Condensed Dendrogram¶

The condensed dendrogram summarizes clusters while preserving the original dendrogram order and branch heights. Labels are generated internally from Results.cluster_labels(...). plot_dendrogram_condensed(...) returns a rendered handle (CondensedDendrogramPlot) with explicit show() and save() methods.

Signature¶

plot_dendrogram_condensed(
    results: Results,
    *,
    rank_by: str = "p",
    label_mode: str = "top_term",
    figsize: Sequence[float] = (10, 10),
    sigbar_cmap: str | Colormap = "YlOrBr",
    sigbar_min_logp: float = 2.0,
    sigbar_max_logp: float = 10.0,
    sigbar_norm: Normalize | None = None,
    sigbar_width: float = 0.06,
    sigbar_height: float = 0.8,
    sigbar_alpha: float = 1.0,
    font: str = "Helvetica",
    fontsize: float = 9,
    max_words: int | None = None,
    wrap_text: bool = True,
    wrap_width: int | None = None,
    overflow: str = "wrap",
    omit_words: Sequence[str] | None = None,
    label_fields: Sequence[str] = ("label", "n", "p"),
    label_overrides: dict[int, str] | None = None,
    label_color: str = "black",
    label_alpha: float = 1.0,
    placeholder_text: str = "—",
    placeholder_color: str | None = None,
    placeholder_alpha: float | None = None,
    skip_unlabeled: bool = False,
    label_fontweight: str = "normal",
    dendrogram_color: str = "black",
    dendrogram_lw: float = 1.0,
    label_left_pad: float = 0.02,
    background_color: str | None = None,
) -> CondensedDendrogramPlot

Parameters¶

Parameter	Type	Default	Description
`results`	`Results`	required	Results exposing `cluster_layout()` and `clusters`.
`rank_by`	`str`	`"p"`	Ranking statistic for representative terms. Must be `"p"` or `"q"`.
`label_mode`	`str`	`"top_term"`	One of `"top_term"` or `"compressed"` for internal label generation.
`figsize`	`Sequence[float]`	`(10, 10)`	Figure size (width, height).
`sigbar_cmap`	`str / Colormap`	`"YlOrBr"`	Colormap for the significance bar.
`sigbar_min_logp`	`float`	`2.0`	Minimum `-log10(score)` for scaling.
`sigbar_max_logp`	`float`	`10.0`	Maximum `-log10(score)` for scaling.
`sigbar_norm`	`Normalize \| None`	`None`	Optional normalization; overrides min/max scaling.
`sigbar_width`	`float`	`0.06`	Significance bar width (axes fraction).
`sigbar_height`	`float`	`0.8`	Height of each significance bar relative to row pitch. Must be in `(0, 1]`.
`sigbar_alpha`	`float`	`1.0`	Significance bar alpha.
`font`	`str`	`"Helvetica"`	Label font family.
`fontsize`	`float`	`9`	Label font size.
`max_words`	`int \| None`	`None`	Word-based truncation limit for labels.
`wrap_text`	`bool`	`True`	Wrap long labels.
`wrap_width`	`int \| None`	`None`	Characters per line when wrapping.
`overflow`	`str`	`"wrap"`	Overflow behavior: `"wrap"` or `"ellipsis"`.
`omit_words`	`Sequence[str] \| None`	`None`	Words to omit from label text.
`label_fields`	`Sequence[str]`	`("label", "n", "p")`	Label fields to include; allowed values are `"label"`, `"n"`, `"p"`, `"q"`.
`label_overrides`	`dict[int, str] \| None`	`None`	Mapping `cluster_id -> label` for custom names.
`label_color`	`str`	`"black"`	Label text color.
`label_alpha`	`float`	`1.0`	Label text alpha for regular (non-placeholder) labels.
`placeholder_text`	`str`	`"—"`	Placeholder label for unlabeled clusters.
`placeholder_color`	`str \| None`	`None`	Placeholder text color. Falls back to `label_color`, then default.
`placeholder_alpha`	`float \| None`	`None`	Placeholder text alpha. Falls back to `label_alpha`, then default.
`skip_unlabeled`	`bool`	`False`	Skip clusters with no label.
`label_fontweight`	`str`	`"normal"`	Label font weight.
`dendrogram_color`	`str`	`"black"`	Dendrogram line color.
`dendrogram_lw`	`float`	`1.0`	Dendrogram line width.
`label_left_pad`	`float`	`0.02`	Left padding for labels (axes fraction).
`background_color`	`str \| None`	`None`	Figure and axes background color.

Use label_overrides to specify custom labels per cluster.

Behavior notes:

Labels are always generated from the attached Results object.
Unknown keyword arguments in plot_dendrogram_condensed(...) raise TypeError.
label_fields values outside { "label", "n", "p", "q" } raise ValueError.
Cluster significance bars are scaled from -log10(score), where score is selected by rank_by.
If skip_unlabeled=True, clusters without labels are omitted instead of receiving placeholder text.
Placeholder style resolves as placeholder_color -> label_color, and placeholder_alpha -> label_alpha.
Placeholder styling applies only to unlabeled/placeholder cluster labels.
Condensed plotting requires at least two clusters; single-cluster inputs raise a ValueError.

Return Handle¶

CondensedDendrogramPlot

Property / Method	Type	Description
`fig`	`matplotlib.figure.Figure`	Rendered figure handle.
`ax_den`	`matplotlib.axes.Axes`	Dendrogram axis.
`ax_sig`	`matplotlib.axes.Axes`	Significance bar axis.
`ax_txt`	`matplotlib.axes.Axes`	Label text axis.
`show()`	`() -> None`	Displays the rendered figure. If the backing figure was closed, it is rebuilt automatically from the stored render specification.
`save(path, **kwargs)`	`(str \| PathLike[str], **kwargs) -> None`	Saves the rendered figure with current facecolor. If the backing figure was closed, it is rebuilt automatically before saving.

Example¶

from himalayas.plot import plot_dendrogram_condensed

condensed = plot_dendrogram_condensed(
    results,
    rank_by="q",
    figsize=(6, 10),
    sigbar_min_logp=2.0,
    sigbar_max_logp=10.0,
    label_fields=("label", "n", "p", "q"),
    wrap_text=True,
    wrap_width=34,
)

condensed.show()

Example (Zoom Results)¶

After a zoom analysis, you can summarize the zoomed hierarchy with the same function:

condensed = plot_dendrogram_condensed(
    zoom_results,
    figsize=(4, 8),
    sigbar_min_logp=0.0,
    sigbar_max_logp=30.0,
    label_fields=("label", "n"),
    wrap_text=True,
    wrap_width=30,
)

condensed.save("zoom_condensed_dendrogram.png", dpi=300)

Notes¶

Labels are generated internally from the attached Results object.
label_fields must be a subset of { "label", "n", "p", "q" }.
Use Results.cluster_labels(...) only for inspection, export, or external workflows, not as required input.