dq-summary-llm 0.0.3

No project description provided

Data Quality Summary using LLM

This Python package generates a data quality report for a dataset using a Large Language Model (LLM) such as OpenAI's GPT. It takes a Pandas DataFrame as input and returns a human-readable report about potential data issues. The package also includes full LLM responses and logs for transparency.

Features

• Accepts any Pandas DataFrame as input.
• Connects to an LLM using OpenAI API credentials.
• Generates intelligent data quality reports and explains the data.
• Optionally allows custom prompts to guide the LLM.

Class: DataQualitySummery

__init__(api_key: str, base_url: str, model: str, temprature: float, max_token: int)

Initializes the LLM connection.

Parameters:

• api_key: OpenAI or compatible API key.
• base_url: Base URL of the LLM API.
• model: The model name, e.g., "gpt-4".
• temprature: Float to control randomness.
• max_token: Maximum tokens allowed in the LLM response.

---

data_quality_details(data: pd.DataFrame, response: str = None) -> str

Generates a data quality report using the provided DataFrame.

Parameters:

• data: A pandas DataFrame.
• response: Optional custom instruction for the LLM.

Returns:

• A string containing the data quality report, including raw LLM responses.

---

Example Usage

from your_module import DataQualitySummery
import pandas as pd

# Create a sample DataFrame
df = pd.DataFrame({
    'Name': ['Alice', None, 'Charlie'],
    'Age': [25, 30, None]
})

# Initialize the class
dq = DataQualitySummery(
    api_key='your-api-key',
    base_url='https://api.openai.com/v1',
    model='gpt-4',
    temprature=0.7,
    max_token=500
)

# Generate the report
report = dq.data_quality_details(df)
print(report)

---

data_quality_summary(response) -> str

Generates a data quality summary using the provided response. Pass the data_quality_details

Parameters:

• response: Previous data_quality_details instruction for the LLM.

Returns:

• A string containing the data quality summary, including raw LLM responses.

---

Example Usage

from your_module import DataQualitySummery
import pandas as pd

# Create a sample DataFrame
df = pd.DataFrame({
    'Name': ['Alice', None, 'Charlie'],
    'Age': [25, 30, None]
})

# Initialize the class
dq = DataQualitySummery(
    api_key='your-api-key',
    base_url='https://api.openai.com/v1',
    model='gpt-4',
    temprature=0.7,
    max_token=500
)

# Generate the report
report = dq.data_quality_details(df)
print(report)

summary= dq.data_quality_summary(report)
print(summary)

data_explainer(data, response:str= None, sample_size: int=10) -> str

Generates a explanation base on the dataset make sure the column name is proper not abbreviation of actual column name.

Parameters:

• data: A pandas DataFrame.
• response: Optional custom instruction for the LLM.
• sample_size: Number of data points as an example for the LLM.

Returns:

• A string containing the data Expalantion what is the data represent, including raw LLM responses.

---

Example Usage

from your_module import DataQualitySummery
import pandas as pd

# Create a sample DataFrame
df = pd.DataFrame({
    'Name': ['Alice', None, 'Charlie'],
    'Age': [25, 30, None]
})

# Initialize the class
dq = DataQualitySummery(
    api_key='your-api-key',
    base_url='https://api.openai.com/v1',
    model='gpt-4',
    temprature=0.7,
    max_token=500
)

# Generate the explanation

explanation= dq.data_explainer(data, 5)
print()