Minor fixes, CI should build

.github/workflows/test.yml

@@ -66,6 +66,8 @@ jobs:
           key: huggingface-${{ runner.os }}-qwen-tokenizer-v1
       - name: Pre-download tokenizer
         run: uv run python -c "from transformers import AutoTokenizer; AutoTokenizer.from_pretrained('Qwen/Qwen3-Embedding-0.6B')"
+      - name: Pre-download cross-encoder test model
+        run: uv run python -c "from sentence_transformers import CrossEncoder; CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2')"
       - name: Run tests with coverage
         run: uv run pytest -m "not integration" --cov=haiku --cov-report=xml
       - name: Upload coverage to Codecov

haiku_rag_slim/haiku/rag/client/downloads.py

@@ -58,6 +58,7 @@ async def download_models(
 
             model_name = config.embeddings.model.name
             yield DownloadProgress(model=model_name, status="start")
+            # Wrap in lambda: ty loses ParamSpec inference on third-party __init__.
             await asyncio.to_thread(lambda: SentenceTransformer(model_name))
             yield DownloadProgress(model=model_name, status="done")
         except ImportError:

haiku_rag_slim/haiku/rag/reranking/__init__.py

@@ -71,7 +71,12 @@ def get_reranker(config: AppConfig = Config) -> RerankerBase | None:
         try:
             from haiku.rag.reranking.cross_encoder import CrossEncoderReranker
 
-            return CrossEncoderReranker(config.reranking.model.name)
+            name = config.reranking.model.name
+            if not name:
+                raise ValueError(
+                    "cross-encoder reranker requires name in reranking.model"
+                )
+            return CrossEncoderReranker(name)
         except ImportError:  # pragma: no cover
             return None
 

haiku_rag_slim/haiku/rag/reranking/cross_encoder.py

@@ -33,8 +33,7 @@ class CrossEncoderReranker(RerankerBase):
             return []
 
         documents = [chunk.content for chunk in chunks]
-        loop = asyncio.get_running_loop()
+        rankings = await asyncio.to_thread(
-        rankings = await loop.run_in_executor(
+            lambda: self._reranker.rank(query, documents, top_k=top_n)
-            None, lambda: self._reranker.rank(query, documents, top_k=top_n)
         )
         return [(chunks[r["corpus_id"]], float(r["score"])) for r in rankings]

haiku_rag_slim/haiku/rag/reranking/jina_local.py

@@ -1,3 +1,5 @@
+import asyncio
+
 try:
     from transformers import (
         AutoModel,  # pyright: ignore[reportMissingImports]
@@ -32,6 +34,8 @@ class JinaLocalReranker(RerankerBase):  # pragma: no cover
 
         documents = [chunk.content for chunk in chunks]
 
-        results = self._reranker.rerank(query, documents, top_n=top_n)
+        results = await asyncio.to_thread(
+            lambda: self._reranker.rerank(query, documents, top_n=top_n)
+        )
 
         return [(chunks[r["index"]], float(r["relevance_score"])) for r in results]

haiku_rag_slim/haiku/rag/reranking/mxbai.py

@@ -1,3 +1,5 @@
+import asyncio
+
 from mxbai_rerank import MxbaiRerankV2  # pyright: ignore[reportMissingImports]
 
 from haiku.rag.config import Config
@@ -22,7 +24,9 @@ class MxBAIReranker(RerankerBase):
 
         documents = [chunk.content for chunk in chunks]
 
-        results = self._client.rank(query=query, documents=documents, top_k=top_n)
+        results = await asyncio.to_thread(
+            lambda: self._client.rank(query=query, documents=documents, top_k=top_n)
+        )
         reranked_chunks = []
         for result in results:
             original_chunk = chunks[result.index]

tests/test_reranker.py

@@ -126,6 +126,15 @@ class TestGetReranker:
         with pytest.raises(ValueError, match="vLLM reranker requires base_url"):
             get_reranker(config)
 
+    def test_cross_encoder_provider_without_name_raises_error(self):
+        config = AppConfig(
+            reranking=RerankingConfig(
+                model=ModelConfig(provider="cross-encoder", name="")
+            )
+        )
+        with pytest.raises(ValueError, match="cross-encoder reranker requires name"):
+            get_reranker(config)
+
     @pytest.mark.parametrize(
         "provider, model_name, class_module, class_name, extra_model_kwargs, expected_attrs, env_vars",
         [