Implement backwards compatibility for text/texts arguments to vectorizers

Author: vishal-bala

redisvl/utils/vectorize/base.py

@@ -10,6 +10,7 @@
 from redisvl.extensions.cache.embeddings import EmbeddingsCache
 from redisvl.redis.utils import array_to_buffer
 from redisvl.schema.fields import VectorDataType
+from redisvl.utils.utils import deprecated_argument
 
 try:
     from PIL.Image import Image
@@ -67,9 +68,11 @@ def check_dtype(cls, dtype):
             )
         return dtype
 
+    @deprecated_argument("text", "content")
     def embed(
         self,
-        content: Any,
+        content: Any = None,
+        text: Any = None,
         preprocess: Optional[Callable] = None,
         as_buffer: bool = False,
         skip_cache: bool = False,
@@ -79,6 +82,7 @@ def embed(
 
         Args:
             content: The content to convert to a vector embedding
+            text: The text to convert to a vector embedding (deprecated - use `content` instead)
             preprocess: Function to apply to the content before embedding
             as_buffer: Return the embedding as a binary buffer instead of a list
             skip_cache: Bypass the cache for this request
@@ -91,6 +95,10 @@ def embed(
             >>> embedding = text_vectorizer.embed("Hello world")
             >>> embedding = image_vectorizer.embed(Image.open("test.png"))
         """
+        content = content or text
+        if not content:
+            raise ValueError("No content provided to embed.")
+
         # Apply preprocessing if provided
         if preprocess is not None:
             content = preprocess(content)
@@ -128,9 +136,11 @@ def embed(
         # Process and return result
         return self._process_embedding(embedding, as_buffer, self.dtype)
 
+    @deprecated_argument("texts", "contents")
     def embed_many(
         self,
-        contents: List[Any],
+        contents: Optional[List[Any]] = None,
+        texts: Optional[List[Any]] = None,
         preprocess: Optional[Callable] = None,
         batch_size: int = 10,
         as_buffer: bool = False,
@@ -141,6 +151,7 @@ def embed_many(
 
         Args:
             contents: List of content to convert to vector embeddings
+            texts: List of texts to convert to vector embeddings (deprecated - use `contents` instead)
             preprocess: Function to apply to each item before embedding
             batch_size: Number of items to process in each API call
             as_buffer: Return embeddings as binary buffers instead of lists
@@ -153,6 +164,7 @@ def embed_many(
         Examples:
             >>> embeddings = vectorizer.embed_many(["Hello", "World"], batch_size=2)
         """
+        contents = contents or texts
         if not contents:
             return []
 
@@ -186,9 +198,11 @@ def embed_many(
         # Process and return results
         return [self._process_embedding(emb, as_buffer, self.dtype) for emb in results]
 
+    @deprecated_argument("text", "content")
     async def aembed(
         self,
-        content: Any,
+        content: Any = None,
+        text: Any = None,
         preprocess: Optional[Callable] = None,
         as_buffer: bool = False,
         skip_cache: bool = False,
@@ -198,6 +212,7 @@ async def aembed(
 
         Args:
             content: The content to convert to a vector embedding
+            text: The text to convert to a vector embedding (deprecated - use `content` instead)
             preprocess: Function to apply to the content before embedding
             as_buffer: Return the embedding as a binary buffer instead of a list
             skip_cache: Bypass the cache for this request
@@ -209,6 +224,10 @@ async def aembed(
         Examples:
             >>> embedding = await vectorizer.aembed("Hello world")
         """
+        content = content or text
+        if not content:
+            raise ValueError("No content provided to embed.")
+
         # Apply preprocessing if provided
         if preprocess is not None:
             content = preprocess(content)
@@ -250,9 +269,11 @@ async def aembed(
         # Process and return result
         return self._process_embedding(embedding, as_buffer, self.dtype)
 
+    @deprecated_argument("texts", "contents")
     async def aembed_many(
         self,
-        contents: List[Any],
+        contents: Optional[List[Any]] = None,
+        texts: Optional[List[Any]] = None,
         preprocess: Optional[Callable] = None,
         batch_size: int = 10,
         as_buffer: bool = False,
@@ -263,6 +284,7 @@ async def aembed_many(
 
         Args:
             contents: List of content to convert to vector embeddings
+            texts: List of texts to convert to vector embeddings (deprecated - use `contents` instead)
             preprocess: Function to apply to each item before embedding
             batch_size: Number of texts to process in each API call
             as_buffer: Return embeddings as binary buffers instead of lists
@@ -275,6 +297,7 @@ async def aembed_many(
         Examples:
             >>> embeddings = await vectorizer.aembed_many(["Hello", "World"], batch_size=2)
         """
+        contents = contents or texts
         if not contents:
             return []
 
@@ -308,31 +331,45 @@ async def aembed_many(
         # Process and return results
         return [self._process_embedding(emb, as_buffer, self.dtype) for emb in results]
 
-    def _embed(self, content: Any, **kwargs) -> List[float]:
+    @deprecated_argument("text", "content")
+    def _embed(self, text: Any = "", content: Any = "", **kwargs) -> List[float]:
         """Generate a vector embedding for a single item."""
         raise NotImplementedError
 
+    @deprecated_argument("texts", "contents")
     def _embed_many(
-        self, contents: List[Any], batch_size: int = 10, **kwargs
+        self,
+        contents: Optional[List[Any]] = None,
+        texts: Optional[List[Any]] = None,
+        batch_size: int = 10,
+        **kwargs,
     ) -> List[List[float]]:
         """Generate vector embeddings for a batch of items."""
         raise NotImplementedError
 
-    async def _aembed(self, content: Any, **kwargs) -> List[float]:
+    @deprecated_argument("text", "content")
+    async def _aembed(self, content: Any = "", text: Any = "", **kwargs) -> List[float]:
         """Asynchronously generate a vector embedding for a single item."""
         logger.warning(
             "This vectorizer has no async embed method. Falling back to sync."
         )
-        return self._embed(content, **kwargs)
+        return self._embed(content=content or text, **kwargs)
 
+    @deprecated_argument("texts", "contents")
     async def _aembed_many(
-        self, contents: List[Any], batch_size: int = 10, **kwargs
+        self,
+        contents: Optional[List[Any]] = None,
+        texts: Optional[List[Any]] = None,
+        batch_size: int = 10,
+        **kwargs,
     ) -> List[List[float]]:
         """Asynchronously generate vector embeddings for a batch of items."""
         logger.warning(
             "This vectorizer has no async embed_many method. Falling back to sync."
         )
-        return self._embed_many(contents, batch_size, **kwargs)
+        return self._embed_many(
+            contents=contents or texts, batch_size=batch_size, **kwargs
+        )
 
     def _get_from_cache_batch(
         self, contents: List[Any], skip_cache: bool

redisvl/utils/vectorize/text/azureopenai.py

@@ -214,17 +214,19 @@ def _set_model_dims(self) -> int:
             # fall back (TODO get more specific)
             raise ValueError(f"Error setting embedding model dimensions: {str(e)}")
 
+    @deprecated_argument("text", "content")
     @retry(
         wait=wait_random_exponential(min=1, max=60),
         stop=stop_after_attempt(6),
         retry=retry_if_not_exception_type(TypeError),
     )
-    def _embed(self, content: str, **kwargs) -> List[float]:
+    def _embed(self, content: str = "", text: str = "", **kwargs) -> List[float]:
         """
         Generate a vector embedding for a single text using the AzureOpenAI API.
 
         Args:
             content: Text to embed
+            text: Text to embed (deprecated - use `content` instead)
             **kwargs: Additional parameters to pass to the AzureOpenAI API
 
         Returns:
@@ -234,6 +236,7 @@ def _embed(self, content: str, **kwargs) -> List[float]:
             TypeError: If text is not a string
             ValueError: If embedding fails
         """
+        content = content or text
         if not isinstance(content, str):
             raise TypeError("Must pass in a str value to embed.")
 
@@ -245,19 +248,25 @@ def _embed(self, content: str, **kwargs) -> List[float]:
         except Exception as e:
             raise ValueError(f"Embedding text failed: {e}")
 
+    @deprecated_argument("texts", "contents")
     @retry(
         wait=wait_random_exponential(min=1, max=60),
         stop=stop_after_attempt(6),
         retry=retry_if_not_exception_type(TypeError),
     )
     def _embed_many(
-        self, contents: List[str], batch_size: int = 10, **kwargs
+        self,
+        contents: Optional[List[str]] = None,
+        texts: Optional[List[str]] = None,
+        batch_size: int = 10,
+        **kwargs,
     ) -> List[List[float]]:
         """
         Generate vector embeddings for a batch of texts using the AzureOpenAI API.
 
         Args:
             contents: List of texts to embed
+            texts: List of texts to embed (deprecated - use `contents` instead)
             batch_size: Number of texts to process in each API call
             **kwargs: Additional parameters to pass to the AzureOpenAI API
 
@@ -268,6 +277,7 @@ def _embed_many(
             TypeError: If contents is not a list of strings
             ValueError: If embedding fails
         """
+        contents = contents or texts
         if not isinstance(contents, list):
             raise TypeError("Must pass in a list of str values to embed.")
         if contents and not isinstance(contents[0], str):
@@ -284,17 +294,19 @@ def _embed_many(
         except Exception as e:
             raise ValueError(f"Embedding texts failed: {e}")
 
+    @deprecated_argument("text", "content")
     @retry(
         wait=wait_random_exponential(min=1, max=60),
         stop=stop_after_attempt(6),
         retry=retry_if_not_exception_type(TypeError),
     )
-    async def _aembed(self, content: str, **kwargs) -> List[float]:
+    async def _aembed(self, content: str = "", text: str = "", **kwargs) -> List[float]:
         """
         Asynchronously generate a vector embedding for a single text using the AzureOpenAI API.
 
         Args:
             content: Text to embed
+            text: Text to embed (deprecated - use `content` instead)
             **kwargs: Additional parameters to pass to the AzureOpenAI API
 
         Returns:
@@ -304,6 +316,7 @@ async def _aembed(self, content: str, **kwargs) -> List[float]:
             TypeError: If content is not a string
             ValueError: If embedding fails
         """
+        content = content or text
         if not isinstance(content, str):
             raise TypeError("Must pass in a str value to embed.")
 
@@ -315,19 +328,25 @@ async def _aembed(self, content: str, **kwargs) -> List[float]:
         except Exception as e:
             raise ValueError(f"Embedding text failed: {e}")
 
+    @deprecated_argument("texts", "contents")
     @retry(
         wait=wait_random_exponential(min=1, max=60),
         stop=stop_after_attempt(6),
         retry=retry_if_not_exception_type(TypeError),
     )
     async def _aembed_many(
-        self, contents: List[str], batch_size: int = 10, **kwargs
+        self,
+        contents: Optional[List[str]] = None,
+        texts: Optional[List[str]] = None,
+        batch_size: int = 10,
+        **kwargs,
     ) -> List[List[float]]:
         """
         Asynchronously generate vector embeddings for a batch of texts using the AzureOpenAI API.
 
         Args:
             contents: List of texts to embed
+            texts: List of texts to embed (deprecated - use `contents` instead)
             batch_size: Number of texts to process in each API call
             **kwargs: Additional parameters to pass to the AzureOpenAI API
 
@@ -338,6 +357,7 @@ async def _aembed_many(
             TypeError: If contents is not a list of strings
             ValueError: If embedding fails
         """
+        contents = contents or texts
         if not isinstance(contents, list):
             raise TypeError("Must pass in a list of str values to embed.")
         if contents and not isinstance(contents[0], str):

redisvl/utils/vectorize/text/bedrock.py

@@ -0,0 +1,41 @@
+import warnings
+from typing import Any, List, Optional, Union
+
+from redisvl.utils.utils import (
+    deprecated_argument,
+    deprecated_class,
+    deprecated_function,
+)
+from redisvl.utils.vectorize.bedrock import BedrockVectorizer
+
+
+@deprecated_class(
+    name="BedrockTextVectorizer", replacement="Use BedrockVectorizer instead."
+)
+class BedrockTextVectorizer(BedrockVectorizer):
+    """A backwards-compatible alias for BedrockVectorizer."""
+
+    @deprecated_argument("text", "content")
+    def embed(
+        self, content: Any = "", text: Any = "", **kwargs
+    ) -> Union[List[float], bytes]:
+        """Generate a vector embedding for a single input using the AWS Bedrock API.
+
+        Deprecated: Use `BedrockVectorizer.embed` instead.
+        """
+        content = content or text
+        return super().embed(content=content, **kwargs)
+
+    @deprecated_argument("texts", "contents")
+    def embed_many(
+        self,
+        contents: Optional[List[Any]] = None,
+        texts: Optional[List[Any]] = None,
+        **kwargs,
+    ) -> List[List[float]]:
+        """Generate vector embeddings for a batch of inputs using the AWS Bedrock API.
+
+        Deprecated: Use `BedrockVectorizer.embed_many` instead.
+        """
+        contents = contents or texts
+        return super().embed_many(contents=contents, **kwargs)