Dev oct 2026

src/dimet/config/analysis/method/pca_plot.yaml

@@ -20,3 +20,6 @@ impute_values: # how to impute zeroes
   mean_enrichment: "min"
   isotopologues: "min"
   isotopologue_proportions: "min"
+
+width: !!float 6.2
+height: !!float 4.0

src/dimet/processing/differential_analysis.py

@@ -40,9 +40,13 @@ def compute_span_incomparison(df: pd.DataFrame, groups: List) -> pd.DataFrame:
          the computed differences.
     """
     for i in df.index.values:
-        all_values = list(df.loc[i, groups[0]]) + list(df.loc[i, groups[1]])
+        group1_values = df.loc[i, groups[0]].to_numpy()
+        group2_values = df.loc[i, groups[1]].to_numpy()
+
+        all_values = np.concatenate([group1_values, group2_values])
+
+        interval = np.nanmax(all_values) - np.nanmin(all_values)
 
-        interval = max(all_values) - min(all_values)
         df.loc[i, "span_allsamples"] = interval
 
     return df

src/dimet/visualization/abundance_bars.py

@@ -21,6 +21,7 @@
 
 cs = ConfigStore.instance()
 
+matplotlib.rcParams['svg.fonttype'] = 'none'
 
 def pile_up_abundance(df: pd.DataFrame,
                       metada_sel: pd.DataFrame) -> pd.DataFrame:
@@ -69,6 +70,9 @@ def plot_one_metabolite(df: pd.DataFrame,
     )
     if do_stripplot:
         np.random.seed(123)  # to force the jitter not to randomly change
+        dodge_bool = True
+        if axisx_var == hue_var:
+            dodge_bool = False
         try:
             sns.stripplot(
                 ax=curr_ax,
@@ -77,7 +81,7 @@ def plot_one_metabolite(df: pd.DataFrame,
                 hue=str(hue_var),
                 data=df,
                 palette=palette_choice,
-                dodge=True,
+                dodge=dodge_bool,
                 edgecolor="black",
                 linewidth=1.5,
                 alpha=1,
@@ -132,7 +136,10 @@ def plot_abundance_bars_no_grid(
         axs_k = apply_advanced_param_custom_x_labs(cfg, axs_k)
         if k == 0:  # collect the legend, which is same for any plot
             thehandles, thelabels = axs_k.get_legend_handles_labels()
-        axs_k.legend_.remove()
+        try:
+            axs_k.legend_.remove()
+        except AttributeError:
+            pass  #  no legend is generated when xaxis and barcolor are equal
         plt.tight_layout(pad=0.01, w_pad=-2, h_pad=0.1)
         plt.savefig(output_path_k,
                     bbox_inches="tight",
@@ -141,6 +148,8 @@ def plot_abundance_bars_no_grid(
 
     plt.legend(handles=thehandles, labels=thelabels, loc="upper right")
     plt.axis("off")
+    if hue_var == axisx_var:
+        plt.figtext(0.1, 0.7, "No legend:\n xaxis equal to barcolor")
     plt.savefig(os.path.join(
         output_directory,
         f"legend.{cfg.analysis.method.figure_format}"),
@@ -198,7 +207,10 @@ def plot_as_grid_of_bars(
 
     thehandles, thelabels = axs[-1].get_legend_handles_labels()
     for i in range(len(selected_metabolites)):
-        axs[i].legend_.remove()
+        try:
+            axs[i].legend_.remove()
+        except AttributeError:
+            pass #  no legend is generated when xaxis and barcolor are equal
 
     plt.subplots_adjust(left=0.2, top=0.76, bottom=0.2,
                         hspace=corrector_factor)
@@ -211,7 +223,10 @@ def plot_as_grid_of_bars(
     plt.close()
 
     plt.legend(handles=thehandles, labels=thelabels, loc="upper right")
+    if hue_var == axisx_var:
+        plt.figtext(0.1, 0.7, "No legend:\n xaxis equal to barcolor")
     plt.axis("off")
+
     plt.savefig(os.path.join(
         output_directory,
         f"legend.{cfg.analysis.method.figure_format}"),

src/dimet/visualization/isotopologue_proportions.py

@@ -22,6 +22,8 @@
 
 cs = ConfigStore.instance()
 
+matplotlib.rcParams['svg.fonttype'] = 'none'
+
 
 def isotopologue_proportions_2piled_df(
         compartment_df: pd.DataFrame,

src/dimet/visualization/mean_enrichment_line_plot.py

@@ -16,6 +16,8 @@
 
 cs = ConfigStore.instance()
 
+matplotlib.rcParams['svg.fonttype'] = 'none'
+
 
 def melt_data_metadata_2df(compartment_df: pd.DataFrame,
                            metadata_co_df: pd.DataFrame) -> pd.DataFrame:

src/dimet/visualization/metabologram.py

@@ -28,6 +28,8 @@
 
 logger = logging.getLogger(__name__)
 
+matplotlib.rcParams['svg.fonttype'] = 'none'
+
 
 def get_differential_results_dict(file_name: str,
                                   data_integration: DataIntegration,

src/dimet/visualization/pca_plot.py

@@ -5,6 +5,7 @@
 import os
 from typing import List, Union
 
+import matplotlib
 import matplotlib.figure as figure
 import matplotlib.pyplot as plt
 import numpy as np
@@ -18,6 +19,8 @@
 
 cs = ConfigStore.instance()
 
+matplotlib.rcParams['svg.fonttype'] = 'none'
+
 
 def variance_expl_plot(var_explained_df: pd.DataFrame) -> figure.Figure:
     """
@@ -47,18 +50,28 @@ def pca_scatter_plot(pc_df: pd.DataFrame,
                      var_explained_df: pd.DataFrame,
                      col1: str, col2: str,
                      labels_column: str,
-                     ellipses_column: Union[str, None]) -> figure.Figure:
+                     ellipses_column: Union[str, None],
+                     width: float=6.0, height: float=4.0) -> figure.Figure:
     """
     returns the scatter plot (a seaborn matplotlib figure)
     """
-    fig, ax = plt.subplots()
+    fig, ax = plt.subplots(figsize=(width, height))
     sns.scatterplot(x="PC1", y="PC2",
                     ax=ax,
                     data=pc_df,
                     hue=col1,
                     style=col2,
                     legend=True,
                     s=80, zorder=3)
+    handles, labels = ax.get_legend_handles_labels()
+    ax.legend(
+        handles,
+        labels,
+        bbox_to_anchor=(1.05, 1), # Places legend outside the top-right corner
+        loc='upper left',
+        borderaxespad=0.0
+    )
+    plt.tight_layout()  # Ensures the legend is not clipped
     ax.axhline(0, ls="--", color="gray", zorder=1)
     ax.axvline(0, ls="--", color="gray", zorder=1)
     if labels_column != "":
@@ -149,6 +162,7 @@ def run_pca_plot(pca_results_dict: dict,  cfg: DictConfig,
 
         color_dot = cfg.analysis.method.color
         style_dot = cfg.analysis.method.style
+
         options_labels = {'label-y': "name_to_plot",
                           'label-n': ""}  # when empty string, no dot labels
 
@@ -159,7 +173,10 @@ def run_pca_plot(pca_results_dict: dict,  cfg: DictConfig,
             scatter_fig: figure.Figure = pca_scatter_plot(
                 pc_df,  var_explained_df, color_dot,
                 style_dot, labels_column,
-                ellipses_column=cfg.analysis.method.draw_ellipses)
+                ellipses_column=cfg.analysis.method.draw_ellipses,
+                width=cfg.analysis.method.width,
+                height=cfg.analysis.method.height
+            )
             pca_scatter_2_file(scatter_fig, name_elements, out_plot_dir,
                                cfg.analysis.method.figure_format)
             plt.close()