feat(document-index): introduce is_null filter (#1183)

* feat(document-index): introduce filter `is_null` * style: fix formatting * chore: add release notes --------- Co-authored-by: Sebastian Niehus <[email protected]>
Aleph-Alpha · Dec 17, 2024 · 61d0660 · 61d0660
1 parent 44805cc
commit 61d0660
Show file tree

Hide file tree

Showing 4 changed files with 86 additions and 0 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -10,7 +10,10 @@
 - Add `create_project` bool to `StudioClient.__init__()` to enable users to automatically create their Studio projects
 - Add progressbar to the `Runner` to be able to track the `Run`
 - Add `StudioClient.submit_benchmark_lineages` function and include it in `StudioClient.submit_benchmark_execution`
+
+#### DocumentIndexClient 
 - Add method `DocumentIndexClient.chunks()` for retrieving all text chunks of a document.
+- Add metadata filter `FilterOps.IS_NULL`, that allows to filter fields based on whether their value is null.
 
 ### Fixes
 - The Document Index `SearchQuery` now correctly allows searches with a negative `min_score`.

diff --git a/src/intelligence_layer/connectors/document_index/document_index.py b/src/intelligence_layer/connectors/document_index/document_index.py
@@ -227,6 +227,7 @@ class FilterOps(Enum):
     BEFORE = "before"
     AT_OR_BEFORE = "at_or_before"
     EQUAL_TO = "equal_to"
+    IS_NULL = "is_null"
 
 
 class FilterField(BaseModel):

diff --git a/tests/conftest_document_index.py b/tests/conftest_document_index.py
@@ -158,6 +158,7 @@ def document_contents_with_metadata() -> list[DocumentContents]:
 
     metadata_1: JsonSerializable = {
         "string-field": "example_string_1",
+        "option-field": None,
         "integer-field": 123,
         "float-field": 123.45,
         "boolean-field": True,
@@ -168,6 +169,7 @@ def document_contents_with_metadata() -> list[DocumentContents]:
 
     metadata_2: JsonSerializable = {
         "string-field": "example_string_2",
+        "option-field": "example_string_2",
         "integer-field": 456,
         "float-field": 678.90,
         "boolean-field": False,
@@ -178,6 +180,7 @@ def document_contents_with_metadata() -> list[DocumentContents]:
 
     metadata_3: JsonSerializable = {
         "string-field": "example_string_3",
+        "option-field": "example_string_3",
         "integer-field": 789,
         "float-field": 101112.13,
         "boolean-field": True,
@@ -237,6 +240,10 @@ def filter_index_configs(
             "field-name": "string-field",
             "field-type": "string",
         },
+        random_identifier(): {
+            "field-name": "option-field",
+            "field-type": "string",
+        },
         random_identifier(): {
             "field-name": "integer-field",
             "field-type": "integer",

diff --git a/tests/connectors/document_index/test_document_index.py b/tests/connectors/document_index/test_document_index.py
@@ -408,6 +408,81 @@ def search() -> None:
     search()
 
 
+def test_search_with_null_filter(
+    document_index: DocumentIndexClient,
+    read_only_populated_collection: tuple[CollectionPath, IndexPath],
+) -> None:
+    search_query = SearchQuery(
+        query="Pemberton",
+        max_results=10,
+        min_score=0.5,
+        filters=[
+            Filters(
+                filter_type="with",
+                fields=[
+                    FilterField(
+                        field_name="option-field",
+                        field_value=True,
+                        criteria=FilterOps.IS_NULL,
+                    )
+                ],
+            )
+        ],
+    )
+
+    @retry
+    def search() -> None:
+        collection_path, index_path = read_only_populated_collection
+        results = document_index.search(
+            collection_path,
+            index_path.index,
+            search_query,
+        )
+        assert len(results) == 1
+        assert results[0].document_path.document_name == "document-0"
+
+    search()
+
+
+def test_search_with_null_filter_without(
+    document_index: DocumentIndexClient,
+    read_only_populated_collection: tuple[CollectionPath, IndexPath],
+) -> None:
+    search_query = SearchQuery(
+        query="Pemberton",
+        max_results=10,
+        min_score=0.5,
+        filters=[
+            Filters(
+                filter_type="without",
+                fields=[
+                    FilterField(
+                        field_name="option-field",
+                        field_value=True,
+                        criteria=FilterOps.IS_NULL,
+                    )
+                ],
+            )
+        ],
+    )
+
+    @retry
+    def search() -> None:
+        collection_path, index_path = read_only_populated_collection
+        results = document_index.search(
+            collection_path,
+            index_path.index,
+            search_query,
+        )
+        assert len(results) == 2
+        assert {r.document_path.document_name for r in results} == {
+            "document-1",
+            "document-2",
+        }
+
+    search()
+
+
 def test_search_with_integer_filter(
     document_index: DocumentIndexClient,
     read_only_populated_collection: tuple[CollectionPath, IndexPath],