(feat) working - sync semantic caching

litellm/caching.py

@@ -223,94 +223,161 @@ class RedisCache(BaseCache):
         self.redis_client.delete(key)
 
 
-class RedisSemanticCache(RedisCache):
+class RedisSemanticCache(BaseCache):
-    def __init__(self, host, port, password, **kwargs):
+    def __init__(
-        super().__init__()
+        self,
+        host=None,
+        port=None,
+        password=None,
+        redis_url=None,
+        similarity_threshold=None,
+        **kwargs,
+    ):
+        from redisvl.index import SearchIndex
+        from redisvl.query import VectorQuery
 
-        # from redis.commands.search.field import TagField, TextField, NumericField, VectorField
+        print_verbose(
-        # from redis.commands.search.indexDefinition import IndexDefinition, IndexType
+            "redis semantic-cache initializing INDEX - litellm_semantic_cache_index"
-        # from redis.commands.search.query import Query
+        )
+        if similarity_threshold is None:
+            raise Exception("similarity_threshold must be provided, passed None")
+        self.similarity_threshold = similarity_threshold
+        schema = {
+            "index": {
+                "name": "litellm_semantic_cache_index",
+                "prefix": "litellm",
+                "storage_type": "hash",
+            },
+            "fields": {
+                "text": [{"name": "response"}],
+                "text": [{"name": "prompt"}],
+                "vector": [
+                    {
+                        "name": "litellm_embedding",
+                        "dims": 1536,
+                        "distance_metric": "cosine",
+                        "algorithm": "flat",
+                        "datatype": "float32",
+                    }
+                ],
+            },
+        }
+        self.index = SearchIndex.from_dict(schema)
+        if redis_url is None:
+            # if no url passed, check if host, port and password are passed, if not raise an Exception
+            if host is None or port is None or password is None:
+                raise Exception(f"Redis host, port, and password must be provided")
+            redis_url = "redis://:" + password + "@" + host + ":" + port
+        print_verbose(f"redis semantic-cache redis_url: {redis_url}")
+        self.index.connect(redis_url=redis_url)
+        self.index.create(overwrite=False)  # don't overwrite existing index
 
-        # INDEX_NAME = 'idx:litellm_completion_response_vss'
+    def _get_cache_logic(self, cached_response: Any):
-        # DOC_PREFIX = 'bikes:'
+        """
+        Common 'get_cache_logic' across sync + async redis client implementations
+        """
+        if cached_response is None:
+            return cached_response
 
-        # try:
+        # check if cached_response is bytes
-        #     # check to see if index exists
+        if isinstance(cached_response, bytes):
-        #     client.ft(INDEX_NAME).info()
+            cached_response = cached_response.decode("utf-8")
-        #     print('Index already exists!')
-        # except:
-        #     # schema
-        #     schema = (
-        #         TextField('$.model', no_stem=True, as_name='model'),
-        #         TextField('$.brand', no_stem=True, as_name='brand'),
-        #         NumericField('$.price', as_name='price'),
-        #         TagField('$.type', as_name='type'),
-        #         TextField('$.description', as_name='description'),
-        #         VectorField('$.description_embeddings',
-        #             'FLAT', {
-        #                 'TYPE': 'FLOAT32',
-        #                 'DIM': VECTOR_DIMENSION,
-        #                 'DISTANCE_METRIC': 'COSINE',
-        #             },  as_name='vector'
-        #         ),
-        #     )
 
-        #     # index Definition
+        try:
-        #     definition = IndexDefinition(prefix=[DOC_PREFIX], index_type=IndexType.JSON)
+            cached_response = json.loads(
-
+                cached_response
-        #     # create Index
+            )  # Convert string to dictionary
-        #     client.ft(INDEX_NAME).create_index(fields=schema, definition=definition)
+        except:
+            cached_response = ast.literal_eval(cached_response)
+        return cached_response
 
     def set_cache(self, key, value, **kwargs):
-        ttl = kwargs.get("ttl", None)
+        import numpy as np
-        print_verbose(f"Set Redis Cache: key: {key}\nValue {value}\nttl={ttl}")
-        try:
-            # get text response
-            # print("in redis semantic cache: value: ", value)
-            llm_response = value["response"]
 
-            # if llm_response is a string, convert it to a dictionary
+        print_verbose(f"redis semantic-cache set_cache, kwargs: {kwargs}")
-            if isinstance(llm_response, str):
-                llm_response = json.loads(llm_response)
 
-            # print("converted llm_response: ", llm_response)
+        # get the prompt
-            response = llm_response["choices"][0]["message"]["content"]
+        messages = kwargs["messages"]
+        prompt = ""
+        for message in messages:
+            prompt += message["content"]
 
-            # create embedding response
+        # create an embedding for prompt
+        embedding_response = litellm.embedding(
+            model="text-embedding-ada-002",
+            input=prompt,
+            cache={"no-store": True, "no-cache": True},
+        )
 
-            embedding_response = litellm.embedding(
+        # get the embedding
-                model="text-embedding-ada-002",
+        embedding = embedding_response["data"][0]["embedding"]
-                input=response,
-                cache={"no-store": True},
-            )
 
-            raw_embedding = embedding_response["data"][0]["embedding"]
+        # make the embedding a numpy array, convert to bytes
-            raw_embedding_dimension = len(raw_embedding)
+        embedding_bytes = np.array(embedding, dtype=np.float32).tobytes()
+        value = str(value)
+        assert isinstance(value, str)
 
-            # print("embedding: ", raw_embedding)
+        new_data = [
-            key = "litellm-semantic:" + key
+            {"response": value, "prompt": prompt, "litellm_embedding": embedding_bytes}
-            self.redis_client.json().set(
+        ]
-                name=key,
-                path="$",
-                obj=json.dumps(
-                    {
-                        "response": response,
-                        "embedding": raw_embedding,
-                        "dimension": raw_embedding_dimension,
-                    }
-                ),
-            )
 
-            stored_redis_value = self.redis_client.json().get(name=key)
+        # Add more data
+        keys = self.index.load(new_data)
 
-            # print("Stored Redis Value: ", stored_redis_value)
+        pass
-
-        except Exception as e:
-            # print("Error occurred: ", e)
-            # NON blocking - notify users Redis is throwing an exception
-            logging.debug("LiteLLM Caching: set() - Got exception from REDIS : ", e)
 
     def get_cache(self, key, **kwargs):
+        print_verbose(f"redis semantic-cache get_cache, kwargs: {kwargs}")
+        from redisvl.query import VectorQuery
+        import numpy as np
+
+        # query
+
+        # get the messages
+        messages = kwargs["messages"]
+        prompt = ""
+        for message in messages:
+            prompt += message["content"]
+
+        # convert to embedding
+        embedding_response = litellm.embedding(
+            model="text-embedding-ada-002",
+            input=prompt,
+            cache={"no-store": True, "no-cache": True},
+        )
+
+        # get the embedding
+        embedding = embedding_response["data"][0]["embedding"]
+
+        query = VectorQuery(
+            vector=embedding,
+            vector_field_name="litellm_embedding",
+            return_fields=["response", "prompt", "vector_distance"],
+            num_results=1,
+        )
+
+        results = self.index.query(query)
+
+        vector_distance = results[0]["vector_distance"]
+        vector_distance = float(vector_distance)
+        similarity = 1 - vector_distance
+        cached_prompt = results[0]["prompt"]
+
+        # check similarity, if more than self.similarity_threshold, return results
+        print_verbose(
+            f"semantic cache: similarity threshold: {self.similarity_threshold}, similarity: {similarity}, prompt: {prompt}, closest_cached_prompt: {cached_prompt}"
+        )
+        if similarity > self.similarity_threshold:
+            # cache hit !
+            cached_value = results[0]["response"]
+            print_verbose(
+                f"got a cache hit, similarity: {similarity}, Current prompt: {prompt}, cached_prompt: {cached_prompt}"
+            )
+            return self._get_cache_logic(cached_response=cached_value)
+        else:
+            # cache miss !
+            return None
+
         pass
 
     async def async_set_cache(self, key, value, **kwargs):
@@ -527,6 +594,7 @@ class Cache:
         host: Optional[str] = None,
         port: Optional[str] = None,
         password: Optional[str] = None,
+        similarity_threshold: Optional[float] = None,
         supported_call_types: Optional[
             List[Literal["completion", "acompletion", "embedding", "aembedding"]]
         ] = ["completion", "acompletion", "embedding", "aembedding"],
@@ -547,10 +615,12 @@ class Cache:
         Initializes the cache based on the given type.
 
         Args:
-            type (str, optional): The type of cache to initialize. Can be "local" or "redis". Defaults to "local".
+            type (str, optional): The type of cache to initialize. Can be "local", "redis", "redis-semantic", or "s3". Defaults to "local".
             host (str, optional): The host address for the Redis cache. Required if type is "redis".
             port (int, optional): The port number for the Redis cache. Required if type is "redis".
             password (str, optional): The password for the Redis cache. Required if type is "redis".
+            similarity_threshold (float, optional): The similarity threshold for semantic-caching, Required if type is "redis-semantic"
+
             supported_call_types (list, optional): List of call types to cache for. Defaults to cache == on for all call types.
             **kwargs: Additional keyword arguments for redis.Redis() cache
 
@@ -563,7 +633,13 @@ class Cache:
         if type == "redis":
             self.cache: BaseCache = RedisCache(host, port, password, **kwargs)
         elif type == "redis-semantic":
-            self.cache = RedisSemanticCache(host, port, password, **kwargs)
+            self.cache = RedisSemanticCache(
+                host,
+                port,
+                password,
+                similarity_threshold=similarity_threshold,
+                **kwargs,
+            )
         elif type == "local":
             self.cache = InMemoryCache()
         elif type == "s3":
@@ -743,6 +819,7 @@ class Cache:
             The cached result if it exists, otherwise None.
         """
         try:  # never block execution
+            messages = kwargs.get("messages", [])
             if "cache_key" in kwargs:
                 cache_key = kwargs["cache_key"]
             else:
@@ -752,7 +829,7 @@ class Cache:
                 max_age = cache_control_args.get(
                     "s-max-age", cache_control_args.get("s-maxage", float("inf"))
                 )
-                cached_result = self.cache.get_cache(cache_key)
+                cached_result = self.cache.get_cache(cache_key, messages=messages)
                 return self._get_cache_logic(
                     cached_result=cached_result, max_age=max_age
                 )