encyclopedia.py 42.3 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
# Copyright 2018 Markus Scheidgen
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an"AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

15
"""
16
The encyclopedia API of the nomad@FAIRDI APIs.
17
"""
18
import re
19
import math
20
import json
21
from typing import List, Dict, Union, Sequence
22

23
24
from flask_restplus import Resource, abort, fields, marshal
from flask import request
25
from elasticsearch_dsl import Search, Q, A
26
from elasticsearch_dsl.utils import AttrDict
27

28
from nomad import config, files
29
from nomad.archive import ArchiveObject
30
from nomad.units import ureg
31
from nomad.metainfo import MSection
Lauri Himanen's avatar
Lauri Himanen committed
32
from nomad.atomutils import get_hill_decomposition
33
from .api import api
34

35
ns = api.namespace("encyclopedia", description="Access encyclopedia metadata.")
36
37
re_formula = re.compile(r"([A-Z][a-z]?)(\d*)")

38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
material_prop_map = {
    # General
    "material_id": "encyclopedia.material.material_id",
    "formula": "encyclopedia.material.formula",
    "formula_reduced": "encyclopedia.material.formula_reduced",
    "system_type": "encyclopedia.material.material_type",
    # Bulk only
    "has_free_wyckoff_parameters": "encyclopedia.material.bulk.has_free_wyckoff_parameters",
    "strukturbericht_designation": "encyclopedia.material.bulk.strukturbericht_designation",
    "material_name": "encyclopedia.material.material_name",
    "bravais_lattice": "encyclopedia.material.bulk.bravais_lattice",
    "crystal_system": "encyclopedia.material.bulk.crystal_system",
    "point_group": "encyclopedia.material.bulk.point_group",
    "space_group_number": "encyclopedia.material.bulk.space_group_number",
    "space_group_international_short_symbol": "encyclopedia.material.bulk.space_group_international_short_symbol",
    "structure_prototype": "encyclopedia.material.bulk.structure_prototype",
    "structure_type": "encyclopedia.material.bulk.structure_type",
}
56
57


58
def get_es_doc_values(es_doc, mapping, keys=None):
59
60
    """Used to form a material definition for "materials/<material_id>" from
    the given ElasticSearch root document.
61
    """
62
63
64
    if keys is None:
        keys = mapping.keys()

65
    result = {}
66
    for key in keys:
67
        es_key = mapping[key]
68
69
70
71
72
73
74
        try:
            value = es_doc
            for part in es_key.split("."):
                value = getattr(value, part)
        except AttributeError:
            value = None
        result[key] = value
75
76
77
78
79

    return result


material_query = api.parser()
80
81
82
83
84
85
86
87
material_query.add_argument(
    "property",
    type=str,
    choices=tuple(material_prop_map.keys()),
    help="Optional single property to retrieve for the given material. If not specified, all properties will be returned.",
    location="args"
)
material_result = api.model("material_result", {
88
89
    # General
    "material_id": fields.String,
90
91
    "formula": fields.String,
    "formula_reduced": fields.String,
92
    "system_type": fields.String,
93
    "n_matches": fields.Integer,
94
    # Bulk only
95
    "has_free_wyckoff_parameters": fields.Boolean,
96
    "strukturbericht_designation": fields.String,
97
    "material_name": fields.String,
98
99
    "bravais_lattice": fields.String,
    "crystal_system": fields.String,
100
    "point_group": fields.String,
101
102
103
    "space_group_number": fields.Integer,
    "space_group_international_short_symbol": fields.String,
    "structure_prototype": fields.String,
104
105
    "structure_type": fields.String,
})
106
107


108
@ns.route("/materials/<string:material_id>")
109
class EncMaterialResource(Resource):
110
111
112
    @api.response(404, "The material does not exist")
    @api.response(200, "Metadata send", fields.Raw)
    @api.doc("material/<material_id>")
113
    @api.expect(material_query)
114
    @api.marshal_with(material_result, skip_none=True)
115
116
117
    def get(self, material_id):
        """Used to retrive basic information related to the specified material.
        """
118
119
120
121
122
123
124
125
126
127
        # Parse request arguments
        args = material_query.parse_args()
        prop = args.get("property", None)
        if prop is not None:
            keys = [prop]
            es_keys = [material_prop_map[prop]]
        else:
            keys = list(material_prop_map.keys())
            es_keys = list(material_prop_map.values())

128
129
130
131
132
133
134
135
136
        # Find the first public entry with this material id and take
        # information from there. In principle all other entries should have
        # the same information.
        s = Search(index=config.elastic.index_name)

        # Since we are looking for an exact match, we use filter context
        # together with term search for speed (instead of query context and
        # match search)
        query = Q(
137
            "bool",
138
            filter=[
139
140
141
                Q("term", published=True),
                Q("term", with_embargo=False),
                Q("term", encyclopedia__material__material_id=material_id),
142
143
144
            ]
        )
        s = s.query(query)
145

146
        # The query is collapsed already on the ES side so we don"t need to
147
148
149
        # transfer so much data.
        s = s.extra(**{
            "collapse": {"field": "encyclopedia.material.material_id"},
150
            "_source": {"includes": es_keys},
151
152
        })

153
154
        response = s.execute()

155
        # No such material
156
        if len(response) == 0:
157
            abort(404, message="There is no material {}".format(material_id))
158

159
        # Create result JSON
160
        entry = response[0]
161
        result = get_es_doc_values(entry, material_prop_map, keys)
162

163
164
165
        return result, 200


166
range_query = api.model("range_query", {
167
168
169
    "max": fields.Float,
    "min": fields.Float,
})
170
171
172
materials_after = api.model("materials_after", {
    "materials": fields.String,
})
173
174
materials_query = api.model("materials_input", {
    "search_by": fields.Nested(api.model("search_query", {
175
176
        "exclusive": fields.Boolean(default=False),
        "formula": fields.String,
Lauri Himanen's avatar
Lauri Himanen committed
177
        "element": fields.String,
178
        "page": fields.Integer(default=1),
179
        "after": fields.Nested(materials_after, allow_null=True),
180
181
        "per_page": fields.Integer(default=25),
        "pagination": fields.Boolean,
182
        "mode": fields.String(default="aggregation"),
183
    })),
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
    "material_name": fields.List(fields.String),
    "structure_type": fields.List(fields.String),
    "space_group_number": fields.List(fields.Integer),
    "system_type": fields.List(fields.String),
    "crystal_system": fields.List(fields.String),
    "band_gap": fields.Nested(range_query, description="Band gap range in eV."),
    "band_gap_direct": fields.Boolean,
    "has_band_structure": fields.Boolean,
    "has_dos": fields.Boolean,
    "has_fermi_surface": fields.Boolean,
    "has_thermal_properties": fields.Boolean,
    "functional_type": fields.List(fields.String),
    "basis_set_type": fields.List(fields.String),
    "code_name": fields.List(fields.String),
    "mass_density": fields.Nested(range_query, description="Mass density range in kg / m ** 3."),
199
})
200
201
202
203
204
pages_result = api.model("page_info", {
    "per_page": fields.Integer,
    "total": fields.Integer,
    "page": fields.Integer,
    "pages": fields.Integer,
205
    "after": fields.Nested(materials_after),
206
207
})

208
209
210
materials_result = api.model("materials_result", {
    "total_results": fields.Integer(allow_null=False),
    "results": fields.List(fields.Nested(material_result)),
211
    "pages": fields.Nested(pages_result),
212
    "es_query": fields.String(allow_null=False),
213
214
215
})


216
@ns.route("/materials")
217
class EncMaterialsResource(Resource):
218
219
220
    @api.response(404, "No materials found")
    @api.response(400, "Bad request")
    @api.response(200, "Metadata send", fields.Raw)
221
    @api.expect(materials_query, validate=False)
222
    @api.marshal_with(materials_result, skip_none=True)
223
    @api.doc("materials")
224
225
226
227
228
229
230
231
232
233
234
235
236
237
    def post(self):
        """Used to query a list of materials with the given search options.
        """
        # Get query parameters as json
        try:
            data = marshal(request.get_json(), materials_query)
        except Exception as e:
            abort(400, message=str(e))

        filters = []
        must_nots = []
        musts = []

        # Add term filters
238
239
        filters.append(Q("term", published=True))
        filters.append(Q("term", with_embargo=False))
240
241
242
243
244
245
246

        def add_terms_filter(source, target, query_type="terms"):
            if data[source]:
                filters.append(Q(query_type, **{target: data[source]}))

        add_terms_filter("material_name", "encyclopedia.material.material_name")
        add_terms_filter("structure_type", "encyclopedia.material.bulk.structure_type")
247
        add_terms_filter("space_group_number", "encyclopedia.material.bulk.space_group_number")
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
        add_terms_filter("system_type", "encyclopedia.material.material_type")
        add_terms_filter("crystal_system", "encyclopedia.material.bulk.crystal_system")
        add_terms_filter("band_gap_direct", "encyclopedia.properties.band_gap_direct", query_type="term")
        add_terms_filter("functional_type", "encyclopedia.method.functional_type")
        add_terms_filter("basis_set_type", "dft.basis_set")
        add_terms_filter("code_name", "dft.code_name")

        # Add exists filters
        def add_exists_filter(source, target):
            param = data[source]
            if param is not None:
                query = Q("exists", field=target)
                if param is True:
                    filters.append(query)
                elif param is False:
                    must_nots.append(query)

        add_exists_filter("has_thermal_properties", "encyclopedia.properties.thermodynamical_properties")
        add_exists_filter("has_band_structure", "encyclopedia.properties.electronic_band_structure")
        add_exists_filter("has_dos", "encyclopedia.properties.electronic_dos")
        add_exists_filter("has_fermi_surface", "encyclopedia.properties.fermi_surface")

        # Add range filters
        def add_range_filter(source, target, source_unit=None, target_unit=None):
            param = data[source]
            query_dict = {}
            if param["min"] is not None:
                if source_unit is None and target_unit is None:
                    gte = param["min"]
                else:
                    gte = (param["min"] * source_unit).to(target_unit).magnitude
                query_dict["gte"] = gte
            if param["max"] is not None:
                if source_unit is None and target_unit is None:
                    lte = param["max"]
                else:
                    lte = (param["max"] * source_unit).to(target_unit).magnitude
                query_dict["lte"] = lte
            if len(query_dict) != 0:
                query = Q("range", **{target: query_dict})
                filters.append(query)

        add_range_filter("band_gap", "encyclopedia.properties.band_gap", ureg.eV, ureg.J)
        add_range_filter("mass_density", "encyclopedia.properties.mass_density")
292

293
294
        # Create query for elements or formula
        search_by = data["search_by"]
295
        mode = search_by["mode"]
296
        formula = search_by["formula"]
Lauri Himanen's avatar
Lauri Himanen committed
297
        elements = search_by["element"]
298
299
300
        exclusive = search_by["exclusive"]

        if formula is not None:
Lauri Himanen's avatar
Lauri Himanen committed
301
302
            # Here we determine a list of atom types. The types may occur
            # multiple times and at multiple places.
303
304
305
306
307
308
309
310
311
312
            element_list = []
            matches = re_formula.finditer(formula)
            for match in matches:
                groups = match.groups()
                symbol = groups[0]
                count = groups[1]
                if symbol != "":
                    if count == "":
                        element_list.append(symbol)
                    else:
Lauri Himanen's avatar
Lauri Himanen committed
313
314
315
316
317
                        element_list += [symbol] * int(count)

            # The given list of species is reformatted with the Hill system
            # into a query string. The counts are reduced by the greatest
            # common divisor.
318
            names, reduced_counts = get_hill_decomposition(element_list, reduced=True)
Lauri Himanen's avatar
Lauri Himanen committed
319
320
321
322
323
            query_string = []
            for name, count in zip(names, reduced_counts):
                if count == 1:
                    query_string.append(name)
                else:
324
                    query_string.append("{}{}".format(name, int(count)))
Lauri Himanen's avatar
Lauri Himanen committed
325
            query_string = " ".join(query_string)
326
327
328

            # With exclusive search we look for exact match
            if exclusive:
Lauri Himanen's avatar
Lauri Himanen committed
329
                filters.append(Q("term", **{"encyclopedia.material.species_and_counts.keyword": query_string}))
330
331
332
333
334
            # With non-exclusive search we look for match that includes at
            # least all parts of the formula, possibly even more.
            else:
                musts.append(Q(
                    "match",
Lauri Himanen's avatar
Lauri Himanen committed
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
                    encyclopedia__material__species_and_counts={"query": query_string, "operator": "and"}
                ))
        elif elements is not None:
            # The given list of species is reformatted with the Hill system into a query string
            species, _ = get_hill_decomposition(elements.split(","))
            query_string = " ".join(species)

            # With exclusive search we look for exact match
            if exclusive:
                filters.append(Q("term", **{"encyclopedia.material.species.keyword": query_string}))
            # With non-exclusive search we look for match that includes at
            # least all species, possibly even more.
            else:
                musts.append(Q(
                    "match",
                    encyclopedia__material__species={"query": query_string, "operator": "and"}
351
352
                ))

353
354
        page = search_by["page"]
        per_page = search_by["per_page"]
355
        after = search_by["after"]
356
        bool_query = Q(
357
            "bool",
358
359
360
361
            filter=filters,
            must_not=must_nots,
            must=musts,
        )
362

363
364
365
        # 1: The paginated approach: No way to know the amount of matches,
        # but can return aggregation results in a quick fashion including
        # the number of matches entries per material.
366
        if mode == "aggregation":
367

368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
            # The top query filters out entries based on the user query
            s = Search(index=config.elastic.index_name)
            s = s.query(bool_query)

            # The materials are grouped by using three aggregations:
            # "Composite" to enable scrolling, "Terms" to enable selecting
            # by material_id and "Top Hits" to fetch a single
            # representative material document. Unnecessary fields are
            # filtered to reduce data transfer.
            terms_agg = A("terms", field="encyclopedia.material.material_id")
            composite_kwargs = {"sources": {"materials": terms_agg}, "size": per_page}
            if after is not None:
                composite_kwargs["after"] = after
            composite_agg = A("composite", **composite_kwargs)
            composite_agg.metric("representative", A(
                "top_hits",
                size=1,
                _source={"includes": list(material_prop_map.values())},
            ))
            s.aggs.bucket("materials", composite_agg)

            # We ignore the top level hits
            s = s.extra(**{
                "size": 0,
            })
393

394
395
396
397
398
            response = s.execute()
            materials = response.aggs.materials.buckets
            if len(materials) == 0:
                abort(404, message="No materials found for the given search criteria or pagination.")
            after = response.aggs.materials["after_key"]
399
400
401
402

            # Gather results from aggregations
            result_list = []
            materials = response.aggs.materials.buckets
403
            keys = list(material_prop_map.keys())
404
405
            for material in materials:
                representative = material["representative"][0]
406
407
                mat_dict = get_es_doc_values(representative, material_prop_map, keys)
                mat_dict["n_matches"] = material.doc_count
408
409
410
411
412
413
                result_list.append(mat_dict)

            # Page information is incomplete for aggregations
            pages = {
                "page": page,
                "per_page": per_page,
414
                "after": after,
415
416
            }
        # 2. Collapse approach. Quickly provides a list of materials
417
        # corresponding to the query, offers full pagination, doesn"t include
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
        # the number of matches per material.
        elif mode == "collapse":
            s = Search(index=config.elastic.index_name)
            s = s.query(bool_query)
            s = s.extra(**{
                "collapse": {"field": "encyclopedia.material.material_id"},
                "size": per_page,
                "from": (page - 1) * per_page,
            })

            # Execute query
            response = s.execute()

            # No matches
            if len(response) == 0:
433
                abort(404, message="No materials found for the given search criteria or pagination.")
434
435
436

            # Loop over materials
            result_list = []
437
            keys = list(material_prop_map.keys())
438
            for material in response:
439
                mat_result = get_es_doc_values(material, material_prop_map, keys)
440
441
442
443
444
445
446
447
448
                result_list.append(mat_result)

            # Full page information available for collapse
            pages = {
                "page": page,
                "per_page": per_page,
                "pages": math.ceil(response.hits.total / per_page),
                "total": response.hits.total,
            }
449
450
451

        result = {
            "results": result_list,
452
            "pages": pages,
453
        }
454
        return result, 200
455
456


457
group_result = api.model("group_result", {
458
459
460
    "calculations": fields.List(fields.String),
    "energies": fields.List(fields.Float),
    "volumes": fields.List(fields.Float),
461
462
463
464
    "energy_minimum": fields.Float,
    "group_hash": fields.String,
    "group_type": fields.String,
    "nr_of_calculations": fields.Integer,
465
    "representative_calc_id": fields.String,
466
})
467
468
469
groups_result = api.model("groups_result", {
    "total_groups": fields.Integer(allow_null=False),
    "groups": fields.List(fields.Nested(group_result)),
470
})
471
472
473
474
group_source = {
    "includes": [
        "calc_id",
        "encyclopedia.properties.energies.energy_total",
475
        "encyclopedia.material.idealized_structure.cell_volume",
476
477
    ]
}
478
479


480
@ns.route("/materials/<string:material_id>/groups")
Lauri Himanen's avatar
Lauri Himanen committed
481
class EncGroupsResource(Resource):
482
483
484
    @api.response(404, "Material not found")
    @api.response(400, "Bad request")
    @api.response(200, "Metadata send", fields.Raw)
Lauri Himanen's avatar
Lauri Himanen committed
485
    @api.expect(material_query, validate=False)
486
    @api.marshal_with(groups_result)
487
    @api.doc("enc_materials")
Lauri Himanen's avatar
Lauri Himanen committed
488
489
    def get(self, material_id):

490
491
492
        # Find entries for the given material, which have EOS or parameter
        # variation hashes set.
        bool_query = Q(
493
            "bool",
494
            filter=[
495
496
497
                Q("term", published=True),
                Q("term", with_embargo=False),
                Q("term", encyclopedia__material__material_id=material_id),
498
499
500
            ],
            must=[
                Q("exists", field="encyclopedia.properties.energies.energy_total"),
501
                Q("exists", field="encyclopedia.material.idealized_structure.cell_volume"),
502
503
504
505
506
507
508
            ],
            should=[
                Q("exists", field="encyclopedia.method.group_eos_hash"),
                Q("exists", field="encyclopedia.method.group_parametervariation_hash"),
            ],
            minimum_should_match=1,  # At least one of the should query must match
        )
Lauri Himanen's avatar
Lauri Himanen committed
509
510

        s = Search(index=config.elastic.index_name)
511
512
513
514
515
516
517
518
519
520
521
        s = s.query(bool_query)

        # Bucket the calculations by the group hashes. Only create a bucket if an
        # above-minimum number of documents are found.
        group_eos_bucket = A("terms", field="encyclopedia.method.group_eos_hash", min_doc_count=4)
        group_param_bucket = A("terms", field="encyclopedia.method.group_parametervariation_hash", min_doc_count=2)

        # calc_id and energy should be extracted for each matched document. The
        # documents are sorted by energy so that the minimum energy one can be
        # easily extracted. A maximum request size is set in order to limit the
        # result size. ES also has an index-level property
522
        # "index.max_inner_result_window" that limits the number of results
523
524
525
526
527
528
529
530
531
532
533
        # that an inner result can contain.
        energy_aggregation = A(
            "top_hits",
            _source=group_source,
            sort=[{"encyclopedia.properties.energies.energy_total": {"order": "asc"}}],
            size=100,
        )
        group_eos_bucket.bucket("energies", energy_aggregation)
        group_param_bucket.bucket("energies", energy_aggregation)
        s.aggs.bucket("groups_eos", group_eos_bucket)
        s.aggs.bucket("groups_param", group_param_bucket)
534

535
536
537
538
        # We ignore the top level hits
        s = s.extra(**{
            "size": 0,
        })
539

540
541
        # No hits on the top query level
        response = s.execute()
542
        groups = []
543
544
545
546
547
548
549
550

        # Collect information for each group from the aggregations
        groups_eos = response.aggs.groups_eos.buckets
        groups_param = response.aggs.groups_param.buckets

        def get_group(group, group_type, group_hash):
            hits = group.energies.hits
            calculations = [doc.calc_id for doc in hits]
551
552
            energies = [doc.encyclopedia.properties.energies.energy_total for doc in hits]
            volumes = [doc.encyclopedia.material.idealized_structure.cell_volume for doc in hits]
553
554
555
556
            group_dict = {
                "group_hash": group_hash,
                "group_type": group_type,
                "nr_of_calculations": len(calculations),
557
558
559
560
                "representative_calc_id": hits[0].calc_id,
                "calculations": calculations,
                "energies": energies,
                "volumes": volumes,
561
562
563
                "energy_minimum": hits[0].encyclopedia.properties.energies.energy_total,
            }
            return group_dict
Lauri Himanen's avatar
Lauri Himanen committed
564

565
566
567
568
569
570
571
572
573
574
575
        for group in groups_eos:
            groups.append(get_group(group, "equation of state", group.key))
        for group in groups_param:
            groups.append(get_group(group, "parameter variation", group.key))

        # Return results
        result = {
            "groups": groups,
            "total_groups": len(groups),
        }
        return result, 200
576
577


578
579
580
581
suggestions_map = {
    "code_name": "dft.code_name",
    "structure_type": "encyclopedia.material.bulk.structure_type",
}
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
suggestions_query = api.parser()
suggestions_query.add_argument(
    "property",
    type=str,
    choices=("code_name", "structure_type"),
    help="The property name for which suggestions are returned.",
    location="args"
)
suggestions_result = api.model("suggestions_result", {
    "code_name": fields.List(fields.String),
    "structure_type": fields.List(fields.String),
})


@ns.route("/suggestions")
class EncSuggestionsResource(Resource):
    @api.response(404, "Suggestion not found")
    @api.response(400, "Bad request")
    @api.response(200, "Metadata send", fields.Raw)
    @api.expect(suggestions_query, validate=False)
    @api.marshal_with(suggestions_result, skip_none=True)
    @api.doc("enc_suggestions")
    def get(self):

        # Parse request arguments
        args = suggestions_query.parse_args()
        prop = args.get("property", None)

610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
        # Use aggregation to return all unique terms for the requested field.
        # Without using composite aggregations there is a size limit for the
        # number of aggregation buckets. This should, however, not be a problem
        # since the number of unique values is low for all supported properties.
        s = Search(index=config.elastic.index_name)
        query = Q(
            "bool",
            filter=[
                Q("term", published=True),
                Q("term", with_embargo=False),
            ]
        )
        s = s.query(query)
        s = s.extra(**{
            "size": 0,
        })

        terms_agg = A("terms", field=suggestions_map[prop])
        s.aggs.bucket("suggestions", terms_agg)

        # Gather unique values into a list
        response = s.execute()
        suggestions = [x.key for x in response.aggs.suggestions.buckets]

        return {prop: suggestions}, 200
635
636


637
638
calcs_query = api.parser()
calcs_query.add_argument(
639
640
641
642
643
644
    "page",
    default=0,
    type=int,
    help="The page number to return.",
    location="args"
)
645
calcs_query.add_argument(
646
647
648
649
650
651
652
653
654
655
656
657
    "per_page",
    default=25,
    type=int,
    help="The number of results per page",
    location="args"
)
calc_prop_map = {
    "calc_id": "calc_id",
    "code_name": "dft.code_name",
    "code_version": "dft.code_version",
    "functional_type": "encyclopedia.method.functional_type",
    "basis_set_type": "dft.basis_set",
658
    "core_electron_treatment": "encyclopedia.method.core_electron_treatment",
659
660
661
662
663
664
665
666
667
668
669
    "run_type": "encyclopedia.calculation.calculation_type",
    "has_dos": "encyclopedia.properties.electronic_dos",
    "has_band_structure": "encyclopedia.properties.electronic_band_structure",
    "has_thermal_properties": "encyclopedia.properties.thermodynamical_properties",
}
calculation_result = api.model("calculation_result", {
    "calc_id": fields.String,
    "code_name": fields.String,
    "code_version": fields.String,
    "functional_type": fields.String,
    "basis_set_type": fields.String,
670
    "core_electron_treatment": fields.String,
671
672
673
674
675
676
677
678
679
680
681
682
683
    "run_type": fields.String,
    "has_dos": fields.Boolean,
    "has_band_structure": fields.Boolean,
    "has_thermal_properties": fields.Boolean,
})
calculations_result = api.model("calculations_result", {
    "total_results": fields.Integer,
    "pages": fields.Nested(pages_result),
    "results": fields.List(fields.Nested(calculation_result)),
})


@ns.route("/materials/<string:material_id>/calculations")
684
class EncCalculationsResource(Resource):
685
686
687
    @api.response(404, "Suggestion not found")
    @api.response(400, "Bad request")
    @api.response(200, "Metadata send", fields.Raw)
688
    @api.expect(calcs_query, validate=False)
689
690
691
692
    @api.doc("enc_calculations")
    def get(self, material_id):
        """Used to return all calculations related to the given material.
        """
693
        args = calcs_query.parse_args()
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
        page = args["page"]
        per_page = args["per_page"]

        s = Search(index=config.elastic.index_name)
        query = Q(
            "bool",
            filter=[
                Q("term", published=True),
                Q("term", with_embargo=False),
                Q("term", encyclopedia__material__material_id=material_id),
            ]
        )
        s = s.query(query)

        # The query is filtered already on the ES side so we don"t need to
        # transfer so much data.
        s = s.extra(**{
            "_source": {"includes": list(calc_prop_map.values())},
            "size": per_page,
            "from": page,
        })

        response = s.execute()

        # No such material
        if len(response) == 0:
            abort(404, message="There is no material {}".format(material_id))

        # Create result JSON
        results = []
        for entry in response:
            calc_dict = get_es_doc_values(entry, calc_prop_map)
            calc_dict["has_dos"] = calc_dict["has_dos"] is not None
727
            calc_dict["has_band_structure"] = calc_dict["has_band_structure"] is not None
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
            calc_dict["has_thermal_properties"] = calc_dict["has_thermal_properties"] is not None
            results.append(calc_dict)

        result = {
            "total_results": len(results),
            "results": results,
            "pages": {
                "per_page": per_page,
                "page": page,
            }
        }

        return result, 200


743
744
745
746
histogram = api.model("histogram", {
    "occurrences": fields.List(fields.Integer),
    "values": fields.List(fields.Float),
})
747
748
statistics_query = api.model("statistics_query", {
    "calculations": fields.List(fields.String),
749
    "properties": fields.List(fields.String),
750
    "n_histogram_bins": fields.Integer,
751
752
753
754
755
})
statistics = api.model("statistics", {
    "min": fields.Float,
    "max": fields.Float,
    "avg": fields.Float,
756
    "histogram": fields.Nested(histogram)
757
758
759
})
statistics_result = api.model("statistics_result", {
    "cell_volume": fields.Nested(statistics),
760
761
762
763
764
765
766
767
    "atomic_density": fields.Nested(statistics),
    "mass_density": fields.Nested(statistics),
    "lattice_a": fields.Nested(statistics),
    "lattice_b": fields.Nested(statistics),
    "lattice_c": fields.Nested(statistics),
    "alpha": fields.Nested(statistics),
    "beta": fields.Nested(statistics),
    "gamma": fields.Nested(statistics),
768
})
769
770
771
772
773
774
775
776
777
778
779
property_map = {
    "cell_volume": "encyclopedia.material.idealized_structure.cell_volume",
    "atomic_density": "encyclopedia.properties.atomic_density",
    "mass_density": "encyclopedia.properties.mass_density",
    "lattice_a": "encyclopedia.material.idealized_structure.lattice_parameters.a",
    "lattice_b": "encyclopedia.material.idealized_structure.lattice_parameters.b",
    "lattice_c": "encyclopedia.material.idealized_structure.lattice_parameters.c",
    "alpha": "encyclopedia.material.idealized_structure.lattice_parameters.alpha",
    "beta": "encyclopedia.material.idealized_structure.lattice_parameters.beta",
    "gamma": "encyclopedia.material.idealized_structure.lattice_parameters.gamma",
}
780
781
782


@ns.route("/materials/<string:material_id>/statistics")
783
class EncStatisticsResource(Resource):
784
785
786
    @api.response(404, "Suggestion not found")
    @api.response(400, "Bad request")
    @api.response(200, "Metadata send", fields.Raw)
787
788
    @api.expect(statistics_query, validate=False)
    @api.marshal_with(statistics_result, skip_none=True)
789
790
    @api.doc("enc_statistics")
    def post(self, material_id):
791
792
        """Used to return statistics related to the specified material and
        calculations.
793
        """
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
        # Get query parameters as json
        try:
            data = marshal(request.get_json(), statistics_query)
        except Exception as e:
            abort(400, message=str(e))

        # Find entries for the given material.
        bool_query = Q(
            "bool",
            filter=[
                Q("term", published=True),
                Q("term", with_embargo=False),
                Q("term", encyclopedia__material__material_id=material_id),
                Q("terms", calc_id=data["calculations"]),
            ]
        )

        s = Search(index=config.elastic.index_name)
        s = s.query(bool_query)
        s = s.extra(**{
            "size": 0,
        })

817
818
819
820
821
822
        # Add statistics aggregations for each requested property
        properties = data["properties"]
        for prop in properties:
            stats_agg = A("stats", field=property_map[prop])
            s.aggs.bucket("{}_stats".format(prop), stats_agg)

823
824
825
826
827
        # No hits on the top query level
        response = s.execute()
        if response.hits.total == 0:
            abort(404, message="Could not find matching calculations.")

828
829
830
831
832
833
834
835
        # Run a second query that creates histograms with fixed size buckets
        # based on the min and max from previous query. Might make more sense
        # to use the mean and sigma to define the range?
        s = Search(index=config.elastic.index_name)
        s = s.query(bool_query)
        s = s.extra(**{
            "size": 0,
        })
836
        n_bins = data["n_histogram_bins"]
837
838
        for prop in properties:
            stats = getattr(response.aggs, "{}_stats".format(prop))
839
            interval = (stats.max * 1.001 - stats.min) / n_bins
840
841
            if interval == 0:
                interval = 1
842
            hist_agg = A("histogram", field=property_map[prop], interval=interval, offset=stats.min, min_doc_count=0)
843
844
845
            s.aggs.bucket("{}_hist".format(prop), hist_agg)
        response_hist = s.execute()

846
        # Return results
847
848
849
850
851
852
853
854
855
856
        result = {}
        for prop in properties:
            stats = getattr(response.aggs, "{}_stats".format(prop))
            hist = getattr(response_hist.aggs, "{}_hist".format(prop))
            occurrences = [x.doc_count for x in hist.buckets]
            values = [x.key for x in hist.buckets]
            result[prop] = {
                "min": stats.min,
                "max": stats.max,
                "avg": stats.avg,
857
                "histogram": {
858
859
860
                    "occurrences": occurrences,
                    "values": values,
                }
861
            }
862

863
        return result, 200
864
865
866
867
868
869
870
871
872
873
874
875
876


wyckoff_variables_result = api.model("wyckoff_variables_result", {
    "x": fields.Float,
    "y": fields.Float,
    "z": fields.Float,
})
wyckoff_set_result = api.model("wyckoff_set_result", {
    "wyckoff_letter": fields.String,
    "indices": fields.List(fields.Integer),
    "element": fields.String,
    "variables": fields.List(fields.Nested(wyckoff_variables_result)),
})
877
878
879
880
881
882
883
884
lattice_parameters = api.model("lattice_parameters", {
    "a": fields.Float,
    "b": fields.Float,
    "c": fields.Float,
    "alpha": fields.Float,
    "beta": fields.Float,
    "gamma": fields.Float,
})
885
886
887
888
889
890

idealized_structure_result = api.model("idealized_structure_result", {
    "atom_labels": fields.List(fields.String),
    "atom_positions": fields.List(fields.List(fields.Float)),
    "lattice_vectors": fields.List(fields.List(fields.Float)),
    "lattice_vectors_primitive": fields.List(fields.List(fields.Float)),
891
    "lattice_parameters": fields.Nested(lattice_parameters),
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
    "periodicity": fields.List(fields.Boolean),
    "number_of_atoms": fields.Integer,
    "cell_volume": fields.Float,
    "wyckoff_sets": fields.List(fields.Nested(wyckoff_set_result)),
})


@ns.route("/materials/<string:material_id>/idealized_structure")
class EncIdealizedStructureResource(Resource):
    @api.response(404, "Suggestion not found")
    @api.response(400, "Bad request")
    @api.response(200, "Metadata send", fields.Raw)
    @api.marshal_with(idealized_structure_result, skip_none=True)
    @api.doc("enc_material_idealized_structure")
    def get(self, material_id):
        """Specialized path for returning a representative idealized structure
        that is displayed in the gui for this material.
        """
        # The representative idealized structure simply comes from the first
        # calculation when the calculations are alphabetically sorted by their
        # calc_id. Coming up with a good way to select the representative one
        # is pretty tricky in general, there are several options:
        # - Lowest energy: This would be most intuitive, but the energy scales
        #   between codes do not match, and the energy may not have been
        #   reported.
        # - Volume that is closest to mean volume: how to calculate volume for
        #   molecules, surfaces, etc...
        # - Random: We would want the representative visualization to be
        #   relatively stable.
        s = Search(index=config.elastic.index_name)
        query = Q(
            "bool",
            filter=[
                Q("term", published=True),
                Q("term", with_embargo=False),
                Q("term", encyclopedia__material__material_id=material_id),
            ]
        )
        s = s.query(query)

        # The query is filtered already on the ES side so we don"t need to
        # transfer so much data.
        s = s.extra(**{
            "sort": [{"calc_id": {"order": "asc"}}],
            "_source": {"includes": ["upload_id", "calc_id"]},
            "size": 1,
        })

        response = s.execute()

        # No such material
        if len(response) == 0:
            abort(404, message="There is no material {}".format(material_id))

        # Read the idealized_structure from the Archive. The structure can be
        # quite large and no direct search queries are performed against it, so
        # it is not in the ES index.
        entry = response[0]
        upload_id = entry.upload_id
        calc_id = entry.calc_id
952
953
        ideal_struct_path = "section_metadata/encyclopedia/material/idealized_structure"
        idealized_structure = read_archive(upload_id, calc_id, ideal_struct_path)[ideal_struct_path]
954
955
956
957

        return idealized_structure, 200


958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
calculation_property_map = {
    "lattice_parameters": {
        "es_source": "encyclopedia.material.idealized_structure.lattice_parameters"
    },
    "energies": {
        "es_source": "encyclopedia.properties.energies",
    },
    "mass_density": {
        "es_source": "encyclopedia.properties.mass_density",
    },
    "atomic_density": {
        "es_source": "encyclopedia.properties.atomic_density",
    },
    "cell_volume": {
        "es_source": "encyclopedia.material.idealized_structure.cell_volume"
    },
    "electronic_band_structure": {
        "es_source": "encyclopedia.properties.electronic_band_structure"
    },
    "electronic_dos": {
        "es_source": "encyclopedia.properties.electronic_dos"
    },
    "wyckoff_sets": {
        "arch_source": "section_metadata/encyclopedia/material/idealized_structure/wyckoff_sets"
    },
}

calculation_property_query = api.model("calculation_query", {
    "properties": fields.List(fields.String),
})
energies = api.model("energies", {
    "energy_total": fields.Float,
    "energy_total_T0": fields.Float,
    "energy_free": fields.Float,
})
calculation_property_result = api.model("calculation_query", {
    "lattice_parameters": fields.Nested(lattice_parameters),
    "energies": fields.Nested(energies),
    "mass_density": fields.Float,
    "atomic_density": fields.Float,
    "cell_volume": fields.Float,
    "wyckoff_sets": fields.Nested(wyckoff_set_result),
    # "electronic_band_structure": fields.Nested(electronic_band_structure),
    # "electronic_dos": fields.Nested(electronic_dos),
})


1005
1006
1007
1008
1009
@ns.route("/materials/<string:material_id>/calculations/<string:calc_id>")
class EncCalculationResource(Resource):
    @api.response(404, "Material or calculation not found")
    @api.response(400, "Bad request")
    @api.response(200, "Metadata send", fields.Raw)
1010
    @api.expect(calculation_property_query, validate=False)
1011
    # @api.marshal_with(calculation_property_result, skip_none=True)
1012
    @api.doc("enc_calculation")
1013
1014
1015
1016
    def post(self, material_id, calc_id):
        """Used to return calculation details. Some properties are not
        available in the ES index and are instead read from the Archive
        directly.
1017
        """
1018
1019
1020
1021
1022
1023
        # Get query parameters as json
        try:
            data = marshal(request.get_json(), calculation_property_query)
        except Exception as e:
            abort(400, message=str(e))

1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
        s = Search(index=config.elastic.index_name)
        query = Q(
            "bool",
            filter=[
                Q("term", published=True),
                Q("term", with_embargo=False),
                Q("term", encyclopedia__material__material_id=material_id),
                Q("term", calc_id=calc_id),
            ]
        )
        s = s.query(query)

1036
        # Create dictionaries for requested properties
1037
        references = []
1038
1039
1040
1041
1042
1043
1044
        properties = data["properties"]
        arch_properties = {}
        es_properties = {}
        for prop in properties:
            es_source = calculation_property_map[prop].get("es_source")
            if es_source is not None:
                es_properties[prop] = es_source
1045
1046
                if prop in set(("electronic_dos", "electronic_band_structure")):
                    references.append(prop)
1047
1048
1049
1050
            arch_source = calculation_property_map[prop].get("arch_source")
            if arch_source is not None:
                arch_properties[prop] = arch_source

1051
        # The query is filtered already on the ES side so we don't need to
1052
        # transfer so much data.
1053
1054
1055
        sources = [
            "upload_id",
            "calc_id",
1056
            "encyclopedia",
1057
1058
1059
        ]
        sources += list(es_properties.values())

1060
        s = s.extra(**{
1061
            "_source": {"includes": sources},
1062
1063
1064
1065
1066
1067
1068
1069
1070
            "size": 1,
        })

        response = s.execute()

        # No such material
        if len(response) == 0:
            abort(404, message="There is no material {} with calculation {}".format(material_id, calc_id))

1071
1072
1073
1074
1075
1076
1077
1078
        # Add references that are to be read from the archive
        for ref in references:
            arch_path = response[0]
            for attr in es_properties[ref].split("."):
                arch_path = arch_path[attr]
            arch_properties[ref] = arch_path
            del es_properties[ref]

1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
        # If any of the requested properties require data from the Archive, the
        # file is opened and read.
        result = {}
        if len(arch_properties) != 0:
            arch_paths = set(arch_properties.values())
            entry = response[0]
            upload_id = entry.upload_id
            calc_id = entry.calc_id
            data = read_archive(
                upload_id,
                calc_id,
                arch_paths,
            )

            # Add results from archive
            for key, value in arch_properties.items():
                value = data[value]
1096
1097
1098
1099
1100
1101
1102
1103

                # DOS results are simplified
                if key == "electronic_dos":
                    del value["dos_energies"]
                    del value["dos_values"]
                    del value["dos_integrated_values"]
                    del value["dos_fermi_energy"]

1104
1105
1106
                result[key] = value

        # Add results from ES
1107
1108
1109
1110
1111
1112
1113
        for prop, es_source in es_properties.items():
            value = response[0]
            for attr in es_source.split("."):
                value = value[attr]
            if isinstance(value, AttrDict):
                value = value.to_dict()
            result[prop] = value
1114
1115
1116
1117
1118

        return result, 200


def read_archive(upload_id: str, calc_id: str, paths: List[str]) -> Dict[str, MSection]:
1119
    """Used to read data from the archive.
1120
1121
1122
1123
1124
1125
1126
1127
1128

    Args:
        upload_id: Upload id.
        calc_id: Calculation id.
        paths: List of metainfo paths to read and return.

    Returns:
        For each path, a dictionary containing the path as key and the returned
        section as value.
1129
    """
1130
1131
1132
1133
    if isinstance(paths, str):
        paths = [paths]

    result = {}
1134
1135
1136
    upload_files = files.UploadFiles.get(upload_id)
    with upload_files.read_archive(calc_id) as archive:
        data = archive[calc_id]
1137
        for path in paths:
1138
1139
1140
1141
            i_path = path
            if i_path .startswith("/"):
                i_path = i_path[1:]
            parts: Sequence[Union[str, int]] = i_path.split("/")
1142
            for part in parts:
1143
1144
1145
1146
                try:
                    part = int(part)
                except Exception:
                    pass
1147
                data = data[part]
1148
            if isinstance(data, ArchiveObject):
1149
1150
                data = data.to_dict()
            result[path] = data
1151

1152
    return result