[Question]: how to get RetrieverQueryEngine full prompt

[whoo9112]

Question Validation

• I have searched both the documentation and discord for an answer.

Question

When using RetrieverQueryEngine, can you check how the Retriever results I extracted are applied to the prompt and delivered to LLM?

For example
"""
you are helpful assistant
{context_str}

{query_str}
"""

I'd like to check the full prompt with {context_str}, {query_str} completely filled.

[dosubot]

To view the full prompt with {context_str} and {query_str} completely filled when using RetrieverQueryEngine, you can use the get_prompts method to access the set of prompts used by the query engine, including those in sub-modules. This allows you to see how the context_str and query_str are applied in the prompt templates. Here's how you can do it:

1. Access Prompts: Use the get_prompts method on your query engine to retrieve all the prompts, including those used in nested modules.

   prompts_dict = query_engine.get_prompts()

2. Display Prompts: Define a function to display the prompts in a readable format.

   def display_prompt_dict(prompts_dict):
       for k, p in prompts_dict.items():
           text_md = f"**Prompt Key**: {k}<br>" f"**Text:** <br>"
           display(Markdown(text_md))
           print(p.get_template())
           display(Markdown("<br><br>"))

3. View Filled Prompts: Call the display function to see the prompts with {context_str} and {query_str} filled in.

   display_prompt_dict(prompts_dict)

This approach allows you to centralize and view the prompts used by the RetrieverQueryEngine, providing insight into how the retriever results are formatted and delivered to the LLM ^[1][2].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}

[whoo9112]

To view the full prompt with {context_str} and {query_str} completely filled when using RetrieverQueryEngine, you can use the get_prompts method to access the set of prompts used by the query engine, including those in sub-modules. This allows you to see how the context_str and query_str are applied in the prompt templates. Here's how you can do it:

1. Access Prompts: Use the get_prompts method on your query engine to retrieve all the prompts, including those used in nested modules.
   prompts_dict = query_engine.get_prompts()
2. Display Prompts: Define a function to display the prompts in a readable format.
   def display_prompt_dict(prompts_dict):
   for k, p in prompts_dict.items():
   text_md = f"Prompt Key: {k}
   " f"Text:
   "
   display(Markdown(text_md))
   print(p.get_template())
   display(Markdown("
   
   "))
3. View Filled Prompts: Call the display function to see the prompts with {context_str} and {query_str} filled in.
   display_prompt_dict(prompts_dict)

This approach allows you to centralize and view the prompts used by the RetrieverQueryEngine, providing insight into how the retriever results are formatted and delivered to the LLM [1][2].

To continue talking to Dosu, mention @dosu.

Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other

Dosubot always gives me different answers than I want.. it's awful