diff --git a/megaradrp/recipes/calibration/trace.py b/megaradrp/recipes/calibration/trace.py
index cdfe4c0..664997d 100644
--- a/megaradrp/recipes/calibration/trace.py
+++ b/megaradrp/recipes/calibration/trace.py
@@ -133,11 +133,11 @@ def check_qc_tracemap(self, tracemap):
                 continue
             if trace.start > start_min:
                 cost_ids[idx] += 1
-                msg = 'In fiber {}, trace start > {}'.format(trace.fibid, start_min)
+                msg = f'In fiber {trace.fibid}, trace start > {start_min}'
                 warnings.warn(msg)
             if trace.stop < end_max:
                 cost_ids[idx] += 1
-                msg = 'In fiber {}, trace end < {}'.format(trace.fibid, end_max)
+                msg = f'In fiber {trace.fibid}, trace end < {end_max}'
                 warnings.warn(msg)
 
         total = sum(cost_ids)
@@ -149,7 +149,7 @@ def check_qc_tracemap(self, tracemap):
         else:
             tracemap.quality_control = qc.QC.GOOD
 
-    #@numina.core.validator.validate
+    # @numina.core.validator.validate
     def run(self, rinput):
         """Execute the recipe.
 
@@ -295,8 +295,8 @@ def obtain_boxes_from_image(self, reduced, expected, npeaks, cstart=2000):
         res = numpy.fft.ifft(yv)
         final = res.real
         plt.plot(final)
-        #trend = detrend(final)
-        #plt.plot(final - trend)
+        # trend = detrend(final)
+        # plt.plot(final - trend)
         plt.show()
 
         idx = find_peaks_indexes(final, window_width=3, threshold=0.3, fpeak=1)
@@ -311,7 +311,7 @@ def obtain_boxes_from_image(self, reduced, expected, npeaks, cstart=2000):
         nidxs = numpy.sort(nidx)
 
         plt.plot(final)
-        #plt.scatter(idx, [0.9 for m in idx])
+        # plt.scatter(idx, [0.9 for m in idx])
         plt.scatter(nidx, [0.95 for m in nidx], c='r')
         plt.scatter(expected, [1.0 for m in expected])
         plt.show()
@@ -379,14 +379,15 @@ def refine_boxes_from_image(self, reduced, expected, cstart=2000, nsearch=20):
         ioffset = xcorr[ycorr.index(max(ycorr))]
 
         # auxiliary plot showing the cross-correlation work
-        fig, ax = plt.subplots(ncols=1, nrows=1)
-        ax.plot(xcorr, ycorr, 'o-')
-        ax.set_xlabel('ioffset')
-        ax.set_ylabel('peak of corrrelation function')
-        ax.axvline(ioffset, linestyle=':', color='C1')
-        ax.set_title('optimal initial offset: {}'.format(ioffset))
-        plt.savefig('offset_borders_cross.png')
-        plt.close()
+        if self.intermediate_results:
+            fig, ax = plt.subplots(ncols=1, nrows=1)
+            ax.plot(xcorr, ycorr, 'o-')
+            ax.set_xlabel('ioffset')
+            ax.set_ylabel('peak of corrrelation function')
+            ax.axvline(ioffset, linestyle=':', color='C1')
+            ax.set_title('optimal initial offset: {}'.format(ioffset))
+            plt.savefig('offset_borders_cross.png')
+            plt.close()
 
         # using the initial offset, refine the peak search around the new
         # expected location, looking for a maximum in +/- nsearch pixels
@@ -398,21 +399,22 @@ def refine_boxes_from_image(self, reduced, expected, cstart=2000, nsearch=20):
             refined[ibox] = iargmax + box - nsearch + ioffset
 
         # auxiliary plot showing the initial and final frontier locations
-        fig, ax = plt.subplots(ncols=1, nrows=1)
-        ax.plot(final, label='cross section at x={}'.format(cstart))
-        ax.plot(xwave, sp_comb_lines0, label='expected location of frontiers')
-        for idum, item in enumerate(expected):
-            if idum == 0:
-                label = 'refined frontier location'
-            else:
-                label = None
-            ax.plot(refined, final[refined], 'ro', label=label)
-        ax.legend()
-        ax.set_title('optimal initial offset: {}'.format(ioffset))
-        ax.set_xlabel('pixel number - 1')
-        ax.set_ylabel('inverted normalized signal')
-        plt.savefig('frontiers_between_pseudoslits.png')
-        plt.close()
+        if self.intermediate_results:
+            fig, ax = plt.subplots(ncols=1, nrows=1)
+            ax.plot(final, label='cross section at x={}'.format(cstart))
+            ax.plot(xwave, sp_comb_lines0, label='expected location of frontiers')
+            for idum, item in enumerate(expected):
+                if idum == 0:
+                    label = 'refined frontier location'
+                else:
+                    label = None
+                ax.plot(refined, final[refined], 'ro', label=label)
+            ax.legend()
+            ax.set_title('optimal initial offset: {}'.format(ioffset))
+            ax.set_xlabel('pixel number - 1')
+            ax.set_ylabel('inverted normalized signal')
+            plt.savefig('frontiers_between_pseudoslits.png')
+            plt.close()
 
         return refined, cstart
 
@@ -480,10 +482,10 @@ def search_traces(self, reduced, boxes, box_borders, inactive_fibers=None, cstar
 
                     if debug_plot:
                         plt.plot(mm[:, 0], mm[:, 1], '.')
-                        plt.savefig('trace-xy-{:03d}.png'.format(dtrace.fibid))
+                        plt.savefig(f'trace-xy-{dtrace.fibid:03d}.png')
                         plt.close()
                         plt.plot(mm[:, 0], mm[:, 2], '.')
-                        plt.savefig('trace-xz-{:03d}.png'.format(dtrace.fibid))
+                        plt.savefig(f'trace-xz-{dtrace.fibid:03d}.png')
                         plt.close()
                     if len(mm) < poldeg + 1:
                         self.logger.warning('in fibid %d, only %d points to fit pol of degree %d',
@@ -621,8 +623,8 @@ def init_traces(image, center, hs, boxes, box_borders, tol=1.5, threshold=0.37,
             plt.plot(ipeaks_int+b1, region[ipeaks_int], 'r*')
             plt.xlabel('pixel number - 1')
             plt.ylabel('number of counts')
-            plt.title('pseudoslit box: {}, id: {}'.format(box['name'], boxid))
-            plt.savefig('central_cut_{:02d}.png'.format(boxid))
+            plt.title(f'pseudoslit box: {box["name"]}, id: {boxid}')
+            plt.savefig(f'central_cut_{boxid:02d}.png')
             plt.close()
 
         startid = lastid + 1