Files
langchain/docs/extras/modules/callbacks/integrations/infino.ipynb
T
os1maandBagatur 2667ddc686 Fix make docs_build and related scripts (#7276)
**Description: a description of the change**

Fixed `make docs_build` and related scripts which caused errors. There
are several changes.

First, I made the build of the documentation and the API Reference into
two separate commands. This is because it takes less time to build. The
commands for documents are `make docs_build`, `make docs_clean`, and
`make docs_linkcheck`. The commands for API Reference are `make
api_docs_build`, `api_docs_clean`, and `api_docs_linkcheck`.

It looked like `docs/.local_build.sh` could be used to build the
documentation, so I used that. Since `.local_build.sh` was also building
API Rerefence internally, I removed that process. `.local_build.sh` also
added some Bash options to stop in error or so. Futher more added `cd
"${SCRIPT_DIR}"` at the beginning so that the script will work no matter
which directory it is executed in.

`docs/api_reference/api_reference.rst` is removed, because which is
generated by `docs/api_reference/create_api_rst.py`, and added it to
.gitignore.

Finally, the description of CONTRIBUTING.md was modified.

**Issue: the issue # it fixes (if applicable)**

https://github.com/hwchase17/langchain/issues/6413

**Dependencies: any dependencies required for this change**

`nbdoc` was missing in group docs so it was added. I installed it with
the `poetry add --group docs nbdoc` command. I am concerned if any
modifications are needed to poetry.lock. I would greatly appreciate it
if you could pay close attention to this file during the review.

**Tag maintainer**
- General / Misc / if you don't know who to tag: @baskaryan

If this PR needs any additional changes, I'll be happy to make them!

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
2023-07-11 22:05:14 -04:00

186 KiB

Infino - LangChain LLM Monitoring Example

This example shows how one can track the following while calling OpenAI models via LangChain and Infino:

  • prompt input,
  • response from chatgpt or any other LangChain model,
  • latency,
  • errors,
  • number of tokens consumed
In [9]:
# Install necessary dependencies.
!pip install infinopy
!pip install matplotlib

# Remove the (1) import sys and sys.path.append(..) and (2) uncomment `!pip install langchain` after merging the PR for Infino/LangChain integration.
import sys

sys.path.append("../../../../../langchain")
#!pip install langchain


import datetime as dt
from infinopy import InfinoClient
import json
from langchain.llms import OpenAI
from langchain.callbacks import InfinoCallbackHandler
import matplotlib.pyplot as plt
import matplotlib.dates as md
import os
import time
import sys
Requirement already satisfied: matplotlib in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (3.7.1)
Requirement already satisfied: contourpy>=1.0.1 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (1.0.7)
Requirement already satisfied: cycler>=0.10 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (0.11.0)
Requirement already satisfied: fonttools>=4.22.0 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (4.39.4)
Requirement already satisfied: kiwisolver>=1.0.1 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (1.4.4)
Requirement already satisfied: numpy>=1.20 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (1.24.3)
Requirement already satisfied: packaging>=20.0 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (23.1)
Requirement already satisfied: pillow>=6.2.0 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (9.5.0)
Requirement already satisfied: pyparsing>=2.3.1 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (3.0.9)
Requirement already satisfied: python-dateutil>=2.7 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from matplotlib) (2.8.2)
Requirement already satisfied: six>=1.5 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from python-dateutil>=2.7->matplotlib) (1.16.0)
Requirement already satisfied: infinopy in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (0.0.1)
Requirement already satisfied: docker in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from infinopy) (6.1.3)
Requirement already satisfied: requests in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from infinopy) (2.31.0)
Requirement already satisfied: packaging>=14.0 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from docker->infinopy) (23.1)
Requirement already satisfied: urllib3>=1.26.0 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from docker->infinopy) (2.0.2)
Requirement already satisfied: websocket-client>=0.32.0 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from docker->infinopy) (1.5.2)
Requirement already satisfied: charset-normalizer<4,>=2 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from requests->infinopy) (3.1.0)
Requirement already satisfied: idna<4,>=2.5 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from requests->infinopy) (3.4)
Requirement already satisfied: certifi>=2017.4.17 in /Users/vinaykakade/.pyenv/versions/3.10.11/lib/python3.10/site-packages (from requests->infinopy) (2023.5.7)

Start Infino server, initialize the Infino client

In [10]:
# Start server using the Infino docker image.
!docker run --rm --detach --name infino-example -p 3000:3000 infinohq/infino:latest

# Create Infino client.
client = InfinoClient()
497a621125800abdd19f57ce7e033349b3cf83ca8cea6a74e8e28433a42ecadd

Read the questions dataset

In [17]:
# These are a subset of questions from Stanford's QA dataset -
# https://rajpurkar.github.io/SQuAD-explorer/
data = """In what country is Normandy located?
When were the Normans in Normandy?
From which countries did the Norse originate?
Who was the Norse leader?
What century did the Normans first gain their separate identity?
Who gave their name to Normandy in the 1000's and 1100's
What is France a region of?
Who did King Charles III swear fealty to?
When did the Frankish identity emerge?
Who was the duke in the battle of Hastings?
Who ruled the duchy of Normandy
What religion were the Normans
What type of major impact did the Norman dynasty have on modern Europe?
Who was famed for their Christian spirit?
Who assimilted the Roman language?
Who ruled the country of Normandy?
What principality did William the conquerer found?
What is the original meaning of the word Norman?
When was the Latin version of the word Norman first recorded?
What name comes from the English words Normans/Normanz?"""

questions = data.split("\n")

LangChain OpenAI Q&A; Publish metrics and logs to Infino

In [18]:
# Set your key here.
# os.environ["OPENAI_API_KEY"] = "YOUR_API_KEY"

# Create callback handler. This logs latency, errors, token usage, prompts as well as prompt responses to Infino.
handler = InfinoCallbackHandler(
    model_id="test_openai", model_version="0.1", verbose=False
)

# Create LLM.
llm = OpenAI(temperature=0.1)

# Number of questions to ask the OpenAI model. We limit to a short number here to save $$ while running this demo.
num_questions = 10

questions = questions[0:num_questions]
for question in questions:
    print(question)

    # We send the question to OpenAI API, with Infino callback.
    llm_result = llm.generate([question], callbacks=[handler])
    print(llm_result)
In what country is Normandy located?
generations=[[Generation(text='\n\nNormandy is located in France.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 16, 'completion_tokens': 9, 'prompt_tokens': 7}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('8de21639-acec-4bd1-a12d-8124de1e20da'))
When were the Normans in Normandy?
generations=[[Generation(text='\n\nThe Normans first settled in Normandy in the late 9th century.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 24, 'completion_tokens': 16, 'prompt_tokens': 8}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('cf81fc86-250b-4e6e-9d92-2df3bebb019a'))
From which countries did the Norse originate?
generations=[[Generation(text='\n\nThe Norse originated from Scandinavia, which includes modern-day Norway, Sweden, and Denmark.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 29, 'completion_tokens': 21, 'prompt_tokens': 8}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('50f42f5e-b4a4-411a-a049-f92cb573a74f'))
Who was the Norse leader?
generations=[[Generation(text='\n\nThe most famous Norse leader was the legendary Viking king Ragnar Lodbrok. He is believed to have lived in the 9th century and is renowned for his exploits in England and France.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 45, 'completion_tokens': 39, 'prompt_tokens': 6}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('e32f31cb-ddc9-4863-8e6e-cb7a281a0ada'))
What century did the Normans first gain their separate identity?
generations=[[Generation(text='\n\nThe Normans first gained their separate identity in the 11th century.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 28, 'completion_tokens': 16, 'prompt_tokens': 12}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('da9d8f73-b3b3-4bc5-8495-da8b11462a51'))
Who gave their name to Normandy in the 1000's and 1100's
generations=[[Generation(text='\n\nThe Normans, a people from northern France, gave their name to Normandy in the 1000s and 1100s. The Normans were descended from Viking settlers who had come to the region in the late 800s.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 58, 'completion_tokens': 45, 'prompt_tokens': 13}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('bb5829bf-b6a6-4429-adfa-414ac5be46e5'))
What is France a region of?
generations=[[Generation(text='\n\nFrance is a region of Europe.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 16, 'completion_tokens': 9, 'prompt_tokens': 7}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('6943880b-b4e4-4c74-9ca1-8c03c10f7e9c'))
Who did King Charles III swear fealty to?
generations=[[Generation(text='\n\nKing Charles III swore fealty to Pope Innocent III.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 23, 'completion_tokens': 13, 'prompt_tokens': 10}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('c91fd663-09e6-4d00-b746-4c7fd96f9ceb'))
When did the Frankish identity emerge?
generations=[[Generation(text='\n\nThe Frankish identity began to emerge in the late 5th century, when the Franks began to expand their power and influence in the region. The Franks were a Germanic tribe that had migrated to the area from the east and had established a kingdom in what is now modern-day France. The Franks were eventually able to establish a powerful kingdom that lasted until the 10th century.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 86, 'completion_tokens': 78, 'prompt_tokens': 8}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('23f86775-e592-4cb8-baa3-46ebe74305b2'))
Who was the duke in the battle of Hastings?
generations=[[Generation(text='\n\nThe Duke of Normandy, William the Conqueror, was the leader of the Norman forces at the Battle of Hastings in 1066.', generation_info={'finish_reason': 'stop', 'logprobs': None})]] llm_output={'token_usage': {'total_tokens': 39, 'completion_tokens': 28, 'prompt_tokens': 11}, 'model_name': 'text-davinci-003'} run=RunInfo(run_id=UUID('ad5b7984-8758-4d95-a5eb-ee56e0218f6b'))

Create Metric Charts

We now use matplotlib to create graphs of latency, errors and tokens consumed.

In [19]:
# Helper function to create a graph using matplotlib.
def plot(data, title):
    data = json.loads(data)

    # Extract x and y values from the data
    timestamps = [item["time"] for item in data]
    dates = [dt.datetime.fromtimestamp(ts) for ts in timestamps]
    y = [item["value"] for item in data]

    plt.rcParams["figure.figsize"] = [6, 4]
    plt.subplots_adjust(bottom=0.2)
    plt.xticks(rotation=25)
    ax = plt.gca()
    xfmt = md.DateFormatter("%Y-%m-%d %H:%M:%S")
    ax.xaxis.set_major_formatter(xfmt)

    # Create the plot
    plt.plot(dates, y)

    # Set labels and title
    plt.xlabel("Time")
    plt.ylabel("Value")
    plt.title(title)

    plt.show()


response = client.search_ts("__name__", "latency", 0, int(time.time()))
plot(response.text, "Latency")

response = client.search_ts("__name__", "error", 0, int(time.time()))
plot(response.text, "Errors")

response = client.search_ts("__name__", "prompt_tokens", 0, int(time.time()))
plot(response.text, "Prompt Tokens")

response = client.search_ts("__name__", "completion_tokens", 0, int(time.time()))
plot(response.text, "Completion Tokens")

response = client.search_ts("__name__", "total_tokens", 0, int(time.time()))
plot(response.text, "Total Tokens")

Full text query on prompt or prompt outputs.

In [20]:
# Search for a particular prompt text.
query = "normandy"
response = client.search_log(query, 0, int(time.time()))
print("Results for", query, ":", response.text)

print("===")

query = "king charles III"
response = client.search_log("king charles III", 0, int(time.time()))
print("Results for", query, ":", response.text)
Results for normandy : [{"time":1686821979,"fields":{"prompt":"In what country is Normandy located?"},"text":"In what country is Normandy located?"},{"time":1686821982,"fields":{"prompt_response":"\n\nNormandy is located in France."},"text":"\n\nNormandy is located in France."},{"time":1686821984,"fields":{"prompt_response":"\n\nThe Normans first settled in Normandy in the late 9th century."},"text":"\n\nThe Normans first settled in Normandy in the late 9th century."},{"time":1686821993,"fields":{"prompt":"Who gave their name to Normandy in the 1000's and 1100's"},"text":"Who gave their name to Normandy in the 1000's and 1100's"},{"time":1686821997,"fields":{"prompt_response":"\n\nThe Normans, a people from northern France, gave their name to Normandy in the 1000s and 1100s. The Normans were descended from Viking settlers who had come to the region in the late 800s."},"text":"\n\nThe Normans, a people from northern France, gave their name to Normandy in the 1000s and 1100s. The Normans were descended from Viking settlers who had come to the region in the late 800s."}]
===
Results for king charles III : [{"time":1686821998,"fields":{"prompt":"Who did King Charles III swear fealty to?"},"text":"Who did King Charles III swear fealty to?"},{"time":1686822000,"fields":{"prompt_response":"\n\nKing Charles III swore fealty to Pope Innocent III."},"text":"\n\nKing Charles III swore fealty to Pope Innocent III."}]

Step 5: Stop infino server

In [21]:
!docker rm -f infino-example
infino-example
In [ ]: