Support client-side schema validation using Pydantic

.github/workflows/test.yml

@@ -133,6 +133,11 @@ jobs:
         with:
           credentials_json: ${{ secrets.GOOGLE_CREDENTIALS }}
 
+      - name: Set HuggingFace token
+        run: |
+          mkdir -p ~/.huggingface
+          echo '{"token":"${{ secrets.HF_TOKEN }}"}' > ~/.huggingface/token 
+
       - name: Run tests
         if: matrix.connection == 'plain' && matrix.redis-version == 'latest'
         env:
@@ -149,6 +154,7 @@ jobs:
           OPENAI_API_VERSION: ${{ secrets.OPENAI_API_VERSION }}
           AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
           AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
         run: |
           make test-all
 
@@ -173,6 +179,7 @@ jobs:
           OPENAI_API_VERSION: ${{ secrets.OPENAI_API_VERSION }}
           AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
           AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
         run: |
           docker run -d --name redis -p 6379:6379 redis/redis-stack-server:latest
           make test-notebooks

pyproject.toml

@@ -33,6 +33,8 @@ tenacity = ">=8.2.2"
 tabulate = "^0.9.0"
 ml-dtypes = "^0.4.0"
 python-ulid = "^3.0.0"
+jsonpath-ng = "^1.5.0"
+
 openai = { version = "^1.13.0", optional = true }
 sentence-transformers = { version = "^3.4.0", optional = true }
 scipy = [

redisvl/exceptions.py

@@ -1,10 +1,32 @@
-class RedisVLException(Exception):
-    """Base RedisVL exception"""
+"""
+RedisVL Exception Classes
 
+This module defines all custom exceptions used throughout the RedisVL library.
+"""
 
-class RedisModuleVersionError(RedisVLException):
-    """Invalid module versions installed"""
 
+class RedisVLError(Exception):
+    """Base exception for all RedisVL errors."""
 
-class RedisSearchError(RedisVLException):
-    """Error while performing a search or aggregate request"""
+    pass
+
+
+class RedisModuleVersionError(RedisVLError):
+    """Error raised when required Redis modules are missing or have incompatible versions."""
+
+    pass
+
+
+class RedisSearchError(RedisVLError):
+    """Error raised for Redis Search specific operations."""
+
+    pass
+
+
+class SchemaValidationError(RedisVLError):
+    """Error when validating data against a schema."""
+
+    def __init__(self, message, index=None):
+        if index is not None:
+            message = f"Validation failed for object at index {index}: {message}"
+        super().__init__(message)

redisvl/extensions/llmcache/semantic.py

@@ -95,12 +95,8 @@ def __init__(
         }
 
         # Use the index name as the key prefix by default
-        if "prefix" in kwargs:
-            prefix = kwargs["prefix"]
-        else:
-            prefix = name
-
-        dtype = kwargs.get("dtype")
+        prefix = kwargs.pop("prefix", name)
+        dtype = kwargs.pop("dtype", None)
 
         # Validate a provided vectorizer or set the default
         if vectorizer:
@@ -111,7 +107,10 @@ def __init__(
                     f"Provided dtype {dtype} does not match vectorizer dtype {vectorizer.dtype}"
                 )
         else:
-            vectorizer_kwargs = {"dtype": dtype} if dtype else {}
+            vectorizer_kwargs = kwargs
+
+            if dtype:
+                vectorizer_kwargs.update(**{"dtype": dtype})
 
             vectorizer = HFTextVectorizer(
                 model="sentence-transformers/all-mpnet-base-v2",

redisvl/extensions/router/semantic.py

@@ -72,7 +72,7 @@ def __init__(
             connection_kwargs (Dict[str, Any]): The connection arguments
                 for the redis client. Defaults to empty {}.
         """
-        dtype = kwargs.get("dtype")
+        dtype = kwargs.pop("dtype", None)
 
         # Validate a provided vectorizer or set the default
         if vectorizer:
@@ -83,8 +83,15 @@ def __init__(
                     f"Provided dtype {dtype} does not match vectorizer dtype {vectorizer.dtype}"
                 )
         else:
-            vectorizer_kwargs = {"dtype": dtype} if dtype else {}
-            vectorizer = HFTextVectorizer(**vectorizer_kwargs)
+            vectorizer_kwargs = kwargs
+
+            if dtype:
+                vectorizer_kwargs.update(**{"dtype": dtype})
+
+            vectorizer = HFTextVectorizer(
+                model="sentence-transformers/all-mpnet-base-v2",
+                **vectorizer_kwargs,
+            )
 
         if routing_config is None:
             routing_config = RoutingConfig()

redisvl/extensions/session_manager/semantic_session.py

@@ -71,7 +71,7 @@ def __init__(
         super().__init__(name, session_tag)
 
         prefix = prefix or name
-        dtype = kwargs.get("dtype")
+        dtype = kwargs.pop("dtype", None)
 
         # Validate a provided vectorizer or set the default
         if vectorizer:
@@ -82,10 +82,13 @@ def __init__(
                     f"Provided dtype {dtype} does not match vectorizer dtype {vectorizer.dtype}"
                 )
         else:
-            vectorizer_kwargs = {"dtype": dtype} if dtype else {}
+            vectorizer_kwargs = kwargs
+
+            if dtype:
+                vectorizer_kwargs.update(**{"dtype": dtype})
 
             vectorizer = HFTextVectorizer(
-                model="sentence-transformers/msmarco-distilbert-cos-v5",
+                model="sentence-transformers/all-mpnet-base-v2",
                 **vectorizer_kwargs,
             )