A simple script for extracting numerical data from plots: load an image and click on it to record point coordinates. The first click sets the origin, the second sets the unit scale, the subsequent clicks sample the data points.

trace_image.py

#!/usr/bin/env python3

"""
A simple script for extracting numerical data from plots: load an image and click on it to record point coordinates.
Usage example:

    ./trace_image.py my_plot.jpg

Copyright (C) 2025 Pavel Kirienko <[email protected]>
"""

import sys
from pathlib import Path
import matplotlib.pyplot as plt
import matplotlib.image as mpimg

THRESHOLD = 10

origin = None  # Will hold (x0, y0) in pixel coords
scale_ref = None  # Will hold (dx_pixels, dy_pixels) to define 1 unit in each axis
points = []  # List of collected points in the new coordinate system

fig, ax = plt.subplots()

# Load and display the image
img = mpimg.imread(sys.argv[1])
ax.imshow(img)
ax.set_title("Click to set the origin")


def onclick(event):
    """Handle mouse click events."""
    global origin, scale_ref, points
    if event.xdata is None or event.ydata is None:
        return

    if origin is None:
        origin = (event.xdata, event.ydata)
        ax.plot(origin[0], origin[1], 'k+')
    elif scale_ref is None or (scale_ref[0] is None or scale_ref[1] is None):
        ax.plot(event.xdata, event.ydata, 'g+')
        if scale_ref is None:
            dx = event.xdata - origin[0]
            dy = event.ydata - origin[1]
            if abs(dx) < THRESHOLD and abs(dy) < THRESHOLD:
                print("Scale reference click too close to the origin in both axes, try again.", file=sys.stderr)
            elif abs(dx) < THRESHOLD and abs(dy) >= THRESHOLD:
                scale_ref = (None, dy)
            elif abs(dy) < THRESHOLD and abs(dx) >= THRESHOLD:
                scale_ref = (dx, None)
            else:
                scale_ref = (dx, dy)
        else:
            if scale_ref[0] is None:
                dx = event.xdata - origin[0]
                if abs(dx) < THRESHOLD:
                    print("X scale reference too close to the origin, click again for X scale.", file=sys.stderr)
                else:
                    scale_ref = (dx, scale_ref[1])
            elif scale_ref[1] is None:
                dy = event.ydata - origin[1]
                if abs(dy) < THRESHOLD:
                    print("Y scale reference too close to the origin, click again for Y scale.", file=sys.stderr)
                else:
                    scale_ref = (scale_ref[0], dy)
    else:
        pixel_x = event.xdata
        pixel_y = event.ydata
        ax.plot(pixel_x, pixel_y, 'rx')
        dx_from_origin = pixel_x - origin[0]
        dy_from_origin = pixel_y - origin[1]
        new_x = dx_from_origin / scale_ref[0]
        new_y = dy_from_origin / scale_ref[1]
        points.append((new_x, new_y))
        print(f"{new_x:+012.6f}\t{new_y:+012.6f}")

    title = ""
    if scale_ref is None:
        title = "Click a unit distance away from the origin on one or both axes to set the scale"
    elif scale_ref[0] is None or scale_ref[1] is None:
        axis = "X" if scale_ref[0] is None else "Y"
        title = f"Click a unit distance away from the origin along the {axis} axis to set the {axis} scale"
    elif not points:
        title = "Click to record data points"
    else:
        title = "Add more data points or close the window if done"

    ax.set_title(title)
    plt.draw()


fig.canvas.mpl_connect("button_press_event", onclick)
plt.tight_layout()
try:
    plt.show()
except KeyboardInterrupt:
    pass

if points:
    out_file = "points.tab"
    print("Saving the data into", out_file, file=sys.stderr)
    Path(out_file).write_text("\n".join(f"{x:+012.6f}\t{y:+012.6f}" for x, y in points))