PLMBR

create data processing pipelines

Installation

Installing plmbr, which is available as a Python package, can be done with pip, preferably in a virtual environment. Launch a terminal and use the following command to install plmbr.

pip install plmbr

Introduction

It is a Python library for building predictable and reproducible data pipelines with ease. It wraps pandas with a clean, modular API inspired by TensorFlow’s Keras. plmbr simplifies complex data workflows into composable steps. This works with Jupyter notebooks as well.

Creating Pipes

Pipe objects are simple python classes that implements flow method which takes a pandas DataFrame as input and returns a pandas DataFrame as output. Pipe objects are mean to perform a specific operation on the data. Following is an example of a Pipe object that doubles the values in a column.

from pandas import DataFrame
from plmbr import Pipe, Pipeline

class DoubleIt(Pipe):
    def __init__(self, col: str):
        self.col = col

    def flow(self, data: DataFrame) -> DataFrame:
        data[self.col] = data[self.col] * 2
        return data

Building Pipeline

Pipeline is similar to Sequential in Keras. It takes a list of Pipe objects as input and builds a pipeline of operations. Each Pipe object is called in sequence to process the data.

pipeline = Pipeline([
    DoubleIt(col="A"),
])

Processing Data

df = DataFrame({ "A": [1, 2, 3] }, columns=["A"])
doubled_df = pipeline(df) # or pipeline.flow(df)
doubled_df

	A
0	2
1	4
2	6

Contribute

Data transformation is a complex task, and plmbr is designed to be extensible and flexible to accommodate a wide range of use cases. If you have ideas for new features, find bugs or usability issues, or want to submit a pull request, your input is appreciated. Contributions of all kinds code, documentation, or feedback help improve the library and shape its future.