Ragas Wrapper
This file is used to wrap the LLM and embeddings for Ragas.
RagasEmbeddingsWrapper(em_invoker)
Bases: BaseRagasEmbeddings
Ragas embeddings wrapper.
This wrapper is used to wrap the embeddings for Ragas.
Attributes:
| Name | Type | Description |
|---|---|---|
em_invoker |
BaseEMInvoker
|
The embeddings invoker. |
Initialize the Ragas embeddings wrapper.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
em_invoker
|
BaseEMInvoker
|
The embeddings invoker. |
required |
aembed_documents(texts)
async
Embed documents asynchronously.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
texts
|
list[str]
|
The texts to embed. |
required |
Returns:
| Type | Description |
|---|---|
list[list[float]]
|
list[list[float]]: The embeddings of the documents. |
aembed_query(query)
async
Embed a query asynchronously.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
query
|
str
|
The query to embed. |
required |
Returns:
| Type | Description |
|---|---|
list[float]
|
list[float]: The embedding of the query. |
embed_documents(texts)
Embed documents synchronously.
This method is not implemented because our embeddings invoker does not support synchronous embedding.
Raises:
| Type | Description |
|---|---|
NotImplementedError
|
Always. |
embed_query(query)
Embed a query synchronously.
This method is not implemented because our embeddings invoker does not support synchronous embedding.
Raises:
| Type | Description |
|---|---|
NotImplementedError
|
Always. |
RagasLLMWrapper(lm_invoker, fallback_lms=None)
Bases: UsageRecordingMixin, BaseRagasLLM, LMComponent
Ragas LLM wrapper backed by LMComponent.
Provides the BaseRagasLLM interface with automatic fallback chain support. When the primary invoker raises BaseInvokerError or TimeoutError during agenerate_text, the next invoker in fallback_lms is tried in order.
Attributes:
| Name | Type | Description |
|---|---|---|
lm_invoker |
BaseLMInvoker
|
Primary LM invoker. |
fallback_lms |
list[BaseLMInvoker]
|
Ordered fallback invokers. |
Initialize the Ragas LLM wrapper.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
lm_invoker
|
BaseLMInvoker
|
The LLM invoker. |
required |
fallback_lms
|
list[BaseLMInvoker] | None
|
Ordered fallback invokers. Defaults to None. |
None
|
winning_invoker
property
The invoker that handled the most recent successful request.
agenerate_text(prompt, n=1, temperature=None, stop=None, callbacks=None)
async
Generate text using the LLM with automatic fallback chain.
Converts the Ragas PromptValue to gllm Messages and iterates through the fallback chain until a successful response is returned.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
prompt
|
PromptValue
|
The prompt to generate text from. |
required |
n
|
int
|
The number of responses to generate. Default is 1. |
1
|
temperature
|
float | None
|
The temperature to use for the generation. If an invoker rejects it (e.g. reasoning models), the call is retried without it and the invoker is remembered. Default is None. |
None
|
stop
|
list[str] | None
|
The stop tokens to use for the generation. Default is None. |
None
|
callbacks
|
Callbacks | None
|
The callbacks to use for the generation. Default is None. |
None
|
Returns:
| Name | Type | Description |
|---|---|---|
LLMResult |
LLMResult
|
The result of the generation. |
Raises:
| Type | Description |
|---|---|
InvokerRuntimeError
|
If n != 1 (only single-generation is supported). |
generate_text(prompt, n=1, temperature=None, stop=None, callbacks=None)
Generate text using the LLM.
This method is not implemented because our LM invoker does not support synchronous generation.
Raises:
| Type | Description |
|---|---|
NotImplementedError
|
Always. |
is_finished(response)
Check if the response is finished.
Since we didn't do any modification to the response, we assume that the response is finished.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
response
|
LLMResult
|
The response to check. |
required |