From ec1280980f844eb6a6dbab65e78afde324abce64 Mon Sep 17 00:00:00 2001
From: Markus Scheidgen <markus.scheidgen@gmail.com>
Date: Mon, 10 Feb 2020 14:32:38 +0100
Subject: [PATCH] Search aggregation without min_doc_count in API.
---
gui/src/components/api.js | 25 +++++++++++++++++++++++++
nomad/datamodel/base.py | 6 +-----
nomad/datamodel/dft.py | 2 +-
nomad/search.py | 5 +----
4 files changed, 28 insertions(+), 10 deletions(-)
diff --git a/gui/src/components/api.js b/gui/src/components/api.js
index 3d4019b54a..9305c2f432 100644
--- a/gui/src/components/api.js
+++ b/gui/src/components/api.js
@@ -207,6 +207,8 @@ class Api {
this.onStartLoading = () => null
this.onFinishLoading = () => null
+ this.statistics = {}
+
this._swaggerClient = Swagger(`${apiBase}/swagger.json`)
this.keycloak = keycloak
@@ -389,6 +391,29 @@ class Api {
...search}))
.catch(handleApiError)
.then(response => response.body)
+ .then(response => {
+ // fill absent statistics values with values from prior searches
+ // this helps to keep consistent values, e.g. in the metadata search view
+ if (response.statistics) {
+ const empty = {}
+ Object.keys(response.statistics.total.all).forEach(metric => empty[metric] = 0)
+ Object.keys(response.statistics)
+ .filter(key => !['total', 'authors', 'atoms'].includes(key))
+ .forEach(key => {
+ if (!this.statistics[key]) {
+ this.statistics[key] = new Set()
+ }
+ const values = this.statistics[key]
+ Object.keys(response.statistics[key]).forEach(value => values.add(value))
+ values.forEach(value => {
+ if (!response.statistics[key][value]) {
+ response.statistics[key][value] = empty
+ }
+ })
+ })
+ }
+ return response
+ })
.finally(this.onFinishLoading)
}
diff --git a/nomad/datamodel/base.py b/nomad/datamodel/base.py
index c49716feb7..88e76268ad 100644
--- a/nomad/datamodel/base.py
+++ b/nomad/datamodel/base.py
@@ -193,8 +193,6 @@ class DomainQuantity:
0 (the default) means no aggregations.
metric: Indicates that this quantity should be used as search metric. Values need
to be tuples with metric name and elastic aggregation (e.g. sum, cardinality)
- zero_aggs: Return aggregation values for values with zero hits in the search. Default
- is with zero aggregations.
elastic_mapping: An optional elasticsearch_dsl mapping. Default is ``Keyword``.
elastic_search_type: An optional elasticsearch search type. Default is ``term``.
elastic_field: An optional elasticsearch key. Default is the name of the quantity.
@@ -206,8 +204,7 @@ class DomainQuantity:
def __init__(
self, description: str = None, multi: bool = False, aggregations: int = 0,
order_default: bool = False, metric: Tuple[str, str] = None,
- zero_aggs: bool = True, metadata_field: str = None,
- elastic_mapping: type = None,
+ metadata_field: str = None, elastic_mapping: type = None,
elastic_search_type: str = 'term', elastic_field: str = None,
elastic_value: Callable[[Any], Any] = None,
argparse_action: str = 'append'):
@@ -218,7 +215,6 @@ class DomainQuantity:
self.order_default = order_default
self.aggregations = aggregations
self.metric = metric
- self.zero_aggs = zero_aggs
self.elastic_mapping = elastic_mapping
self.elastic_search_type = elastic_search_type
self.metadata_field = metadata_field
diff --git a/nomad/datamodel/dft.py b/nomad/datamodel/dft.py
index d2ca7cb930..c1d9f85d34 100644
--- a/nomad/datamodel/dft.py
+++ b/nomad/datamodel/dft.py
@@ -274,7 +274,7 @@ Domain(
order_default=True),
atoms=DomainQuantity(
'The atom labels of all atoms in the simulated system.',
- aggregations=len(ase.data.chemical_symbols), multi=True, zero_aggs=False),
+ aggregations=len(ase.data.chemical_symbols), multi=True),
only_atoms=DomainQuantity(
'The atom labels concatenated in species-number order. Used with keyword search '
'to facilitate exclusive searches.',
diff --git a/nomad/search.py b/nomad/search.py
index 39cdbe4527..a913775b46 100644
--- a/nomad/search.py
+++ b/nomad/search.py
@@ -440,10 +440,7 @@ class SearchRequest:
The basic doc_count metric ``code_runs`` is always given.
"""
quantity = quantities[quantity_name]
- min_doc_count = 0 if quantity.zero_aggs else 1
- terms = A(
- 'terms', field=quantity.elastic_field, size=size, min_doc_count=min_doc_count,
- order=dict(_key='asc'))
+ terms = A('terms', field=quantity.elastic_field, size=size, order=dict(_key='asc'))
buckets = self._search.aggs.bucket('statistics:%s' % quantity_name, terms)
self._add_metrics(buckets, metrics_to_use)
--
GitLab