-
|
I am using base_query_engine like below instead of Can you hlep? I like to add self-correcting logic like the link suggested below, which is my final goal. |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 9 replies
-
|
To integrate self-correcting logic into your
This setup ensures that your |
Beta Was this translation helpful? Give feedback.
-
|
@dosu I am getting this error |
Beta Was this translation helpful? Give feedback.
-
|
To resolve the Here is an example of how you might implement the class RAGStringQueryEngine(BaseQueryEngine):
def __init__(self, retriever, response_synthesizer, llm, qa_prompt):
self.retriever = retriever
self.response_synthesizer = response_synthesizer
self.llm = llm
self.qa_prompt = qa_prompt
def _query(self, query_bundle: QueryBundle) -> RESPONSE_TYPE:
# Implement the logic to handle the query and return a response
nodes = self.retriever.retrieve(query_bundle.query_str)
context_str = "\n\n".join([n.node.get_content() for n in nodes])
response = self.llm.complete(
self.qa_prompt.format(context_str=context_str, query_str=query_bundle.query_str)
)
return str(response)
async def _aquery(self, query_bundle: QueryBundle) -> RESPONSE_TYPE:
"""Not supported."""
return self._query(query_bundle)Make sure that the |
Beta Was this translation helpful? Give feedback.
-
|
@dosu thank you. Seems like I am getting close. I modified as per your suggestions. However, I am getting attribute error. seems like I need to change my class object. Please help. My customized class: My error: |
Beta Was this translation helpful? Give feedback.
-
|
To resolve the Here's how you can modify your
By returning a |
Beta Was this translation helpful? Give feedback.
-
|
@dosu Fixed one but getting a new error :-( |
Beta Was this translation helpful? Give feedback.
-
|
The To resolve this, ensure that all required arguments are provided when initializing any class that might be raising a
By ensuring that all required fields are correctly initialized and validated, and by applying |
Beta Was this translation helpful? Give feedback.
-
|
@dosu no luck... |
Beta Was this translation helpful? Give feedback.
-
|
To resolve the Here is the corrected initialization of from llama_index.legacy.evaluation.guideline import GuidelineEvaluator
from llama_index.legacy.indices.query.query_transform.feedback_transform import FeedbackQueryTransformation
from llama_index.legacy.core.base_query_engine import BaseQueryEngine
# Assuming RAGStringQueryEngine is a subclass of BaseQueryEngine
class RAGStringQueryEngine(BaseQueryEngine):
# Implementation of RAGStringQueryEngine
pass
# Initialize your query engine and guideline evaluator
query_engine = RAGStringQueryEngine()
guideline_evaluator = GuidelineEvaluator(model='your_model') # Ensure 'model' argument is provided
# Initialize RetryGuidelineQueryEngine
retry_guideline_query_engine = RetryGuidelineQueryEngine(
query_engine=query_engine,
guideline_evaluator=guideline_evaluator,
resynthesize_query=False,
max_retries=3,
callback_manager=None,
query_transformer=FeedbackQueryTransformation(resynthesize_query=False)
)Ensure that the |
Beta Was this translation helpful? Give feedback.
-
|
@dosu Error again: :-) Where I add your suggestion below??? Look like GuidelineEvaluator won't take model |
Beta Was this translation helpful? Give feedback.
-
|
The Here is the corrected setup:
This setup ensures that your |
Beta Was this translation helpful? Give feedback.
To resolve the
TypeError: ValidationError.__init__() missing 1 required positional argument: 'model'error and ensure that theRetryGuidelineQueryEngineworks correctly with yourRAGStringQueryEngine, you need to ensure that all required arguments are provided when initializing theRetryGuidelineQueryEngine.Here is the corrected initialization of
RetryGuidelineQueryEngine: