Renaming, bugfix

src/xia2/Modules/SSX/batch_scale.py

@@ -28,10 +28,12 @@
 from xia2.Modules.SSX.data_reduction_definitions import FilePair
 
 
-class BatchScale(object):
+class BatchScaleReindex(object):
     def __init__(
         self, batches: List[FilePair], reference: Path, space_group: sgtbx.space_group
     ):
+        # The purpose here is to do quick KB scaling of batches of datasets,
+        # for the purpose of testing reindexing against a reference.
         phil_scope = phil.parse(
             """
     include scope dials.command_line.scale.phil_scope
@@ -48,9 +50,11 @@ def __init__(
         params.reflection_selection.method = "intensity_ranges"
         params.reflection_selection.Isigma_range = [2, 0]
         params.cut_data.partiality_cutoff = 0.25
+        params.output.html = None
         self.params = params
         self.input_sg = space_group
         self.reference = reference
+        assert self.reference is not None
         self.input_batches = batches
         self._experiments = ExperimentList([])
         self._reflections = []
@@ -60,7 +64,9 @@ def __init__(
         # create a single table and expt per batch
         all_expts = ExperimentList([])
 
-        class SANoStats(ScalingAlgorithm):
+        class ScalingAlgorithmNoStats(ScalingAlgorithm):
+            # merging stats calculation is very slow for lots of data.
+            # avoid as not needed here.
             def calculate_merging_stats(self):
                 pass
 
@@ -72,8 +78,9 @@ def calculate_merging_stats(self):
                 table["intensity.sum.value"] / (table["intensity.sum.variance"] ** 0.5)
             ) > 2.0
             n_sel = sel.count(True)
-            if (n_sel / sel.size()) < 0.05 and n_sel > 10:
-                params.reflection_selection.Isigma_range = [0, 0]
+            if (n_sel / sel.size()) < 0.05 or n_sel < 10:  # revert to safe defaults
+                self.params.reflection_selection.Isigma_range = [-5, 0]
+                self.params.reflection_selection.method = "quasi_random"
             else:
                 table = table.select(
                     (
@@ -98,37 +105,42 @@ def calculate_merging_stats(self):
             table.experiment_identifiers()[i] = str(i)
             self._reflections.append(table)
 
-        wavelength = np.mean([expt.beam.get_wavelength() for expt in expts])
+        self.wavelength = np.mean([expt.beam.get_wavelength() for expt in expts])
         best_unit_cell = determine_best_unit_cell(all_expts)
         for expt in self._experiments:
             expt.crystal.set_unit_cell(best_unit_cell)
-            expt.beam.set_wavelength(wavelength)
+            expt.beam.set_wavelength(self.wavelength)
 
-        self.algorithm = SANoStats(params, self._experiments, self._reflections)
+        self.algorithm = ScalingAlgorithmNoStats(
+            self.params, self._experiments, self._reflections
+        )
 
     def run(self):
         self.algorithm.run()
         del self._reflections
-        if self.reference:
-            wavelength = np.mean(
-                [expt.beam.get_wavelength() for expt in self._experiments]
-            )
-            reference_miller_set = intensities_from_reference_file(
-                os.fspath(self.reference), wavelength=wavelength
-            )
-            test_miller_set = self.algorithm.scaled_miller_array
-            change_of_basis_op = determine_reindex_operator_against_reference(
-                test_miller_set, reference_miller_set
-            )
+        reference_miller_set = intensities_from_reference_file(
+            os.fspath(self.reference), wavelength=self.wavelength
+        )
+        test_miller_set = self.algorithm.scaled_miller_array
+        self.change_of_basis_op = determine_reindex_operator_against_reference(
+            test_miller_set, reference_miller_set
+        )
+        if self.change_of_basis_op.as_abc() != "a,b,c":
             for i, fp in enumerate(self.input_batches):
                 expts = load.experiment_list(fp.expt, check_format=False)
                 refl = flex.reflection_table.from_file(fp.refl)
                 for expt in expts:
-                    expt.crystal = expt.crystal.change_basis(change_of_basis_op)
+                    expt.crystal = expt.crystal.change_basis(self.change_of_basis_op)
                     expt.crystal.set_space_group(self.input_sg)
                 expts.as_file(f"processed_{i}.expt")
 
-                refl["miller_index"] = change_of_basis_op.apply(refl["miller_index"])
+                refl["miller_index"] = self.change_of_basis_op.apply(
+                    refl["miller_index"]
+                )
                 refl.as_file(f"processed_{i}.refl")
                 self._output_expt_files.append(f"processed_{i}.expt")
                 self._output_refl_files.append(f"processed_{i}.refl")
+        else:  # don't need to do anything
+            for i, fp in enumerate(self.input_batches):
+                self._output_expt_files.append(fp.expt)
+                self._output_refl_files.append(fp.refl)

src/xia2/Modules/SSX/data_reduction_programs.py

@@ -39,7 +39,7 @@
 from xia2.Driver.timing import record_step
 from xia2.Handlers.Files import FileHandler
 from xia2.Modules.SSX.batch_cosym import BatchCosym
-from xia2.Modules.SSX.batch_scale import BatchScale
+from xia2.Modules.SSX.batch_scale import BatchScaleReindex
 from xia2.Modules.SSX.data_reduction_definitions import FilePair, ReductionParams
 from xia2.Modules.SSX.reporting import condensed_unit_cell_info
 from xia2.Modules.SSX.util import log_to_file, run_in_directory
@@ -888,7 +888,7 @@ def scale_reindex(
     with run_in_directory(working_directory), record_step("scale_reindex"), log_to_file(
         logfile
     ):
-        s = BatchScale(batches_to_scale, reference, space_group)
+        s = BatchScaleReindex(batches_to_scale, reference, space_group)
         s.run()
     xia2_logger.info("Reindexed against reference file")
     outfiles = []