uqmodels.visualization package

Submodules

uqmodels.visualization.aux_visualization module

uqmodels.visualization.aux_visualization.aux_adjust_axes(ax, x, y_list, ylim=None, x_lim=None, margin=0.05, x_margin=0.5)

Adjust x/y axis limits based on data and optional explicit limits.

Parameters:

• ax (Axes)

• x (array-like)

• y_list (array-like or list of array-like) – One or several y-series to consider for limits.

• ylim (tuple or None) – If not None, force (ymin, ymax).

• x_lim (tuple or None) – If not None, force (xmin, xmax).

• margin (float) – Relative margin applied to inferred y-limits (ignored if ylim is set).

• x_margin (float) – Margin added around min/max of x (ignored if x_lim is set).

uqmodels.visualization.aux_visualization.aux_build_data_colors(data, list_anom_ind=None)

Construit la palette de couleurs par canal, en surlignant éventuellement certains indices anormaux.

uqmodels.visualization.aux_visualization.aux_compute_layout_params(n_score, dim_score, true_label, score2, data, grid_spec=None)

Calcule n_fig, sharey, grid_spec.

uqmodels.visualization.aux_visualization.aux_create_score_figure(n_fig, sharey, grid_spec, figsize)

Crée la figure et les axes pour la matrice d’anomalies.

uqmodels.visualization.aux_visualization.aux_fill_area(ax, x, env_bot, env_top, config=None)

Fill an envelope between two curves on an Axes.

Parameters:

• ax (Axes)

• x (array-like) – Coordinates and envelope bounds.

• env_bot (array-like) – Coordinates and envelope bounds.

• env_top (array-like) – Coordinates and envelope bounds.

• config (dict, optional) – Style overrides (color, alpha, label).

uqmodels.visualization.aux_visualization.aux_fill_between(ax, x, y1, y2, where=None, config=None)

Wrapper around ax.fill_between with configurable style.

Parameters:

• ax (Axes)

• x (array-like)

• y1 (array-like)

• y2 (array-like)

• where (array-like or None) – Boolean mask for conditional fill.

• config (dict, optional) – Style overrides (color, facecolor, alpha, label, interpolate, zorder).

uqmodels.visualization.aux_visualization.aux_finalize_figure(fig, show_plot=True)

Finalise la figure (tight_layout + show optionnel).

uqmodels.visualization.aux_visualization.aux_format_time_axis(ax, x_flag, x_date)

Configure l’axe des x comme temporel et éventuellement applique un formatter.

uqmodels.visualization.aux_visualization.aux_norm_score_inputs(score, f_obs=None, cmap=None)

Normalise score en liste, infère len_score, dim_score, n_score et f_obs.

uqmodels.visualization.aux_visualization.aux_overlay_score_anoms_on_data(ax, x, data, score, f_obs, dim, threshold=1.0)

Superpose les points où |score| > threshold sur les séries de données.

uqmodels.visualization.aux_visualization.aux_overlay_setup_grid(ax, setup, n_points)

Superpose la grille channels/sensors sur une matrice de score. setup = (n_chan, n_sensor)

uqmodels.visualization.aux_visualization.aux_overlay_true_label_on_data(ax, x, data, true_label, f_obs, color='purple')

Superpose les labels vrais sur les séries temporelles.

uqmodels.visualization.aux_visualization.aux_plot_PIs(ax, x, list_PIs, list_alpha_PIs, list_colors_PIs=['lightblue', 'lightgreen'], list_alpha_fig_PIs=[0.3, 0.15], list_label_PIs=None, config=None)

Plot multiple prediction interval envelopes on an Axes.

Parameters:

• ax (Axes)

• x (array-like)

• list_PIs (list of array-like) – [low_1, …, low_k, high_k, …, high_1].

• list_alpha_PIs (list of float) – Quantile levels for each bound.

• list_colors_PIs (list, optional) – Per-interval colors overriding config.

• list_alpha_fig_PIs (list, optional) – Per-interval alphas overriding config.

• list_label_PIs (list, optional) – Per-interval labels.

• config (dict, optional) – Global style config: {“line”: {…}, “fill”: {…}}.

uqmodels.visualization.aux_visualization.aux_plot_anom(ax, x, y, config=None)

Plot anomalous observations on an Axes using aux_plot_line.

Parameters:

• ax (Axes)

• x (array-like) – Coordinates and anomalous values.

• y (array-like) – Coordinates and anomalous values.

• config (dict, optional) – Style overrides for anomalous points.

uqmodels.visualization.aux_visualization.aux_plot_conf_score(ax, x, pred, confidence_lvl, label, mode_res=False, config=None)

Plot confidence scores as colored markers on an Axes.

Parameters:

• ax (Axes)

• x (array-like) – Coordinates and predictions.

• pred (array-like) – Coordinates and predictions.

• confidence_lvl (array-like) – Discrete confidence levels (int).

• label (list of str) – Legend labels per confidence level.

• mode_res (bool, default False) – If True, plot residual scores around zero.

• config (dict, optional) – Style overrides (marker, s, edgecolors, linewidth, cmap, zorder_base).

uqmodels.visualization.aux_visualization.aux_plot_confiance(ax, y, pred, var_A, var_E, born=None, born_bis=None, ylim=None, split_values=-1, x=None, mode_res=False, min_A=0.08, min_E=0.02, env=[0.95, 0.68], config=None, **kwarg)

Plot prediction, uncertainty intervals and anomaly regions on an Axes.

uqmodels.visualization.aux_visualization.aux_plot_data_timeseries(ax, x, data, f_obs, dim, colors, lw=0.9)

Trace les séries temporelles multicanal.

uqmodels.visualization.aux_visualization.aux_plot_line(ax, x, y, config=None)

Plot a line (or markers only) on an Axes with configurable style.

Parameters:

• ax (Axes)

• x (array-like)

• y (array-like)

• config (dict, optional) – Keys: color, linestyle, linewidth, marker, markersize, label, zorder.

uqmodels.visualization.aux_visualization.aux_plot_pred(ax, x, y, pred, config=None)

Plot observations and predictions on a given Axes with optional style config.

Parameters:

• ax (Axes)

• x (array-like) – Coordinates, observations, and predictions.

• y (array-like) – Coordinates, observations, and predictions.

• pred (array-like) – Coordinates, observations, and predictions.

• config (dict, optional) – Style overrides for truth line, prediction line, and observation scatter. Keys: “truth_line”, “pred_line”, “obs_scatter”.

uqmodels.visualization.aux_visualization.aux_plot_score_matrix(ax, score_mat, f_obs, extent, vmin, vmax, cmap, title='score')

Affiche une matrice de score via imshow.

uqmodels.visualization.aux_visualization.aux_plot_true_label_matrix(ax, true_label, f_obs, extent)

Affiche la matrice de labels vrais.

uqmodels.visualization.aux_visualization.aux_prepare_x_extent(x, f_obs, dim_score)

Prépare x (échelle) et la liste d’extent pour imshow. Retourne x, x_flag (datetime ou non), list_extent.

uqmodels.visualization.aux_visualization.provide_cmap(mode='bluetored')

Generate a bluetored or a cyantopurple cutsom cmap

Parameters:

mode (str, optional) – Values: bluetored’ or ‘cyantopurple ‘

Returns:

Colormap matplotlib

uqmodels.visualization.old_visualisation module

uqmodels.visualization.old_visualisation.plot_prediction_interval(y: array, y_pred_lower: array, y_pred_upper: array, X: array | None = None, y_pred: array | None = None, save_path: str | None = None, sort_X: bool = False, **kwargs) → None

Plot prediction intervals whose bounds are given by y_pred_lower and y_pred_upper. True values and point estimates are also plotted if given as argument.

Parameters:

• y – label true values.

• y_pred_lower – lower bounds of the prediction interval.

• y_pred_upper – upper bounds of the prediction interval.

• <optionnal> (y_pred) – abscisse vector.

• <optionnal> – predicted values.

• kwargs – plot parameters.

uqmodels.visualization.old_visualisation.plot_sorted_pi(y: array, y_pred_lower: array, y_pred_upper: array, X: array | None = None, y_pred: array | None = None, **kwargs) → None

Plot prediction intervals in an ordered fashion (lowest to largest width), showing the upper and lower bounds for each prediction. :param y: label true values. :param y_pred_lower: lower bounds of the prediction interval. :param y_pred_upper: upper bounds of the prediction interval. :param X <optionnal>: abscisse vector. :param y_pred <optionnal>: predicted values. :param kwargs: plot parameters.

uqmodels.visualization.old_visualisation.show_dUQ_refinement(UQ, y=None, d=0, f_obs=None, max_cut_A=0.99, q_Eratio=2, E_cut_in_var_nominal=False, A_res_in_var_atypic=False)

uqmodels.visualization.old_visualisation.visu_latent_space(grid_dim, embedding, f_obs, context_grid, context_grid_name=None)

uqmodels.visualization.visualization module

Visualization module.

uqmodels.visualization.visualization.aux_get_var_color_sets()

Return the color sets used for percentile envelope visualization in plot_var.

Returns:

• color_full (list of tuple) – Colors for filled regions between percentile curves.

• color_full2 (list of tuple) – Colors for percentile boundary lines.

uqmodels.visualization.visualization.plot_anom_matrice(score, score2=None, f_obs=None, true_label=None, data=None, x=None, vmin=-3, vmax=3, cmap=None, list_anom_ind=None, figsize=(15, 6), grid_spec=None, x_date=False, show_plot=True, setup=None)

Visualize anomaly score matrices and optional ground-truth labels or data.

This function plots one or several anomaly score matrices (e.g., per model or per transformation), an optional secondary anomaly score matrix, optional ground-truth anomaly labels, and optional multichannel time series data. It supports contextual segmentation, date-based x-axes, sensor/channel structural overlays, and anomaly highlighting. The function preserves its original API while delegating rendering to modular helpers.

Parameters:

• score (array-like or list of array-like) – Primary anomaly score matrix or list of matrices. Each matrix must be of shape (n_samples, n_features).

• score2 (array-like, optional) – Secondary anomaly score matrix of shape (n_samples, n_features).

• f_obs (array-like, optional) – Indices of samples to visualize; defaults to all.

• true_label (array-like, optional) – Ground-truth anomaly labels of shape (n_samples, n_features).

• data (array-like, optional) – Multichannel time series of shape (n_samples, n_features), used for overlaying raw data and score-based anomaly markers.

• x (array-like, optional) – X-axis values. If None, integer indices are used. If datetime-like, the function automatically switches to a date axis.

• vmin (float, default=(-3, 3)) – Color limits for the anomaly score colormap.

• vmax (float, default=(-3, 3)) – Color limits for the anomaly score colormap.

• cmap (Colormap, optional) – Colormap for score matrices. If None, a default diverging map is used.

• list_anom_ind (list of int, optional) – Indices of features/sensors to highlight in the time-series panel.

• figsize (tuple, default=(15, 6)) – Figure size in inches.

• grid_spec (array-like, optional) – Height ratios for subplot layout. If None, all subplots have equal height.

• x_date (bool, default=False) – If True, the x-axis is formatted as a date axis (dd/mm HH:MM).

• show_plot (bool, default=True) – Whether to display the resulting figure.

• setup (tuple, optional) – Tuple (n_channel_per_sensor, n_sensor) enabling structural overlays (horizontal grid lines) on score matrices for multi-sensor setups.

Notes

• The function supports:

  • multiple score matrices displayed in stacked subplots,

  • contextual slicing when x contains datetime values,

  • ground-truth anomaly maps,

  • multichannel data with anomaly highlighting,

  • optional highlighting of anomalous sensor indices.

• Rendering is internally modularized via helper functions to improve clarity and maintainability, while keeping the public API identical.

Returns:

The function creates the figure and optionally displays it.

Return type:

None

uqmodels.visualization.visualization.plot_pi(y, y_pred, y_pred_lower, y_pred_upper, mode_res=False, f_obs=None, X=None, size=(12, 2), name=None, show_plot=True, config=None, ylim=None, xlim=None, **kwargs)

Plot prediction intervals (PI) together with observations and predictions.

Displays observed values, predicted values, and their prediction interval (upper/lower bounds). Optionally plots residuals instead of absolute values (mode_res=True). Observations falling outside the PI are highlighted. Parameter f_obs selects which observation indices to display.

uqmodels.visualization.visualization.plot_var(Y, data_full, variance, impact_anom=None, anom=None, f_obs=None, dim=(400, 20, 3), g=0, res_flag=False, fig_s=(20, 3), title=None, ylim=None)

Plot empirical variance envelopes around a univariate time series.

This function builds a set of percentile-based envelopes from the provided variance and overlays them on the original (or residual) series together with anomaly markers. It visualizes how the variance translates into coverage levels for a given component of a multivariate signal.

Parameters:

• Y (array-like) – Ground-truth series of shape (n_samples, n_dim).

• data_full (array-like) – Reference series used to construct the envelopes, same shape as Y.

• variance (array-like) – Point-wise variance of shape (n_samples, n_dim) for the selected component.

• impact_anom (array-like, optional) – Anomaly impact indicator of shape (n_samples, n_dim). Non-zero entries are flagged as anomalies.

• anom (array-like, optional) – Unused placeholder kept for backward compatibility.

• f_obs (array-like, optional) – Indices of samples to visualize. If None, all samples are used.

• dim (tuple, default=(400, 20, 3)) – Unused placeholder describing (n_samples, n_time, n_groups). Kept for backward compatibility.

• g (int, default=0) – Index of the dimension (component) to plot.

• res_flag (bool, default=False) – If True, envelopes are computed around data_full - data_full (i.e. residuals), otherwise around data_full.

• fig_s (tuple, default=(20, 3)) – Figure size in inches.

• title (str, optional) – Figure title.

• ylim (tuple, optional) – Manual y-axis limits (ymin, ymax). If None, limits are inferred from the outer envelopes.

Returns:

• per (list of np.ndarray) – List of envelope curves (one array per percentile in per_list).

• per_list (list of float) – Percentile levels used to build the envelopes.

uqmodels.visualization.visualization.uncertainty_plot(y, output, context=None, size=(15, 5), f_obs=None, name='UQplot', mode_res=False, born=None, born_bis=None, dim=0, confidence_lvl=None, list_percent=[0.8, 0.9, 0.99, 0.999, 1], env=[0.95, 0.65], type_UQ='old', show_plot=True, with_colorbar=False, **kwarg)

Visualize uncertainty diagnostics for multivariate predictive models.

This function plots observations, predictions, prediction intervals, aleatoric/epistemic uncertainty contributions, confidence-level scores, optional anomaly bounds, and context-based segmentations. It supports multi-output signals, multiple contextual partitions, residual mode, and both full UQ views and data-only views. The function preserves the original API and integrates with modular visualization helpers (aux_*).

Parameters:

• y (array-like) – Ground-truth observations of shape (n_samples, n_dim).

• output (tuple or None) – UQ model outputs. Either (pred, var_A, var_E) or (pred, (var_A, var_E)), depending on type_UQ. Set to None in data-only mode.

• context (array-like, optional) – Context matrix used for splitting the plot by contextual dimension or highlighting contextual regions.

• size (tuple, default=(15, 5)) – Figure size in inches.

• f_obs (array-like, optional) – Indices of samples to display; defaults to all.

• name (str, default="UQplot") – Figure suptitle.

• mode_res (bool, default=False) – If True, plot residuals instead of raw values.

• born (tuple of array-like, optional) – Lower and upper anomaly bounds for each dimension.

• born_bis (tuple of array-like, optional) – Secondary set of anomaly bounds.

• dim (int or list of int, default=0) – Target output dimensions to visualize.

• confidence_lvl (array-like, optional) – Precomputed confidence-level matrix. If None, it is computed internally.

• list_percent (list of float, default=[0.8, 0.9, 0.99, 0.999, 1]) – Confidence thresholds used to compute epistemic confidence levels.

• env (list of float, default=[0.95, 0.65]) – Default uncertainty envelopes for plotting.

• type_UQ ({"old", "var_A&E"}, default="old") – Format specification of output.

• show_plot (bool, default=True) – Whether to display the final figure.

• with_colorbar (bool, default=False) – Whether to add a confidence-level colorbar.

• **kwarg – Additional parameters, including: - “ind_ctx”: context values to include, - “split_ctx”: context dimension used for splitting subplots, - “ylim”: vertical limits, - “var_min”: minimum (var_A, var_E) values, - “only_data”: disable UQ & plot observations only, - “x”: explicit x-axis values, - “ctx_attack”: tuple defining contextual highlight rules, - “list_name_subset”: labels for contextual annotations.

Notes

• This function acts as a high-level orchestrator and delegates rendering to modular visualization helpers (aux_plot_confiance, aux_plot_conf_score, aux_plot_line, aux_fill_between, etc.).

• The input API is preserved for backward compatibility.

Returns:

The function creates a figure and optionally displays it.

Return type:

None

uqmodels.visualization.visualization_mutisource module

Visualization_multisource module.

uqmodels.visualization.visualization_mutisource.apply_cmap(val, vmin, vmax, cmap)

Transform valyes array into color values array using cmap and considering bound [vmin,vmax]

Parameters:

• val (array) – Values to turn to color

• vmin (float) – min_val

• vmax (float) – max_val

• cmap (cmap) – matplotlib cmap

Returns:

Array of color

Return type:

_type_

uqmodels.visualization.visualization_mutisource.compute_dev_score(val, y, vmin, vmax)

Compute signed relative errors

Parameters:

• val (array) – Prediction

• y (array) – Target

• vmin (float) – negative minimum sensitivité

• vmax (float) – positive minimun sensitivity

Returns:

signed relative errors.

Return type:

r

uqmodels.visualization.visualization_mutisource.load_and_select(storing, keys, x_min, x_max)

uqmodels.visualization.visualization_mutisource.load_from_metadata(storing, str_keys_metadata)

uqmodels.visualization.visualization_mutisource.plot_analysis(storing_data, storing_res, sensors_mask, x_min, x_max, figsize=(20, 4), matplot=True)

uqmodels.visualization.visualization_mutisource.plot_anom_mat(storing_data, storing_res, sensors_mask, x_min, x_max, figsize=(20, 12), metadata=None)

uqmodels.visualization.visualization_mutisource.plot_channel(storing_data, storing_res, sensors_mask, x_min, x_max, mode=None, figsize=(20, 4), canal_nature=['Mean', 'Std', 'ExT'])

uqmodels.visualization.visualization_mutisource.plot_row_series(storing, sensors_mask, x_min, x_max, figsize=(12, 10), date_init=None, anom=None)

uqmodels.visualization.visualization_mutisource.plot_row_series_statistics(storing, sensors_mask, x_min, x_max, figsize=(12, 10), interpolate_data=False, date_init=None)

Module contents