LlamaIndex Stream¶

This notebook demonstrates how to monitor Llama-index streaming 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
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 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"] = "..."

import os 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)

chat_engine = index.as_chat_engine()

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

In [ ]:

stream = chat_engine.stream_chat("What did the author do growing up?")

for chunk in stream.response_gen:
    print(chunk, end="")

stream = chat_engine.stream_chat("What did the author do growing up?") for chunk in stream.response_gen: print(chunk, end="")

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_chat_engine_recorder = TruLlama(chat_engine, feedbacks=[f_qa_relevance])

tru_chat_engine_recorder = TruLlama(chat_engine, feedbacks=[f_qa_relevance])

Run Async Application with TruLens¶

In [ ]:

with tru_chat_engine_recorder as recording:
    stream = chat_engine.stream_chat("What did the author do growing up?")

    for chunk in stream.response_gen:
        print(chunk, end="")

record = recording.get()

with tru_chat_engine_recorder as recording: stream = chat_engine.stream_chat("What did the author do growing up?") for chunk in stream.response_gen: print(chunk, end="") record = recording.get()

In [ ]:

# Check recorded input and output:

print(record.main_input)
print(record.main_output)

# Check recorded input and output: print(record.main_input) print(record.main_output)

In [ ]:

# Check costs

record.cost

# Check costs record.cost

In [ ]:

# Check feedback results:

record.feedback_results[0].result()

# Check feedback results: record.feedback_results[0].result()

In [ ]:

from trulens.dashboard import run_dashboard

run_dashboard(session)

from trulens.dashboard import run_dashboard run_dashboard(session)