MicroFrame is a lightweight educational data manipulation library designed to provide a pandas-like interface for students learning to work with real-world data. It is optimized for toy datasets and aims to introduce users to data analysis concepts without the overhead of pandas.

• Efficient CSV Reading: Quickly and easily read CSV files to create MicroFrame objects, optimized for educational purposes and smaller datasets.
• Data Type Handling: Advanced data type inference and explicit type setting offer both convenience and control over the structure of your data.
• Flexible MicroFrame Objects: Utilize MicroFrame objects that mimic pandas DataFrame for intuitive data manipulation and analysis.
• Clear Tabular Display: Use MicroFrame’s printing capabilities to generate well-formatted tabular representations of your data, making it easier to interpret and present.
• Robust Data Manipulation: Perform a variety of data manipulation tasks with methods similar to pandas, such as filtering, column dtype modification and summarizing data.
• Advanced Indexing: Access data efficiently with advanced indexing options, using the iloc method similar to pandas iloc method.
• Data Conversion Tools: Seamlessly convert your MicroFrame objects to other formats, including NumPy arrays, with the to_numpy method for further numerical computation.
• User-Friendly API: Experience a user-friendly API that mirrors pandas to facilitate the transition from educational projects to real-world data analysis.

Install MicroFrame using pip:

pip install microframe

Import MicroFrame and load a dataset:

import microframe as mf

# Read a CSV file into a MicroFrame object
mframe = mf.read_csv("path_to_your_csv_file.csv")

# Alternatively, create a MicroFrame object manually
data = [[1, "a"], [2, "b"], [3, "c"]]
dtypes = ["int32", "U1"]
columns = ["num", "char"]

mframe = mf.MicroFrame(data, dtypes, columns)
mframe.head() # Display first 5 rows
# Extract data as numpy array
mframe_slice = mframe.iloc[:, 0] # returns all rows, but just col 0
numpy_array = mframe_slice.to_numpy() # returns mframe_slice as a numpy array

MicroFrame simplifies the process of data analysis. Here are some basic operations:

import microframe as mf

mframe = mf.read_csv("path_to_your_csv_file.csv")

import microframe as mf
data = [[1, "a"], [2, "b"], [3, "c"]]
dtypes = ["int32", "U1"]
columns = ["num", "char"]

mframe = mf.MicroFrame(data, dtypes, columns)

MicroFrame provides several methods to manipulate your data:

mframe.rename({"num": "number", "char": "character"})

mframe.change_dtypes({"number": "float64", "character": "U10"})

data = [[1, "a"], [2, "b"], [3, "c"]]
dtypes = ["int32", "U1"]
columns = ["num", "char"]

mframe = mf.MicroFrame(data, dtypes, columns)
first_col = mframe["num"] # Access just num column

first_row = mframe.iloc[0]

The iloc indexer allows for integer-location based indexing:

first_row = mframe.iloc[0]

first_two_rows = mframe.iloc[:2]

cell_value = mframe.iloc[2, 1]

mframe.iloc[2, 1] = "Test"

subset = mframe.iloc[:2, :2]

Similar to pandas, you can display parts of your dataset:

mframe.head(2)

mframe.tail(2)

For times when you need to work with a NumPy array, MicroFrame provides the to_numpy method:

# Convert the MicroFrame to a 2D NumPy array
numpy_array = mframe.to_numpy()

This method will convert the structured data within the MicroFrame to a regular 2D NumPy array.

For scenarios where you need to perform NumPy operations on a subset of your data, you can chain the iloc indexer with the to_numpy method:

# Select the first two rows using iloc and convert them to a NumPy array
numpy_subset = mframe.iloc[:, 1:5].to_numpy()

pip install "microframe[dev]"

python -m pytest tests

For full documentation, visit our MicroFrame Documentation. Here, you will find detailed information on all the functionalities that MicroFrame offers.

MicroFrame is released under the MIT License. Feel free to use it in your projects, and we'd love to hear about what you build!