fix: add unaccent to search

adds a generic filter method that takes either all text based
fields or those defined in _default_search_fields and adds
them to a search combined by OR using unaccent. Also adds
`unaccent` to the default text filters.

resolves #109

Author: sennierer

apis_ontology/filtersets.py

@@ -2,7 +2,52 @@
 from apis_core.generic.filtersets import django_filters
 from apis_core.relations.filtersets import RelationFilterSet
 from django.db import models
+from django.db.models import Q, CharField, TextField
 from apis_ontology.forms import RelationFilterSetForm, EntityFilterSetForm
+from django.contrib.postgres.search import SearchVector, SearchQuery
+
+
+def generic_search_filter(queryset, name, value, fields=None):
+    """
+    A generic filter that searches across specified fields using unaccent__icontains with OR logic.
+
+    Priority for fields selection:
+    1. Explicitly provided fields parameter
+    2. _default_search_fields attribute on the model
+    3. All CharField and TextField fields from the model
+
+    Args:
+        queryset: The queryset to filter
+        name: The name of the filter (not used)
+        value: The search value
+        fields: Optional list of specific field names to search in
+
+    Returns:
+        Filtered queryset
+    """
+    if not value:
+        return queryset
+
+    # If no fields specified, check for _default_search_fields or use all text fields
+    if fields is None:
+        model = queryset.model
+
+        # Check if model has _default_search_fields attribute
+        if hasattr(model, "_default_search_fields"):
+            fields = model._default_search_fields
+        else:
+            # Fall back to all CharField and TextField fields
+            fields = []
+            for field in model._meta.get_fields():
+                if isinstance(field, (CharField, TextField)) and not field.primary_key:
+                    fields.append(field.name)
+
+    # Build Q objects for each field with OR logic
+    q_objects = Q()
+    for field in fields:
+        q_objects |= Q(**{f"{field}__unaccent__icontains": value})
+
+    return queryset.filter(q_objects)
 
 
 class NomanslandMixinFilterSet(AbstractEntityFilterSet):
@@ -28,17 +73,30 @@ class Meta(AbstractEntityFilterSet.Meta):
             models.CharField: {
                 "filter_class": django_filters.CharFilter,
                 "extra": lambda f: {
-                    "lookup_expr": "icontains",
+                    "lookup_expr": "unaccent__icontains",
                 },
             },
             models.TextField: {
                 "filter_class": django_filters.CharFilter,
                 "extra": lambda f: {
-                    "lookup_expr": "icontains",
+                    "lookup_expr": "unaccent__icontains",
                 },
             },
         }
 
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        for filter in self.filters.values():
+            if (
+                hasattr(filter, "label")
+                and filter.label
+                and "unaccent contains" in filter.label
+            ):
+                filter.label = filter.label.replace("unaccent contains", "")
+        self.filters["search"] = django_filters.CharFilter(
+            method=generic_search_filter, label="Search"
+        )
+
 
 class NomanslandRelationMixinFilterSet(RelationFilterSet):
     class Meta(RelationFilterSet.Meta):
@@ -67,3 +125,12 @@ class Meta(RelationFilterSet.Meta):
                 },
             },
         }
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        for filter in self.filters.values():
+            if hasattr(filter, "label") and filter.label and "contains" in filter.label:
+                filter.label = filter.label.replace("contains", "")
+        self.filters["search"] = django_filters.CharFilter(
+            method=generic_search_filter, label="Search"
+        )

apis_ontology/models.py

@@ -98,6 +98,7 @@ class Meta:
 
 
 class Person(E21_Person, VersionMixin, NomanslandMixin, AbstractEntity):
+    _default_search_fields = ["forename", "surname"]
     GENDERS = [
         ("male", "Male"),
         ("female", "Female"),
@@ -145,6 +146,7 @@ class Meta:
 class Place(
     E53_Place, VersionMixin, NomanslandDateMixin, NomanslandMixin, AbstractEntity
 ):
+    _default_search_fields = ["label"]
     class_uri = "http://id.loc.gov/ontologies/bibframe/Place"
     kind = models.ForeignKey(
         PlaceType, blank=True, null=True, on_delete=models.SET_NULL
@@ -176,6 +178,7 @@ class Meta:
 
 
 class Institution(VersionMixin, NomanslandDateMixin, NomanslandMixin, AbstractEntity):
+    _default_search_fields = ["name"]
     name = models.CharField(max_length=255)
     kind = models.ForeignKey(
         InstitutionType, blank=True, null=True, on_delete=models.SET_NULL
@@ -201,6 +204,7 @@ class Meta:
 
 
 class Event(VersionMixin, NomanslandDateMixin, NomanslandMixin, AbstractEntity):
+    _default_search_fields = ["name"]
     name = models.CharField(max_length=255)
     kind = models.ForeignKey(
         EventType, blank=True, null=True, on_delete=models.SET_NULL
@@ -237,6 +241,7 @@ class Meta:
 
 
 class Work(VersionMixin, NomanslandDateMixin, NomanslandMixin, AbstractEntity):
+    _default_search_fields = ["name"]
     name = models.CharField(max_length=255)
     kind = models.ForeignKey(WorkType, blank=True, null=True, on_delete=models.SET_NULL)
     subject_heading = models.ManyToManyField(SubjectHeading, blank=True)
@@ -296,6 +301,7 @@ def get_queryset(self):
 
 
 class Expression(VersionMixin, NomanslandDateMixin, NomanslandMixin, AbstractEntity):
+    _default_search_fields = ["title"]
     title = models.CharField(max_length=255, blank=True, null=True)
     locus = models.CharField(max_length=255, blank=True, null=True)
     script_type_title = models.ForeignKey(
@@ -337,6 +343,7 @@ class Meta:
 
 
 class Manuscript(VersionMixin, NomanslandDateMixin, NomanslandMixin, AbstractEntity):
+    _default_search_fields = ["name", "identifier", "description", "notes", "additions"]
     name = models.CharField(max_length=255, blank=True, null=True)
     identifier = models.CharField(max_length=255, blank=True, null=True)
     extent = models.CharField(max_length=255, blank=True, null=True)