not-again-ai 0.5.1__tar.gz → 0.7.0__tar.gz

{not_again_ai-0.5.1 → not_again_ai-0.7.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: not-again-ai
-Version: 0.5.1
+Version: 0.7.0
 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 +21,11 @@ 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: ollama (>=0.1.9,<0.2.0) ; extra == "llm"
+Requires-Dist: openai (>=1.25.1,<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"
@@ -47,9 +48,9 @@ Description-Content-Type: text/markdown
 [ruff-badge]: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json
 [mypy-badge]: https://www.mypy-lang.org/static/mypy_badge.svg
 
-**not-again-ai** is a collection of various functionalities that come up over and over again when developing AI projects as a Data/Applied/Research Scientist. It is designed to be simple and minimize dependencies first and foremost. This is not meant to reinvent the wheel, but instead be a home for functions that don’t belong well elsewhere. Additionally, feel free to **a)** use this as a template for your own Python package. **b)** instead of installing the package, copy and paste functions into your own projects (this is made easier with the limited amount of dependencies and the MIT license).
+**not-again-ai** is a collection of various building blocks that come up over and over again when developing AI products. The key goals of this package are to have simple, but flexible interfaces and to minimize dependencies. Feel free to **a)** use this as a template for your own Python package. **b)** instead of installing the package, copy and paste functions into your own projects (this is made possible with the limited amount of dependencies and the MIT license).
 
-**Documentation** available within the [readmes](readmes) or auto-generated at [DaveCoDev.github.io/not-again-ai/](https://DaveCoDev.github.io/not-again-ai/).
+**Documentation** available within individual **[notebooks](notebooks)**, docstrings within the source, or auto-generated at [DaveCoDev.github.io/not-again-ai/](https://DaveCoDev.github.io/not-again-ai/).
 
 # Installation
 
@@ -61,82 +62,26 @@ Install the entire package from [PyPI](https://pypi.org/project/not-again-ai/) w
 $ pip install not_again_ai[llm,statistics,viz]
 ```
 
-The package is split into subpackages, so you can install only the parts you need.
+The package is split into subpackages, so you can install only the parts you need. See the **[notebooks](notebooks)** for examples.
 * **Base only**: `pip install not_again_ai`
-* **LLM only**: `pip install not_again_ai[llm]`
+* **LLM**: `pip install not_again_ai[llm]`
+    1. If you wish to use OpenAI
+        1. Go to https://platform.openai.com/settings/profile?tab=api-keys to get your API key.
+        1. (Optionally) Set the `OPENAI_API_KEY` and the `OPENAI_ORG_ID` environment variables.
+    1. If you wish to use Ollama:
+        1. follow the instructions to install ollama for your system: https://github.com/ollama/ollama
+        1. [Add Ollama as a startup service (recommended)](https://github.com/ollama/ollama/blob/main/docs/linux.md#adding-ollama-as-a-startup-service-recommended)
+        1. If you'd like to make the ollama service accessible on your local network and it is hosted on Linux, add the following to the `/etc/systemd/system/ollama.service` file:
+            ```bash
+            [Service]
+            ...
+            Environment="OLLAMA_HOST=0.0.0.0"
+            ```
+        Now ollama will be available at `http://<local_address>:11434`
 * **Statistics**: `pip install not_again_ai[statistics]`
 * **Visualization**: `pip install not_again_ai[viz]`
 
 
-# Quick Tour
-
-## Base
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/base.md)
-
-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), [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.
-
-One example: 
-```python
-client = openai_client()
-messages = [{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Hello!"}]
-response = chat_completion(messages=messages, model="gpt-3.5-turbo", max_tokens=100, client=client)["message"]
->>> "Hello! How can I help you today?"
-```
-
-## Statistics
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/statistics.md)
-
-We provide a few helpers for data analysis such as:
-
-```python
-from not_again_ai.statistics.dependence import pearson_correlation
-# quadratic dependence
->>> x = (rs.rand(500) * 4) - 2
->>> y = x**2 + (rs.randn(500) * 0.2)
->>> pearson_correlation(x, y)
-0.05
-```
-
-## Visualization
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/viz.md)
-
-We offer opinionated wrappers around seaborn to make common visualizations easier to create and customize.
-
-```python
->>> import numpy as np
->>> import pandas as pd
->>> from not_again_ai.viz.time_series import ts_lineplot
->>> from not_again_ai.viz.distributions import univariate_distplot
-
-# get some time series data
->>> rs = np.random.RandomState(365)
->>> values = rs.randn(365, 4).cumsum(axis=0)
->>> dates = pd.date_range('1 1 2021', periods=365, freq='D')
-# plot the time series and save it to a file
->>> ts_lineplot(ts_data=values, save_pathname='myplot.png', ts_x=dates, ts_names=['A', 'B', 'C', 'D'])
-
-# get a random distribution
->>> distrib = np.random.beta(a=0.5, b=0.5, size=1000)
-# plot the distribution and save it to a file
->>> univariate_distplot(
-...     data=distrib, 
-...     save_pathname='mydistribution.svg', 
-...     print_summary=False, bins=100, 
-...     title=r'Beta Distribution $\alpha=0.5, \beta=0.5$'
-... )
-```
-
-<p float="center">
-  <img src="https://raw.githubusercontent.com/DaveCoDev/not-again-ai/44c53fb7fb07234aaceea40c90d8cb74e5fa6c15/assets/distributions_test4.svg" width="404" />
-  <img src="https://raw.githubusercontent.com/DaveCoDev/not-again-ai/44c53fb7fb07234aaceea40c90d8cb74e5fa6c15/assets/ts_lineplot5.svg" width="404" /> 
-</p>
-
-
 # Development Information
 
 The following information is relevant if you would like to contribute or use this package as a template for yourself. 

{not_again_ai-0.5.1 → not_again_ai-0.7.0}/README.md

@@ -13,9 +13,9 @@
 [ruff-badge]: https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json
 [mypy-badge]: https://www.mypy-lang.org/static/mypy_badge.svg
 
-**not-again-ai** is a collection of various functionalities that come up over and over again when developing AI projects as a Data/Applied/Research Scientist. It is designed to be simple and minimize dependencies first and foremost. This is not meant to reinvent the wheel, but instead be a home for functions that don’t belong well elsewhere. Additionally, feel free to **a)** use this as a template for your own Python package. **b)** instead of installing the package, copy and paste functions into your own projects (this is made easier with the limited amount of dependencies and the MIT license).
+**not-again-ai** is a collection of various building blocks that come up over and over again when developing AI products. The key goals of this package are to have simple, but flexible interfaces and to minimize dependencies. Feel free to **a)** use this as a template for your own Python package. **b)** instead of installing the package, copy and paste functions into your own projects (this is made possible with the limited amount of dependencies and the MIT license).
 
-**Documentation** available within the [readmes](readmes) or auto-generated at [DaveCoDev.github.io/not-again-ai/](https://DaveCoDev.github.io/not-again-ai/).
+**Documentation** available within individual **[notebooks](notebooks)**, docstrings within the source, or auto-generated at [DaveCoDev.github.io/not-again-ai/](https://DaveCoDev.github.io/not-again-ai/).
 
 # Installation
 
@@ -27,82 +27,26 @@ Install the entire package from [PyPI](https://pypi.org/project/not-again-ai/) w
 $ pip install not_again_ai[llm,statistics,viz]
 ```
 
-The package is split into subpackages, so you can install only the parts you need.
+The package is split into subpackages, so you can install only the parts you need. See the **[notebooks](notebooks)** for examples.
 * **Base only**: `pip install not_again_ai`
-* **LLM only**: `pip install not_again_ai[llm]`
+* **LLM**: `pip install not_again_ai[llm]`
+    1. If you wish to use OpenAI
+        1. Go to https://platform.openai.com/settings/profile?tab=api-keys to get your API key.
+        1. (Optionally) Set the `OPENAI_API_KEY` and the `OPENAI_ORG_ID` environment variables.
+    1. If you wish to use Ollama:
+        1. follow the instructions to install ollama for your system: https://github.com/ollama/ollama
+        1. [Add Ollama as a startup service (recommended)](https://github.com/ollama/ollama/blob/main/docs/linux.md#adding-ollama-as-a-startup-service-recommended)
+        1. If you'd like to make the ollama service accessible on your local network and it is hosted on Linux, add the following to the `/etc/systemd/system/ollama.service` file:
+            ```bash
+            [Service]
+            ...
+            Environment="OLLAMA_HOST=0.0.0.0"
+            ```
+        Now ollama will be available at `http://<local_address>:11434`
 * **Statistics**: `pip install not_again_ai[statistics]`
 * **Visualization**: `pip install not_again_ai[viz]`
 
 
-# Quick Tour
-
-## Base
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/base.md)
-
-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), [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.
-
-One example: 
-```python
-client = openai_client()
-messages = [{"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Hello!"}]
-response = chat_completion(messages=messages, model="gpt-3.5-turbo", max_tokens=100, client=client)["message"]
->>> "Hello! How can I help you today?"
-```
-
-## Statistics
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/statistics.md)
-
-We provide a few helpers for data analysis such as:
-
-```python
-from not_again_ai.statistics.dependence import pearson_correlation
-# quadratic dependence
->>> x = (rs.rand(500) * 4) - 2
->>> y = x**2 + (rs.randn(500) * 0.2)
->>> pearson_correlation(x, y)
-0.05
-```
-
-## Visualization
-[README](https://github.com/DaveCoDev/not-again-ai/blob/main/readmes/viz.md)
-
-We offer opinionated wrappers around seaborn to make common visualizations easier to create and customize.
-
-```python
->>> import numpy as np
->>> import pandas as pd
->>> from not_again_ai.viz.time_series import ts_lineplot
->>> from not_again_ai.viz.distributions import univariate_distplot
-
-# get some time series data
->>> rs = np.random.RandomState(365)
->>> values = rs.randn(365, 4).cumsum(axis=0)
->>> dates = pd.date_range('1 1 2021', periods=365, freq='D')
-# plot the time series and save it to a file
->>> ts_lineplot(ts_data=values, save_pathname='myplot.png', ts_x=dates, ts_names=['A', 'B', 'C', 'D'])
-
-# get a random distribution
->>> distrib = np.random.beta(a=0.5, b=0.5, size=1000)
-# plot the distribution and save it to a file
->>> univariate_distplot(
-...     data=distrib, 
-...     save_pathname='mydistribution.svg', 
-...     print_summary=False, bins=100, 
-...     title=r'Beta Distribution $\alpha=0.5, \beta=0.5$'
-... )
-```
-
-<p float="center">
-  <img src="https://raw.githubusercontent.com/DaveCoDev/not-again-ai/44c53fb7fb07234aaceea40c90d8cb74e5fa6c15/assets/distributions_test4.svg" width="404" />
-  <img src="https://raw.githubusercontent.com/DaveCoDev/not-again-ai/44c53fb7fb07234aaceea40c90d8cb74e5fa6c15/assets/ts_lineplot5.svg" width="404" /> 
-</p>
-
-
 # Development Information
 
 The following information is relevant if you would like to contribute or use this package as a template for yourself. 

{not_again_ai-0.5.1 → not_again_ai-0.7.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "not-again-ai"
-version = "0.5.1"
+version = "0.7.0"
 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"
@@ -28,16 +28,17 @@ 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 }
+ollama = { version = "^0.1.9", optional = true }
+openai = { version = "^1.25.1", 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 }
 
 [tool.poetry.extras]
-llm = ["openai", "python-liquid", "tiktoken"]
+llm = ["ollama", "openai", "python-liquid", "tiktoken"]
 statistics = ["numpy", "scikit-learn", "scipy"]
 viz = ["numpy", "pandas", "seaborn"]
 
@@ -128,6 +129,7 @@ filterwarnings = [
     # Add additional warning supressions as needed here. For example, if a third-party library
     # is throwing a deprecation warning that needs to be fixed upstream:
     # "ignore::DeprecationWarning:typer",
+    "ignore::pytest.PytestUnraisableExceptionWarning"
 ]
 
 [tool.coverage.run]

not_again_ai-0.7.0/src/not_again_ai/base/file_system.py

@@ -0,0 +1,34 @@
+from pathlib import Path
+
+
+def create_file_dir(filepath: str | Path) -> None:
+    """Creates the parent directories for the specified filepath.
+    Does not throw any errors if the directories already exist.
+
+    Args:
+        filepath (str | Path): path to a file
+    """
+    root_path = Path(filepath).parent
+    root_path.mkdir(parents=True, exist_ok=True)
+
+
+def readable_size(size: float) -> str:
+    """Convert a file size given in bytes to a human-readable format.
+
+    Args:
+        size (int): file size in bytes
+
+    Returns:
+        str: human-readable file size
+    """
+    # Define the suffixes for each size unit
+    suffixes = ["B", "KB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB"]
+
+    # Start with bytes
+    count = 0
+    while size >= 1024 and count < len(suffixes) - 1:
+        count += 1
+        size /= 1024
+
+    # Format the size to two decimal places and append the appropriate suffix
+    return f"{size:.2f} {suffixes[count]}"

not_again_ai-0.7.0/src/not_again_ai/llm/ollama/chat_completion.py

@@ -0,0 +1,95 @@
+import contextlib
+import json
+import re
+from typing import Any
+
+from ollama import Client, ResponseError
+
+
+def _convert_duration(nanoseconds: int) -> float:
+    seconds = nanoseconds / 1_000_000_000
+    return round(seconds, 5)
+
+
+def chat_completion(
+    messages: list[dict[str, Any]],
+    model: str,
+    client: Client,
+    max_tokens: int | None = None,
+    context_window: int | None = None,
+    temperature: float = 0.8,
+    json_mode: bool = False,
+    seed: int | None = None,
+    **kwargs: Any,
+) -> dict[str, Any]:
+    """Gets a Ollama chat completion response, see https://github.com/ollama/ollama/blob/main/docs/api.md#generate-a-chat-completion
+    For a full list of valid parameters: https://github.com/ollama/ollama/blob/main/docs/modelfile.md#valid-parameters-and-values
+
+    Args:
+        messages (list[dict[str, Any]]): A list of messages to send to the model.
+        model (str): The model to use.
+        client (Client): The Ollama client.
+        max_tokens (int, optional): The maximum number of tokens to generate. Ollama calls this `num_predict`.
+        context_window (int, optional): The number of tokens to consider as context. Ollama calls this `num_ctx`.
+        temperature (float, optional): The temperature of the model. Increasing the temperature will make the model answer more creatively.
+        json_mode (bool, optional): This will structure the response as a valid JSON object.
+            It is important to instruct the model to use JSON in the prompt. Otherwise, the model may generate large amounts whitespace.
+        seed (int, optional): The seed to use for the model for reproducible outputs. Defaults to None.
+
+    Returns:
+        dict[str, Any]: A dictionary with the following keys
+            message (str | dict): The content of the generated assistant message.
+                If json_mode is True, this will be a dictionary.
+            completion_tokens (int): The number of tokens used by the model to generate the completion.
+            response_duration (float): The time taken to generate the response in seconds.
+    """
+
+    options = {
+        "num_predict": max_tokens,
+        "num_ctx": context_window,
+        "temperature": temperature,
+    }
+    if seed is not None:
+        options["seed"] = seed
+    options.update(kwargs)
+
+    all_args = {
+        "model": model,
+        "messages": messages,
+        "options": options,
+    }
+    if json_mode:
+        all_args["format"] = "json"
+
+    try:
+        response = client.chat(**all_args)
+    except ResponseError as e:
+        # If the error says "model 'model' not found" use regex then raise a more specific error
+        expected_pattern = f"model '{model}' not found"
+        if re.search(expected_pattern, e.error):
+            raise ResponseError(
+                f"Model '{model}' not found. Please use not_again_ai.llm.ollama.service.pull() first."
+            ) from e
+        else:
+            raise ResponseError(e.message) from e
+
+    response_data: dict[str, Any] = {}
+
+    # Handle getting the message returned by the model
+    message = response["message"].get("content", None)
+    if message and json_mode:
+        with contextlib.suppress(json.JSONDecodeError):
+            message = json.loads(message)
+    if message:
+        response_data["message"] = message
+
+    # Get the number of tokens generated
+    response_data["completion_tokens"] = response.get("eval_count", None)
+
+    # Get the latency of the response
+    if response.get("total_duration", None):
+        response_data["response_duration"] = _convert_duration(response["total_duration"])
+    else:
+        response_data["response_duration"] = None
+
+    return response_data

not_again_ai-0.7.0/src/not_again_ai/llm/ollama/ollama_client.py

@@ -0,0 +1,24 @@
+import os
+
+from ollama import Client
+
+
+def ollama_client(host: str | None = None, timeout: float | None = None) -> Client:
+    """Create an Ollama client instance based on the specified host or will read from the OLLAMA_HOST environment variable.
+
+    Args:
+        host (str, optional): The host URL of the Ollama server.
+        timeout (float, optional): The timeout for requests
+
+    Returns:
+        Client: An instance of the Ollama client.
+
+    Examples:
+        >>> client = client(host="http://localhost:11434")
+    """
+    if host is None:
+        host = os.getenv("OLLAMA_HOST")
+        if host is None:
+            raise ValueError("Host must be provided or OLLAMA_HOST environment variable must be set.")
+
+    return Client(host=host, timeout=timeout)

not_again_ai-0.7.0/src/not_again_ai/llm/ollama/service.py

@@ -0,0 +1,81 @@
+from typing import Any
+
+from ollama import Client
+
+from not_again_ai.base.file_system import readable_size
+
+
+def list_models(client: Client) -> list[dict[str, Any]]:
+    """List models that are available locally.
+
+    Args:
+        client (Client): The Ollama client.
+
+    Returns:
+        list[dict[str, Any]]: A list of dictionaries (each corresponding to an available model) with the following keys:
+            name (str): Name of the model
+            model (str): Name of the model. This should be the same as the name.
+            modified_at (str): The date and time the model was last modified.
+            size (int): The size of the model in bytes.
+            size_readable (str): The size of the model in a human-readable format.
+            details (dict[str, Any]): Additional details about the model.
+    """
+    response = client.list().get("models", [])
+
+    response_data = []
+    for model_data in response:
+        curr_model_data = {}
+        curr_model_data["name"] = model_data["name"]
+        curr_model_data["model"] = model_data["model"]
+        curr_model_data["modified_at"] = model_data["modified_at"]
+        curr_model_data["size"] = model_data["size"]
+        curr_model_data["size_readable"] = readable_size(model_data["size"])
+        curr_model_data["details"] = model_data["details"]
+
+        response_data.append(curr_model_data)
+
+    return response_data
+
+
+def is_model_available(model_name: str, client: Client) -> bool:
+    """Check if a model is available locally.
+
+    Args:
+        model_name (str): The name of the model.
+        client (Client): The Ollama client.
+
+    Returns:
+        bool: True if the model is available locally, False otherwise.
+    """
+    # If model_name does not have a ":", append ":latest"
+    if ":" not in model_name:
+        model_name = f"{model_name}:latest"
+    models = list_models(client)
+    return any(model["name"] == model_name for model in models)
+
+
+def show(model_name: str, client: Client) -> dict[str, Any]:
+    """Show information about a model including the modelfile, available parameters, template, and additional details.
+
+    Args:
+        model_name (str): The name of the model.
+        client (Client): The Ollama client.
+    """
+    response = client.show(model_name)
+
+    response_data = {}
+    response_data["modelfile"] = response["modelfile"]
+    response_data["parameters"] = response["parameters"]
+    response_data["template"] = response["template"]
+    response_data["details"] = response["details"]
+    return response_data
+
+
+def pull(model_name: str, client: Client) -> Any:
+    """Pull a model from the Ollama server and returns the status of the pull operation."""
+    return client.pull(model_name)
+
+
+def delete(model_name: str, client: Client) -> Any:
+    """Delete a model from the local filesystem and returns the status of the delete operation."""
+    return client.delete(model_name)

{not_again_ai-0.5.1/src/not_again_ai/llm → not_again_ai-0.7.0/src/not_again_ai/llm/openai}/chat_completion.py

@@ -6,7 +6,7 @@ from openai import OpenAI
 
 
 def chat_completion(
-    messages: list[dict[str, str]],
+    messages: list[dict[str, Any]],
     model: str,
     client: OpenAI,
     tools: list[dict[str, Any]] | None = None,
@@ -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:
@@ -29,8 +33,8 @@ def chat_completion(
         client (OpenAI): An instance of the OpenAI client.
         tools (list[dict[str, Any]], optional): A list of tools the model may generate JSON inputs for.
             Defaults to None.
-        tool_choice (str, optional): The tool choice to use. Can be "auto", "none", or a specific function name.
-            Defaults to "auto".
+        tool_choice (str, optional): The tool choice to use. Can be "auto", "required", "none", or a specific function name.
+            Note the function name cannot be any of "auto", "required", or "none". Defaults to "auto".
         max_tokens (int, optional): The maximum number of tokens to generate in the chat completion.
             Defaults to None, which automatically limits to the model's maximum context length.
         temperature (float, optional): What sampling temperature to use, between 0 and 2.
@@ -83,7 +87,7 @@ def chat_completion(
 
     if tools is not None:
         kwargs["tools"] = tools
-        if tool_choice not in ["none", "auto"]:
+        if tool_choice not in ["none", "auto", "required"]:
             kwargs["tool_choice"] = {"type": "function", "function": {"name": tool_choice}}
         else:
             kwargs["tool_choice"] = tool_choice

{not_again_ai-0.5.1/src/not_again_ai/llm → not_again_ai-0.7.0/src/not_again_ai/llm/openai}/context_management.py

@@ -1,6 +1,6 @@
 import copy
 
-from not_again_ai.llm.tokens import num_tokens_from_messages, truncate_str
+from not_again_ai.llm.openai.tokens import num_tokens_from_messages, truncate_str
 
 
 def _inject_variable(

{not_again_ai-0.5.1/src/not_again_ai/llm → not_again_ai-0.7.0/src/not_again_ai/llm/openai}/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-0.5.1/src/not_again_ai/llm → not_again_ai-0.7.0/src/not_again_ai/llm/openai}/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-0.5.1/src/not_again_ai/base/file_system.py

@@ -1,12 +0,0 @@
-import pathlib
-
-
-def create_file_dir(filepath: str) -> None:
-    """Creates the parent directories for the specified filepath.
-    Does not throw any errors if the directories already exist.
-
-    Args:
-        filepath (str): path to a file
-    """
-    root_path = pathlib.Path(filepath).parent
-    root_path.mkdir(parents=True, exist_ok=True)

not_again_ai-0.5.1/src/not_again_ai/llm/chat_completion_vision.py

@@ -1,87 +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.
-            '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.
-    """
-    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