Added metainfo and normalization for DOS, compressed test files, added regtests for DOS.

nomad/metainfo/encyclopedia.py

@@ -537,6 +537,41 @@ class ElectronicBandStructure(MSection):
     )
 
 
+class ElectronicDOS(MSection):
+    m_def = Section(
+        a_flask=dict(skip_none=True),
+        a_elastic=dict(type=InnerDoc),
+        description="""
+        Store the electronic density of states (DOS).
+        """
+    )
+    fermi_level = Quantity(
+        type=float,
+        unit=units.J,
+        description="""
+        Fermi level reported for the density of states.
+        """
+    )
+    energies = Quantity(
+        type=np.dtype('f8'),
+        shape=["1..*"],
+        unit=units.J,
+        description="""
+        Array containing the set of discrete energy values with respect to the
+        top of the valence band for the density of states (DOS).
+        """
+    )
+    values = Quantity(
+        type=np.dtype('f8'),
+        shape=["1..2", "1..*"],
+        unit=units.J**(-1),
+        description="""
+        Values (number of states for a given energy, the set of discrete energy
+        values is given in dos_energies) of density of states.
+        """
+    )
+
+
 class Properties(MSection):
     m_def = Section(
         a_flask=dict(skip_none=True),
@@ -566,13 +601,8 @@ class Properties(MSection):
         formula unit.
         """
     )
-    dos = Quantity(
-        type=str,
-        description="""
-        A JSON encoded string that contains the density of states.
-        """
-    )
     electronic_band_structure = SubSection(sub_section=ElectronicBandStructure.m_def, repeats=False)
+    electronic_dos = SubSection(sub_section=ElectronicDOS.m_def, repeats=False)
 
 
 class Encyclopedia(MSection):

nomad/normalizing/encyclopedia.py

@@ -26,9 +26,30 @@ import numpy as np
 from matid import SymmetryAnalyzer
 import matid.geometry
 
-from nomad.normalizing.normalizer import Normalizer, s_scc, s_system, s_method, s_frame_sequence, r_frame_sequence_to_sampling, s_sampling_method, r_frame_sequence_local_frames
+from nomad.normalizing.normalizer import (
+    Normalizer,
+    s_scc,
+    s_system,
+    s_method,
+    s_frame_sequence,
+    r_frame_sequence_to_sampling,
+    s_sampling_method,
+    r_frame_sequence_local_frames,
+)
+from nomad.metainfo.encyclopedia import (
+    Encyclopedia,
+    Material,
+    Method,
+    Properties,
+    RunType,
+    WyckoffSet,
+    WyckoffVariables,
+    ElectronicBandStructure,
+    ElectronicDOS,
+    BandGap,
+    BandSegment,
+)
 from nomad.normalizing.settingsbasisset import SettingsBasisSet
-from nomad.metainfo.encyclopedia import Encyclopedia, Material, Method, Properties, RunType, WyckoffSet, WyckoffVariables, ElectronicBandStructure, BandGap, BandSegment
 from nomad.normalizing import structure
 from nomad.utils import hash
 from nomad import config
@@ -1682,7 +1703,27 @@ class PropertiesNormalizer():
                 properties.m_add_sub_section(Properties.electronic_band_structure, band_structure)
 
     def dos(self) -> None:
-        pass
+        dos_sections = self.backend.get('section_dos')
+        if dos_sections:
+            # There should be only one section dos
+            dos_data = dos_sections[0]
+            dos_kind = dos_data.get('dos_kind', None)
+
+            # dos_kind might not be given for electronic DOS
+            if dos_kind is None:
+                self.logger.warning("dos_kind is not specified, assuming electronic DOS")
+            elif dos_kind == 'vibrational':
+                return None
+
+            # Use only normalized data
+            dos_energies = dos_data.get('dos_energies_normalized')
+            dos_values = dos_data.get('dos_values')
+            if dos_energies is not None and dos_values is not None:
+                sec_enc = self.backend.get_mi2_section(Encyclopedia.m_def)
+                properties = sec_enc.properties
+                dos = properties.m_create(ElectronicDOS)
+                dos.energies = dos_energies
+                dos.values = dos_values
 
     def elastic_constants_matrix(self) -> None:
         pass
@@ -1752,3 +1793,4 @@ class PropertiesNormalizer():
         self.context = ctx
         self.band_structure()
         self.energies()
+        self.dos()

tests/normalizing/conftest.py

@@ -191,3 +191,21 @@ def bands_polarized_gap_indirect() -> LocalBackend:
     backend = parse_file((parser_name, filepath))
     backend = run_normalize(backend)
     return backend
+
+
+@pytest.fixture(scope='session')
+def dos_polarized_vasp() -> LocalBackend:
+    parser_name = "parsers/vasp"
+    filepath = "tests/data/normalizers/dos/polarized_vasp/vasprun.xml.relax2.xz"
+    backend = parse_file((parser_name, filepath))
+    backend = run_normalize(backend)
+    return backend
+
+
+@pytest.fixture(scope='session')
+def dos_unpolarized_vasp() -> LocalBackend:
+    parser_name = "parsers/vasp"
+    filepath = "tests/data/normalizers/dos/unpolarized_vasp/vasprun.xml.xz"
+    backend = parse_file((parser_name, filepath))
+    backend = run_normalize(backend)
+    return backend

tests/normalizing/test_encyclopedia.py

@@ -23,7 +23,20 @@ from nomad.utils import hash
 from nomad.parsing import LocalBackend
 from nomad.normalizing import structure
 from nomad.metainfo.encyclopedia import Encyclopedia
-from tests.normalizing.conftest import run_normalize_for_structure, geometry_optimization, molecular_dynamics, phonon, two_d, bulk, bands_unpolarized_no_gap, bands_polarized_no_gap, bands_unpolarized_gap_indirect, bands_polarized_gap_indirect  # pylint: disable=unused-import
+from tests.normalizing.conftest import (  # pylint: disable=unused-import
+    run_normalize_for_structure,
+    geometry_optimization,
+    molecular_dynamics,
+    phonon,
+    two_d,
+    bulk,
+    bands_unpolarized_no_gap,
+    bands_polarized_no_gap,
+    bands_unpolarized_gap_indirect,
+    bands_polarized_gap_indirect,
+    dos_unpolarized_vasp,
+    dos_polarized_vasp,
+)
 
 ureg = UnitRegistry()
 
@@ -508,3 +521,25 @@ def test_band_structure(bands_unpolarized_no_gap, bands_polarized_no_gap, bands_
     assert gap_up_ev == gap_ev
     assert gap_up_ev == pytest.approx(0.956, 0.01)
     assert gap_down_ev == pytest.approx(1.230, 0.01)
+
+
+def test_dos(dos_unpolarized_vasp, dos_polarized_vasp):
+
+    def test_generic(dos, n_channels):
+        """Generic tests for DOS data."""
+        assert len(dos.energies.shape) == 1
+        assert len(dos.values.shape) == 2
+        assert dos.values.shape[1] == dos.energies.shape[0]
+        assert dos.values.shape[0] == n_channels
+
+    # Unpolarized
+    enc = dos_unpolarized_vasp.get_mi2_section(Encyclopedia.m_def)
+    properties = enc.properties
+    dos = properties.electronic_dos
+    test_generic(dos, n_channels=1)
+
+    # Polarized
+    enc = dos_polarized_vasp.get_mi2_section(Encyclopedia.m_def)
+    properties = enc.properties
+    dos = properties.electronic_dos
+    test_generic(dos, n_channels=2)