Added CASTEP, repo normalizer checks multiple sections.

.gitmodules

@@ -58,3 +58,6 @@
 [submodule "dependencies/parsers/orca"]
 	path = dependencies/parsers/orca
 	url = https://gitlab.mpcdf.mpg.de/nomad-lab/parser-orca
+[submodule "dependencies/parsers/castep"]
+	path = dependencies/parsers/castep
+	url = https://gitlab.mpcdf.mpg.de/nomad-lab/parser-castep

.vscode/launch.json

@@ -157,7 +157,16 @@
         "-sv", "tests/test_normalizing.py::test_normalizer[parsers/abinit-tests/data/parsers/abinit/Fe.out]"
       ]
     },
-
+    {
+      "name": "Castep Normalizer",
+      "type": "python",
+      "request": "launch",
+      "cwd": "${workspaceFolder}",
+      "program": "${workspaceFolder}/.pyenv/bin/pytest",
+      "args": [
+        "-sv", "tests/test_normalizing.py::test_normalizer[parsers/castep-tests/data/parsers/castep/BC2N-Pmm2-Raman.castep]"
+      ]
+    },
     {
       "name": "Python: Current File",
       "type": "python",

castep @ b42810de

+Subproject commit b42810dee0ef7eb97cc5de41f4b5b110f21ab374

nomad/normalizing/repository.py

@@ -52,11 +52,40 @@ class RepositoryNormalizer(Normalizer):
         else:
             return match.group(0)
 
-    def get_optional_value(self, key):
-        try:
-            return self._backend.get_value(key, 0)
-        except KeyError:
+    def get_optional_value(self, key, section):
+        # Section is section_system, section_symmetry, etc...
+        val = None  # Initialize to None, so we can compare section values.
+        diff_flag = False  # Flag to store whether vals differ between sections.
+        # Loop over the sections with the name section in the backend.
+        for section_index in self._backend.get_sections(section):
+            try:
+                new_val = self._backend.get_value(key, section_index)
+            except KeyError:
+                continue
+
+            # Compare values from iterations.
+            diff_bool = new_val != val
+
+            if type(diff_bool) is bool:
+                if diff_bool and val is not None:
+                    diff_flag = True
+            elif diff_bool.all() and (val is not None):
+                # Then we have an array, and diff bool has multiple values since
+                # each item in array has been compared item for item.
+                diff_flag = True
+
+            val = new_val
+
+        if diff_flag is True:
+            self.logger.warning(
+                'The values for %s differ between different %s' % (key, section))
+
+        if val is None:
+            self.logger.warning(
+                'The values for %s where not available in any %s' % (key, section))
             return unavailable_label
+        else:
+            return val
 
     def normalize(self, logger=None) -> None:
         super().normalize(logger)
@@ -75,19 +104,33 @@ class RepositoryNormalizer(Normalizer):
             'repository_code_version',
             self.simplify_version(b.get_value('program_version', 0)))
         b.addValue('repository_parser_id', b.get_value('parser_name', 0))
-        b.addValue('repository_chemical_formula', b.get_value('chemical_composition_bulk_reduced', 0))
-        atom_labels = b.get_value('atom_labels', 0)
+        # We add these values as optional as some parser developers may create parsers
+        # where the first section contains only properties that are constant over
+        # several simulations and use the other sections to define simulation specific
+        # parameters. Ex. in first section_system CASTEP parsers defines # of elections
+        # and in subsequent sections it defines atom labels, positions, etc...
+        atom_labels = self.get_optional_value('atom_labels', 'section_system')
         b.addValue('repository_atomic_elements', list(set(atom_labels)))
         b.addValue('repository_atomic_elements_count', len(atom_labels))
-        b.addValue('repository_system_type', b.get_value('system_type', 0))
-
-        b.addValue('repository_crystal_system', self.get_optional_value('crystal_system'))
-        b.addValue('repository_spacegroup_nr', self.get_optional_value('space_group_number'))
-
-        b.addValue('repository_basis_set_type', self.get_optional_value('program_basis_set_type'))
+        b.addValue(
+            'repository_crystal_system',
+            self.get_optional_value('crystal_system', 'section_symmetry'))
+        b.addValue(
+            'repository_spacegroup_nr',
+            self.get_optional_value('space_group_number', 'section_symmetry'))
+        b.addValue(
+            'repository_basis_set_type',
+            self.get_optional_value('program_basis_set_type', 'section_run'))
+        b.addValue(
+            'repository_system_type',
+            self.get_optional_value('system_type', 'section_system'))
+        b.addValue(
+            'repository_chemical_formula',
+            self.get_optional_value('chemical_composition_bulk_reduced', 'section_system'))
         b.addValue(
             'repository_xc_treatment',
-            self.map_functional_name_to_xc_treatment(self.get_optional_value(('XC_functional_name'))))
+            self.map_functional_name_to_xc_treatment(
+                self.get_optional_value('XC_functional_name', 'section_method')))
 
         b.closeNonOverlappingSection('section_repository_parserdata')
         if repository_info_context is None:

nomad/normalizing/system.py

@@ -48,10 +48,10 @@ class SystemNormalizer(SystemBasedNormalizer):
         self.atom_labels = section_system['atom_labels']
         self.atom_species = section_system['atom_atom_numbers']
         self.atom_positions = section_system['atom_positions']
-        # Assume if no config for periodic dimensions that we simply don't have any.
-        # TODO: @dts, this may be something we should change since many DFT programs
-        # have implicit periodicity. Talk with Georg/Claudia.
         self.periodic_dirs = section_system.get('configuration_periodic_dimensions', None)
+        if self.periodic_dirs is None:
+            self.logger.warning(
+                'Unable to get PBCs in this section_system, assume False, False, Fasle')
         # Try to first read the cell information from the renamed metainfo
         # lattice_vectors, if this doesn't work try the depreciated name
         # simulation_cell. Otherwise, if neither are present, assign None.
@@ -119,7 +119,6 @@ class SystemNormalizer(SystemBasedNormalizer):
             self.atom_species = [
                 atom_label_to_num(atom_label) for atom_label in self.atom_labels
             ]
-
         formula = None
 
         if self.atom_species:
@@ -133,7 +132,6 @@ class SystemNormalizer(SystemBasedNormalizer):
                 formula_bulk = formula_reduced
             else:
                 formula_bulk = formula
-
         if self.cell is not None:
             results['lattice_vectors'] = self.cell
 
@@ -149,7 +147,6 @@ class SystemNormalizer(SystemBasedNormalizer):
         # TODO: @dts, might be good to clean this up so it is more readable in the
         # future.
         configuration_id = 's' + addShasOfJson(results).b64digests()[0][0:28]
-
         self._backend.addValue('configuration_raw_gid', configuration_id)
         self._backend.addValue('atom_species', self.atom_species)
         self._backend.addValue('chemical_composition', formula)
@@ -309,4 +306,4 @@ class SystemNormalizer(SystemBasedNormalizer):
         if nomad_classification == 'Atom' and (len(self.atom_labels) > 1):
             nomad_classification = 'Molecule / Cluster'
 
-        return nomad_classification
\ No newline at end of file
+        return nomad_classification

nomad/parsing/__init__.py

@@ -269,6 +269,11 @@ parsers = [
             r'\s+\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\**\s*'
             r'\s*'
             r'\s*--- An Ab Initio, DFT and Semiempirical electronic structure package ---\s*')
+    ),
+    LegacyParser(
+        name='parsers/castep',
+        parser_class_name='castepparser.CastepParser',
+        mainfile_contents_re=(r'\s\|\s*CCC\s*AA\s*SSS\s*TTTTT\s*EEEEE\s*PPPP\s*\|\s*')
     )
 ]
 

tests/test_parsing.py

@@ -43,14 +43,15 @@ parser_examples = [
     ('parsers/gaussian', 'tests/data/parsers/gaussian/aniline.out'),
     ('parsers/abinit', 'tests/data/parsers/abinit/Fe.out'),
     ('parsers/quantumespresso', 'tests/data/parsers/quantum-espresso/benchmark.out'),
-    ('parsers/orca', 'tests/data/parsers/orca/orca3dot2706823.out')
+    ('parsers/orca', 'tests/data/parsers/orca/orca3dot2706823.out'),
+    ('parsers/castep', 'tests/data/parsers/castep/BC2N-Pmm2-Raman.castep')
 ]
 
 faulty_unknown_one_d_matid_example = [
     ('parsers/template', 'tests/data/normalizers/no_sim_cell_boolean_positions.json')
 ]
 
-correct_num_output_files = 21
+correct_num_output_files = 22
 
 
 class TestLocalBackend(object):