dope-plot 0.2.0

Icon sets, an emoji default, and sharper scatter callouts.

⚠️ Breaking: config key change

[paths] birds_dir is gone. Point icons_set at an icon-set directory instead:

[paths]
icons_set = "icons/emoji"   # or "icons/hunt", or your own <dir>/<bird>.png set
output = "charts"

Highlights

Icon sets

• Icons live at icons/<set>/<bird>.png — a set is one artist's collection, and its directory name carries the attribution (icons/hunt credits TICK, icons/emoji credits Noto Emoji, Apache-2.0).
• Both sets ship in the wheel (installed next to the package), so installed users can switch sets with one config line.
• emoji is the new default — the chart gallery in the README shows it.

Scatter callouts, refined

• Labels sit centred on their true point; a label is nudged (minimally, one card per overlap) only when another card would actually cover its text — measured from real rendered extents. Cards may softly overlap.
• Small nudges are left unmarked; only a far-displaced card gets a dot on its true point + dashed leader line.
• Per-card shadow stacking: an overlapping card now casts its shadow over the card beneath it.

Hardening

• Service layer validates radar/comparison profiles and unparseable CSV — clean ValueErrors everywhere (matters for MCP tools fed by an AI client).
• CI runs the MCP tests, builds the wheel, asserts its contents, and smoke-tests a clean install.
• Example gallery regenerated (emoji default) and PNG-optimized; removed 20 orphaned stale charts.

Install (from this release)

pip install https://github.com/soil-dev/dope-plot/releases/download/v0.2.0/dope_plot-0.2.0-py3-none-any.whl
# with the MCP server extra:
pip install 'dope-plot[mcp] @ https://github.com/soil-dev/dope-plot/releases/download/v0.2.0/dope_plot-0.2.0-py3-none-any.whl'

dope-plot 0.1.1

First release under the dope-plot name (previously bird-plot) — radar, quadrant and comparison charts for TICK four-bird personality profiles (Dove, Eagle, Owl, Peacock).

Highlights

Use it from AI assistants (MCP)

• New dope-plot-mcp server (stdio): scatter_chart, radar_chart and comparison_chart tools return the rendered PNG inline — works with Claude Desktop/Code, n8n's MCP node, and any MCP-capable client.
• Ships as an optional extra; a friendly hint tells you to install it if missing.

Readable scatter, honest positions

• Gradient label cards blend each person's primary/secondary bird colours in proportion to their scores.
• Every label sits centred on its true point; when text would be covered, only one card is nudged the minimum distance (cards may softly overlap). A card pushed far off its point gets a dot at the true position plus a dashed leader line.
• Per-card shadow stacking, collision detection on real rendered text extents.

Installable anywhere

• Bundled default config + bird assets resolve via importlib.resources, so the CLI works outside a source checkout — verified by a clean-install smoke test (also wired into CI).
• Input validation across CLI, service layer and MCP: invalid data raises clean ValueErrors instead of deep matplotlib errors.
• PEP 621/639 metadata, Python 3.12/3.13, Apache-2.0.

Install (from this release)

pip install https://github.com/soil-dev/dope-plot/releases/download/v0.1.1/dope_plot-0.1.1-py3-none-any.whl
# with the MCP server extra:
pip install 'dope-plot[mcp] @ https://github.com/soil-dev/dope-plot/releases/download/v0.1.1/dope_plot-0.1.1-py3-none-any.whl'

Then:

dope-plot --data data.csv --graph radar scatter   # charts into ./charts
dope-plot-mcp                                     # MCP server over stdio

v2026.02.04

Release Notes

Functional changes

• Safe coordinate scaling: guard against divide‑by‑zero when all scores are zero, preventing NaN plot coordinates. bird_plot/cli.py
• Sanitized names for output paths to avoid unsafe filesystem characters. bird_plot/cli.py
• Comparison filenames standardized to with_TeamAverage.png for team-average overlays. bird_plot/cli.py
• Radar title formatting now skips empty/NaN notes instead of always appending. bird_plot/plots/radar.py
• Quadrant color keys now map to the correct corners while preserving the existing visual colors. bird_plot/plots/base.py, config.toml
• Team average labeling cleaned up (Name now “Team Average Profile”, Note empty). bird_plot/cli.py

v2025.02.24

I am pleased to introduce the first release of Bird-Plot – a Python-based visualization tool that converts Tick.com.au personality profiling data into engaging, quadrant-based radar charts.

Key Features in this Release:

• Individual Radar Charts:
  Visualize a single person’s scores across the four "bird" dimensions (Eagle, Dove, Owl, and Peacock) using a radar chart.

• Group Distribution Visualization:
  Compare multiple profiles on a quadrant layout to quickly assess team strengths and personality diversity.

• Pairwise Comparison:
  Overlay radar charts to highlight similarities and differences between two profiles, including an overlap percentage.

This initial release lays the groundwork for Bird-Plot. Feedback and suggestions are welcome.