CompressedList Implementation in Python

A Python implementation of the CompressedList class from R/Bioconductor for memory-efficient list-like objects.

CompressedList is a memory-efficient container for list-like objects. Instead of storing each list element separately, it concatenates all elements into a single vector-like object and maintains information about where each original element begins and ends. This approach is significantly more memory-efficient than standard lists, especially when dealing with many list elements.

Install

To get started, install the package from PyPI

pip install compressed-lists

Usage

from compressed_lists import CompressedIntegerList, CompressedStringList

# Create a CompressedIntegerList
int_data = [[1, 2, 3], [4, 5], [6, 7, 8, 9]]
names = ["A", "B", "C"]
int_list = CompressedIntegerList.from_list(int_data, names)

# Access elements
print(int_list[0])      # [1, 2, 3]
print(int_list["B"])    # [4, 5]
print(int_list[1:3])    # Slice of elements

# Apply a function to each element
squared = int_list.lapply(lambda x: [i**2 for i in x])
print(squared[0])       # [1, 4, 9]

# Convert to a regular Python list
regular_list = int_list.to_list()

# Create a CompressedStringList
char_data = [["apple", "banana"], ["cherry", "date", "elderberry"], ["fig"]]
char_list = CompressedStringList.from_list(char_data)

Partitioning

The Partitioning class handles the information about where each element begins and ends in the concatenated data. It allows for efficient extraction of elements without storing each element separately.

from compressed_lists import Partitioning

# Create partitioning from end positions
ends = [3, 5, 10]
names = ["A", "B", "C"]
part = Partitioning(ends, names)

# Get partition range for an element
start, end = part[1]  # Returns (3, 5)

Note

Check out the documentation for extending CompressedLists to custom data types.

Note

This project has been set up using BiocSetup and PyScaffold.