🎨 reformatted by ruff

Author: XyLearningProgramming

slm_server/app.py

@@ -194,9 +194,7 @@ async def create_embeddings(
     """Create embeddings using the dedicated ONNX embedding model."""
     with slm_embedding_span(req) as span:
         inputs = req.input if isinstance(req.input, list) else [req.input]
-        vectors = await asyncio.to_thread(
-            emb_model.encode, inputs, True
-        )
+        vectors = await asyncio.to_thread(emb_model.encode, inputs, True)
         result = EmbeddingResponse(
             data=[
                 EmbeddingData(embedding=vec.tolist(), index=i)

slm_server/embedding.py

@@ -18,6 +18,7 @@
 KEY_ATTENTION_MASK: str = "attention_mask"
 KEY_TOKEN_TYPE_IDS: str = "token_type_ids"
 
+
 class OnnxEmbeddingModel:
     """Lightweight ONNX-based sentence embedding model.
 
@@ -35,9 +36,7 @@ def __init__(self, settings: EmbeddingSettings):
         self.tokenizer.enable_truncation(max_length=settings.max_length)
         self.tokenizer.enable_padding(length=None)
 
-        self.session = InferenceSession(
-            settings.onnx_path, providers=ONNX_PROVIDERS
-        )
+        self.session = InferenceSession(settings.onnx_path, providers=ONNX_PROVIDERS)
 
         elapsed_ms = (time.monotonic() - start) * 1000
         logger.info(
@@ -52,9 +51,7 @@ def encode(self, texts: list[str], normalize: bool = True) -> np.ndarray:
         encodings = self.tokenizer.encode_batch(texts)
 
         input_ids = np.array([e.ids for e in encodings], dtype=np.int64)
-        attention_mask = np.array(
-            [e.attention_mask for e in encodings], dtype=np.int64
-        )
+        attention_mask = np.array([e.attention_mask for e in encodings], dtype=np.int64)
         token_type_ids = np.zeros_like(input_ids)
 
         outputs = self.session.run(