LlamaIndex Async¶

Async is a growing method for LLM applications and can be especially useful for reducing indexing time. This notebook demonstrates how to monitor Llama-index async apps with TruLens.

Import from LlamaIndex and TruLens¶

In [ ]:

# !pip install trulens trulens-apps-llamaindex trulens-providers-openai llama_index==0.10.11 llama-index-readers-web openai

# !pip install trulens trulens-apps-llamaindex trulens-providers-openai llama_index==0.10.11 llama-index-readers-web openai

In [ ]:

from llama_index.core import VectorStoreIndex
from llama_index.readers.web import SimpleWebPageReader
import openai
from trulens.core import Feedback
from trulens.core import TruSession
from trulens.apps.llamaindex import TruLlama
from trulens.providers.openai import OpenAI

session = TruSession()

from llama_index.core import VectorStoreIndex from llama_index.readers.web import SimpleWebPageReader import openai from trulens.core import Feedback from trulens.core import TruSession from trulens.apps.llamaindex import TruLlama from trulens.providers.openai import OpenAI session = TruSession()

Add API keys¶

For this example you need an OpenAI key

In [ ]:

import os

os.environ["OPENAI_API_KEY"] = "..."
openai.api_key = os.environ["OPENAI_API_KEY"]

import os os.environ["OPENAI_API_KEY"] = "..." openai.api_key = os.environ["OPENAI_API_KEY"]

Create Async App¶

In [ ]:

documents = SimpleWebPageReader(html_to_text=True).load_data(
    ["http://paulgraham.com/worked.html"]
)
index = VectorStoreIndex.from_documents(documents)

query_engine = index.as_query_engine()

documents = SimpleWebPageReader(html_to_text=True).load_data( ["http://paulgraham.com/worked.html"] ) index = VectorStoreIndex.from_documents(documents) query_engine = index.as_query_engine()

In [ ]:

response = query_engine.aquery("What did the author do growing up?")

print(response)  # should be awaitable
print(await response)

response = query_engine.aquery("What did the author do growing up?") print(response) # should be awaitable print(await response)

Set up Evaluation¶

In [ ]:

# Initialize OpenAI-based feedback function collection class:
openai = OpenAI()

# Question/answer relevance between overall question and answer.
f_qa_relevance = Feedback(
    openai.relevance, name="QA Relevance"
).on_input_output()

# Initialize OpenAI-based feedback function collection class: openai = OpenAI() # Question/answer relevance between overall question and answer. f_qa_relevance = Feedback( openai.relevance, name="QA Relevance" ).on_input_output()

Create tracked app¶

In [ ]:

tru_query_engine_recorder = TruLlama(query_engine, feedbacks=[f_qa_relevance])

tru_query_engine_recorder = TruLlama(query_engine, feedbacks=[f_qa_relevance])

Run Async Application with TruLens¶

In [ ]:

with tru_query_engine_recorder as recording:
    response = await query_engine.aquery("What did the author do growing up?")

print(response)

with tru_query_engine_recorder as recording: response = await query_engine.aquery("What did the author do growing up?") print(response)

In [ ]:

from trulens.dashboard import run_dashboard

run_dashboard(session)

from trulens.dashboard import run_dashboard run_dashboard(session)