@pikaa-ai/pikaa 0.3.1 → 0.3.5

package/skills/scientific-visualization/references/matplotlib_examples.md

@@ -1,336 +0,0 @@
-# Current Matplotlib, Seaborn, and Plotly Patterns
-
-Verified 2026-07-23 against Matplotlib 3.11.1, Seaborn 0.13.2, Plotly 6.9.0, Kaleido 1.3.0, Pillow 12.3.0, and pypdf 6.14.2. Source IDs resolve in `sources.md`.
-
-Run examples from the skill directory with pinned direct dependencies:
-
-```bash
-uv run --isolated --no-project --python 3.13 \
-  --with "matplotlib==3.11.1" \
-  --with "seaborn==0.13.2" \
-  --with "plotly==6.9.0" \
-  --with "kaleido==1.3.0" \
-  --with "pillow==12.3.0" \
-  --with "pypdf==6.14.2" \
-  python your_figure.py
-```
-
-These pins are a dated direct-dependency snapshot, not a lock of all transitive artifacts. Keep a project lock when exact environment replay is required.
-
-## Rendering and hardcopy backends
-
-Matplotlib separates interactive display backends from hardcopy renderers. The current built-ins include PDF (`pdf`), PS (`ps`, `eps`), SVG (`svg`), PGF (`pgf`, `pdf` through TeX), and optional Cairo (`png`, `ps`, `pdf`, `svg`); Agg is the common raster renderer [MPL-BACKENDS]. JPEG, TIFF, and WebP saving uses Pillow through the raster path.
-
-Available output depends on the active backend/build:
-
-```python
-supported = fig.canvas.get_supported_filetypes()
-print(supported)
-```
-
-`Figure.savefig(..., backend="cairo")` or `backend="pgf"` can select another renderer, but Matplotlib documents the default as normally sufficient [MPL-SAVE]. PGF requires a working TeX setup; Cairo requires pycairo or cairocffi. Inspect output because a vector container can still contain rasterized artists.
-
-## Scoped style and exact dimensions
-
-Prefer temporary style contexts to global state:
-
-```python
-import matplotlib.pyplot as plt
-
-from style_presets import style_context
-
-with style_context("default", palette_name="okabe_ito_on_white"):
-    fig, ax = plt.subplots(
-        figsize=(89 / 25.4, 60 / 25.4),
-        layout="constrained",
-    )
-    ax.plot([0, 1, 2], [1, 3, 2], marker="o", label="Observed")
-    ax.set(xlabel="Time (hours)", ylabel="Response (unit)")
-    ax.legend()
-```
-
-`layout="constrained"` handles labels, legends, nested layouts, and colorbars more flexibly than `tight_layout`; calling `tight_layout()` turns constrained layout off [MPL-LAYOUT].
-
-If exact page dimensions matter, do not export with `bbox_inches="tight"`; it recalculates the bounding box and changes the physical output size [MPL-SAVE].
-
-You can also use the bundled parseable style:
-
-```python
-from pathlib import Path
-import matplotlib.pyplot as plt
-
-skill_root = Path("skills/scientific-visualization")
-with plt.style.context(skill_root / "assets" / "publication.mplstyle"):
-    fig, ax = plt.subplots(layout="constrained")
-```
-
-## Preserve raw observations and define uncertainty
-
-```python
-import numpy as np
-import matplotlib.pyplot as plt
-
-rng = np.random.default_rng(20260723)
-groups = {
-    "Control": rng.normal(0.0, 1.0, 24),
-    "Treatment": rng.normal(0.7, 1.1, 24),
-}
-
-fig, ax = plt.subplots(figsize=(3.5, 2.8), layout="constrained")
-for position, (label, values) in enumerate(groups.items()):
-    jitter = rng.uniform(-0.08, 0.08, len(values))
-    ax.scatter(
-        position + jitter,
-        values,
-        alpha=0.65,
-        label=label,
-    )
-    mean = values.mean()
-    sem = values.std(ddof=1) / np.sqrt(len(values))
-    ax.errorbar(position, mean, yerr=sem, color="black", capsize=3)
-
-ax.set(
-    xticks=range(len(groups)),
-    xticklabels=list(groups),
-    ylabel="Response (unit)",
-)
-```
-
-Caption the error bars as mean ± one SEM and state `n=24` independent observations per group. If independence is false, use an analysis and interval that respects the design.
-
-## Missing data and no silent interpolation
-
-```python
-import numpy as np
-
-time = np.arange(8)
-signal = np.array([1.0, 1.4, np.nan, np.nan, 2.1, 2.0, 2.4, 2.7])
-
-fig, ax = plt.subplots(layout="constrained")
-ax.plot(time, signal, marker="o", label="Observed")  # gaps remain gaps
-ax.scatter([2, 3], [0.9, 0.9], marker="x", color="0.35", label="Missing")
-ax.set(xlabel="Time (days)", ylabel="Signal (unit)")
-ax.legend()
-```
-
-If a model estimates the missing interval, plot the model separately with its uncertainty and identify it as modeled, not observed.
-
-## Log axes and explicit nonpositive policy
-
-```python
-import numpy as np
-import matplotlib.pyplot as plt
-
-concentration = np.array([0.1, 1.0, 10.0, 100.0])
-response = np.array([0.4, 0.9, 2.1, 4.3])
-
-fig, ax = plt.subplots(layout="constrained")
-ax.plot(concentration, response, marker="o")
-ax.set_xscale("log", base=10)
-ax.set(xlabel="Concentration (µM; log10 axis)", ylabel="Response (unit)")
-```
-
-Do not silently omit zeros or negatives. State the measurement-domain rule or use another representation.
-
-## Centered heatmap and missing-value color
-
-```python
-import matplotlib as mpl
-import matplotlib.pyplot as plt
-import numpy as np
-
-values = np.array([
-    [-2.0, -0.5, 0.1],
-    [-1.1, np.nan, 1.8],
-    [-0.2, 0.7, 3.0],
-])
-norm = mpl.colors.TwoSlopeNorm(vmin=-2, vcenter=0, vmax=3)
-cmap = mpl.colormaps["RdBu_r"].with_extremes(bad="#777777")
-
-fig, ax = plt.subplots(layout="constrained")
-image = ax.imshow(values, norm=norm, cmap=cmap, interpolation="nearest")
-colorbar = fig.colorbar(image, ax=ax)
-colorbar.set_label("Change from baseline (unit)")
-ax.set(xlabel="Sample", ylabel="Feature")
-```
-
-`TwoSlopeNorm` gives each side of the center a different linear mapping. Use `CenteredNorm` when symmetric treatment around a center is appropriate, `LogNorm` for strictly positive orders of magnitude, and `BoundaryNorm` for declared classes [MPL-NORM].
-
-## Multi-panel layout
-
-```python
-import matplotlib.pyplot as plt
-
-fig = plt.figure(figsize=(7.0, 4.0), layout="constrained")
-subfigures = fig.subfigures(1, 2, width_ratios=[2, 1])
-left_axes = subfigures[0].subplots(2, 1, sharex=True)
-right_ax = subfigures[1].subplots()
-
-for label, ax in zip("ABC", [*left_axes, right_ax]):
-    ax.text(
-        -0.12,
-        1.05,
-        label,
-        transform=ax.transAxes,
-        fontweight="bold",
-        va="top",
-    )
-```
-
-`GridSpec`, subgrids, `subplot_mosaic`, and subfigures all work with constrained layout [MPL-LAYOUT] [MPL-GRIDSPEC].
-
-## Selective rasterization in vector output
-
-Dense point clouds can make PDF/SVG huge. Rasterize only the dense artist:
-
-```python
-fig, ax = plt.subplots(layout="constrained")
-ax.scatter(x, y, s=2, alpha=0.25, rasterized=True)
-ax.set(xlabel="Predictor (unit)", ylabel="Outcome (unit)")
-
-from figure_export import export_figure
-
-report = export_figure(
-    fig,
-    "outputs/figure1",
-    formats=["pdf", "png"],
-    dpi=600,  # controls PNG and rasterized artists embedded in PDF
-    provenance={
-        "raw_data": "data/observations.csv",
-        "transformations": ["rows filtered by predeclared QC flag"],
-        "uncertainty": "none displayed",
-        "missing_data": "retained as gaps",
-    },
-    write_manifest=True,
-)
-```
-
-The exporter:
-
-- refuses implicit overwrite;
-- preserves page dimensions by default;
-- passes DPI to vector backends for embedded raster artists;
-- writes TIFF with LZW compression;
-- can keep TrueType text editable in PDF/PS;
-- can write an explicit provenance manifest.
-
-It does not inspect data truth or certify submission compliance.
-
-## Raster image export and inspection
-
-```python
-report = export_figure(
-    fig,
-    "outputs/microscopy_panel",
-    formats=["tiff"],
-    dpi=300,
-    facecolor="white",
-    overwrite=False,
-)
-```
-
-Then inspect:
-
-```bash
-uv run --isolated --no-project --python 3.13 \
-  --with "pillow==12.3.0" \
-  python scripts/image_metadata.py outputs/microscopy_panel.tiff \
-  --format tiff --mode RGB --min-dpi 300 --target-width-mm 85 \
-  --alpha-policy forbid
-```
-
-Effective DPI is pixel width divided by final width in inches. Changing only the TIFF DPI tag does not create detail.
-
-## Seaborn 0.13.2
-
-Seaborn remains built on Matplotlib. Use axes-level functions for custom multi-panel layouts and figure-level functions for automatic faceting [SEABORN-FAQ].
-
-```python
-import seaborn as sns
-import matplotlib.pyplot as plt
-
-from color_palettes import OKABE_ITO_ON_WHITE
-from style_presets import style_context
-
-sns.set_theme(style="ticks", context="paper", palette=OKABE_ITO_ON_WHITE)
-with style_context("default", palette_name="okabe_ito_on_white"):
-    fig, ax = plt.subplots(figsize=(3.5, 2.8), layout="constrained")
-    sns.lineplot(
-        data=frame,
-        x="time",
-        y="response",
-        hue="treatment",
-        style="treatment",
-        markers=True,
-        errorbar=("ci", 95),
-        n_boot=5000,
-        seed=20260723,
-        ax=ax,
-    )
-    ax.set(xlabel="Time (hours)", ylabel="Response (unit)")
-```
-
-Current `errorbar` choices include `"sd"`, `"se"`, `"pi"`, `"ci"`, tuples, callables, or `None`. The old `ci=` interface is not the current general API [SEABORN-ERROR].
-
-For categorical axes whose numeric/datetime values must retain their real spacing, use supported functions with `native_scale=True`. Do not assume every categorical plot uses native coordinates by default.
-
-## Plotly 6.9 and Kaleido 1.3
-
-Interactive HTML:
-
-```python
-fig.write_html(
-    "outputs/exploration.html",
-    include_plotlyjs=True,  # self-contained, larger file
-    full_html=True,
-)
-```
-
-Static image:
-
-```python
-fig.write_image(
-    "outputs/figure.svg",
-    width=700,
-    height=450,
-    scale=1,
-)
-```
-
-Batch export is faster with Kaleido v1:
-
-```python
-import plotly.io as pio
-
-pio.write_images(
-    fig=[figure_a, figure_b],
-    file=["outputs/a.pdf", "outputs/b.pdf"],
-)
-```
-
-Current facts [PLOTLY-STATIC] [KALEIDO]:
-
-- Kaleido v1 requires a compatible Chrome/Chromium installation; Chrome is no longer bundled.
-- Plotly `write_image` supports PNG, JPEG, WebP, SVG, and PDF.
-- EPS was supported only by Kaleido versions earlier than 1.0.
-- `engine=` and Orca are deprecated; do not use them in new code.
-- `plotly.io.kaleido.scope` is deprecated; use `plotly.io.defaults`.
-- Width/height are logical pixels and `scale` multiplies output pixels; `scale=3` is **not inherently “300 DPI.”**
-- WebGL traces embed raster content inside vector exports.
-- Fully offline MathJax/topojson use requires local resources; do not assume a network-independent export when a figure references external assets.
-
-An interactive HTML file does not replace a static fallback, caption, alt text, keyboard review, or accessible data table.
-
-## Font and transparency checks
-
-Matplotlib 3.11.1 defaults PDF/PS to Type 3 and SVG text to paths. The bundled presets instead use PDF/PS Type 42 and leave SVG text as text [MPL-STYLE]. Verify the actual PDF:
-
-```bash
-uv run --isolated --no-project --python 3.13 \
-  --with "pypdf==6.14.2" \
-  python scripts/image_metadata.py outputs/figure1.pdf
-```
-
-SVG text is not an embedded font; its appearance depends on the renderer’s installed fonts. If portability matters more than editable/searchable text, use paths and retain an editable source separately.
-
-Use opaque white submission output unless transparency is explicitly required. Transparent artists blend with the destination and can change apparent contrast [MPL-SAVE].

package/skills/scientific-visualization/references/publication_guidelines.md

@@ -1,196 +0,0 @@
-# Publication Figure Principles
-
-Reviewed 2026-07-23. These are general scientific-communication principles, not publisher requirements. Date-sensitive rules belong in `journal_requirements.md`. Source IDs resolve in `sources.md`.
-
-## Preserve evidence before styling
-
-Keep three layers separate:
-
-1. **Raw/source data**: immutable originals, acquisition metadata, exclusions, and missing-value codes.
-2. **Transformation record**: executable code or a machine-readable log of filtering, normalization, aggregation, statistical estimation, image processing, and random seeds.
-3. **Presentation output**: the figure and an export manifest recording dimensions, format, package versions, and source references.
-
-Do not overwrite raw images or tabular data. Keep native-resolution images. Upsampling changes pixel count, not information; PLOS, Science, Nature, Cell Press, Elsevier, and IEEE explicitly warn against treating it as improved quality [PLOS-FIG] [SCIENCE-REVISED] [NATURE-FINAL] [CELL-FIG] [ELSEVIER-SIZE] [IEEE-SIZE].
-
-For experimental/observational images:
-
-- Apply necessary brightness, contrast, and color adjustments consistently to the whole image unless a disclosed scientific method requires otherwise.
-- Never selectively erase, obscure, clone, or enhance features.
-- Preserve background and nonspecific signal.
-- Mark and explain splices, omitted lanes, composites, or stitched fields.
-- Retain originals and the exact processing steps. Some publishers request originals during review or production [CELL-FIG] [PLOS-FIG].
-
-The bundled exporter can write a provenance manifest, but it cannot confirm that supplied provenance is complete.
-
-## Avoid visual deception
-
-### Baselines and context
-
-- **Bars and filled areas encode length/area from a baseline**: normally show the zero baseline. If a nonzero reference is scientifically meaningful, make that reference explicit and avoid implying absolute magnitude.
-- **Points and lines encode position**: a nonzero axis limit can be valid, but show enough context, disclose breaks, and avoid choosing limits solely to magnify a small effect.
-- Use common limits for panels intended for direct comparison. If limits differ, make the difference unmistakable.
-- Do not extend axes far beyond observed data merely to suppress visible variation; Science explicitly advises that scales not extend beyond plotted data [SCIENCE-INITIAL].
-
-### Uncertainty and raw observations
-
-- Do not add error bars mechanically. Show uncertainty when an estimate is displayed, and show spread when the distribution is the question.
-- Name the interval precisely: SD, SE, percentile interval, parametric CI, bootstrap CI, posterior interval, or another definition.
-- State sample size, unit of replication, estimator, interval level, and dependence/repeated-measure handling.
-- Use deterministic seeds for bootstrap displays. Seaborn 0.13.2 supports `errorbar="sd"`, `"se"`, `"pi"`, `"ci"`, tuples such as `("ci", 95)`, or a callable; bootstrap results vary unless `seed` is set [SEABORN-ERROR].
-- Show raw observations when feasible. Do not jitter points so far that their category or value becomes ambiguous.
-- Significance stars are not uncertainty. Add them only for a reported analysis, identify the test and multiplicity handling, and provide exact values where practical.
-
-### Missing, excluded, and censored data
-
-- Keep missing values distinct from zero, below-detection-limit values, and excluded observations.
-- Do not silently connect across missing time points. Use a gap, explicit interpolation style, or a model curve whose status is stated.
-- Give missing values a dedicated legend entry or neutral `bad` colormap color.
-- Record exclusions and their rationale outside the plotting code as well as in the caption/methods.
-
-### Area, volume, and 3D encodings
-
-- Prefer position on a common scale.
-- If area represents magnitude, scale **area**, not radius. If volume represents magnitude, scale volume, not diameter.
-- Avoid perspective 3D bars, pies, and surfaces for simple comparisons; occlusion and perspective distort values.
-- If a true 3D scientific structure is necessary, add orthogonal views, scale/orientation cues, and accessible alternatives.
-
-### Logarithms and other transforms
-
-- Label the transformed scale and base. Say whether values, axes, or model outputs were transformed.
-- A logarithmic axis requires a declared policy for zero and negative values; never silently discard them.
-- Interpret equal distances as ratios, not additive differences.
-- For signed data around zero, consider `SymLogNorm`/a symmetric log axis with a disclosed linear region; for unequal ranges around a meaningful center, consider `TwoSlopeNorm` [MPL-NORM].
-- Power-law or arbitrary transforms need strong justification and conspicuous disclosure. Matplotlib itself notes that viewers are less familiar with power normalization [MPL-NORM].
-
-### Binning, smoothing, and aggregation
-
-- Record bin edges, inclusion convention, bandwidth/window, smoothing method, and whether choices were made before seeing the result.
-- Show sensitivity to reasonable bin or bandwidth choices when conclusions depend on them.
-- Do not use interpolation or smoothing to imply observations between measured points.
-- Preserve and, where practical, expose underlying observations.
-
-### Normalization and color limits
-
-- State the formula and reference: per-capita, percent of baseline, z-score axis, library-size factor, min-max range, or other transformation.
-- Fit normalization parameters on the appropriate data partition; avoid information leakage.
-- Use the same normalization and color limits across directly compared panels unless the difference is explicit.
-- A diverging map needs a scientifically meaningful center. `Normalize`, `LogNorm`, `CenteredNorm`, `SymLogNorm`, `TwoSlopeNorm`, and `BoundaryNorm` encode different assumptions [MPL-NORM].
-- Always label colorbars with units and transformed scale.
-
-### Dual axes
-
-Prefer aligned panels or normalized/common-unit displays. Dual y-axes can make unrelated series appear correlated because each range can be tuned independently. If unavoidable:
-
-- justify the shared x-domain and distinct units;
-- label each axis and series directly;
-- avoid matching colors as the only association cue;
-- choose limits independently of the desired visual relationship;
-- provide the underlying data.
-
-### Image contrast and channels
-
-- Inspect histograms and clipped-pixel counts before and after adjustment.
-- Apply comparable processing to images being compared.
-- State channel assignment, lookup table, projection, denoising, deconvolution, thresholding, and contrast limits.
-- Use scale bars based on calibration, not magnification text.
-- Do not rely on red/green channel identity alone; use accessible channel combinations, outlines, labels, or separate grayscale panels.
-
-## Encoding and color
-
-Match the palette to data:
-
-- **Qualitative** for unordered categories; do not imply order.
-- **Sequential** for ordered magnitude.
-- **Diverging** only when a meaningful midpoint exists.
-- **Cyclic** for periodic variables such as direction or phase.
-
-Use hue consistently across a manuscript. Avoid rainbow maps for ordered data unless there is a documented scientific reason and the map has been evaluated for perceptual artifacts. Paul Tol explains why ordinary rainbow schemes create false transitions and fail for some color-vision conditions [TOL].
-
-Color is not a sufficient encoding:
-
-- combine it with marker shape, line style, hatching, direct labels, or panel separation;
-- audit contrast against the actual background;
-- inspect grayscale, but do not treat grayscale conversion as a complete color-vision simulation;
-- keep legends ordered like the data or direct-label series.
-
-See `color_palettes.md` and run `scripts/palette_audit.py`.
-
-## Accessibility
-
-WCAG 2.2 is a web-content standard, not a journal-print specification. It provides useful targets for figures delivered on the web [WCAG22]:
-
-- SC 1.4.1 (Level A): color is not the only visual means of conveying information.
-- SC 1.4.3 (Level AA): normal text has at least 4.5:1 contrast; large text has at least 3:1, with stated exceptions.
-- SC 1.4.11 (Level AA): graphical objects required to understand content have at least 3:1 contrast against adjacent colors, with an essential-presentation exception.
-- SC 1.1.1 (Level A): non-text content has an equivalent text alternative.
-- SC 1.4.5 (Level AA): use actual text rather than images of text when the technology can provide it, subject to exceptions.
-
-For web/interactive figures also provide:
-
-- a concise alt text naming chart type, variables, main pattern, and important exception;
-- a long description or nearby narrative for complex figures;
-- the underlying data in an accessible table/download;
-- keyboard-operable interactions, visible focus, and non-hover access to values;
-- a static fallback that preserves the scientific message.
-
-Passing a palette ratio audit does not prove WCAG conformance; applicability depends on rendered context and alternatives.
-
-## Layout, typography, and annotation
-
-- Design at final physical size. Judge labels, symbols, and line weights at that size.
-- Use one legible font family and a restrained size hierarchy.
-- Include units in axis/colorbar labels. Define abbreviations.
-- Keep panel labels consistent and outside dense data regions.
-- Use layout engines intentionally: `layout="constrained"` handles nested grids and colorbars; calling `tight_layout()` disables constrained layout [MPL-LAYOUT].
-- Check all labels and legends after export. `bbox_inches="tight"` can alter physical page dimensions, so do not use it when exact page width is required [MPL-SAVE].
-- Keep decorative ink subordinate to data, uncertainty, and annotations. Gridlines can help value lookup when light and sparse; removing them is not a universal rule.
-
-## Static, vector, raster, and interactive output
-
-### Vector
-
-PDF/SVG/EPS are useful for text and line art, but a vector container may include rasterized artists. DPI still controls those raster elements [MPL-SAVE]. Dense scatter plots can be selectively rasterized to control file size while preserving vector text/axes.
-
-For Matplotlib:
-
-```python
-import matplotlib as mpl
-
-mpl.rcParams["pdf.fonttype"] = 42
-mpl.rcParams["ps.fonttype"] = 42
-mpl.rcParams["svg.fonttype"] = "none"
-```
-
-PDF/PS Type 42 embeds TrueType fonts. `svg.fonttype="none"` leaves text as text and therefore depends on font availability; `svg.fonttype="path"` trades editability/searchability for appearance portability [MPL-STYLE]. Inspect the delivered file rather than assuming font behavior.
-
-### Raster
-
-Required pixel width is:
-
-```text
-pixels = final width (inches) × target pixels per inch
-```
-
-Embedded DPI metadata alone does not add detail. TIFF/PNG are lossless choices; JPEG can be accepted by some publishers for photographs but is a poor choice for line art or text because it is lossy. Preserve original bit depth and color profile when scientifically important.
-
-Transparency can reveal an unintended background or change apparent contrast. Matplotlib's `transparent=True` makes axes patches transparent and, unless explicitly overridden, the figure patch too [MPL-SAVE]. Prefer an explicit opaque background for submission unless the destination requires transparency.
-
-### Plotly
-
-- `write_html()` preserves interaction and is self-contained by default; that embeds Plotly.js and creates a large file. `include_plotlyjs` and `full_html=False` change portability and embedding behavior [PLOTLY-HTML].
-- Static `write_image()` uses Kaleido and supports PNG, JPEG, WebP, SVG, and PDF. Width/height are logical pixels; `scale` changes physical output pixel count, not a journal DPI declaration [PLOTLY-STATIC].
-- WebGL traces are partly rasterized inside SVG/PDF output [PLOTLY-STATIC].
-- Export a static, captioned fallback and accessible data alongside interactive output.
-
-## Final scientific review
-
-- [ ] Raw data and native images are preserved.
-- [ ] Transformations, exclusions, normalization, bins, and random seeds are recorded.
-- [ ] Missing/censored values are explicit.
-- [ ] Baselines, limits, log scales, and breaks are scientifically justified.
-- [ ] Uncertainty and sample size are defined.
-- [ ] Area/volume and color normalization encode magnitude correctly.
-- [ ] Color is redundant; foreground/background contrast was reviewed.
-- [ ] Alt text/long description and underlying data are available for web delivery.
-- [ ] Physical size, raster pixels, fonts, transparency, and file format were inspected after export.
-- [ ] The current target-journal instructions were checked for the correct submission phase.

package/skills/scientific-visualization/references/sources.md

@@ -1,76 +0,0 @@
-# Sources and Version Snapshot
-
-Research refreshed 2026-07-23 with `parallel-cli search` and `parallel-cli extract`. API and publisher requirements below use current official project, standards-body, publisher, or journal sources only. “Accessed” is 2026-07-23 unless another date is stated.
-
-## Tested direct package snapshot
-
-- **Matplotlib 3.11.1**, released 2026-07-18; Python >=3.11 [MPL-PYPI].
-- **Seaborn 0.13.2**, released 2024-01-25; Python >=3.8 [SEABORN-PYPI].
-- **Plotly 6.9.0**, released 2026-07-09; Python >=3.8 [PLOTLY-PYPI].
-- **Kaleido 1.3.0**, released 2026-05-04 [KALEIDO-PYPI].
-- **Pillow 12.3.0**, released 2026-07-01; Python >=3.10 [PIL-PYPI].
-- **pypdf 6.14.2**, released 2026-06-23; Python >=3.9 [PYPDF-PYPI].
-
-These are pinned direct-dependency snapshots used for smoke tests, not a transitive lock.
-
-## Matplotlib
-
-- **[MPL-PYPI]** [matplotlib on PyPI](https://pypi.org/project/matplotlib/) — current package version and release history; page dated 2026-07-18.
-- **[MPL-RELEASE]** [Matplotlib release notes](https://matplotlib.org/stable/release/release_notes.html) — 3.11 release/API changes.
-- **[MPL-SAVE]** [`matplotlib.figure.Figure.savefig`](https://matplotlib.org/stable/api/_as_gen/matplotlib.figure.Figure.savefig.html) — 3.11.1 signature; format inference, DPI, metadata, bounding boxes, transparency, backends, Pillow kwargs; built 2026-07-18.
-- **[MPL-BACKENDS]** [Backends](https://matplotlib.org/stable/users/explain/figure/backends.html) — interactive versus static renderers; PDF/PS/SVG/PGF/Cairo formats.
-- **[MPL-STYLE]** [Customizing Matplotlib with style sheets and rcParams](https://matplotlib.org/stable/users/explain/customizing.html) — `rc_context`, style composition, save settings, PDF/PS/SVG font types; built 2026-07-18.
-- **[MPL-LAYOUT]** [Constrained layout guide](https://matplotlib.org/stable/users/explain/axes/constrainedlayout_guide.html) — `layout="constrained"`, colorbars, subfigures, GridSpec, interaction with `tight_layout`.
-- **[MPL-GRIDSPEC]** [`matplotlib.gridspec`](https://matplotlib.org/stable/api/gridspec_api.html) — current grid layout API.
-- **[MPL-NORM]** [Colormap normalization](https://matplotlib.org/stable/users/explain/colors/colormapnorms.html) — `Normalize`, `LogNorm`, `CenteredNorm`, `SymLogNorm`, `PowerNorm`, `BoundaryNorm`, `TwoSlopeNorm`; built 2026-07-18.
-- **[MPL-CMAP]** [Choosing colormaps](https://matplotlib.org/stable/users/explain/colors/colormaps.html) — data classes and perceived lightness.
-
-## Seaborn
-
-- **[SEABORN-PYPI]** [seaborn on PyPI](https://pypi.org/project/seaborn/) — 0.13.2 package metadata and release history.
-- **[SEABORN-ERROR]** [Statistical estimation and error bars](https://seaborn.pydata.org/tutorial/error_bars.html) — current `errorbar` methods, callable intervals, bootstrapping, `seed`, and `n_boot`.
-- **[SEABORN-FAQ]** [Frequently asked questions](https://seaborn.pydata.org/faq.html) — axes-level versus figure-level functions, Matplotlib object-oriented integration, DPI/SVG notes.
-- **[SEABORN-PALETTE]** [Choosing color palettes](https://seaborn.pydata.org/tutorial/color_palettes.html) — qualitative, sequential, and diverging palette APIs.
-- **[SEABORN-THEME]** [`seaborn.set_theme`](https://seaborn.pydata.org/generated/seaborn.set_theme.html) — style, context, palette, font, scale, and rc parameters.
-
-## Plotly and Kaleido
-
-- **[PLOTLY-PYPI]** [plotly on PyPI](https://pypi.org/project/plotly/) — 6.9.0 package metadata; released 2026-07-09.
-- **[PLOTLY-STATIC]** [Static image export in Python](https://plotly.com/python/static-image-export/) — Kaleido/Chrome setup, formats, `write_image`, `write_images`, dimensions/scale, WebGL rasterization, offline assets, defaults, EPS/Orca/engine deprecations; page dated 2026.
-- **[PLOTLY-HTML]** [Interactive HTML export](https://plotly.com/python/interactive-html-export/) — `write_html`, `to_html`, `include_plotlyjs`, `full_html`; page dated 2026.
-- **[PLOTLY-CHANGES]** [Static image generation changes in Plotly.py 6.1](https://plotly.com/python/static-image-generation-changes/) — Kaleido v1 migration and deprecations.
-- **[KALEIDO]** [Plotly Kaleido repository](https://github.com/plotly/Kaleido) — Chrome requirement, v1 migration, direct APIs, and offline/page behavior.
-- **[KALEIDO-PYPI]** [kaleido on PyPI](https://pypi.org/project/kaleido/) — 1.3.0 package metadata; released 2026-05-04.
-
-## Accessibility and color
-
-- **[WCAG22]** [Web Content Accessibility Guidelines (WCAG) 2.2](https://www.w3.org/TR/WCAG22/) — W3C Recommendation; normative SC 1.1.1, 1.4.1, 1.4.3, 1.4.5, and 1.4.11.
-- **[WCAG-NONTEXT]** [Understanding SC 1.4.11: Non-text Contrast](https://www.w3.org/WAI/WCAG22/Understanding/non-text-contrast.html) — informative chart/graph examples and testing principles; not itself normative.
-- **[WCAG-COLOR]** [Understanding SC 1.4.1: Use of Color](https://www.w3.org/WAI/WCAG22/Understanding/use-of-color.html) — informative non-color cue guidance.
-- **[COLORBREWER]** [ColorBrewer 2.0](https://colorbrewer2.org/) — Cynthia Brewer, Mark Harrower, and Penn State; scheme type, data-class count, colorblind/print/photocopy filters, and exports.
-- **[TOL-HOME]** [Paul Tol’s Notes](https://sronpersonalpages.nl/~pault/) — canonical site; page states the move from SRON on 2026-07-07.
-- **[TOL]** [Paul Tol, “Colour Schemes”](https://sronpersonalpages.nl/~pault/data/colourschemes.pdf) — SRON/EPS/TN/09-002, issue 3.2, 2021-08-18; exact sRGB palettes, intended uses, color-vision checks, and grayscale analysis.
-- **[WONG]** [Bang Wong, “Color blindness”](https://www.nature.com/articles/nmeth.1618) — Nature Methods 8, 441 (2011); source commonly used for the eight-color palette.
-
-## Publishers and journals
-
-All rules were accessed 2026-07-23. Pages without a displayed update date are labeled by access date rather than assigning an invented publication date.
-
-- **[NATURE-FINAL]** [`Nature` final submission](https://www.nature.com/nature/for-authors/final-submission) — flagship final files, dimensions, fonts, formats, raster resolution, RGB/CMYK, Extended Data distinctions.
-- **[NATURE-FIG]** [`Nature` research figure specifications](https://research-figure-guide.nature.com/figures/preparing-figures-our-specifications) — graphs, accessibility, RGB, 300/450 dpi discussion, editable Type 42 text, export.
-- **[SCIENCE-INITIAL]** [`Science` initial manuscript instructions](https://www.science.org/content/page/instructions-preparing-initial-manuscript) — initial figure embedding, 300 dpi, widths, fonts, color/contrast, source data.
-- **[SCIENCE-REVISED]** [`Science` revised manuscript instructions](https://www.science.org/content/page/instructions-preparing-revised-manuscript) — separate files, formats, minimum resolution, dimensions, no upsampling.
-- **[CELL-FIG]** [Cell Press figure guidelines](https://www.cell.com/information-for-authors/figure-guidelines) — initial versus final stages, formats, widths, file size, DPI, RGB, fonts, image integrity, AI-assisted image policy.
-- **[PLOS-FIG]** [PLOS Computational Biology figures](https://journals.plos.org/ploscompbiol/s/figures) — provisional-accept waiver, TIFF/EPS, dimensions, 300-600 dpi, RGB/grayscale, file size, image integrity, 2026-04-01 blot/gel requirement.
-- **[ELSEVIER-FORMAT]** [Elsevier artwork formats checklist](https://www.elsevier.com/about/policies-and-standards/author/artwork-and-media-instructions/artwork-formats-checklist) — general formats, RGB preference, separate files, journal override.
-- **[ELSEVIER-SIZE]** [Elsevier artwork sizing](https://www.elsevier.com/about/policies-and-standards/author/artwork-and-media-instructions/artwork-sizing) — general widths, 300/500/1,000 dpi, typography, and explicit journal variability.
-- **[IEEE-SIZE]** [IEEE Resolution and Size](https://journals.ieeeauthorcenter.ieee.org/create-your-ieee-journal-article/create-graphics-for-your-article/resolution-and-size/) — modified 2025-02-25; PS/EPS/PDF, >300/>600 dpi, 88.9/182 mm.
-- **[BMC-BIOINFO]** [BMC Bioinformatics: preparing your manuscript](https://bmcbioinformatics.biomedcentral.com/submission-guidelines/preparing-your-manuscript) — journal-specific formats, 85/170 mm, approximately 300 dpi, 10 MB, embedded fonts.
-- **[ACS-GRAPHICS]** [ACS Preparing Manuscript Graphics](https://pubs.acs.org/page/4authors/submission/graphics_prep.html) — general dimensions and typography; no page update date displayed.
-
-## Optional inspection backends
-
-- **[PIL-PYPI]** [Pillow on PyPI](https://pypi.org/project/Pillow/) — 12.3.0 package metadata; released 2026-07-01.
-- **[PYPDF-PYPI]** [pypdf on PyPI](https://pypi.org/project/pypdf/) — 6.14.2 package metadata; released 2026-06-23.
-
-No Parallel JSON research artifacts are stored in this skill.