https://github.com/miku/pgrk

Command line pagerank computation with Go.
https://github.com/miku/pgrk

pagerank

Last synced: 28 days ago
JSON representation

Command line pagerank computation with Go.

• Host: GitHub
• URL: https://github.com/miku/pgrk
• Owner: miku
• License: mit
• Created: 2013-11-11T11:51:39.000Z (about 11 years ago)
• Default Branch: master
• Last Pushed: 2020-06-12T09:47:32.000Z (over 4 years ago)
• Last Synced: 2024-06-19T05:40:55.560Z (6 months ago)
• Topics: pagerank
• Language: Go
• Homepage:
• Size: 22.5 KB
• Stars: 19
• Watchers: 5
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
README

======

Simple command line [PageRank](http://infolab.stanford.edu/~backrub/google.html) calculator.

The input file represents a directed graph as adjacency list, where each node

must be identified by an integer (translation from and to your domain must

happen elsewhere).

Installation

------------

    $ go get github.com/miku/pgrk/cmd/{pgrk,pgrk-gen,pgrk-dot}

Example

-------

![Shiny graph](http://i.imgur.com/0ZQYLFl.png)

Example input file:

    $ cat example.in

    0   1

    1   2

    2   3   4   5

    4   2

    5   1

Usage:

    $ pgrk example.in 2> /dev/null | sort -k2,2 -nr

    2   0.33477170103317816

    1   0.20154082325712538

    5   0.13963495553328562

    4   0.13963495553328562

    3   0.13963495553328562

    0   0.0447826091098397

Options:

    $ pgrk

    Usage: pgrk [OPTIONS] FILE

    File format: TSV (NODE OUTBOUND [OUTBOUND, ...])

      -c=0.0001: convergence criteron

      -w=0.85: walk probability

Benchmark

---------

* Computing the pagerank of an 10,004,750 node sparse graph takes about 30s (tested on i5-3470, 4 core, 3.2GHz machine, 16G RAM).

* `{"p": 0.8, "nodes": 30000000, "edges": 46735510}`: 1m55.010s

* `{"p": 0.8, "nodes": 50000000, "edges": 77879276}`: 3m6.173s

* `{"p": 0.99, "nodes": 50000000, "edges": 268891019}`: 4m10.350s

Utils

-----

Generate and visualize random directed graphs, with `pgrk-gen`

and `pgrk-dot`:

    $ pgrk-gen -n 30 -p 0.75 > test.in

    $ cat test.in | pgrk-dot | dot -Tpdf -o test.pdf && open test.pdf

    $ pgrk test.in 2> /dev/null | sort -nrk2,2 | head -5

    20  0.10566870915778451

    18  0.10238092345567337

    6   0.10012509291487659

    28  0.06610747553255986

    16  0.06485242652853762

![Another Shiny graph](http://i.imgur.com/hzzKtzq.png)

Credits

-------

* [Thomas Dimson](https://web.archive.org/web/20150408134637/https://github.com/cosbynator) for [WikiRank](https://github.com/cosbynator/WikiRank)