im-openai 0.15.2

Wrapper library for openai to send events to the Imaginary Programming monitor

DEPRECATED: Imaginary Dev OpenAI wrapper

This package has been deprecated, please migrate to https://pypi.org/project/libretto-openai.

Wrapper library for openai to send events to the Imaginary Programming monitor

Features

• Patches the openai library to allow user to set an ip_api_key and ip_prompt_template_name for each request
• Works out of the box with langchain

Get Started

To send events to Imaginary Programming, you'll need to create a project. From the project you'll need two things:

1. API key: (api_key) This is generated for the project and is used to identify the project and environment (dev, staging, prod) that the event is coming from.
2. Template Name: (prompt_template_name) This uniquely identifies a particular prompt that you are using and allows projects to have multiple prompts. This can be in any format but we recommend using a dash-separated format, e.g. my-prompt-name.

Note: Prompt template names can be auto-generated if the allow_unnamed_prompts configuration option is set (see below). However, if you rely on auto-generated names, new revisions of the same prompt will show up as different prompt templates in Templatest.

OpenAI

You can use the patched_openai context manager to patch your code that uses the existing OpenAI client library:

To allow our tools to separate the "prompt" from the "prompt parameters", use TemplateChat and TemplateText to create templates.

Use TemplateChat For the ChatCompletion APIs:

from im_openai import patched_openai, TemplateChat

with patched_openai(api_key="...", prompt_template_name="sport-emoji"):
    import openai

    completion = openai.ChatCompletion.create(
        # Standard OpenAI parameters
        model="gpt-3.5-turbo",
        messages=TemplateChat(
            [{"role": "user", "content": "Show me an emoji that matches the sport: {sport}"}],
            {"sport": "soccer"},
        ),
    )

Use TemplateText for the Completion API:

from im_openai import patched_openai, TemplateText

with patched_openai(api_key="...", prompt_template_name="sport-emoji"):
    import openai

    completion = openai.Completion.create(
        # Standard OpenAI parameters
        model="text-davinci-003",
        prompt=TemplateText("Show me an emoji that matches the sport: {sport}", {"sport": "soccer"}),
    )

Advanced usage

Patching at startup

Rather than using a context manager, you can patch the library once at startup:

from im_openai import patch_openai
patch_openai(api_key="...", prompt_template_name="...")

Then, you can use the patched library as normal:

import openai

completion = openai.ChatCompletion.create(
    # Standard OpenAI parameters
    ...)

Manually passing parameters

While the use of TemplateText and TemplateChat are preferred, Most of the parameters passed during patch can also be passed directly to the create(), with an ip_ prefix.

from im_openai import patch_openai
patch_openai()

completion = openai.ChatCompletion.create(
    model="gpt-3.5-turbo",

    # Note we are passing the raw chat object here
    messages=[{"role": "user", "content": "Show me an emoji that matches the sport: soccer"}],

    # call configuration
    ip_api_key="...",
    ip_prompt_template_name="sport-emoji",

    # Here the prompt and parameters is passed seperately
    ip_template_params={"sport": "soccer"},
    ip_template_chat=[
        {"role": "user", "content": "Show me an emoji that matches the sport: {sport}"}
    ],
)

Langchain

For langchain, you can directly patch, or use a context manager before setting up a chain:

Using a context manager: (recommended)

from langchain import LLMChain, PromptTemplate, OpenAI
from im_openai.langchain import prompt_watch_tracing

with prompt_watch_tracing(api_key="4b2a6608-86cd-4819-aba6-479f9edd8bfb", prompt_template_name="sport-emoji"):
    chain = LLMChain(llm=OpenAI(),
        prompt=PromptTemplate.from_template("What is the capital of {country}?"))
    capital = chain.run({"country": "Sweden"})

The api_key parameter is visible from your project's settings page.

the prompt_template_name parameter can also be passed directly to a template when you create it, so that it can be tracked separately from other templates:

from langchain import OpenAI, PromptTemplate, LLMChain
from im_openai.langchain import prompt_watch_tracing

# The default prompt_template_name is "default-questions"
with prompt_watch_tracing(api_key="4b2a6608-86cd-4819-aba6-479f9edd8bfb", prompt_template_name="default-questions"):
    prompt = PromptTemplate.from_template("""
Please answer the following question: {question}.
""")
    llm = LLMChain(prompt=prompt, llm=OpenAI())
    llm.run(question="What is the meaning of life?")

    # Track user greetings separately under the `user-greeting` api name
    greeting_prompt = PromptTemplate.from_template("""
Please greet our newest forum member, {user}.
Be nice and enthusiastic but not overwhelming.
""",
        additional_kwargs={"ip_prompt_template_name": "user-greeting"})
    llm = LLMChain(prompt=greeting_prompt, llm=OpenAI(openai_api_key=...))
    llm.run(user="Bob")

Advanced usage

You can patch directly:

from im_openai.langchain import enable_prompt_watch_tracing, disable_prompt_watch_tracing

old_tracer = enable_prompt_watch_tracing(api_key="4b2a6608-86cd-4819-aba6-479f9edd8bfb", prompt_template_name="sport-emoji")

prompt = PromptTemplate.from_template("""
Please answer the following question: {question}.
""")
llm = LLMChain(prompt=prompt, llm=OpenAI())
llm.run(question="What is the meaning of life?")

# Track user greetings separately under the `user-greeting` api name
greeting_prompt = PromptTemplate.from_template("""
Please greet our newest forum member, {user}. Be nice and enthusiastic but not overwhelming.
""",
    additional_kwargs={"ip_prompt_template_name": "user-greeting"})
llm = LLMChain(prompt=greeting_prompt, llm=OpenAI(openai_api_key=...))
llm.run(user="Bob")

# optional, if you need to disable tracing later
disable_prompt_watch_tracing(old_tracer)

Configuration

The following options may be passed as kwargs when patching:

• prompt_template_name: A default name to associate with prompts. If provided, this is the name that will be associated with any create call that's made without an ip_prompt_template_name parameter.
• allow_unnamed_prompts: When set to True, every prompt will be sent to Templatest even if no prompt template name as been provided (either via the prompt_template_name kwarg or via the ip_prompt_template_name parameter on create). False by default.
• redact_pii: When True, certain personally identifying information (PII) will be attempted to be redacted before being sent to the Templatest backend. See the pii package for details about the types of PII being detected/redacted. False by default.

Additional Parameters

The following parameters are available in both the patched OpenAI client and the Langchain wrapper.

• For OpenAI, pass these to the create() methods.
• For Langchain, pass these to the prompt_watch_tracing() context manager or the enable_prompt_watch_tracing() function.

Parameters:

• chat_id / ip_chat_id: The id of a "chat session" - if the chat API is being used in a conversational context, then the same chat id can be provided so that the events are grouped together, in order. If not provided, this will be left blank.

OpenAI-only parameters:

These parameters can only be passed to the create() methods of the patched OpenAI client.

• ip_template_chat: The chat template to record for chat requests. This is a list of dictionaries with the following keys:

  • role: The role of the speaker. Either "system", "user" or "ai".
  • content: The content of the message. This can be a string or a template string with {} placeholders.

  For example:

  completion = openai.ChatCompletion.create(
      ...,
      ip_template_chat=[
          {"role": "ai", "content": "Hello, I'm {system_name}!"},
          {"role": "user", "content": "Hi {system_name}, I'm {user_name}!"}
      ])
  

  To represent an array of chat messages, use the artificial role "chat_history" with content set to the variable name in substitution format: [{"role": "chat_history", "content": "{prev_messages}"}}]

• ip_template_text: The text template to record for completion requests. This is a string or a template string with {} placeholders.

  For example:

  completion = openai.Completion.create(
      ...,
      ip_template_text="Please welcome the user to {system_name}!")
  

• ip_template_params: The parameters to use for template strings. This is a dictionary of key-value pairs.

  For example:

  completion = openai.Completion.create(
      ...,
      ip_template_text="Please welcome the user to {system_name}!"),
      ip_template_params={"system_name": "Awesome Comics Incorporated"})
  

• ip_event_id: A unique UUID for a specific call. If not provided, one will be generated. Note: In the langchain wrapper, this value is inferred from the chain run_id.

  For example:

  import uuid
  
  completion = openai.Completion.create(
      ...,
      ip_event_id=uuid.uuid4())
  

• ip_parent_event_id: The UUID of the parent event. All calls with the same parent id are grouped as a "Run Group". Note: In the langchain wrapper, this value is inferred from the chain parent_run_id.

  For example:

  import uuid
  
  parent_id = uuid.uuid4()
  # First call in the run group
  completion = openai.Completion.create(
      ...,
      ip_parent_event_id=parent_id)
  
  # Another call in the same group
  completion = openai.Completion.create(
      ...,
      ip_parent_event_id=parent_id)
  

Sending Feedback

Sometimes the answer provided by the LLM is not ideal, and your users may be able to help you find better responses. There are a few common cases:

• You might use the LLM to suggest the title of a news article, but let the user edit it. If they change the title, you can send feedback to Templatest that the answer was not ideal.
• You might provide a chatbot that answers questions, and the user can rate the answers with a thumbs up (good) or thumbs down (bad).

You can send this feedback to Tepmlatest by calling send_feedback(). This will send a feedback event to Templatest about a prompt that was previously called, and let you review this feedback in the Templatest dashboard. You can use this feedback to develop new tests and improve your prompts.

from im_openai import patch_openai, client
patch_openai()

completion = openai.ChatCompletion.create(
    ...)


# Maybe the user didn't like the answer, so ask them for a better one.
better_response = askUserForBetterResult(completion["choices"][0]["text"])

# If the user provided a better answer, send feedback to Templatest
if better_response !== completion["choices"][0]["text"]:
# feedback key is automatically injected into OpenAI response object.
feedback_key = completion.ip_feedback_key
client.send_feedback(
    api_key=api_key,
    feedback_key=feedback_key,
    # Better answer from the user
    better_response=better_response,
    # Rating of existing answer, from 0 to 1
    rating=0.2)

Note that feedback can include either rating, better_response, or both.

Parameters:

• rating - a value from 0 (meaning the result was completely wrong) to 1 (meaning the result was correct)
• better_response - the better response from the user

Credits

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.

• Cookiecutter: https://github.com/audreyr/cookiecutter
• audreyr/cookiecutter-pypackage: https://github.com/audreyr/cookiecutter-pypackage