Pandas_llm 0.0.6 (conversational dataframe queries)

This release contains a bug fix in the sandbox

Pandas_llm 0.0.5 (conversational dataframe queries)

This version fixes an installation problem.

Pandas_llm 0.0.4 (conversational dataframe queries)

Release Notes for pandas-llm v0.0.4:

PandasLLM Class Constructor

The constructor for the PandasLLM class has been enhanced in this release to provide more flexibility and control over the language model interaction. The constructor accepts the following arguments:

data (mandatory): The data to be used. It can be a Pandas DataFrame, a list of lists, tuples, dictionaries, a dictionary, a string, or a list.

llm_engine (optional): The name of the LLM engine to use. Currently, only OpenAI is supported. Defaults to "openai".

llm_params (optional): A dictionary of parameters to be used with the OpenAI API. This allows customization of the LLM behavior. Defaults to model=gpt-3.5-turbo and temperature=0.2.

prompt_override (optional): A boolean that determines whether or not the prompt is overridden. If set to True, the custom prompt becomes the main prompt. Defaults to False.

custom_prompt (optional): A string that can be provided if prompt_override is False. The custom prompt will be added to the default pandas_llm prompt. Defaults to an empty string.

path (optional): The path to the file where the debug data will be saved. If not specified, debug data files will not be generated.

verbose (optional): A boolean determines whether debugging information will be printed. If set to True, additional debugging info will be displayed. Defaults to False.

data_privacy (optional): A boolean determines whether the data is treated as private. If set to True, the function will not send the data content to OpenAI. Defaults to True.

llm_api_key (optional): The OpenAI API key to be used. The library will attempt to use the default API key configured if not provided.

force_sandbox (optional): A boolean determining the fallback behaviour if the sandbox environment fails. If set to False and the sandbox fails, the library will retry using eval, which is less safe. Defaults to False.

The constructor also calls the parent class's constructor to ensure proper initialization.

These enhancements offer more control and customization options for utilizing the PandasLLM class in your projects.

Pandas_llm 0.0.3 (conversational dataframe queries)

This version is more requirements friendly

Pandas_llm 0.0.2 (conversational dataframe queries)

This version is more restrictive on the dependencies.

Pandas_llm 0.0.1 (conversational dataframe queries)

Release Notes for pandas_llm v0.0.1

We are pleased to announce the release of pandas_llm version 0.0.1. This library introduces the powerful capabilities of OpenAI's language models to your Pandas dataframes. With pandas_llm, you can now make data queries using natural language prompts.

New Features in v0.0.1

Pandas DataFrame Extension: The PandasLLM class extends the standard pandas DataFrame, allowing OpenAI language models to query and manipulate data.

Natural Language Queries: Use natural language prompts to interact with your dataframes. The prompt method takes a request string, constructs a prompt for the OpenAI API, and returns the resulting code execution.

Code Execution in a Sandbox: Execute generated Python code in a secure environment. The execInSandbox method runs the Python code received from the OpenAI API in a sandboxed environment, protecting your system from potentially harmful code.

Privacy Protection: Avoid sending data to OpenAI

Customizable Configuration: Customize your PandasLLM instance with various configuration options, such as the OpenAI model to use, the API key, the temperature for the model and more.

Verbose Mode: Track the progress of your operations with the verbose mode. When turned on, the library will print detailed logs of its activities.

This release of pandas_llm opens up new possibilities for data exploration and manipulation. By combining the power of OpenAI's language models with the versatility of pandas, we aim to make data analysis more intuitive and accessible.

Please review the official documentation for a detailed guide on using the pandas_llm library. We welcome feedback and contributions from the community.