DM-56184: Various fixes from Rubin

spectractor/extractor/background.py

@@ -38,7 +38,8 @@ def remove_image_background_sextractor(data, sigma=3.0, box_size=(50, 50), filte
         plot_image_simple(ax[0], bkg.background, scale="lin")
         plot_image_simple(ax[1], data_wo_bkg, scale="symlog")
         fig.tight_layout()
-        plt.show()
+        if parameters.DISPLAY:
+            plt.show()
         if parameters.PdfPages:
             parameters.PdfPages.savefig()
     return data_wo_bkg

spectractor/extractor/extractor.py

@@ -1781,4 +1781,4 @@ def plot_comparison_truth(spectrum, w):  # pragma: no cover
         plt.show()
     if parameters.LSST_SAVEFIGPATH:
         fig.savefig(os.path.join(parameters.LSST_SAVEFIGPATH, 'deconvolution_truth.pdf'), transparent=True)
-    plt.show()
+

spectractor/extractor/images.py

@@ -929,7 +929,8 @@ def find_target(image, guess=None, rotated=False, widths=[parameters.XWINDOW, pa
             plt.figure(figsize=(5, 5))
             plot_image_simple(plt.gca(), data=sub_image_subtracted, scale="lin", title="", units=image.units,
                                 target_pixcoords=[theX - x0 + Dx, theX - x0 + Dx])
-            plt.show()
+            if parameters.DISPLAY:
+                plt.show()
         if parameters.PdfPages:
             parameters.PdfPages.savefig()
 

spectractor/extractor/spectroscopy.py

@@ -485,7 +485,7 @@ def build_detected_line_table(self, amplitude_units="", calibration_only=False):
             t = Table(rows=rows, names=(
                 'Line', 'Tabulated', 'Detected', 'Shift', 'Err', 'FWHM', 'Amplitude', 'SNR', 'Chisq', 'Eqwidth_mod',
                 'Eqwidth_data'),
-                      dtype=('a12', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4'))
+                      dtype=('S12', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4', 'f4'))
             for col in t.colnames[1:6]:
                 t[col].unit = 'nm'
             t[t.colnames[5]].unit = amplitude_units

spectractor/extractor/spectrum.py

@@ -622,7 +622,8 @@ def generate_axes(fig):
         fig.subplots_adjust(hspace=0)
         if save_as:
             plt.savefig(save_as)
-        plt.show()
+        if parameters.DISPLAY:
+            plt.show()
 
     def save_spectrum(self, output_file_name, overwrite=False):
         """Save the spectrum into a fits file (data, error and wavelengths).

spectractor/fit/fit_multispectra.py

@@ -790,7 +790,8 @@ def plot_A1s(self):  # pragma: no cover
         plt.legend()
         if parameters.SAVE:
             fig.savefig(self.output_file_name + '_A1s.pdf', dpi=100, bbox_inches='tight')
-        plt.show()
+        if parameters.DISPLAY:
+            plt.show()
 
     def save_transmissions(self):
         aerosols, angstrom_exponent, ozone, pwv, reso, *A1s = self.params.values

spectractor/fit/fitter.py

@@ -1605,7 +1605,8 @@ def simple_newton_minimisation(fit_workspace, niter=10, xtol=1e-3, ftol=1e-3):
         plt.plot(r, js, label="J")
         plt.grid()
         plt.legend()
-        plt.show()
+        if parameters.DISPLAY:
+            plt.show()
 
         if parameters.DISPLAY:
             fig = plt.figure()

spectractor/fit/mcmc.py

@@ -85,13 +85,13 @@ def getMaximum(self):
 
     def getTotal(self):
         if len(self.axes[-1].axis) > 1:
-            self.total = np.trapz(y=self.grid, x=self.axes[-1].axis, axis=self.rangedim[-1])
+            self.total = np.trapezoid(y=self.grid, x=self.axes[-1].axis, axis=self.rangedim[-1])
         else:
             self.total = np.sum(self.grid)
         if self.dim > 1:
             for i in reversed(self.rangedim[:-1]):
                 if len(self.axes[i].axis) > 1:
-                    self.total = np.trapz(y=self.total, x=self.axes[i].axis, axis=i)
+                    self.total = np.trapezoid(y=self.total, x=self.axes[i].axis, axis=i)
                 else:
                     self.total = np.sum(self.total, axis=i)
         return self.total
@@ -106,7 +106,7 @@ def normalize(self):
 
     def marginalizeAlongAxis(self, axis_index):
         # return(np.sum(self.grid,axis=axis_index))
-        return np.trapz(self.grid, self.axes[axis_index].axis, axis=axis_index)
+        return np.trapezoid(self.grid, self.axes[axis_index].axis, axis=axis_index)
 
 
 class PDF(Grid):
@@ -174,7 +174,7 @@ def stats(self, verbose=True):
         xprod = np.zeros_like(self.grid)
         for xindex, x in enumerate(self.axe.axis):
             xprod[xindex] = x * self.grid[xindex]
-        self.mean = np.trapz(xprod, x=self.axe.axis)
+        self.mean = np.trapezoid(xprod, x=self.axe.axis)
         # cumprob = np.cumsum(self.grid)
         cumprob = np.zeros_like(self.grid)
         cumprob[0] = self.grid[0]
@@ -226,7 +226,7 @@ def stats(self, verbose=True):
         xxprod = np.zeros_like(self.grid)
         for xindex, x in enumerate(self.axe.axis):
             xxprod[xindex] = self.grid[xindex] * (x - self.mean) ** 2
-        self.variance = np.trapz(xxprod, self.axe.axis)
+        self.variance = np.trapezoid(xxprod, self.axe.axis)
         txt = "%s: %s +%s -%s (std: %s)" % formatting_numbers(self.mean, self.error_high, self.error_low,
                                                                 std=np.sqrt(self.variance), label=self.label)
         self.title = '$%s^{+%s}_{-%s}$' % formatting_numbers(self.mean, self.error_high, self.error_low)
@@ -254,9 +254,9 @@ def fill_histogram(self, data1, data2, weights=None):
 
     def covariance(self):
         self.normalize()
-        pdf = np.trapz(y=self.grid, x=self.axes[1].axis, axis=1)
+        pdf = np.trapezoid(y=self.grid, x=self.axes[1].axis, axis=1)
         self.pdfs[0].fill(pdf)
-        pdf = np.trapz(y=self.grid, x=self.axes[0].axis, axis=0)
+        pdf = np.trapezoid(y=self.grid, x=self.axes[0].axis, axis=0)
         self.pdfs[1].fill(pdf)
         if self.pdfs[0].max_pdf is None:
             self.pdfs[0].stats(verbose=False)
@@ -270,7 +270,7 @@ def covariance(self):
         for xindex, x in enumerate(self.axes[0].axis):
             for yindex, y in enumerate(self.axes[1].axis):
                 xyprod[xindex][yindex] = (x - self.pdfs[0].mean) * (y - self.pdfs[1].mean) * self.grid[xindex][yindex]
-        self.cov = np.trapz(np.trapz(y=xyprod, x=self.axes[1].axis, axis=1), self.axes[0].axis, axis=0)
+        self.cov = np.trapezoid(np.trapezoid(y=xyprod, x=self.axes[1].axis, axis=1), self.axes[0].axis, axis=0)
         self.rho = self.cov / (np.sqrt(self.pdfs[0].variance * self.pdfs[1].variance))
         return self.cov
 
@@ -346,7 +346,8 @@ def plot(self, plot=False, truth=None):
         if plot:
             plt.xlabel(self.axis_names[0])
             plt.ylabel(self.axis_names[1])
-            plt.show()
+            if parameters.DISPLAY:
+                plt.show()
         if parameters.PdfPages:
             parameters.PdfPages.savefig()
 
@@ -394,12 +395,12 @@ def marginalize(self, pdfonly=False):  # pragma: no cover
                         if len(self.axes[axes[-1]].axis) == 1:
                             pdf = np.sum(self.grid, axis=axes[-1])
                         else:
-                            pdf = np.trapz(y=self.grid, x=self.axes[axes[-1]].axis, axis=axes[-1])
+                            pdf = np.trapezoid(y=self.grid, x=self.axes[axes[-1]].axis, axis=axes[-1])
                         for axe in reversed(axes[:-1]):
                             if len(self.axes[axe].axis) == 1:
                                 pdf = np.sum(pdf, axis=axe)
                             else:
-                                pdf = np.trapz(y=pdf, x=self.axes[axe].axis, axis=axe)
+                                pdf = np.trapezoid(y=pdf, x=self.axes[axe].axis, axis=axe)
                         self.pdfs[i].fill(pdf)
                     else:
                         self.pdfs[i].fill(self.grid)
@@ -409,13 +410,13 @@ def marginalize(self, pdfonly=False):  # pragma: no cover
                             if len(self.axes[axes[-1]].axis) == 1:
                                 self.contours[i][j].grid = np.sum(self.grid, axis=axes[-1])
                             else:
-                                self.contours[i][j].grid = np.trapz(y=self.grid, x=self.axes[axes[-1]].axis,
+                                self.contours[i][j].grid = np.trapezoid(y=self.grid, x=self.axes[axes[-1]].axis,
                                                                     axis=axes[-1])
                             for axe in reversed(axes[:-1]):
                                 if len(self.axes[axe].axis) == 1:
                                     self.contours[i][j].grid = np.sum(self.contours[i][j].grid, axis=axe)
                                 else:
-                                    self.contours[i][j].grid = np.trapz(y=self.contours[i][j].grid,
+                                    self.contours[i][j].grid = np.trapezoid(y=self.contours[i][j].grid,
                                                                         x=self.axes[axe].axis, axis=axe)
                             if i < j:
                                 self.contours[i][j].grid = self.contours[i][j].grid.T
@@ -962,5 +963,5 @@ def run_emcee(mcmc_fit_workspace, ln=lnprob):
             p0 = sampler.get_last_sample()
         for _ in sampler.sample(p0, iterations=max(0, nsamples - backend.iteration), progress=True, store=True):
             continue
-    mcmc_fit_workspace.chains = sampler.chain
-    mcmc_fit_workspace.lnprobs = sampler.lnprobability
+    mcmc_fit_workspace.chains = sampler.get_chain()
+    mcmc_fit_workspace.lnprobs = sampler.get_log_prob()

spectractor/parameters.py

@@ -140,8 +140,8 @@ def __getattr__(name):
 
 # Plots
 DISPLAY = True
-if os.environ.get('DISPLAY', '') == '':
-    mpl.use('agg')
+backend = mpl.get_backend()
+if os.environ.get('DISPLAY', '') == '' or backend.lower() in ["agg", "pdf", "ps", "svg", "pgf", "cairo"]:
     DISPLAY = False
 PLOT_XLABEL = r"$x$ [pixels]"
 PLOT_YLABEL = r"$y$ [pixels]"

ups/spectractor.table

@@ -1 +1,3 @@
+setupRequired(utils)
+
 envPrepend(PYTHONPATH, ${PRODUCT_DIR}/lib/python)