Skip to content
TheChatArchiveThe Chat Archive
o1-preview-2024-09-12· OpenAIAi2 WildChat-4.8M (AI2)

In this code, i specify log tick levels, but this takes no effect. This can be due to…

  1. User

    In this code, i specify log tick levels, but this takes no effect. This can be due to interaction with WCS (see tick fixes in the code). Try to fix the issue.

    '''!!! СДЕЛАТЬ RESHAPE !!!''' from matplotlib.ticker import LogLocator fig = plt.figure(figsize=(8, 4)) plt.rcParams['text.usetex'] = True

    Create the first subplot

    ax1 = fig.add_subplot(121, projection=Ha_cut.wcs) im1 = ax1.imshow(binned_ratio_SII, vmin=1.5, vmax=150, norm='log') ax1.contour(np.log10(Ha), cmap='Reds', linewidths=1, levels=(-10, 2, 2.5, 3, 3.5, 4)) ax1.invert_yaxis() ax1.set_xlabel('RA') ax1.set_ylabel('DEC') ax1.set_title('[O III]/[S II]') ax1.set_title(r'[O,\textsc{iii}] / [S,\textsc{ii}]', pad=35) ax1.invert_yaxis()

    Create consistent colorbar for ax1

    divider1 = make_axes_locatable(ax1) cax1 = divider1.append_axes('top', size='5%', pad=0.05) cax1.tick_params(labelsize=8, axis='y', direction='in', labelleft=False, left=False, right=False) cax1.tick_params(labelsize=8, axis='x', direction='out', labeltop=True, labelbottom=False, top=True)

    cb1 = fig.colorbar(im1, cax=cax1, orientation='horizontal') cb1.set_ticks([2, 10, 30, 100]) # Choose appropriate values on a log scale cb1.ax.xaxis.set_major_locator(LogLocator(base=50, subs='auto')) # Optional: fine-tune

    Create the second subplot

    ax2 = fig.add_subplot(122, projection=Ha_cut.wcs, sharey=ax1, sharex=ax1) # Share y-axis to ensure alignment im2 = ax2.imshow(binned_ratio_Ha, vmin=1, vmax=12) ax2.contour(np.log10(Ha), cmap='Reds', linewidths=1, levels=(-10, 2, 2.5, 3, 3.5, 4)) ax2.set_xlabel('RA')

    ax2.set_title('[O III]/Ha', pad = 30)

    ax2.set_title(r'[O,\textsc{iii}] / Hβ\beta', pad=35)

    Create consistent colorbar for ax2

    divider2 = make_axes_locatable(ax2) cax2 = divider2.append_axes('top', size='5%', pad=0.05) cax2.tick_params(labelsize=8, axis='y', direction='in', labelleft=False, left=False, right=False) cax2.tick_params(labelsize=8, axis='x', direction='out', labeltop=True, labelbottom=False, top=True) cax2.xaxis.get_offset_text().set_visible(False) # Hide offset label (if any) cax1.coords[0].set_auto_axislabel(False) # Disable WCS axis labeling on the x-axis cax1.coords[1].set_auto_axislabel(False)

    cax2.coords[0].set_auto_axislabel(False) # Disable WCS axis labeling on the x-axis cax2.coords[1].set_auto_axislabel(False)
    ax2.coords[1].set_auto_axislabel(False) # Disable WCS axis labeling on the x-axis ax2.coords[1].set_ticklabel_visible(False) # Hide y-axis tick labels ax2.invert_yaxis()

    cax2.coords[1].set_auto_axislabel(False)

    cb2 = fig.colorbar(im2, cax=cax2, orientation='horizontal')

    cb2.set_ticks()

    Align the axes limits explicitly

    ax1.set_xlim(350, 550)

    ax1.set_ylim(700, 500)

    ax2.set_xlim(350, 550)

    ax2.set_ylim(700, 500)

    plt.tight_layout() plt.savefig('OIII_Ha.pdf', dpi=300)

  2. Assistant

    The issue arises because when using make_axes_locatable with a WCSAxes (ax1 and ax2 in your case), the appended axes (cax1 and cax2) inherit the projection and become WCSAxes as well. This can interfere with standard Matplotlib functionalities like setting ticks, especially when dealing with colorbars.

    To fix this, you need to ensure that the colorbar axes (cax1 and cax2) are standard Matplotlib Axes, not WCSAxes. You can achieve this by specifying the axes_class parameter when using append_axes:

    cax1 = divider1.append_axes('top', size='5%', pad=0.05, axes_class=plt.Axes)
    

    Additionally, when you're using a logarithmic normalization (norm='log'), it's important to use LogNorm from matplotlib.colors, as 'log' is not a valid string for the norm parameter in imshow.

    Here's the corrected and updated code:

    import matplotlib.pyplot as plt
    from mpl_toolkits.axes_grid1.axes_divider import make_axes_locatable
    from matplotlib.colors import LogNorm
    from matplotlib.ticker import LogLocator, LogFormatter
    
    plt.rcParams['text.usetex'] = True
    
    fig = plt.figure(figsize=(8, 4))
    
    # Create the first subplot
    ax1 = fig.add_subplot(121, projection=Ha_cut.wcs)
    im1 = ax1.imshow(binned_ratio_SII, vmin=1.5, vmax=150, norm=LogNorm(vmin=1.5, vmax=150))
    ax1.contour(np.log10(Ha), cmap='Reds', linewidths=1, levels=(-10, 2, 2.5, 3, 3.5, 4))
    ax1.invert_yaxis()
    ax1.set_xlabel('RA')
    ax1.set_ylabel('DEC')
    ax1.set_title(r'[O\,\textsc{iii}] / [S\,\textsc{ii}]', pad=35)
    
    # Create consistent colorbar for ax1
    divider1 = make_axes_locatable(ax1)
    cax1 = divider1.append_axes('top', size='5%', pad=0.05, axes_class=plt.Axes)
    cb1 = fig.colorbar(im1, cax=cax1, orientation='horizontal')
    
    # Set logarithmic ticks on the colorbar
    cb1.ax.xaxis.set_major_locator(LogLocator(base=10))
    cb1.ax.xaxis.set_major_formatter(LogFormatter())
    cb1.ax.xaxis.set_minor_locator(LogLocator(base=10, subs='auto', numticks=10))
    cb1.ax.tick_params(labelsize=8, axis='x', direction='out', labeltop=True, labelbottom=False, top=True)
    cb1.ax.xaxis.get_offset_text().set_visible(False)  # Hide offset label if present
    
    # Remove the WCS tick labels if any (since cax1 is now a standard Axes, this should be unnecessary)
    # cax1.coords[0].set_auto_axislabel(False)
    # cax1.coords[1].set_auto_axislabel(False)
    
    # Create the second subplot
    ax2 = fig.add_subplot(122, projection=Ha_cut.wcs, sharex=ax1, sharey=ax1)
    im2 = ax2.imshow(binned_ratio_Ha, vmin=1, vmax=12)
    ax2.contour(np.log10(Ha), cmap='Reds', linewidths=1, levels=(-10, 2, 2.5, 3, 3.5, 4))
    ax2.set_xlabel('RA')
    ax2.set_title(r'[O\,\textsc{iii}] / H$\beta$', pad=35)
    ax2.coords[1].set_auto_axislabel(False)   # Disable WCS axis labeling on the y-axis
    ax2.coords[1].set_ticklabel_visible(False)  # Hide y-axis tick labels
    ax2.invert_yaxis()
    
    # Create consistent colorbar for ax2
    divider2 = make_axes_locatable(ax2)
    cax2 = divider2.append_axes('top', size='5%', pad=0.05, axes_class=plt.Axes)
    cb2 = fig.colorbar(im2, cax=cax2, orientation='horizontal')
    
    # Customize colorbar ticks for the second plot if needed
    cb2.ax.xaxis.set_major_locator(plt.MaxNLocator(integer=True))
    cb2.ax.tick_params(labelsize=8, axis='x', direction='out', labeltop=True, labelbottom=False, top=True)
    cb2.ax.xaxis.get_offset_text().set_visible(False)  # Hide offset label if present
    
    # Adjust layout and save the figure
    plt.tight_layout()
    plt.savefig('OIII_Ha.pdf', dpi=300)
    

    Explanation of Changes:

    1. Use LogNorm for Logarithmic Scaling:

      • Changed norm='log' to norm=LogNorm(vmin=1.5, vmax=150) in imshow. The norm parameter requires an instance of Normalize or its subclasses, and LogNorm is appropriate for logarithmic scaling.
    2. Ensure Colorbar Axes are Standard Axes:

      • Specified axes_class=plt.Axes when creating cax1 and cax2 using append_axes. This prevents the appended axes from inheriting the WCS projection of the main axes.
    3. Set Logarithmic Ticks on the Colorbar:

      • Used LogLocator and LogFormatter to set major and minor ticks on the colorbar axes.
      • Adjusted tick parameters to display the ticks correctly.
    4. Removed Unnecessary WCS Configurations on Colorbar Axes:

      • Since cax1 and cax2 are now standard Axes, there's no need to manipulate coords or WCS-related settings.
    5. Adjusted Ticks for the Second Colorbar:

      • For cb2, customized the ticks using MaxNLocator to display integer ticks, if necessary.
    6. Ensured Consistency in Tick Parameters:

      • Set consistent tick parameters for both colorbars to maintain a uniform appearance.

    Note:

    • By ensuring cax1 and cax2 are standard Axes, standard Matplotlib functions for setting ticks and labels work as expected.
    • The use of LogNorm ensures that the color mapping is correctly logarithmic, and the colorbar reflects this scaling.

    Additional Tips:

    • Always verify that the normalization (norm) is set correctly when working with logarithmic data.
    • When working with WCS projections, be mindful of how appended axes may inherit properties. Specifying the axes class can prevent unexpected behaviors.
    • Customize tick locators and formatters to control the appearance of ticks on both plots and colorbars.