pycbc_plot_bank_corner improvements for live (#4849)

* Add IFO to fits plot title

* allow log parameters in pycbc_plot_bank_corner

* Fix use of log scales in the case that histograms are being used

* Accidentally made some linear parameters log-scaled

* Deal with case where not using any log parameters

* CC

bin/live/pycbc_live_supervise_collated_trigger_fits

@@ -270,7 +270,10 @@ def plot_fits(
     ]
     fits_plot_arguments += sv.dict_to_args(plot_fit_options)
 
-    title = "Fit parameters for pycbc-live, triggers from " + day_title_str
+    title = "Fit parameters for pycbc-live, triggers from {}, {}".format(
+        ifo,
+        day_title_str
+    )
     if smoothed == True:
         title += ', smoothed'
     fits_plot_arguments += ['--title', title]

bin/plotting/pycbc_plot_bank_corner

@@ -72,6 +72,13 @@ parser.add_argument("--parameters",
          "property of that parameter will be used. If not "
          "provided, will plot all of the parameters in the "
          "bank.")
+parser.add_argument(
+    '--log-parameters',
+    nargs='+',
+    default=[],
+    help="Which parameters are to be plotted on a log scale? "
+         "Must be also given in parameters"
+)
 parser.add_argument('--plot-histogram',
     action='store_true',
     help="Plot 1D histograms of parameters on the "
@@ -103,6 +110,11 @@ parser.add_argument("--color-parameter",
     help="Color scatter points according to the parameter given. "
          "May optionally provide a label in the same way as for "
          "--parameters. Default=No scatter point coloring.")
+parser.add_argument(
+    '--log-colormap',
+    action='store_true',
+    help="Should the colorbar be plotted on a log scale?"
+)
 parser.add_argument('--dpi',
     type=int,
     default=200,
@@ -117,6 +129,13 @@ parser.add_argument('--title',
 add_style_opt_to_parser(parser)
 args = parser.parse_args()
 
+for lp in args.log_parameters:
+    if not lp in args.parameters:
+        parser.error(
+            "--log-parameters should be in --parameters. "
+            f"{lp} not in [{', '.join(args.parameters)}]"
+        )
+
 pycbc.init_logging(args.verbose)
 set_style_from_cli(args)
 
@@ -146,7 +165,7 @@ if args.fits_file is not None:
             param = fits_f[p][:].astype(float)
             # We need to check for the cardinal '-1' value which means
             # that the fit is invalid
-            param[param <= 0] = 0 if 'count' in p and 'log' not in p else np.nan
+            param[param <= 0] = np.nan
             bank[p] = param
 
 logging.info("Got %d templates from the bank", banklen)
@@ -227,12 +246,21 @@ if cpar:
 for p in required_minmax:
     minval = np.nanmin(bank_fa[p][bank_fa[p] != -np.inf])
     maxval = np.nanmax(bank_fa[p][bank_fa[p] != np.inf])
-    valrange = maxval - minval
+    if (p in args.log_parameters) or (p == cpar and args.log_colormap):
+        # Extend the range by 10% in log-space
+        logvalrange = np.log(maxval) - np.log(minval)
+        if p not in mins:
+            mins[p] = np.exp(np.log(minval) - 0.05 * logvalrange)
+        if p not in maxs:
+            maxs[p] = np.exp(np.log(maxval) + 0.05 * logvalrange)
+    else:
+        # Extend the range by 10%
+        valrange = maxval - minval
+        if p not in mins:
+            mins[p] = minval - 0.05 * valrange
+        if p not in maxs:
+            maxs[p] = maxval + 0.05 * valrange
 
-    if p not in mins:
-        mins[p] = minval - 0.05 * valrange
-    if p not in maxs:
-        maxs[p] = maxval + 0.05 * valrange
 
 # Deal with non-coloring case:
 zvals = bank_fa[cpar] if cpar else None
@@ -247,6 +275,7 @@ fig, axis_dict = create_multidim_plot(
     plot_scatter=True,
     plot_contours=False,
     scatter_cmap="viridis",
+    scatter_log_cmap=args.log_colormap,
     marginal_title=False,
     marginal_percentiles=[],
     fill_color='g',
@@ -258,6 +287,7 @@ fig, axis_dict = create_multidim_plot(
     hist_color=hist_color,
     mins=mins,
     maxs=maxs,
+    log_parameters=args.log_parameters,
 )
 
 title_text = f"{os.path.basename(args.bank_file)}"
@@ -293,6 +323,19 @@ for i in range(len(args.parameters)):
         for s0, s1 in zip(sharex_axes[:-1], sharex_axes[1:]):
             s0.sharex(s1)
 
+for (p1, p2), ax in axis_dict.items():
+    if p1 == p2 and p1 in args.log_parameters:
+        if p1 == args.parameters[-1] and len(args.parameters) == 2:
+            # This will be turned on its side, so set _y_ axis to log
+            ax[0].semilogy()
+        else:
+            ax[0].semilogx()
+    else:
+        if p1 in args.log_parameters:
+            ax[0].semilogx()
+        if p2 in args.log_parameters:
+            ax[0].semilogy()
+
 logging.info("Plot generated")
 fig.set_dpi(args.dpi)
 

pycbc/results/scatter_histograms.py

@@ -43,7 +43,7 @@
 if 'matplotlib.backends' not in sys.modules:  # nopep8
     matplotlib.use('agg')
 
-from matplotlib import (offsetbox, pyplot, gridspec)
+from matplotlib import (offsetbox, pyplot, gridspec, colors)
 
 from pycbc.results import str_utils
 from pycbc.io import FieldArray
@@ -337,7 +337,7 @@ def create_marginalized_hist(ax, values, label, percentiles=None,
                              linestyle='-', plot_marginal_lines=True,
                              title=True, expected_value=None,
                              expected_color='red', rotated=False,
-                             plot_min=None, plot_max=None):
+                             plot_min=None, plot_max=None, log_scale=False):
     """Plots a 1D marginalized histogram of the given param from the given
     samples.
 
@@ -380,6 +380,8 @@ def create_marginalized_hist(ax, values, label, percentiles=None,
         creates.
     scalefac : {1., float}
         Factor to scale the default font sizes by. Default is 1 (no scaling).
+    log_scale : boolean
+        Should the histogram bins be logarithmically spaced
     """
     if fillcolor is None:
         htype = 'step'
@@ -389,7 +391,19 @@ def create_marginalized_hist(ax, values, label, percentiles=None,
         orientation = 'horizontal'
     else:
         orientation = 'vertical'
-    ax.hist(values, bins=50, histtype=htype, orientation=orientation,
+    if log_scale:
+        bins = numpy.logspace(
+            numpy.log10(numpy.nanmin(values)),
+            numpy.log10(numpy.nanmax(values)),
+            50
+        )
+    else:
+        bins = numpy.linspace(
+            numpy.nanmin(values),
+            numpy.nanmax(values),
+            50,
+        )
+    ax.hist(values, bins=bins, histtype=htype, orientation=orientation,
             facecolor=fillcolor, edgecolor=color, ls=linestyle, lw=2,
             density=True)
     if percentiles is None:
@@ -545,6 +559,7 @@ def create_multidim_plot(parameters, samples, labels=None,
                          marginal_title=True, marginal_linestyle='-',
                          zvals=None, show_colorbar=True, cbar_label=None,
                          vmin=None, vmax=None, scatter_cmap='plasma',
+                         scatter_log_cmap=False, log_parameters=None,
                          plot_density=False, plot_contours=True,
                          density_cmap='viridis',
                          contour_color=None, label_contours=True,
@@ -614,6 +629,10 @@ def create_multidim_plot(parameters, samples, labels=None,
         zvals.
     scatter_cmap : {'plasma', string}
         The color map to use for the scatter points. Default is 'plasma'.
+    scatter_log_cmap : boolean
+        Should the scatter point coloring be on a log scale? Default False
+    log_parameters : list or None
+        Which parameters should be plotted on a log scale
     plot_density : {False, bool}
         Plot the density of points as a color map.
     plot_contours : {True, bool}
@@ -649,6 +668,8 @@ def create_multidim_plot(parameters, samples, labels=None,
     """
     if labels is None:
         labels = {p: p for p in parameters}
+    if log_parameters is None:
+        log_parameters = []
     # set up the figure with a grid of axes
     # if only plotting 2 parameters, make the marginal plots smaller
     nparams = len(parameters)
@@ -732,6 +753,7 @@ def create_multidim_plot(parameters, samples, labels=None,
             create_marginalized_hist(
                 ax, samples[param], label=labels[param],
                 color=hist_color, fillcolor=fill_color,
+                log_scale=param in log_parameters,
                 plot_marginal_lines=plot_marginal_lines,
                 linestyle=marginal_linestyle, linecolor=line_color,
                 title=marginal_title, expected_value=expected_value,
@@ -749,8 +771,13 @@ def create_multidim_plot(parameters, samples, labels=None,
                 alpha = 0.3
             else:
                 alpha = 1.
+            if scatter_log_cmap:
+                cmap_norm = colors.LogNorm(vmin=vmin, vmax=vmax)
+            else:
+                cmap_norm = colors.Normalize(vmin=vmin, vmax=vmax)
+
             plt = ax.scatter(x=samples[px], y=samples[py], c=zvals, s=5,
-                             edgecolors='none', vmin=vmin, vmax=vmax,
+                             edgecolors='none', norm=cmap_norm,
                              cmap=scatter_cmap, alpha=alpha, zorder=2)
 
         if plot_contours or plot_density: