haive.core.engine.retriever.providers.ZepRetrieverConfig

Zep Retriever implementation for the Haive framework.

from typing import Any This module provides a configuration class for the Zep retriever, which retrieves conversation history and memory from Zep’s long-term memory store. Zep is designed for storing, searching, and enriching conversational AI chat histories with metadata, summaries, and semantic search capabilities.

The ZepRetriever works by: 1. Connecting to a Zep memory store 2. Searching conversation history using semantic similarity 3. Retrieving relevant chat messages and context 4. Providing conversation memory for AI applications

This retriever is particularly useful when: - Building conversational AI applications with long-term memory - Need to retrieve relevant conversation history - Want to maintain context across multiple chat sessions - Building customer support or chatbot applications - Creating personalized AI assistants with memory

The implementation integrates with LangChain’s ZepRetriever while providing a consistent Haive configuration interface with secure API key management.

Classes

ZepRetrieverConfig

Configuration for Zep retriever in the Haive framework.

Module Contents

class haive.core.engine.retriever.providers.ZepRetrieverConfig.ZepRetrieverConfig

Bases: haive.core.common.mixins.secure_config.SecureConfigMixin, haive.core.engine.retriever.retriever.BaseRetrieverConfig

Configuration for Zep retriever in the Haive framework.

This retriever searches conversational memory stored in Zep and returns relevant chat history and context for AI applications.

retriever_type

The type of retriever (always ZEP).

Type:

RetrieverType

session_id

Zep session ID for conversation history.

Type:

str

url

Zep server URL.

Type:

str

api_key

Zep API key (auto-resolved from ZEP_API_KEY).

Type:

Optional[SecretStr]

top_k

Number of memory entries to retrieve.

Type:

int

search_type

Type of search to perform.

Type:

str

search_scope

Scope of search (messages, summary, etc.).

Type:

str

Examples

>>> from haive.core.engine.retriever import ZepRetrieverConfig
>>>
>>> # Create the Zep retriever config
>>> config = ZepRetrieverConfig(
...     name="zep_retriever",
...     session_id="user-123-session",
...     url="http://localhost:8000",
...     top_k=10,
...     search_type="similarity",
...     search_scope="messages"
... )
>>>
>>> # Instantiate and use the retriever
>>> retriever = config.instantiate()
>>> docs = retriever.get_relevant_documents("what did we discuss about machine learning?")
>>>
>>> # Example for summary search
>>> summary_config = ZepRetrieverConfig(
...     name="zep_summary_retriever",
...     session_id="user-123-session",
...     url="http://localhost:8000",
...     search_scope="summary",
...     top_k=5
... )

get_input_fields()

Return input field definitions for Zep retriever.

Return type:

dict[str, tuple[type, Any]]

get_output_fields()

Return output field definitions for Zep retriever.

Return type:

dict[str, tuple[type, Any]]

instantiate()

Create a Zep retriever from this configuration.

Returns:

Instantiated retriever ready for memory search.

Return type:

ZepRetriever

Raises:

• ImportError – If required packages are not available.

• ValueError – If session configuration is invalid.