hana_ai.smart_dataframe

hana.ai smart_dataframe is a dataframe agent to perform HANA dataframe operations via natural language.

Smart DataFrame.

The following class is available:

  • :class SmartDataFrame

class hana_ai.smart_dataframe.SmartDataFrame(dataframe: DataFrame)

Smart DataFrame.

Parameters:
dataframeDataFrame

Dataframe.

Examples

>>> from hana_ai.smart_dataframe import SmartDataFrame
>>> sdf = SmartDataFrame(dataframe=hana_df)
>>> sdf.configure(llm=llm, verbose=True)
>>> sdf.ask(question="Show the samples of the dataset")
>>> new_sdf = sdf.transform(question="Get last two rows")
configure(llm: BaseLLM, tools: List[BaseTool] = None, verbose: bool = False, **kwargs)

Configure the Smart DataFrame.

Parameters:
llmBaseLLM

LLM.

toolsList of BaseTool, optional

The dataframe tools to use.

Defaults to df_tools.

ask(question: str)

Ask a question.

Parameters:
questionstr

Question.

transform(question: str, output_key='output')

Transform the dataframe.

The model output must be a single read-only SELECT/WITH statement. Comments, unsafe statements and embedded semicolons are rejected; one trailing terminator is removed before passing SQL to hana-ml. These checks do not restrict the permissions of tools used by ask().

Parameters:
questionstr

Question.

output_keystr, optional

Reserved for compatibility; the executor's 'output' key is used.