From 6208773c77060c6604c267ad51dcc34000984fa9 Mon Sep 17 00:00:00 2001
From: ccurme <chester.curme@gmail.com>
Date: Tue, 10 Sep 2024 09:28:41 -0400
Subject: [PATCH] community[patch]: set protected namespaces on embeddings
 (#26156)

Also fix serdes test for langchain-google-genai.
---
 .../langchain_community/embeddings/baichuan.py       |  4 +---
 .../langchain_community/embeddings/bedrock.py        |  4 +---
 .../langchain_community/embeddings/clarifai.py       |  4 +---
 .../embeddings/cloudflare_workersai.py               |  4 +---
 .../langchain_community/embeddings/deepinfra.py      |  4 +---
 .../langchain_community/embeddings/fastembed.py      |  4 +---
 .../langchain_community/embeddings/huggingface.py    | 12 +++---------
 .../embeddings/huggingface_hub.py                    |  4 +---
 .../langchain_community/embeddings/ipex_llm.py       |  4 +---
 .../langchain_community/embeddings/localai.py        |  4 +---
 .../langchain_community/embeddings/modelscope_hub.py |  4 +---
 .../embeddings/oci_generative_ai.py                  |  4 +---
 .../langchain_community/embeddings/ollama.py         |  4 +---
 .../langchain_community/embeddings/openai.py         |  3 +--
 .../langchain_community/embeddings/openvino.py       |  4 +---
 .../langchain_community/embeddings/ovhcloud.py       |  4 +---
 .../embeddings/sagemaker_endpoint.py                 |  3 +--
 .../embeddings/spacy_embeddings.py                   |  4 +---
 .../langchain_community/embeddings/yandex.py         |  4 +---
 .../tests/unit_tests/load/test_serializable.py       |  5 +++--
 20 files changed, 24 insertions(+), 63 deletions(-)

diff --git a/libs/community/langchain_community/embeddings/baichuan.py b/libs/community/langchain_community/embeddings/baichuan.py
index 68fc6d8d954..082ff369ed7 100644
--- a/libs/community/langchain_community/embeddings/baichuan.py
+++ b/libs/community/langchain_community/embeddings/baichuan.py
@@ -70,9 +70,7 @@ class BaichuanTextEmbeddings(BaseModel, Embeddings):
     chunk_size: int = 16
     """Chunk size when multiple texts are input"""
 
-    model_config = ConfigDict(
-        populate_by_name=True,
-    )
+    model_config = ConfigDict(populate_by_name=True, protected_namespaces=())
 
     @model_validator(mode="after")
     def validate_environment(self) -> Self:
diff --git a/libs/community/langchain_community/embeddings/bedrock.py b/libs/community/langchain_community/embeddings/bedrock.py
index 671073e743f..7287e1acee9 100644
--- a/libs/community/langchain_community/embeddings/bedrock.py
+++ b/libs/community/langchain_community/embeddings/bedrock.py
@@ -75,9 +75,7 @@ class BedrockEmbeddings(BaseModel, Embeddings):
     normalize: bool = False
     """Whether the embeddings should be normalized to unit vectors"""
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     @model_validator(mode="after")
     def validate_environment(self) -> Self:
diff --git a/libs/community/langchain_community/embeddings/clarifai.py b/libs/community/langchain_community/embeddings/clarifai.py
index 4c35fcf75b8..e460020bef1 100644
--- a/libs/community/langchain_community/embeddings/clarifai.py
+++ b/libs/community/langchain_community/embeddings/clarifai.py
@@ -43,9 +43,7 @@ class ClarifaiEmbeddings(BaseModel, Embeddings):
     model: Any = Field(default=None, exclude=True)  #: :meta private:
     api_base: str = "https://api.clarifai.com"
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     @model_validator(mode="before")
     @classmethod
diff --git a/libs/community/langchain_community/embeddings/cloudflare_workersai.py b/libs/community/langchain_community/embeddings/cloudflare_workersai.py
index 45fadb39c68..87d15c60992 100644
--- a/libs/community/langchain_community/embeddings/cloudflare_workersai.py
+++ b/libs/community/langchain_community/embeddings/cloudflare_workersai.py
@@ -43,9 +43,7 @@ class CloudflareWorkersAIEmbeddings(BaseModel, Embeddings):
 
         self.headers = {"Authorization": f"Bearer {self.api_token}"}
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Compute doc embeddings using Cloudflare Workers AI.
diff --git a/libs/community/langchain_community/embeddings/deepinfra.py b/libs/community/langchain_community/embeddings/deepinfra.py
index 05f3ed982b4..d0d2c476011 100644
--- a/libs/community/langchain_community/embeddings/deepinfra.py
+++ b/libs/community/langchain_community/embeddings/deepinfra.py
@@ -54,9 +54,7 @@ class DeepInfraEmbeddings(BaseModel, Embeddings):
     batch_size: int = MAX_BATCH_SIZE
     """Batch size for embedding requests."""
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     @pre_init
     def validate_environment(cls, values: Dict) -> Dict:
diff --git a/libs/community/langchain_community/embeddings/fastembed.py b/libs/community/langchain_community/embeddings/fastembed.py
index fe654aa4459..75dcbda805f 100644
--- a/libs/community/langchain_community/embeddings/fastembed.py
+++ b/libs/community/langchain_community/embeddings/fastembed.py
@@ -67,9 +67,7 @@ class FastEmbedEmbeddings(BaseModel, Embeddings):
 
     _model: Any  # : :meta private:
 
-    model_config = ConfigDict(
-        extra="allow",
-    )
+    model_config = ConfigDict(extra="allow", protected_namespaces=())
 
     @pre_init
     def validate_environment(cls, values: Dict) -> Dict:
diff --git a/libs/community/langchain_community/embeddings/huggingface.py b/libs/community/langchain_community/embeddings/huggingface.py
index f4c2f38f8fd..e6eeadca580 100644
--- a/libs/community/langchain_community/embeddings/huggingface.py
+++ b/libs/community/langchain_community/embeddings/huggingface.py
@@ -93,9 +93,7 @@ class HuggingFaceEmbeddings(BaseModel, Embeddings):
             self.model_name, cache_folder=self.cache_folder, **self.model_kwargs
         )
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Compute doc embeddings using a HuggingFace transformer model.
@@ -209,9 +207,7 @@ class HuggingFaceInstructEmbeddings(BaseModel, Embeddings):
                 )
             self.show_progress = self.encode_kwargs.pop("show_progress_bar")
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Compute doc embeddings using a HuggingFace instruct model.
@@ -350,9 +346,7 @@ class HuggingFaceBgeEmbeddings(BaseModel, Embeddings):
                 )
             self.show_progress = self.encode_kwargs.pop("show_progress_bar")
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Compute doc embeddings using a HuggingFace transformer model.
diff --git a/libs/community/langchain_community/embeddings/huggingface_hub.py b/libs/community/langchain_community/embeddings/huggingface_hub.py
index cea56ad11c7..33711181a19 100644
--- a/libs/community/langchain_community/embeddings/huggingface_hub.py
+++ b/libs/community/langchain_community/embeddings/huggingface_hub.py
@@ -48,9 +48,7 @@ class HuggingFaceHubEmbeddings(BaseModel, Embeddings):
 
     huggingfacehub_api_token: Optional[str] = None
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     @model_validator(mode="before")
     @classmethod
diff --git a/libs/community/langchain_community/embeddings/ipex_llm.py b/libs/community/langchain_community/embeddings/ipex_llm.py
index be18e2f3461..1f53cd7fd7b 100644
--- a/libs/community/langchain_community/embeddings/ipex_llm.py
+++ b/libs/community/langchain_community/embeddings/ipex_llm.py
@@ -106,9 +106,7 @@ class IpexLLMBgeEmbeddings(BaseModel, Embeddings):
         if "-zh" in self.model_name:
             self.query_instruction = DEFAULT_QUERY_BGE_INSTRUCTION_ZH
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Compute doc embeddings using a HuggingFace transformer model.
diff --git a/libs/community/langchain_community/embeddings/localai.py b/libs/community/langchain_community/embeddings/localai.py
index b59433ad2e2..f4ca9fa9723 100644
--- a/libs/community/langchain_community/embeddings/localai.py
+++ b/libs/community/langchain_community/embeddings/localai.py
@@ -166,9 +166,7 @@ class LocalAIEmbeddings(BaseModel, Embeddings):
     model_kwargs: Dict[str, Any] = Field(default_factory=dict)
     """Holds any model parameters valid for `create` call not explicitly specified."""
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     @model_validator(mode="before")
     @classmethod
diff --git a/libs/community/langchain_community/embeddings/modelscope_hub.py b/libs/community/langchain_community/embeddings/modelscope_hub.py
index 878f05a4fe2..a0c78633916 100644
--- a/libs/community/langchain_community/embeddings/modelscope_hub.py
+++ b/libs/community/langchain_community/embeddings/modelscope_hub.py
@@ -39,9 +39,7 @@ class ModelScopeEmbeddings(BaseModel, Embeddings):
             model_revision=self.model_revision,
         )
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Compute doc embeddings using a modelscope embedding model.
diff --git a/libs/community/langchain_community/embeddings/oci_generative_ai.py b/libs/community/langchain_community/embeddings/oci_generative_ai.py
index 72c4059d2be..10bef6d441d 100644
--- a/libs/community/langchain_community/embeddings/oci_generative_ai.py
+++ b/libs/community/langchain_community/embeddings/oci_generative_ai.py
@@ -85,9 +85,7 @@ class OCIGenAIEmbeddings(BaseModel, Embeddings):
     """Batch size of OCI GenAI embedding requests. OCI GenAI may handle up to 96 texts
      per request"""
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     @pre_init
     def validate_environment(cls, values: Dict) -> Dict:  # pylint: disable=no-self-argument
diff --git a/libs/community/langchain_community/embeddings/ollama.py b/libs/community/langchain_community/embeddings/ollama.py
index 13be9a3cd84..b55cff9b6fa 100644
--- a/libs/community/langchain_community/embeddings/ollama.py
+++ b/libs/community/langchain_community/embeddings/ollama.py
@@ -141,9 +141,7 @@ class OllamaEmbeddings(BaseModel, Embeddings):
         """Get the identifying parameters."""
         return {**{"model": self.model}, **self._default_params}
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def _process_emb_response(self, input: str) -> List[float]:
         """Process a response from the API.
diff --git a/libs/community/langchain_community/embeddings/openai.py b/libs/community/langchain_community/embeddings/openai.py
index a7abde76ea3..770bb765e81 100644
--- a/libs/community/langchain_community/embeddings/openai.py
+++ b/libs/community/langchain_community/embeddings/openai.py
@@ -255,8 +255,7 @@ class OpenAIEmbeddings(BaseModel, Embeddings):
     """Optional httpx.Client."""
 
     model_config = ConfigDict(
-        populate_by_name=True,
-        extra="forbid",
+        populate_by_name=True, extra="forbid", protected_namespaces=()
     )
 
     @model_validator(mode="before")
diff --git a/libs/community/langchain_community/embeddings/openvino.py b/libs/community/langchain_community/embeddings/openvino.py
index 8a83cfd1679..d58bf92d255 100644
--- a/libs/community/langchain_community/embeddings/openvino.py
+++ b/libs/community/langchain_community/embeddings/openvino.py
@@ -254,9 +254,7 @@ class OpenVINOEmbeddings(BaseModel, Embeddings):
 
         return all_embeddings
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def embed_documents(self, texts: List[str]) -> List[List[float]]:
         """Compute doc embeddings using a HuggingFace transformer model.
diff --git a/libs/community/langchain_community/embeddings/ovhcloud.py b/libs/community/langchain_community/embeddings/ovhcloud.py
index 51b0dd2f34a..5786a761f49 100644
--- a/libs/community/langchain_community/embeddings/ovhcloud.py
+++ b/libs/community/langchain_community/embeddings/ovhcloud.py
@@ -23,9 +23,7 @@ class OVHCloudEmbeddings(BaseModel, Embeddings):
     """ OVHcloud AI Endpoints region"""
     region: str = "kepler"
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     def __init__(self, **kwargs: Any):
         super().__init__(**kwargs)
diff --git a/libs/community/langchain_community/embeddings/sagemaker_endpoint.py b/libs/community/langchain_community/embeddings/sagemaker_endpoint.py
index d5fa48f3b03..d69cbd92ea8 100644
--- a/libs/community/langchain_community/embeddings/sagemaker_endpoint.py
+++ b/libs/community/langchain_community/embeddings/sagemaker_endpoint.py
@@ -111,8 +111,7 @@ class SagemakerEndpointEmbeddings(BaseModel, Embeddings):
     """
 
     model_config = ConfigDict(
-        arbitrary_types_allowed=True,
-        extra="forbid",
+        arbitrary_types_allowed=True, extra="forbid", protected_namespaces=()
     )
 
     @pre_init
diff --git a/libs/community/langchain_community/embeddings/spacy_embeddings.py b/libs/community/langchain_community/embeddings/spacy_embeddings.py
index 63c71e07c3f..0413741c4df 100644
--- a/libs/community/langchain_community/embeddings/spacy_embeddings.py
+++ b/libs/community/langchain_community/embeddings/spacy_embeddings.py
@@ -22,9 +22,7 @@ class SpacyEmbeddings(BaseModel, Embeddings):
     model_name: str = "en_core_web_sm"
     nlp: Optional[Any] = None
 
-    model_config = ConfigDict(
-        extra="forbid",
-    )
+    model_config = ConfigDict(extra="forbid", protected_namespaces=())
 
     @model_validator(mode="before")
     @classmethod
diff --git a/libs/community/langchain_community/embeddings/yandex.py b/libs/community/langchain_community/embeddings/yandex.py
index fdc6dc6873e..59637bf1001 100644
--- a/libs/community/langchain_community/embeddings/yandex.py
+++ b/libs/community/langchain_community/embeddings/yandex.py
@@ -71,9 +71,7 @@ class YandexGPTEmbeddings(BaseModel, Embeddings):
     If you provide personal data, confidential information, disable logging."""
     grpc_metadata: Sequence
 
-    model_config = ConfigDict(
-        populate_by_name=True,
-    )
+    model_config = ConfigDict(populate_by_name=True, protected_namespaces=())
 
     @pre_init
     def validate_environment(cls, values: Dict) -> Dict:
diff --git a/libs/community/tests/unit_tests/load/test_serializable.py b/libs/community/tests/unit_tests/load/test_serializable.py
index 35ea4b169d7..819d47d9a73 100644
--- a/libs/community/tests/unit_tests/load/test_serializable.py
+++ b/libs/community/tests/unit_tests/load/test_serializable.py
@@ -112,8 +112,9 @@ def test_serializable_mapping() -> None:
             "chat_models",
             "ChatGroq",
         ),
-        # TODO(0.3): For now we're skipping this test. Need to fix
-        # so that it only runs when langchain-aws is installed.
+        # TODO(0.3): For now we're skipping the below two tests. Need to fix
+        # so that it only runs when langchain-aws, langchain-google-genai
+        # are installed.
         ("langchain", "chat_models", "bedrock", "ChatBedrock"): (
             "langchain_aws",
             "chat_models",