDataSource, p.wedge() glyph, export_png + output_file + save pattern; idiomatic Bokeh throughout
• LM-02: Distinctive Features (2/5) — Vectorized fill_alpha="alphas" via ColumnDataSource is Bokeh-characteristic (data-driven visual properties), but no HoverTool, callbacks, or other features that truly distinguish Bokeh from other libraries

Score Caps Applied

• None — DE-01=5 (>2) and DE-02=4 (>2) so the "correct but boring" cap does not apply; no other caps triggered

Strengths

• Alpha-based data encoding elegantly communicates rainfall intensity while preserving a single cohesive color story
• Excellent data quality: realistic monthly rainfall values with full seasonal variation, 15/15
• Perfect code quality: clean, deterministic, idiomatic structure
• Radial gridlines with dashed style at low alpha are aesthetically refined and functional

Weaknesses

• Scale labels at 15pt are below the 18pt minimum for pixel-based libraries — increase to 18pt
• Non-standard title format "Monthly Rainfall · {spec-id} · {library} · anyplot.ai" instead of the required {spec-id} · {library} · anyplot.ai
• Summer months produce very small petals, leaving bottom canvas underutilized; consider squarer canvas (3600×3600) which would better suit the symmetric polar layout and reduce wasted whitespace
• LM-02 limited: no Bokeh-distinctive interactive features (HoverTool would add genuine value here)

Issues Found

1. VQ-01 MINOR: Scale labels at 15pt (text_font_size="15pt") are below the 18pt minimum for pixel-based libraries
   • Fix: Change scale label text_font_size to "18pt"
2. SC-04 MINOR: Title has non-standard prefix "Monthly Rainfall · " before "rose-basic · bokeh · anyplot.ai"
   • Fix: Use title format rose-basic · bokeh · anyplot.ai exactly (move the topic into a subtitle or drop it)
3. VQ-05 MODERATE: Landscape canvas leaves bottom third empty due to summer petals being very small
   • Fix: Switch to 3600×3600 square canvas which is more appropriate for symmetric polar charts

AI Feedback for Next Attempt

If repair is needed: (1) Switch to 3600×3600 square canvas for better polar chart proportions. (2) Change title to exactly "rose-basic · bokeh · anyplot.ai". (3) Increase scale label font to 18pt. (4) Consider adding a HoverTool showing month and rainfall value to use Bokeh's interactive strength. The alpha encoding and data choice are strong — keep them.

Verdict: APPROVED