not-again-ai 0.5.0__tar.gz → 0.6.0__tar.gz

{not_again_ai-0.5.0 → not_again_ai-0.6.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: not-again-ai
-Version: 0.5.0
+Version: 0.6.0
 Summary: Designed to once and for all collect all the little things that come up over and over again in AI projects and put them in one place.
 Home-page: https://github.com/DaveCoDev/not-again-ai
 License: MIT
@@ -21,10 +21,10 @@ Provides-Extra: llm
 Provides-Extra: statistics
 Provides-Extra: viz
 Requires-Dist: numpy (>=1.26.4,<2.0.0) ; extra == "statistics" or extra == "viz"
-Requires-Dist: openai (>=1.16.2,<2.0.0) ; extra == "llm"
-Requires-Dist: pandas (>=2.2.1,<3.0.0) ; extra == "viz"
+Requires-Dist: openai (>=1.23.2,<2.0.0) ; extra == "llm"
+Requires-Dist: pandas (>=2.2.2,<3.0.0) ; extra == "viz"
 Requires-Dist: python-liquid (>=1.12.1,<2.0.0) ; extra == "llm"
-Requires-Dist: scikit-learn (>=1.4.1.post1,<2.0.0) ; extra == "statistics"
+Requires-Dist: scikit-learn (>=1.4.2,<2.0.0) ; extra == "statistics"
 Requires-Dist: scipy (>=1.13.0,<2.0.0) ; extra == "statistics"
 Requires-Dist: seaborn (>=0.13.2,<0.14.0) ; extra == "viz"
 Requires-Dist: tiktoken (>=0.6.0,<0.7.0) ; extra == "llm"
@@ -76,7 +76,7 @@ The package is split into subpackages, so you can install only the parts you nee
 The base package includes only functions that have minimal external dependencies and are useful in a variety of situations such as parallelization and filesystem operations.
 
 ## LLM (Large Language Model)
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/llm.md)
+[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/llm.md), [Example Notebooks](https://github.com/DaveCoDev/not-again-ai/blob/main/notebooks/llm/)
 
 Supports OpenAI chat completions and text embeddings. Includes functions for creating chat completion prompts, token management, and context management.
 

{not_again_ai-0.5.0 → not_again_ai-0.6.0}/README.md

@@ -42,7 +42,7 @@ The package is split into subpackages, so you can install only the parts you nee
 The base package includes only functions that have minimal external dependencies and are useful in a variety of situations such as parallelization and filesystem operations.
 
 ## LLM (Large Language Model)
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/llm.md)
+[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/llm.md), [Example Notebooks](https://github.com/DaveCoDev/not-again-ai/blob/main/notebooks/llm/)
 
 Supports OpenAI chat completions and text embeddings. Includes functions for creating chat completion prompts, token management, and context management.
 

{not_again_ai-0.5.0 → not_again_ai-0.6.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "not-again-ai"
-version = "0.5.0"
+version = "0.6.0"
 description = "Designed to once and for all collect all the little things that come up over and over again in AI projects and put them in one place."
 authors = ["DaveCoDev <dave.co.dev@gmail.com>"]
 license = "MIT"
@@ -28,11 +28,11 @@ python = "^3.11, <3.13"
 
 # Optional dependencies are defined here, and groupings are defined below.
 numpy = { version = "^1.26.4", optional = true }
-openai = { version = "^1.16.2", optional = true }
-pandas = { version = "^2.2.1", optional = true }
+openai = { version = "^1.23.2", optional = true }
+pandas = { version = "^2.2.2", optional = true }
 python-liquid = { version = "^1.12.1", optional = true }
 scipy = { version = "^1.13.0", optional = true }
-scikit-learn = { version = "^1.4.1.post1", optional = true }
+scikit-learn = { version = "^1.4.2", optional = true }
 seaborn = { version = "^0.13.2", optional = true }
 tiktoken = { version = "^0.6.0", optional = true }
 
@@ -41,6 +41,9 @@ llm = ["openai", "python-liquid", "tiktoken"]
 statistics = ["numpy", "scikit-learn", "scipy"]
 viz = ["numpy", "pandas", "seaborn"]
 
+[tool.poetry.dev-dependencies]
+ipykernel = "*"
+
 [tool.poetry.group.nox.dependencies]
 nox-poetry = "*"
 

{not_again_ai-0.5.0 → not_again_ai-0.6.0}/src/not_again_ai/llm/chat_completion.py

@@ -21,6 +21,10 @@ def chat_completion(
 ) -> dict[str, Any]:
     """Get an OpenAI chat completion response: https://platform.openai.com/docs/api-reference/chat/create
 
+    NOTE: Depending on the model, certain parameters may not be supported,
+    particularly for older vision-enabled models like gpt-4-1106-vision-preview.
+    Be sure to check the documentation: https://platform.openai.com/docs/models/gpt-4-turbo-and-gpt-4
+
     Args:
         messages (list): A list of messages comprising the conversation so far.
         model (str): ID of the model to use. See the model endpoint compatibility table:
@@ -64,9 +68,8 @@ def chat_completion(
                 this will be a list of dictionaries containing the token, logprob, and bytes for each token in the message.
             'choices' (list[dict], optional): A list of chat completion choices if n > 1 where each dict contains the above fields.
             'completion_tokens' (int): The number of tokens used by the model to generate the completion.
-                NOTE: If n > 1 this is the sum of all completions and thus will be same value in each dict.
-            'prompt_tokens' (int): The number of tokens in the generated response.
-                NOTE: If n > 1 this is the sum of all completions and thus will be same value in each dict.
+                NOTE: If n > 1 this is the sum of all completions.
+            'prompt_tokens' (int): The number of tokens in the messages sent to the model.
             'system_fingerprint' (str, optional): If seed is set, a unique identifier for the model used to generate the response.
     """
     response_format = {"type": "json_object"} if json_mode else None

{not_again_ai-0.5.0 → not_again_ai-0.6.0}/src/not_again_ai/llm/prompts.py

@@ -7,69 +7,13 @@ from typing import Any
 from liquid import Template
 
 
-def _validate_message(message: dict[str, str]) -> bool:
-    """Valides that a message has valid fields and if the role is valid.
-    See https://platform.openai.com/docs/api-reference/chat/create#chat-create-messages
-    """
-    valid_fields = ["role", "content", "name", "tool_call_id", "tool_calls"]
-    # Check if the only keys in the message are in valid_fields
-    if not all(key in valid_fields for key in message):
-        raise ValueError(f"Message contains invalid fields: {message.keys()}")
-
-    # Check if the only roles in the message are in valid_fields
-    valid_roles = ["system", "user", "assistant", "tool"]
-    if message["role"] not in valid_roles:
-        raise ValueError(f"Message contains invalid role: {message['role']}")
-
-    return True
-
-
-def chat_prompt(messages_unformatted: list[dict[str, str]], variables: dict[str, str]) -> list[dict[str, str]]:
-    """
-    Formats a list of messages for OpenAI's chat completion API using Liquid templating.
-
-    Args:
-        messages_unformatted: A list of dictionaries where each dictionary
-            represents a message. Each message must have 'role' and 'content'
-            keys with string values, where content is a Liquid template.
-        variables: A dictionary where each key-value pair represents a variable
-            name and its value for template rendering.
-
-    Returns:
-        A list of dictionaries with the same structure as `messages_unformatted`,
-        but with the 'content' of each message with the provided `variables`.
-
-    Examples:
-        >>> messages = [
-        ...     {"role": "system", "content": "You are a helpful assistant."},
-        ...     {"role": "user", "content": "Help me {{task}}"}
-        ... ]
-        >>> vars = {"task": "write Python code for the fibonnaci sequence"}
-        >>> chat_prompt(messages, vars)
-        [
-            {"role": "system", "content": "You are a helpful assistant."},
-            {"role": "user", "content": "Help me write Python code for the fibonnaci sequence"}
-        ]
-    """
-
-    messages_formatted = deepcopy(messages_unformatted)
-    for message in messages_formatted:
-        if not _validate_message(message):
-            raise ValueError()
-
-        liquid_template = Template(message["content"])
-        message["content"] = liquid_template.render(**variables)
-
-    return messages_formatted
-
-
 def _validate_message_vision(message: dict[str, list[dict[str, Path | str]] | str]) -> bool:
     """Validates that a message for a vision model is valid"""
-    valid_fields = ["role", "content"]
+    valid_fields = ["role", "content", "name", "tool_call_id", "tool_calls"]
     if not all(key in valid_fields for key in message):
         raise ValueError(f"Message contains invalid fields: {message.keys()}")
 
-    valid_roles = ["system", "user", "assistant"]
+    valid_roles = ["system", "user", "assistant", "tool"]
     if message["role"] not in valid_roles:
         raise ValueError(f"Message contains invalid role: {message['role']}")
 
@@ -126,13 +70,13 @@ def create_image_url(image_path: Path) -> str:
     return f"data:{mime_type};base64,{image_data}"
 
 
-def chat_prompt_vision(messages_unformatted: list[dict[str, Any]], variables: dict[str, str]) -> list[dict[str, Any]]:
-    """Formats a list of messages for OpenAI's chat completion API for vision models only using Liquid templating.
+def chat_prompt(messages_unformatted: list[dict[str, Any]], variables: dict[str, str]) -> list[dict[str, Any]]:
+    """Formats a list of messages for OpenAI's chat completion API,
+    including special syntax for vision models, using Liquid templating.
 
     Args:
         messages_unformatted (list[dict[str, list[dict[str, Path | str]] | str]]):
             A list of dictionaries where each dictionary represents a message.
-            Each message must have 'role' and 'content' keys. `role` must be 'system', 'user', or 'assistant'.
             `content` can be a Liquid template string or a list of dictionaries where each dictionary
             represents a content part. Each content part can be a string or a dictionary with 'image' and 'detail' keys.
             The 'image' key must be a Path or a string representing a URL. The 'detail' key is optional and must be 'low' or 'high'.

{not_again_ai-0.5.0 → not_again_ai-0.6.0}/src/not_again_ai/llm/tokens.py

@@ -80,6 +80,8 @@ def num_tokens_from_messages(messages: list[dict[str, str]], model: str = "gpt-3
         "gpt-4-1106-preview",
         "gpt-4-turbo-preview",
         "gpt-4-0125-preview",
+        "gpt-4-turbo",
+        "gpt-4-turbo-2024-04-09",
     }:
         tokens_per_message = 3  # every message follows <|start|>{role/name}\n{content}<|end|>\n
         tokens_per_name = 1  # if there's a name, the role is omitted

not_again_ai-0.5.0/src/not_again_ai/llm/chat_completion_vision.py

@@ -1,88 +0,0 @@
-from typing import Any
-
-from openai import OpenAI
-
-
-def chat_completion_vision(
-    messages: list[dict[str, Any]],
-    model: str,
-    client: OpenAI,
-    max_tokens: int | None = None,
-    temperature: float = 0.7,
-    seed: int | None = None,
-    n: int = 1,
-    **kwargs: Any,
-) -> dict[str, Any]:
-    """Get an OpenAI chat completion response for vision models only: https://platform.openai.com/docs/guides/vision
-
-    Args:
-        messages (list): A list of messages comprising the conversation so far.
-            See https://platform.openai.com/docs/api-reference/chat/create for details on the format
-        model (str): ID of the model to use for generating chat completions. Refer to OpenAI's documentation
-            for details on available models.
-        client (OpenAI): An instance of the OpenAI client, used to make requests to the API.
-        max_tokens (int | None, optional): The maximum number of tokens to generate in the chat completion.
-            If None, defaults to the model's maximum context length. Defaults to None.
-        temperature (float, optional): Controls the randomness of the output. A higher temperature produces
-            more varied results, whereas a lower temperature results in more deterministic and predictable text.
-            Must be between 0 and 2. Defaults to 0.7.
-        seed (int | None, optional): A seed used for deterministic generation. Providing a seed ensures that
-            the same input will produce the same output across different runs. Defaults to None.
-        n (int, optional): The number of chat completion choices to generate for each input message.
-            Defaults to 1.
-        **kwargs (Any): Additional keyword arguments to pass to the OpenAI client chat completion method.
-
-    Returns:
-        dict[str, Any]: A dictionary containing the generated responses and metadata. Key components include:
-            'finish_reason' (str): The reason the model stopped generating further tokens.
-                Can be 'stop' or 'length'
-            'tool_names' (list[str], optional): The names of the tools called by the model.
-            'tool_args_list' (list[dict], optional): The arguments of the tools called by the model.
-            'message' (str | dict): The content of the generated assistant message.
-            'choices' (list[dict], optional): A list of chat completion choices if n > 1 where each dict contains the above fields.
-            'completion_tokens' (int): The number of tokens used by the model to generate the completion.
-                NOTE: If n > 1 this is the sum of all completions and thus will be same value in each dict.
-            'prompt_tokens' (int): The number of tokens in the generated response.
-                NOTE: If n > 1 this is the sum of all completions and thus will be same value in each dict.
-            'system_fingerprint' (str, optional): If seed is set, a unique identifier for the model used to generate the response.
-    """
-    kwargs.update(
-        {
-            "messages": messages,
-            "model": model,
-            "max_tokens": max_tokens,
-            "temperature": temperature,
-            "n": n,
-        }
-    )
-
-    if seed is not None:
-        kwargs["seed"] = seed
-
-    response = client.chat.completions.create(**kwargs)
-
-    response_data: dict[str, Any] = {"choices": []}
-    for response_choice in response.choices:
-        response_data_curr = {}
-        finish_reason = response_choice.finish_reason
-        response_data_curr["finish_reason"] = finish_reason
-
-        if finish_reason == "stop" or finish_reason == "length":
-            message = response_choice.message.content
-            response_data_curr["message"] = message
-
-        response_data["choices"].append(response_data_curr)
-
-    usage = response.usage
-    if usage is not None:
-        response_data["completion_tokens"] = usage.completion_tokens
-        response_data["prompt_tokens"] = usage.prompt_tokens
-
-    if seed is not None and response.system_fingerprint is not None:
-        response_data["system_fingerprint"] = response.system_fingerprint
-
-    if len(response_data["choices"]) == 1:
-        response_data.update(response_data["choices"][0])
-        del response_data["choices"]
-
-    return response_data