backend/retrieval_augmented_generation.py

from time import perf_counter
from jinja2 import Template

from backend.query_llm import generate
from backend.semantic_search import qd_retriever

template_string = """
Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer.
Context:
---
{% for doc in documents %}
    {{ doc.content }}
    ---
{% endfor %}
Query: {{ query }}
"""

template = Template(template_string)


def rag(query, top_k=5):

    # Retrieve documents relevant to query
    document_start = perf_counter()
    documents = qd_retriever.retrieve(query, top_k=top_k)
    document_time = document_start - perf_counter()

    # Create Prompt
    prompt = template.render(documents=documents, query=query)

    # Query LLM with prompt based on relevant documents
    llm_start = perf_counter()
    result = generate(prompt=prompt, history='')
    llm_time = llm_start - perf_counter()

    times = (document_time, llm_time)
    return prompt, result