SentenceTransformerEmbed support additional args passed to the library.

docs/docs/ops/functions.md

@@ -33,6 +33,7 @@ Return type: `Table`, each row represents a chunk, with the following sub fields
 The spec takes the following fields:
 
 *   `model` (type: `str`, required): The name of the SentenceTransformer model to use.
+*   `args` (type: `dict[str, Any]`, optional): Additional arguments to pass to the SentenceTransformer constructor. e.g. `{"trust_remote_code": True}`
 
 Input data:
 

python/cocoindex/functions.py

@@ -1,5 +1,5 @@
 """All builtin functions."""
-from typing import Annotated
+from typing import Annotated, Any
 
 import json
 import sentence_transformers
@@ -13,8 +13,16 @@ class SplitRecursively(op.FunctionSpec):
     language: str | None = None
 
 class SentenceTransformerEmbed(op.FunctionSpec):
-    """Run the sentence transformer"""
+    """
+    `SentenceTransformerEmbed` embeds a text into a vector space using the [SentenceTransformer](https://huggingface.co/sentence-transformers) library.
+
+    Args:
+
+        model: The name of the SentenceTransformer model to use.
+        args: Additional arguments to pass to the SentenceTransformer constructor. e.g. {"trust_remote_code": True}
+    """
     model: str
+    args: dict[str, Any] | None = None
 
 @op.executor_class(gpu=True, cache=True, behavior_version=1)
 class SentenceTransformerEmbedExecutor:
@@ -24,7 +32,8 @@ class SentenceTransformerEmbedExecutor:
     _model: sentence_transformers.SentenceTransformer
 
     def analyze(self, text = None):
-        self._model = sentence_transformers.SentenceTransformer(self.spec.model, 3)
+        args = self.spec.args or {}
+        self._model = sentence_transformers.SentenceTransformer(self.spec.model, **args)
         dim = self._model.get_sentence_embedding_dimension()
         return Annotated[list[Float32], Vector(dim=dim), TypeAttr("cocoindex.io/vector_origin_text", json.loads(text.analyzed_value))]