Fix legend and annotation placement in fitting & merging plots

Author: ahmadomira

interface_DBA_dye_to_host_fitting.py

@@ -6,6 +6,8 @@
 import tkinter as tk
 from tkinter import filedialog
 from pltstyle import create_plots
+from matplotlib.transforms import Bbox
+from plot_replica import place_annotation_safely, place_annotation_opposite_legend
 
 def run_dba_dye_to_host_fitting(file_path, results_dir, h0_in_M, rmse_threshold_factor, r2_threshold, save_plots, display_plots, plots_dir, save_results, results_save_dir, number_of_fit_trials):
     # Convert initial concentration to µM units
@@ -224,7 +226,7 @@ def calculate_fit_metrics(Signal_observed, Signal_computed):
         ax.plot(d0_values, Signal_observed, 'o', label='Observed Signal')
         ax.plot(fitting_curve_x, fitting_curve_y, '--', color='blue', alpha=0.6, label='Simulated Fitting Curve')
         ax.set_title(f'Observed vs. Simulated Fitting Curve for Replica {replica_index}')
-        ax.legend(loc='best', bbox_to_anchor=(0.02, 0.98))
+        ax.legend(loc='best')
 
         # TODO: double check whether x 10^6 is needed for Id and Ihd
         # TODO: should Kd be multiplied by 10^6 or 10^-6?
@@ -235,8 +237,7 @@ def calculate_fit_metrics(Signal_observed, Signal_computed):
                       f"$RMSE$: {rmse:.3f}\n"
                       f"$R^2$: {r_squared:.3f}")
 
-        ax.annotate(param_text, xy=(0.8, 0.04), xycoords='axes fraction', fontsize=10,
-                    ha='left', va='bottom', bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5))
+        place_annotation_opposite_legend(ax, param_text)
 
         if save_plots:
             plot_file = os.path.join(plots_dir, f"fit_plot_replica_{replica_index}.png")

interface_DBA_host_to_dye_fitting.py

@@ -6,6 +6,8 @@
 import matplotlib.pyplot as plt
 from datetime import datetime
 from pltstyle import create_plots
+from plot_replica import place_annotation_safely, place_annotation_opposite_legend
+from matplotlib.transforms import Bbox
 
 # Add number_of_fit_trials to function parameters
 def run_dba_host_to_dye_fitting(file_path, results_dir, d0_in_M, rmse_threshold_factor, r2_threshold, save_plots, display_plots, plots_dir, save_results, results_save_dir, number_of_fit_trials):
@@ -229,7 +231,7 @@ def calculate_fit_metrics(Signal_observed, Signal_computed):
         ax.plot(h0_values, Signal_observed, 'o', label='Observed Signal')
         ax.plot(fitting_curve_x, fitting_curve_y, '--', color='blue', alpha=0.6, label='Simulated Fitting Curve')
         ax.set_title(f'Observed vs. Simulated Fitting Curve for Replica {replica_index}')
-        ax.legend(loc='best', bbox_to_anchor=(0.02, 0.98))
+        ax.legend(loc='best')
 
         param_text = (f"$K_d$: {median_params[1] * 1e6:.2e} $M^{{-1}}$\n"
                       f"$I_0$: {median_params[0]:.2e}\n"
@@ -238,8 +240,7 @@ def calculate_fit_metrics(Signal_observed, Signal_computed):
                       f"$RMSE$: {rmse:.3f}\n"
                       f"$R^2$: {r_squared:.3f}")
 
-        ax.annotate(param_text, xy=(0.8, 0.04), xycoords='axes fraction', fontsize=10,
-                    ha='left', va='bottom', bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5))
+        place_annotation_opposite_legend(ax, param_text)
 
         if save_plots:
             plot_file = os.path.join(plots_dir, f"fit_plot_replica_{replica_index}.png")

interface_DyeAlone_fitting.py

@@ -9,6 +9,7 @@
 from scipy.stats import linregress, ttest_1samp, t
 import matplotlib.pyplot as plt
 from matplotlib.ticker import FuncFormatter
+from matplotlib.transforms import Bbox
 
 # Local imports
 from pltstyle import create_plots
@@ -131,6 +132,52 @@ def round_to_sigfigs(value, sigfigs=4):
         return f"{value:.{sigfigs}g}"
     return value
 
+def place_annotation_safely(ax, text, fontsize=10, margin=10, **kwargs):
+    """
+    Place annotation in a corner that avoids overlap with legend and data points.
+    Tries all corners, prefers the first non-overlapping, else picks the one with least overlap.
+    """
+    renderer = ax.figure.canvas.get_renderer()
+    legend = ax.get_legend()
+    legend_bbox = legend.get_window_extent(renderer) if legend else None
+    # Get data points in display coords
+    data_disp = ax.transData.transform(np.column_stack((ax.get_lines()[0].get_xdata(), ax.get_lines()[0].get_ydata())))
+    data_bboxes = [Bbox.from_bounds(x - margin, y - margin, 2*margin, 2*margin) for x, y in data_disp]
+    # Candidate positions: (x, y, ha, va)
+    candidates = [
+        (0.01, 0.99, 'left', 'top'),    # upper left
+        (0.99, 0.99, 'right', 'top'),   # upper right
+        (0.01, 0.01, 'left', 'bottom'), # lower left
+        (0.99, 0.01, 'right', 'bottom') # lower right
+    ]
+    best_ann = None
+    min_overlap = float('inf')
+    for x, y, ha, va in candidates:
+        ann = ax.annotate(
+            text, xy=(x, y), xycoords='axes fraction',
+            ha=ha, va=va, fontsize=fontsize,
+            bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5),
+            **kwargs
+        )
+        plt.draw()  # Needed to get the bbox
+        ann_bbox = ann.get_window_extent(renderer)
+        overlap = 0
+        # Check overlap with legend
+        if legend_bbox is not None and ann_bbox.overlaps(legend_bbox):
+            overlap += ann_bbox.intersection(legend_bbox).area
+        # Check overlap with data points
+        for db in data_bboxes:
+            if ann_bbox.overlaps(db):
+                overlap += ann_bbox.intersection(db).area
+        if overlap == 0:
+            return ann  # Found a good spot
+        if overlap < min_overlap:
+            min_overlap = overlap
+            best_ann = ann
+        else:
+            ann.remove()
+    return best_ann  # Return the least-overlapping position if all overlap
+
 # Main function to perform the fitting and plotting
 def perform_fitting(input_file_path, output_file_path, save_plots, display_plots, plots_dir):
     if not output_file_path.endswith(".txt"):
@@ -150,7 +197,7 @@ def perform_fitting(input_file_path, output_file_path, save_plots, display_plots
     Id_mean, Id_lower_bound, Id_upper_bound, Id_stdev = prediction_interval(retained_slopes, avg_slope)
     I0_mean, I0_lower_bound, I0_upper_bound, I0_stdev = prediction_interval(retained_intercepts, avg_intercept)
 
-    if len(retained_slopes) == 1:
+    if len(retained_slopes == 1):
         Id_mean = retained_slopes[0]
         I0_mean = retained_intercepts[0]
         Id_lower_bound = "not applicable"
@@ -204,7 +251,10 @@ def scientific_notation(val, pos=0):
     handles.append(avg_fit_plot[0])
     labels.append(rf'Average Fit: $Y = {formatter(avg_slope)}X + {formatter(avg_intercept)}$')
 
-    ax.legend(handles, labels, loc='upper left', bbox_to_anchor=(0, 1))
+    ax.legend(handles, labels, loc='best')
+
+    param_text = f"Id: {Id_mean:.3e}\nI0: {I0_mean:.3e}"
+    place_annotation_safely(ax, param_text)
 
     if save_plots:
         # Create a unique plot filename based on the input file

interface_GDA_fitting.py

@@ -6,12 +6,14 @@
 import numpy as np
 from scipy.optimize import brentq, minimize
 import matplotlib.pyplot as plt
+from matplotlib.transforms import Bbox
 
 from pltstyle import create_plots
+from plot_replica import place_annotation_safely, place_annotation_opposite_legend
 
 def format_value(value):
     return f"{value:.0f}" if value > 10 else f"{value:.2f}"
-    
+
 # Function to run the fitting process
 def run_fitting(file_path, results_file_path, Kd_in_M, h0_in_M, g0_in_M, number_of_fit_trials, rmse_threshold_factor, r2_threshold, save_plots, display_plots, plots_dir, save_results, results_save_dir):
     try:
@@ -149,7 +151,7 @@ def run_fitting(file_path, results_file_path, Kd_in_M, h0_in_M, g0_in_M, number_
             ax.plot(d0_values, Signal_observed, 'o', label='Observed Signal')
             ax.plot(fitting_curve_x, fitting_curve_y, '--', color='blue', alpha=0.6, label='Simulated Fitting Curve')
             ax.set_title(f'Observed vs. Simulated Fitting Curve for Replica {replica_index}')
-            ax.legend(loc='best', bbox_to_anchor=(0.02, 0.98))
+            ax.legend(loc='best')
 
             # Annotate plot with median parameter values and fit metrics
             param_text = (f"$K_g$: {median_params[1] * 1e6:.2e} $M^{{-1}}$\n"
@@ -159,8 +161,7 @@ def run_fitting(file_path, results_file_path, Kd_in_M, h0_in_M, g0_in_M, number_
                         f"$RMSE$: {format_value(rmse)}\n"
                         f"$R^2$: {r_squared:.3f}")
 
-            ax.annotate(param_text, xy=(0.8, 0.04), xycoords='axes fraction', fontsize=10,
-                ha='left', va='bottom', bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5))
+            place_annotation_opposite_legend(ax, param_text)
 
             if save_plots:
                 plot_file = os.path.join(plots_dir, f"fit_plot_replica_{replica_index}.png")

interface_IDA_fitting.py

@@ -6,7 +6,9 @@
 import matplotlib.pyplot as plt
 from scipy.optimize import brentq, minimize
 from pltstyle import create_plots
-    
+from matplotlib.transforms import Bbox
+from plot_replica import place_annotation_safely, place_annotation_opposite_legend
+
 def run_ida_fitting(file_path, results_file_path, Kd_in_M, h0_in_M, g0_in_M, number_of_fit_trials, rmse_threshold_factor, r2_threshold, save_plots, display_plots, plots_dir, save_results, results_save_dir):
 
 
@@ -190,7 +192,7 @@ def split_replicas(data):
         ax.plot(g0_values, Signal_observed, 'o', label='Observed Signal')
         ax.plot(fitting_curve_x, fitting_curve_y, '--', color='blue', alpha=0.6, label='Simulated Fitting Curve')
         ax.set_title(f'Observed vs. Simulated Fitting Curve for Replica {replica_index}')
-        ax.legend()
+        ax.legend(loc='best')
 
         param_text = (f"$K_g$: {median_params[1] * 1e6:.2e} $M^{{-1}}$\n"
                       f"$I_0$: {median_params[0]:.2e}\n"
@@ -199,9 +201,7 @@ def split_replicas(data):
                       f"$RMSE$: {rmse:.3f}\n"
                       f"$R^2$: {r_squared:.3f}")
 
-        ax.annotate(param_text, xy=(0.97, 0.95), xycoords='axes fraction', fontsize=10,
-            ha='right', va='top', bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5),
-            multialignment='left')
+        place_annotation_opposite_legend(ax, param_text)
 
         if save_plots:
             plot_file = os.path.join(plots_dir, f"fit_plot_replica_{replica_index}.png")

interface_dba_merge_fits.py

@@ -6,6 +6,8 @@
 import matplotlib.pyplot as plt
 from datetime import datetime
 from pltstyle import create_plots  # Import the create_plots function
+from matplotlib.transforms import Bbox
+from plot_replica import place_annotation_safely, place_annotation_opposite_legend
 
 
 def format_value(value):
@@ -306,10 +308,9 @@ def run_dba_merge_fits(results_dir, outlier_relative_threshold, rmse_threshold_f
                   f"$I_{{hd}}$: {avg_params[3]:.2e} $M^{{-1}}$ (STDEV: {stdev_params[3]:.2e})\n"
                   f"$RMSE$: {format_value(rmse)}\n"
                   f"$R^2$: {r_squared:.3f}")
-    ax2.annotate(param_text, xy=(0.95, 0.05), xycoords='axes fraction', fontsize=10,
-                ha='right', va='bottom', bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5), multialignment='left')
+    place_annotation_opposite_legend(ax2, param_text)
 
-    ax2.legend()
+    ax2.legend(loc='best')
     fig2.tight_layout()
     if save_plots:
         save_plot(fig2, "averaged_fitting_plot.png", results_dir)

interface_gda_merge_fits.py

@@ -7,6 +7,8 @@
 from scipy.optimize import brentq
 from datetime import datetime
 from pltstyle import create_plots  # Import the create_plots function
+from matplotlib.transforms import Bbox
+from plot_replica import place_annotation_safely, place_annotation_opposite_legend
 
 def format_value(value):
     return f"{value:.0f}" if value > 10 else f"{value:.2f}"
@@ -226,7 +228,7 @@ def run_gda_merge_fits(results_dir, outlier_relative_threshold, rmse_threshold_f
         if len(outlier_indices) > 0:
             ax1.plot(concentrations[outlier_indices], signals[outlier_indices], 'x', color=colors[idx], markersize=8, label=f"Replica {idx + 1} Outliers")
 
-    ax1.legend(loc='best', bbox_to_anchor=(0.02, 0.98))
+    ax1.legend(loc='best')
     fig1.tight_layout()
     if save_plots:
         save_plot(fig1, "all_replicas_fitting_plot_with_outliers.png", results_dir)
@@ -310,10 +312,9 @@ def run_gda_merge_fits(results_dir, outlier_relative_threshold, rmse_threshold_f
                   f"$I_{{hd}}$: {avg_params[3]:.2e} $M^{{-1}}$ (STDEV: {stdev_params[3]:.2e})\n"
                   f"$RMSE$: {format_value(rmse)}\n"
                   f"$R^2$: {r_squared:.3f}")
-    ax2.annotate(param_text, xy=(0.97, 0.04), xycoords='axes fraction', fontsize=10,
-                ha='right', va='bottom', bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5), multialignment='left')
+    place_annotation_opposite_legend(ax2, param_text)
 
-    ax2.legend(loc='best', bbox_to_anchor=(0.02, 0.98))
+    ax2.legend(loc='best')
     fig2.tight_layout()
     if save_plots:
         save_plot(fig2, "averaged_fitting_plot.png", results_dir)

interface_ida_merge_fits.py

@@ -7,6 +7,8 @@
 from datetime import datetime
 from scipy.optimize import brentq
 from pltstyle import create_plots
+from matplotlib.transforms import Bbox
+from plot_replica import place_annotation_safely, place_annotation_opposite_legend
 
 def format_value(value):
     return f"{value:.0f}" if value > 10 else f"{value:.2f}"
@@ -329,11 +331,9 @@ def export_averaged_data(avg_concentrations, avg_signals, avg_fitting_curve_x, a
                   f"$I_{{hd}}$: {avg_params[3]:.2e} $M^{{-1}}$ (STDEV: {stdev_params[3]:.2e})\n"
                   f"$RMSE$: {format_value(rmse)}\n"
                   f"$R^2$: {r_squared:.3f}")
-    ax2.annotate(param_text, xy=(0.97, 0.95), xycoords='axes fraction', fontsize=10,
-                 ha='right', va='top', bbox=dict(boxstyle="round,pad=0.3", edgecolor="black", facecolor="lightgrey", alpha=0.5),
-                 multialignment='left')
+    place_annotation_opposite_legend(ax2, param_text)
 
-    ax2.legend()
+    ax2.legend(loc='best')
     fig2.tight_layout()
     if save_plots:
         save_plot(fig2, "averaged_fitting_plot.png", results_dir)

plot_replica.py

@@ -5,6 +5,8 @@
 import pltstyle
 import tkinter as tk
 import re
+import numpy as np
+from matplotlib.transforms import Bbox
 
 from pathlib import Path
 from tkinter import filedialog, messagebox
@@ -13,6 +15,80 @@
 from bmg_to_txt import read_bmg_xlsx, extract_concentration_vector
 
 
+def place_annotation_safely(ax, text, data_x, data_y, line2d=None, marker_size=12, margin=8, **annotate_kwargs):
+    """
+    Place annotation box in a location that avoids overlap with legend, data points, and optionally a line.
+    - data_x, data_y: arrays of data points (for scatter/markers)
+    - line2d: optional, a matplotlib Line2D object to check for overlap with the line itself
+    - marker_size: pixel size of marker to avoid (default 12)
+    - margin: extra pixels to add around annotation box (default 8)
+    """
+    renderer = ax.figure.canvas.get_renderer()
+    corners = [
+        (0.01, 0.99, {'xycoords': 'axes fraction', 'va': 'top', 'ha': 'left'}),
+        (0.99, 0.99, {'xycoords': 'axes fraction', 'va': 'top', 'ha': 'right'}),
+        (0.01, 0.01, {'xycoords': 'axes fraction', 'va': 'bottom', 'ha': 'left'}),
+        (0.99, 0.01, {'xycoords': 'axes fraction', 'va': 'bottom', 'ha': 'right'}),
+    ]
+    legend = ax.get_legend()
+    legend_bbox = legend.get_window_extent(renderer) if legend else None
+    # Data points as bboxes (expanded for marker size)
+    data_disp = ax.transData.transform(np.column_stack([data_x, data_y]))
+    data_bboxes = [Bbox.from_bounds(x-marker_size, y-marker_size, 2*marker_size, 2*marker_size) for x, y in data_disp]
+    # If a line is provided, rasterize it to points and add bboxes
+    line_bboxes = []
+    if line2d is not None:
+        line_x, line_y = line2d.get_data()
+        line_disp = ax.transData.transform(np.column_stack([line_x, line_y]))
+        for x, y in line_disp:
+            line_bboxes.append(Bbox.from_bounds(x-2, y-2, 4, 4))
+    best = None
+    min_overlap = float('inf')
+    for x, y, opts in corners:
+        ann = ax.annotate(text, (x, y), bbox=dict(boxstyle='round', fc='w', ec='k', alpha=0.8), annotation_clip=False, **opts, **annotate_kwargs)
+        ax.figure.canvas.draw()
+        ann_bbox = ann.get_window_extent(renderer).expanded(1.05, 1.1).padded(margin)
+        overlap = 0
+        if legend_bbox and ann_bbox.overlaps(legend_bbox):
+            overlap += 1e6
+        for db in data_bboxes:
+            if ann_bbox.overlaps(db):
+                overlap += 1
+        for lb in line_bboxes:
+            if ann_bbox.overlaps(lb):
+                overlap += 1
+        if overlap == 0:
+            return ann
+        if overlap < min_overlap:
+            min_overlap = overlap
+            best = ann
+        ann.remove()
+    return best  # fallback: least overlap
+
+
+def place_annotation_opposite_legend(ax, text, offset_frac=0.03, **annotate_kwargs):
+    """
+    Place annotation box in the corner diagonally opposite to the legend, with a margin from axes.
+    offset_frac: fraction of axes width/height to offset from the edge (default 0.03 = 3%)
+    """
+    legend = ax.get_legend()
+    loc = getattr(legend, '_loc', 'upper right') if legend else 'upper right'
+    loc_map = {
+        'upper right':  (offset_frac, offset_frac, {'xycoords': 'axes fraction', 'va': 'bottom', 'ha': 'left'}),
+        'upper left':   (1-offset_frac, offset_frac, {'xycoords': 'axes fraction', 'va': 'bottom', 'ha': 'right'}),
+        'lower left':   (1-offset_frac, 1-offset_frac, {'xycoords': 'axes fraction', 'va': 'top', 'ha': 'right'}),
+        'lower right':  (offset_frac, 1-offset_frac, {'xycoords': 'axes fraction', 'va': 'top', 'ha': 'left'}),
+        'best':         (offset_frac, offset_frac, {'xycoords': 'axes fraction', 'va': 'bottom', 'ha': 'left'}),
+    }
+    x, y, opts = loc_map.get(loc, loc_map['upper right'])
+    return ax.annotate(
+        text, (x, y),
+        bbox=dict(boxstyle='round', fc='w', ec='k', alpha=0.8),
+        annotation_clip=False,
+        **opts, **annotate_kwargs
+    )
+
+
 def plot_all_replica(raw_data_path : str, robot_file_path : str, save_dir : str):
     """this is for ONE excel file"""
     raw_data_path, robot_file_path, save_dir = map(Path, (raw_data_path, robot_file_path, save_dir))
@@ -25,7 +101,7 @@ def plot_all_replica(raw_data_path : str, robot_file_path : str, save_dir : str)
     fig, ax = pltstyle.create_plots(plot_title=plot_title)
     # ax.boxplot(data, tick_labels=concentration_vector, patch_artist=True)
     ax.plot(concentration_vector, data.values.T, label=[f'Replica {i+1}' for i in range(data.shape[0])])#
-    ax.legend()
+    ax.legend(loc='best')
 
     # save plot of different analytes to separate folders
     save_dir_analyte = (save_dir / ' '.join(raw_data_path.stem.split('_')[1:4]))