makeitsample 1.0.2

A Toolkit for Generating Typological Language Samples Based on the Diversity Value

makeitsample

makeitsample is a Python library for generating typological language samples using the diversity value (DV) metric (Rijkhoff et al., 1993; Rijkhoff and Bakker, 1998; Bakker, 2010).

It provides tools to build language family trees from CSV data, compute diversity values for each node, and select a representative set of languages that reflect genealogical and typological diversity.

---

📚 What It Does

makeitsample is designed to support researchers and linguists in the creation of typologically diverse language samples. It consists of two main modules:

• language_family_tree.py — defines tree structures and computes diversity values (DV).
• forest.py — manages a forest of language families and handles sampling logic across multiple trees.

---

🚀 Features

• Build hierarchical language family trees from CSV input.
• Handle both nested genealogies and isolated languages.
• Calculate diversity values at each node in a tree.
• Select representative languages based on weighted sampling.
• Minimal dependencies and easy integration into other projects.

---

📦 Installation

pip install makeitsample

🛠️ Usage

Prepare the input files

makeitsample requires a set of input files (representing language families) in CSV format. The CSV files should contain the following columns:

• id: the id of the node
• name: the name of the node
• parent_id: the id of the parent node
• type: the type of the node (the only allowed values are "family", "group" or "language")

The user can also add any other columns to the CSV files.

As a library

Create a language family tree from CSV data

import makeitsample.language_family_tree as lft

# Create a language family tree from CSV data
family = lft.LanguageFamilyTree("path/to/csv/file.csv")

# Print the tree structure
print(family)

Calculate diversity values for the language family trees

from makeitsample.forest import Forest

# Create a forest of language families
language_families = Forest(dir="path/to/directory/with/csv/files")

# Update the trees with diversity values
language_families.dv()

# Export the updated trees to CSV
language_families.export_forest(dir="path/to/output/dir", format="csv")

Sample languages from the language family trees

from makeitsample.forest import Forest

# Create a forest of language families
language_families = Forest("path/to/directory/with/csv/files")

# Sample languages from the forest
language_families.make_sample(n=100)

# Export the sampled languages to CSV
language_families.export_sample(dir="path/to/output/dir", format="csv")

# Export the sampled languages to JSON
language_families.export_sample(dir="path/to/output/dir", format="json")

As a command-line tool

Sample languages from the language family trees

makeitsample [-h] [-n N] [-i INPUT] [-o OUTPUT] [-f {csv,json}] [-s SAMPLENAME] [-r RANDOM_SEED]

Arguments

• -h, --help: Show this help message and exit.
• -n N, --number N: Number of languages to sample.
• -i INPUT, --input INPUT: Path to the input directory containing CSV files.
• -o OUTPUT, --output OUTPUT: Path to the output directory for sampled languages.
• -f {csv,json}, --format {csv,json}: Output format for sampled languages (default: csv).
• -s SAMPLENAME, --sample_name SAMPLENAME: Name of the sample (default: sample).
• -r RANDOM_SEED, --random_seed RANDOM_SEED: Random seed for reproducibility.

Example usage

makeitsample -n 100 -i data -o out -f csv -s test_sample

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.

📄 Citation

If you use this library in your research, please cite the following paper: