invenio-vocabularies 5.0.3__tar.gz → 5.1.0__tar.gz

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/CHANGES.rst

@@ -8,6 +8,12 @@
 Changes
 =======
 
+Version v5.1.0 (released 2024-09-25)
+
+- funders: tune search boost for acronyms
+    * Add and `acronym.keyword` field to the funders mapping.
+    * Apply to funders the same field boosting as in affiliations.
+
 Version v5.0.3 (released 2024-09-06)
 
 - services: skip index rebuilding

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: invenio-vocabularies
-Version: 5.0.3
+Version: 5.1.0
 Summary: Invenio module for managing vocabularies.
 Home-page: https://github.com/inveniosoftware/invenio-vocabularies
 Author: CERN
@@ -50,6 +50,12 @@ Description: ..
         Changes
         =======
         
+        Version v5.1.0 (released 2024-09-25)
+        
+        - funders: tune search boost for acronyms
+            * Add and `acronym.keyword` field to the funders mapping.
+            * Apply to funders the same field boosting as in affiliations.
+        
         Version v5.0.3 (released 2024-09-06)
         
         - services: skip index rebuilding

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/invenio_vocabularies/__init__.py

@@ -10,6 +10,6 @@
 
 from .ext import InvenioVocabularies
 
-__version__ = "5.0.3"
+__version__ = "5.1.0"
 
 __all__ = ("__version__", "InvenioVocabularies")

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/invenio_vocabularies/contrib/funders/config.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 #
-# Copyright (C) 2022 CERN.
+# Copyright (C) 2022-2024 CERN.
 #
 # Invenio-Vocabularies is free software; you can redistribute it and/or
 # modify it under the terms of the MIT License; see LICENSE file for more
@@ -9,6 +9,7 @@
 """Vocabulary funders configuration."""
 
 from flask import current_app
+from invenio_i18n import get_locale
 from invenio_i18n import lazy_gettext as _
 from invenio_records_resources.services import SearchOptions
 from invenio_records_resources.services.records.components import DataComponent
@@ -22,6 +23,7 @@ funder_schemes = LocalProxy(lambda: current_app.config["VOCABULARIES_FUNDER_SCHE
 funder_fundref_doi_prefix = LocalProxy(
     lambda: current_app.config["VOCABULARIES_FUNDER_DOI_PREFIX"]
 )
+localized_title = LocalProxy(lambda: f"title.{get_locale()}^20")
 
 
 class FundersSearchOptions(SearchOptions):
@@ -30,9 +32,12 @@ class FundersSearchOptions(SearchOptions):
     suggest_parser_cls = SuggestQueryParser.factory(
         fields=[
             "name^100",
+            "acronym.keyword^100",
+            "acronym^40",
+            localized_title,
+            "id^20",
+            "aliases^20",
             "identifiers.identifier^10",
-            "acronym^10",
-            "aliases^10",
         ],
         type="most_fields",  # https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-multi-match-query.html#multi-match-types
         fuzziness="AUTO",  # https://www.elastic.co/guide/en/elasticsearch/reference/current/common-options.html#fuzziness

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/invenio_vocabularies/contrib/funders/mappings/os-v1/funders/funder-v2.0.0.json

@@ -29,6 +29,16 @@
           ]
         }
       },
+      "normalizer": {
+        "accent_normalizer": {
+          "type": "custom",
+          "char_filter": ["strip_special_chars"],
+          "filter": [
+            "lowercase",
+            "asciifolding"
+          ]
+        }
+      },
       "filter": {
         "lowercase": {
           "type": "lowercase",
@@ -112,7 +122,13 @@
       "acronym": {
         "type": "text",
         "analyzer": "accent_edge_analyzer",
-        "search_analyzer": "accent_analyzer"
+        "search_analyzer": "accent_analyzer",
+        "fields": {
+          "keyword": {
+            "type": "keyword",
+            "normalizer": "accent_normalizer"
+          }
+        }
       },
       "status": {
         "type": "keyword"

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/invenio_vocabularies/contrib/funders/mappings/os-v2/funders/funder-v2.0.0.json

@@ -29,6 +29,16 @@
           ]
         }
       },
+      "normalizer": {
+        "accent_normalizer": {
+          "type": "custom",
+          "char_filter": ["strip_special_chars"],
+          "filter": [
+            "lowercase",
+            "asciifolding"
+          ]
+        }
+      },
       "filter": {
         "lowercase": {
           "type": "lowercase",
@@ -112,7 +122,13 @@
       "acronym": {
         "type": "text",
         "analyzer": "accent_edge_analyzer",
-        "search_analyzer": "accent_analyzer"
+        "search_analyzer": "accent_analyzer",
+        "fields": {
+          "keyword": {
+            "type": "keyword",
+            "normalizer": "accent_normalizer"
+          }
+        }
       },
       "status": {
         "type": "keyword"

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/invenio_vocabularies.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: invenio-vocabularies
-Version: 5.0.3
+Version: 5.1.0
 Summary: Invenio module for managing vocabularies.
 Home-page: https://github.com/inveniosoftware/invenio-vocabularies
 Author: CERN
@@ -50,6 +50,12 @@ Description: ..
         Changes
         =======
         
+        Version v5.1.0 (released 2024-09-25)
+        
+        - funders: tune search boost for acronyms
+            * Add and `acronym.keyword` field to the funders mapping.
+            * Apply to funders the same field boosting as in affiliations.
+        
         Version v5.0.3 (released 2024-09-06)
         
         - services: skip index rebuilding

{invenio-vocabularies-5.0.3 → invenio-vocabularies-5.1.0}/tests/contrib/funders/test_funders_resource.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
 #
 # This file is part of Invenio.
-# Copyright (C) 2022 CERN.
+# Copyright (C) 2022-2024 CERN.
 #
 # Invenio-Vocabularies is free software; you can redistribute it and/or
 # modify it under the terms of the MIT License; see LICENSE file for more
@@ -127,9 +127,12 @@ def test_funders_suggest_sort(client, h, prefix, example_funders):
     # Should show 2 results, and id=cern as first due to name
     res = client.get(f"{prefix}?suggest=CERN", headers=h)
     assert res.status_code == 200
-    assert res.json["hits"]["total"] == 2  # should be 2
+    assert res.json["hits"]["total"] == 3
     assert res.json["hits"]["hits"][0]["name"] == "CERN"
     assert res.json["hits"]["hits"][1]["name"] == "CERT"
+    # Matches lower, since title is boosted less
+    assert res.json["hits"]["hits"][2]["name"] == "OTHER"
+    assert res.json["hits"]["hits"][2]["title"]["en"] == "CERN"
 
     res = client.get(f"{prefix}?suggest=N%C3%B5rthw%C3%AAst", headers=h)  # Nõrthwêst
     assert res.status_code == 200