Merge branch 'feature/default-field-optimization' into 'main'

Do not search in all fields by default

# 📌 Summary

With this change we will remove searching in all fields by default. `*:` should be used if searching in all fields is still required.

## 🔗 Issue Reference

See merge request swiss-armed-forces/cyber-command/cea/loom!223

Author: 736-c41-2c1-e464fc974

.vscode/launch.json

@@ -22,7 +22,7 @@
           "remoteRoot": "/code"
         }
       ],
-      "justMyCode": true
+      "justMyCode": false
     },
     {
       "name": "Attach Worker",
@@ -82,7 +82,7 @@
           "remoteRoot": "/code"
         }
       ],
-      "justMyCode": true
+      "justMyCode": false
     },
     {
       "name": "Attach Crawler",
@@ -98,7 +98,7 @@
           "remoteRoot": "/code"
         }
       ],
-      "justMyCode": true
+      "justMyCode": false
     },
     {
       "name": "Attach by Port (Prompt)",
@@ -114,7 +114,7 @@
           "remoteRoot": "/code"
         }
       ],
-      "justMyCode": true
+      "justMyCode": false
     }
   ],
   "inputs": [

Frontend/src/features/search/container/ResultCard.tsx

@@ -68,7 +68,12 @@ export const ResultCard: React.FC<ResultCardProps> = React.memo(
                     inView: inView,
                 }),
             );
-        }, [inView, fileId, dispatch]);
+        }, [
+            inView,
+            fileId,
+            searchQuery, // this is required here as we want to re-run this every time the user changes query. The same file might be in the old and new query results
+            dispatch,
+        ]);
 
         const handleViewDetail = () => {
             dispatch(

backend/common/common/file/file_repository.py

@@ -299,7 +299,6 @@ class Meta:  # pylint: disable=too-few-public-methods
         )
 
     class Index:  # pylint: disable=too-few-public-methods
-        """The index."""
 
         name = "file"
         settings = {
@@ -519,8 +518,12 @@ def get_embedding_generator_by_knn(
         filter_query = Q(
             "query_string",
             query=self._query_builder.build(query),
-            default_field="*",
             default_operator="AND",
+            # Be lenient because the index-level default fields are set dynamically
+            # (index.query.default_field). Those fields can span multiple data types
+            # (text, numeric, boolean, date), and query_string parsing would otherwise
+            # fail when a term is incompatible with a field's type.
+            lenient=True,
         )
 
         for q in embedding_vectors:

backend/common/common/models/es_repository.py

@@ -114,6 +114,51 @@ def to_es_dict(self) -> dict:
 
         return document_dict
 
+    @classmethod
+    def get_default_fields(cls) -> list[str]:
+        """Extract all field names from the document mapping, including nested fields
+        and multi-fields.
+
+        Returns a list of field names that can be used for index.query.default_field
+        setting to limit default search scope and improve performance.
+        """
+
+        def extract_fields(mapping: dict, prefix: str = "") -> list[str]:
+            """Recursively extract field names from mapping dictionary.
+
+            Args:
+                mapping: The mapping dictionary (can contain 'properties' or 'fields')
+                prefix: The current field path prefix (for nested fields)
+
+            Returns:
+                List of field names including nested paths
+                (e.g., 'parent.child', 'title.keyword')
+            """
+            fields = []
+            for field_name, field_config in mapping.items():
+                # Build the full field path
+                full_path = f"{prefix}.{field_name}" if prefix else field_name
+                fields.append(full_path)
+
+                if isinstance(field_config, dict):
+                    # Recurse into nested object properties
+                    if "properties" in field_config:
+                        nested_fields = extract_fields(
+                            field_config["properties"], full_path
+                        )
+                        fields.extend(nested_fields)
+
+                    # Recurse into multi-fields
+                    if "fields" in field_config:
+                        multi_fields = extract_fields(field_config["fields"], full_path)
+                        fields.extend(multi_fields)
+
+            return fields
+
+        mapping_dict = cls._doc_type.mapping.to_dict()
+        field_names = extract_fields(mapping_dict.get("properties", {}))
+        return field_names
+
 
 EsRepositoryObjectT = TypeVar("EsRepositoryObjectT", bound=EsRepositoryObject)
 EsRepositoryDocumentT = TypeVar("EsRepositoryDocumentT", bound=_EsRepositoryDocument)
@@ -307,8 +352,12 @@ def _get_search_by_query(
         search = search.query(
             "query_string",
             query=query_string,
-            default_field="*",
             default_operator="AND",
+            # Be lenient because the index-level default fields are set dynamically
+            # (index.query.default_field). Those fields can span multiple data types
+            # (text, numeric, boolean, date), and query_string parsing would otherwise
+            # fail when a term is incompatible with a field's type.
+            lenient=True,
         )
 
         return search
@@ -522,6 +571,10 @@ def init(self):
             using=self._elasticsearch
         ):
             self._index.close(using=self._elasticsearch)
+        # set settings before initializing
+        self._index.settings(
+            query={"default_field": self._document_type.get_default_fields()}
+        )
         # initialize
         self._document_type.init(using=self._elasticsearch)
         # after initialization: open the index

charts/templates/elasticsearch/statefulset.yaml

@@ -43,20 +43,20 @@ spec:
             httpGet:
               path: /_cluster/health?wait_for_status=yellow&timeout=10s
               port: {{ .Values.elasticsearch.service.port }}
-            initialDelaySeconds: {{ default 0 .Values.elasticsearch.readinessProbe.initialDelaySeconds }}
+            initialDelaySeconds: {{ default 5 .Values.elasticsearch.readinessProbe.initialDelaySeconds }}
             periodSeconds: {{ default 15 .Values.elasticsearch.readinessProbe.periodSeconds }}
             timeoutSeconds: {{ default 12 .Values.elasticsearch.readinessProbe.timeoutSeconds }}
-            failureThreshold: {{ default 3 .Values.elasticsearch.readinessProbe.failureThreshold }}
+            failureThreshold: {{ default 10 .Values.elasticsearch.readinessProbe.failureThreshold }}
           {{- end }}
           {{- if .Values.elasticsearch.livenessProbe.enabled }}
           livenessProbe:
             httpGet:
               path: /_cluster/health?wait_for_status=yellow&timeout=10s
               port: {{ .Values.elasticsearch.service.port }}
-            initialDelaySeconds: {{ default 0 .Values.elasticsearch.livenessProbe.initialDelaySeconds }}
+            initialDelaySeconds: {{ default 5 .Values.elasticsearch.livenessProbe.initialDelaySeconds }}
             periodSeconds: {{ default 15 .Values.elasticsearch.livenessProbe.periodSeconds }}
             timeoutSeconds: {{ default 12 .Values.elasticsearch.livenessProbe.timeoutSeconds }}
-            failureThreshold: {{ default 6 .Values.elasticsearch.livenessProbe.failureThreshold }}
+            failureThreshold: {{ default 15 .Values.elasticsearch.livenessProbe.failureThreshold }}
           {{- end }}
           {{- if .Values.elasticsearch.resources }}
           resources: