• Lightweight PageRank: A clean implementation of the PageRank algorithm using power iteration.
• TextRank for SEO: Extract meaningful keywords from text documents to understand topics and improve SEO.
• Graph Flexibility: Works with graphs represented as dictionaries or lists of lists.
• Customizable: Tweak parameters like damping factor, convergence tolerance, and max iterations.
• Well-Tested: High test coverage to ensure reliability.
• Typed: Fully type-hinted for better code quality and editor support.

• Installation
• Quick Start
  • PageRank Example
  • TextRank Example
• Interactive Demo
• API Reference
• Development
• Contributing
• License

This project uses uv for dependency management.

# Install only the required dependencies for production use
uv sync

# Install all development dependencies for contributing
uv sync --all-extras

You will also need to download NLTK data for TextRank. A helper script is provided:

uv run python download_nltk_data.py

Calculate PageRank scores for a simple graph. The scores represent the "importance" of each node.

from pagerank import power_iteration

# Define a graph where keys are nodes and values are outgoing links
graph = {
    "A": {"B": 1, "C": 1},
    "B": {"C": 1},
    "C": {"A": 1},
}

# Calculate PageRank scores
scores = power_iteration(graph)
print(scores)
# A    0.443029
# C    0.354423
# B    0.202548
# dtype: float64

Extract the most relevant keywords from a piece of text.

from textrank import textrank

document = """
Natural language processing is a subfield of linguistics, computer science,
and artificial intelligence concerned with the interactions between computers
and human language.
"""

# Extract top keywords
keyword_scores = textrank(document)
print(keyword_scores.head())
# language       0.239396
# computer       0.177059
# intelligence   0.155705
# subfield       0.134703
# linguistics    0.134703
# dtype: float64

For a more detailed walkthrough with visualizations, check out the Jupyter Notebook demo.

uv run jupyter notebook demo.ipynb

This demo covers:

• Basic and advanced PageRank examples.
• Keyword extraction with TextRank.
• Visualizing results with matplotlib and seaborn.

power_iteration(transition_weights, rsp=0.15, epsilon=0.00001, max_iterations=1000)

Returns: pandas.Series with nodes as keys and PageRank scores as values.

textrank(document, window_size=2, rsp=0.15, relevant_pos_tags=["NN", "NNP", "ADJ"])

Returns: pandas.Series with words as keys and TextRank scores as values, sorted descending.

To set up a development environment, clone the repo and install the dependencies.

git clone https://github.com/ashkonf/PageRank.git
cd PageRank
uv sync --all-extras

This project uses pre-commit hooks for quality checks. Install them with:

uv run pre-commit install

• Run tests: uv run pytest
• Check formatting: uv run ruff format --check .
• Check for linting errors: uv run ruff check .
• Run type checks: uv run pyright .

Contributions are welcome! Please feel free to open an issue or submit a pull request.

1. Fork the repository.
2. Create your feature branch (git checkout -b feature/my-new-feature).
3. Commit your changes (git commit -am 'Add some feature').
4. Push to the branch (git push origin feature/my-new-feature).
5. Open a new Pull Request.

This project is licensed under the Apache License 2.0. See the LICENSE file for details.

For more on the TextRank algorithm, see the original paper by Mihalcea and Tarau.