Updated ES term index creation

bel/db/elasticsearch.py

@@ -9,8 +9,8 @@
 log = logging.getLogger(__name__)
 
 cur_dir_name = os.path.dirname(os.path.realpath(__file__))
-mapping_terms_fn = f'{cur_dir_name}/es_mapping_terms.yml'
-
+mappings_terms_fn = f'{cur_dir_name}/es_mappings_terms.yml'
+settings_terms_fn = f'{cur_dir_name}/es_settings_terms.yml'
 terms_alias = 'terms'
 
 
@@ -51,11 +51,19 @@ def delete_index(es, index_name: str):
 def create_terms_index(es, index_name: str):
     """Create terms index"""
 
-    with open(mapping_terms_fn, 'r') as f:
-        mapping_terms = yaml.load(f)
+    with open(mappings_terms_fn, 'r') as f:
+        mappings_terms = yaml.load(f)
+
+    with open(settings_terms_fn, 'r') as f:
+        settings_terms = yaml.load(f)
 
     try:
-        es.indices.create(index=index_name, body=mapping_terms)
+        es.indices.create(index=index_name, body=mappings_terms)
+
+        # Update settings - have to close before and then open after
+        es.indices.close(index=index_name)
+        es.indices.put_settings(index=index_name, body=settings_terms)
+        es.indices.open(index=index_name)
 
     except Exception as e:
         log.error(f'Could not create elasticsearch terms index: {e}')

bel/db/es_mapping_terms.yml → bel/db/es_mappings_terms.yml

@@ -1,37 +1,3 @@
-settings:
-  analysis:
-    filter:
-      autocomplete_filter:
-          type: edge_ngram
-          min_gram: 1
-          max_gram: 20
-    analyzer:
-      autocomplete:
-        type: custom
-        tokenizer: autocomplete_tokenizer
-        filter:
-          - lowercase
-          - autocomplete_filter
-      autocomplete_search:
-        type: custom
-        tokenizer: autocomplete_tokenizer
-        filter:
-          - lowercase
-
-    # Added simple pattern split to allow SP:H4_HUMAN in alt_ids to be matched
-    # This will also allow IL-6 to be completed against.  The standard tokenizer/analyzer
-    # broke it into IL and 6
-    tokenizer:
-      autocomplete_tokenizer:
-        type: simple_pattern_split
-        pattern: " |:"
-    normalizer:
-      lowercase:
-        type: custom
-        char_filter: []
-        filter:
-          - lowercase
-
 mappings:
   term:
     properties:

bel/db/es_settings_terms.yml

@@ -0,0 +1,33 @@
+settings:
+  analysis:
+    filter:
+      autocomplete_filter:
+          type: edge_ngram
+          min_gram: 1
+          max_gram: 20
+    analyzer:
+      autocomplete:
+        type: custom
+        tokenizer: autocomplete_tokenizer
+        filter:
+          - lowercase
+          - autocomplete_filter
+      autocomplete_search:
+        type: custom
+        tokenizer: autocomplete_tokenizer
+        filter:
+          - lowercase
+
+    # Added simple pattern split to allow SP:H4_HUMAN in alt_ids to be matched
+    # This will also allow IL-6 to be completed against.  The standard tokenizer/analyzer
+    # broke it into IL and 6
+    tokenizer:
+      autocomplete_tokenizer:
+        type: simple_pattern_split
+        pattern: " |:"
+    normalizer:
+      lowercase:
+        type: custom
+        char_filter: []
+        filter:
+          - lowercase