Struct ConversationRetrievalChain 

pub struct ConversationRetrievalChain<M: BaseChatModel> { /* private fields */ }

ConversationRetrievalChain

Retrieval-augmented conversation chain with memory that automatically:

1. Loads conversation history
2. Retrieves relevant documents
3. Combines history + context + question
4. LLM generates answer
5. Saves to conversation memory

Implementations

impl<M: BaseChatModel + 'static> ConversationRetrievalChain<M>

pub fn new( llm: M, retriever: Arc<dyn RetrieverTrait>, memory: ConversationBufferMemory, ) -> Self

pub fn from_memory( llm: M, retriever: Arc<dyn RetrieverTrait>, memory: Arc<Mutex<dyn BaseMemory>>, ) -> Self

Create from any BaseMemory implementation (window / summary / vector-store / persistent), mirroring ConversationChain::from_memory.

pub fn with_system_prompt(self, prompt: impl Into<String>) -> Self

pub fn with_input_key(self, key: impl Into<String>) -> Self

pub fn with_output_key(self, key: impl Into<String>) -> Self

pub fn with_name(self, name: impl Into<String>) -> Self

pub fn with_k(self, k: usize) -> Self

pub fn with_verbose(self, verbose: bool) -> Self

pub fn with_return_source_documents(self, return_source: bool) -> Self

pub fn memory(&self) -> &Arc<Mutex<dyn BaseMemory>>

pub async fn clear_memory(&self) -> Result<(), ChainError>

pub async fn query( &self, question: impl Into<String>, ) -> Result<String, ChainError>

Simplified query interface.

pub fn build_messages( &self, history: &[Message], context: &str, question: &str, ) -> Vec<Message>

Build structured messages for a given history, context, and question.

Trait Implementations

impl<M: BaseChatModel + Send + Sync + 'static> BaseChain for ConversationRetrievalChain<M>

where <M as Runnable<Vec<Message>, LLMResult>>::Error: Display,

fn stream<'life0, 'async_trait>( &'life0 self, inputs: HashMap<String, Value>, ) -> Pin<Box<dyn Future<Output = Result<ChainStream, ChainError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Stream execution for ConversationRetrievalChain – token by token output.

P2-2: real streaming — loads history, retrieves and assembles the prompt, then pushes LLM tokens via stream_chat. The full answer is accumulated through an unbounded channel (the P1-4 pattern) and written to memory once the stream completes, matching the invoke path’s save_context.

fn input_keys(&self) -> Vec<&str>

Get input keys.

fn output_keys(&self) -> Vec<&str>

Get output keys.

fn invoke<'life0, 'async_trait>( &'life0 self, inputs: HashMap<String, Value>, ) -> Pin<Box<dyn Future<Output = Result<ChainResult, ChainError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Execute the Chain.

fn name(&self) -> &str

Get Chain name.

fn invoke_with_config<'life0, 'async_trait>( &'life0 self, inputs: HashMap<String, Value>, config: Option<RunnableConfig>, ) -> Pin<Box<dyn Future<Output = Result<ChainResult, ChainError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Execute the Chain with a RunnableConfig.

fn stream_with_config<'life0, 'async_trait>( &'life0 self, inputs: HashMap<String, Value>, config: Option<RunnableConfig>, ) -> Pin<Box<dyn Future<Output = Result<ChainStream, ChainError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Stream execute the Chain with a RunnableConfig.

fn validate_inputs( &self, inputs: &HashMap<String, Value>, ) -> Result<(), ChainError>

Validate inputs.