babylon.rag.pre_embeddings.cache_manager

Embedding cache management for the RAG system.

This module provides functionality for caching embeddings to reduce duplicate operations and API costs.

Classes

CacheConfig(**data)	Configuration for embedding cache management.
EmbeddingCacheManager([config])	Manages caching of embeddings to reduce duplicate operations.

class babylon.rag.pre_embeddings.cache_manager.CacheConfig(**data)

Bases: BaseModel

Configuration for embedding cache management.

Parameters:

• max_cache_size (int)

• persist_cache (bool)

• cache_file_path (str)

• hash_algorithm (str)

max_cache_size

Maximum number of embeddings to keep in cache

persist_cache

Whether to persist cache to disk

cache_file_path

Path to cache file for persistence

hash_algorithm

Algorithm to use for content hashing

model_config: ClassVar[ConfigDict] = {'frozen': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

max_cache_size: int

persist_cache: bool

cache_file_path: str

hash_algorithm: str

class babylon.rag.pre_embeddings.cache_manager.EmbeddingCacheManager(config=None)

Bases: object

Manages caching of embeddings to reduce duplicate operations.

This class handles caching of embeddings based on content hashes, with support for LRU eviction, persistence, and metrics collection.

Parameters:

config (CacheConfig | None)

__init__(config=None)

Initialize with configuration options.

Parameters:

config (CacheConfig | None) – Configuration for cache behavior

hash_content(content)

Generate a hash for the given content.

Parameters:

content (str) – Content to hash

Return type:

str

Returns:

Hash string for the content

get_from_cache(content_hash)

Retrieve an embedding from cache by content hash.

Parameters:

content_hash (str) – Hash of the content to retrieve embedding for

Return type:

list[float] | None

Returns:

Cached embedding or None if not found

add_to_cache(content_hash, embedding)

Add or update an embedding in the cache.

Parameters:

• content_hash (str) – Hash of the content

• embedding (list[float]) – Embedding vector to cache

Return type:

None

clear_cache()

Clear all entries from the cache.

Return type:

None

save_cache()

Save the cache to disk if persistence is enabled.

Return type:

None

load_cache()

Load the cache from disk if persistence is enabled.

Return type:

None

get_cache_stats()

Get statistics about the current cache state.

Return type:

dict[str, Any]

Returns:

Dictionary of cache statistics