From dbf6b62379d3e05eacc6c4a755e1ee55c5694804 Mon Sep 17 00:00:00 2001
From: Martin Kuban <kuban@physik.hu-berlin.de>
Date: Mon, 15 Jun 2020 12:57:54 +0200
Subject: [PATCH] added meta-info section for DOS fingerprint
---
nomad/datamodel/metainfo/public.py | 57 ++++++++++++++++++++++++++----
nomad/normalizing/dos.py | 10 ++++--
2 files changed, 58 insertions(+), 9 deletions(-)
diff --git a/nomad/datamodel/metainfo/public.py b/nomad/datamodel/metainfo/public.py
index 48b8aa013e..1a59bea80d 100644
--- a/nomad/datamodel/metainfo/public.py
+++ b/nomad/datamodel/metainfo/public.py
@@ -1201,6 +1201,56 @@ class section_calculation_to_folder_refs(MSection):
a_legacy=LegacyDefinition(name='calculation_to_folder_kind'))
+class section_dos_fingerprint(MSection):
+ '''
+ Section for the fingerprint of the electronic density-of-states (DOS).
+ DOS fingerprints are a modification of the D-Fingerprints reported in Chem. Mater. 2015, 27, 3, 735–743
+ (doi:10.1021/cm503507h). The fingerprint consists of a binary representation of the DOS,
+ that is used to evaluate the similarity of materials based on their electronic structure.
+ '''
+
+ m_def = Section(validate=False, a_legacy=LegacyDefinition(name='section_dos_fingerprint'))
+
+ bins = Quantity(
+ type=str,
+ description='''
+ Byte representation of the DOS fingerprint.
+ ''',
+ a_legacy=LegacyDefinition(name='bins'))
+
+ indices = Quantity(
+ type=int,
+ shape=['first_index_of_DOS_grid', 'last_index_of_DOS_grid'],
+ description='''
+ Indices used to compare DOS fingerprints of different energy ranges.
+ ''',
+ a_legacy=LegacyDefinition(name='indices'))
+
+ stepsize = Quantity(
+ type=np.dtype(np.float64),
+ shape=[],
+ description='''
+ Stepsize of interpolation in the first step of the generation of DOS fingerprints.
+ ''',
+ a_legacy=LegacyDefinition(name='stepsize'))
+
+ filling_factor = Quantity(
+ type=np.dtype(np.float64),
+ shape=[],
+ description='''
+ Proportion of '1' bins in the DOS fingerprint.
+ ''',
+ a_legacy=LegacyDefinition(name='filling_factor'))
+
+ grid_id = Quantity(
+ type=str,
+ description='''
+ Identifier of the DOS grid that was used for the creation of the fingerprint.
+ Similarity can only be calculated if the same grid was used for both fingerprints.
+ ''',
+ a_legacy=LegacyDefinition(name='grid_id'))
+
+
class section_dos(MSection):
'''
Section collecting information of a (electronic-energy or vibrational-energy) density
@@ -1333,12 +1383,7 @@ class section_dos(MSection):
''',
a_legacy=LegacyDefinition(name='number_of_dos_values'))
- dos_fingerprint = Quantity(
- type=str,
- description='''
- The DOS fingerprint.
- ''',
- a_legacy=LegacyDefinition(name='dos_fingerprint'))
+ dos_fingerprint = SubSection(section_dos_fingerprint, repeats=False)
class section_eigenvalues(MSection):
diff --git a/nomad/normalizing/dos.py b/nomad/normalizing/dos.py
index 58570fc2ad..ab2cf27f8d 100644
--- a/nomad/normalizing/dos.py
+++ b/nomad/normalizing/dos.py
@@ -67,12 +67,16 @@ class DosNormalizer(Normalizer):
# Data for DOS fingerprint
dos_energies = dos.dos_energies
- print(type(dos_energies))
- dos_fingerprint = json.dumps(DOSFingerprint().calculate(np.array(dos_energies), dos_normed).to_dict())
+ dos_fingerprint = DOSFingerprint().calculate(np.array(dos_energies), dos_normed)
# Add quantities to NOMAD's Metainfo
scc_url = '/section_run/0/section_single_configuration_calculation/%d/section_dos/0' % scc.m_parent_index
self._backend.openContext(scc_url)
dos.dos_values_normalized = dos_normed
- dos.dos_fingerprint = dos_fingerprint
+ sec_dos_fingerprint = dos.m_create("section_dos_fingerprint")
+ sec_dos_fingerprint.bins = dos_fingerprint.bins
+ sec_dos_fingerprint.indices = dos_fingerprint.indices
+ sec_dos_fingerprint.stepsize = dos_fingerprint.stepsize
+ sec_dos_fingerprint.grid_id = dos_fingerprint.grid_id
+ sec_dos_fingerprint.filling_factor = dos_fingerprint.filling_factor
self._backend.closeContext(scc_url)
--
GitLab