fastembed-haystack 2.2.0__tar.gz → 2.3.0__tar.gz

{fastembed_haystack-2.2.0 → fastembed_haystack-2.3.0}/CHANGELOG.md

@@ -1,20 +1,34 @@
 # Changelog
 
-## [integrations/fastembed-v2.1.0] - 2026-02-13
+## [integrations/fastembed-v2.2.0] - 2026-04-16
+
+### 🚀 Features
+
+- Update `FastembedRanker` to auto call `warm_up` on first run (#2834)
+- Add FastembedLateInteractionRanker for ColBERT late-interaction reranking (#3135)
+
+### 🐛 Bug Fixes
+
+- Replace in-place dataclass mutations with dataclasses.replace() (#3112)
 
 ### 📚 Documentation
 
 - *(fastembed)* Remove explicit warm_up from examples (#2830)
+- Simplify pydoc configs (#2855)
+
+### 🧪 Testing
+
+- Track test coverage for all integrations (#3065)
+
+### 🧹 Chores
+
+- Add ANN type annotations to faiss, fastembed, firecrawl, github, google_ai (#2988)
+- Enforce ruff docstring rules (D102/D103/D205/D209/D213/D417/D419) in integrations 11-20 (#3009)
 
 ### 🌀 Miscellaneous
 
-- Automatically call warm_up() if model is not loaded in FastembedRanker
-- Remove explicit call to warm_up() in ranker_example.py
-- Add unit tests
 - Fix typing issue
 - Fix tying
-- Update docstrings
-- Remove explicit warm_up calls from integration tests
 
 ## [integrations/fastembed-v2.0.0] - 2026-01-12
 

{fastembed_haystack-2.2.0 → fastembed_haystack-2.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fastembed-haystack
-Version: 2.2.0
+Version: 2.3.0
 Summary: Haystack 2.x component to embed strings and Documents using fastembed embedding model
 Project-URL: Source, https://github.com/deepset-ai/haystack-core-integrations
 Project-URL: Documentation, https://github.com/deepset-ai/haystack-core-integrations/blob/main/integrations/fastembed/README.md

{fastembed_haystack-2.2.0 → fastembed_haystack-2.3.0}/src/haystack_integrations/components/rankers/fastembed/ranker.py

@@ -48,6 +48,7 @@ class FastembedRanker:
         local_files_only: bool = False,
         meta_fields_to_embed: list[str] | None = None,
         meta_data_separator: str = "\n",
+        score_threshold: float | None = None,
     ) -> None:
         """
         Creates an instance of the 'FastembedRanker'.
@@ -68,6 +69,8 @@ class FastembedRanker:
             with the document content for reranking.
         :param meta_data_separator: Separator used to concatenate the meta fields
             to the Document content.
+        :param score_threshold: If provided, only documents with a score above the threshold are returned.
+            Applied after `top_k`, so the output may contain fewer than `top_k` documents.
         """
         if top_k <= 0:
             msg = f"top_k must be > 0, but got {top_k}"
@@ -82,6 +85,7 @@ class FastembedRanker:
         self.local_files_only = local_files_only
         self.meta_fields_to_embed = meta_fields_to_embed or []
         self.meta_data_separator = meta_data_separator
+        self.score_threshold = score_threshold
         self._model: TextCrossEncoder | None = None
 
     def to_dict(self) -> dict[str, Any]:
@@ -102,6 +106,7 @@ class FastembedRanker:
             local_files_only=self.local_files_only,
             meta_fields_to_embed=self.meta_fields_to_embed,
             meta_data_separator=self.meta_data_separator,
+            score_threshold=self.score_threshold,
         )
 
     @classmethod
@@ -201,4 +206,8 @@ class FastembedRanker:
 
         top_k_documents = [replace(doc, score=score) for doc, score in sorted_doc_scores[:top_k]]
 
+        score_threshold = self.score_threshold
+        if score_threshold is not None:
+            top_k_documents = [doc for doc in top_k_documents if doc.score is not None and doc.score >= score_threshold]
+
         return {"documents": top_k_documents}

{fastembed_haystack-2.2.0 → fastembed_haystack-2.3.0}/tests/test_fastembed_ranker.py

@@ -28,6 +28,7 @@ class TestFastembedRanker:
         assert not ranker.local_files_only
         assert ranker.meta_fields_to_embed == []
         assert ranker.meta_data_separator == "\n"
+        assert ranker.score_threshold is None
 
     def test_init_with_parameters(self):
         """
@@ -43,6 +44,7 @@ class TestFastembedRanker:
             local_files_only=True,
             meta_fields_to_embed=["test_field"],
             meta_data_separator=" | ",
+            score_threshold=0.5,
         )
         assert ranker.model_name == "BAAI/bge-reranker-base"
         assert ranker.top_k == 64
@@ -53,6 +55,7 @@ class TestFastembedRanker:
         assert ranker.local_files_only
         assert ranker.meta_fields_to_embed == ["test_field"]
         assert ranker.meta_data_separator == " | "
+        assert ranker.score_threshold == 0.5
 
     def test_init_with_incorrect_input(self):
         """
@@ -88,6 +91,7 @@ class TestFastembedRanker:
                 "local_files_only": False,
                 "meta_fields_to_embed": [],
                 "meta_data_separator": "\n",
+                "score_threshold": None,
             },
         }
 
@@ -105,6 +109,7 @@ class TestFastembedRanker:
             local_files_only=True,
             meta_fields_to_embed=["test_field"],
             meta_data_separator=" | ",
+            score_threshold=0.5,
         )
         ranker_dict = ranker.to_dict()
         assert ranker_dict == {
@@ -119,6 +124,7 @@ class TestFastembedRanker:
                 "local_files_only": True,
                 "meta_fields_to_embed": ["test_field"],
                 "meta_data_separator": " | ",
+                "score_threshold": 0.5,
             },
         }
 
@@ -138,6 +144,7 @@ class TestFastembedRanker:
                 "local_files_only": False,
                 "meta_fields_to_embed": [],
                 "meta_data_separator": "\n",
+                "score_threshold": None,
             },
         }
         ranker = default_from_dict(FastembedRanker, ranker_dict)
@@ -150,6 +157,7 @@ class TestFastembedRanker:
         assert not ranker.local_files_only
         assert ranker.meta_fields_to_embed == []
         assert ranker.meta_data_separator == "\n"
+        assert ranker.score_threshold is None
 
     def test_from_dict_with_custom_init_parameters(self):
         """
@@ -167,6 +175,7 @@ class TestFastembedRanker:
                 "local_files_only": True,
                 "meta_fields_to_embed": ["test_field"],
                 "meta_data_separator": " | ",
+                "score_threshold": 0.5,
             },
         }
         ranker = default_from_dict(FastembedRanker, ranker_dict)
@@ -179,6 +188,7 @@ class TestFastembedRanker:
         assert ranker.local_files_only
         assert ranker.meta_fields_to_embed == ["test_field"]
         assert ranker.meta_data_separator == " | "
+        assert ranker.score_threshold == 0.5
 
     def test_run_incorrect_input_format(self):
         """
@@ -257,6 +267,20 @@ class TestFastembedRanker:
             parallel=None,
         )
 
+    def test_run_with_score_threshold(self):
+        """
+        Tests that documents below score_threshold are filtered out.
+        """
+        ranker = FastembedRanker(model_name="model_name", score_threshold=0.5)
+        ranker._model = MagicMock()
+        ranker._model.rerank.return_value = [0.9, 0.3, 0.7]
+
+        documents = [Document(content=f"doc {i}") for i in range(3)]
+        result = ranker.run(query="test", documents=documents)
+
+        assert len(result["documents"]) == 2
+        assert all(doc.score >= 0.5 for doc in result["documents"])
+
     def test_run_calls_warm_up(self):
         """
         Unit test to check that warm_up is called when run is called for the first time.