Skip to main content

A package to mesure diversity of log files

Project description

LogDiv: A Python Module for Computing Diversity in Transaction Logs

LogDiv is a Python module for the computation of the diversity of items requested by users in transaction logs.

It takes two inputs:

  1. A log file with transactions.
  2. A file with item atributes.

Computing the diversity of items requested by users is a task of interest in many fields, such as sociology, recommender systems, e-commerce, and media studies. Check the example below.

Getting Started

Prerequisites

LogDiv requires:

  • Python
  • Numpy - Essential
  • Pandas - Essential
  • Matplotlib - Essential
  • tqdm - Optionnal: progression bar, only one function requires it
  • Graph-tool - Optionnal: only one function requires it
$ python3 -m pip install numpy
$ python3 -m pip install panda
$ python3 -m pip install matplotlib 
$ python3 -m pip install tqdm 

Installing Graph-tool is more complicated: https://git.skewed.de/count0/graph-tool/wikis/installation-instructions

Installing

To install LogDiv, you need to execute:

$ pip install logdiv

Specification

Entries format

LogDiv needs a specific format of entries to run:

  • A file describing all requests under a table format, whose fields are:
  • user ID
  • timestamp
  • requested item ID
  • referrer item ID
  • A file describing all pages visited under a table format, whose fields are:
  • item ID
  • classification 1
  • classification 2
  • ...

YAML file

Codes that use LogDiv are directed by a YAML file: if you want to modify entry files, or the features you want to compute, you just need to modify the YAML file, not the code itself. This file is self-explanatory.

Documentation

If you want precision on a function of LogDiv:

  • what is the purpose of the function
  • what these functions take in entry
  • what they return

you need to run in a Console Python:

>>> help(function)

Examples

You dispose of two examples to familiarize yourself with LogDiv:

  • Example 1 uses a short dataset to show how to use LogDiv
  • Example 2 uses a dataset of more than 100 thousands of requests to show what kind of results can be obtained

These examples (dataset, script and yaml file) can be found in datasets directory.

Example 1

The following example illustrates the entries format of the package.

user timestamp requested_item referrer_item
user1 2019-07-03 00:00:00 v1 v4
user1 2019-07-03 00:01:00 v4 v2
user1 2019-07-03 00:01:10 v4 v6
user1 2019-07-03 00:01:20 v4 v6
user1 2019-07-03 00:02:00 v6 v9
user1 2019-07-03 03:00:00 v8 v10
user1 2019-07-03 03:01:00 v8 v5
user2 2019-07-05 12:00:00 v3 v5
user2 2019-07-05 12:00:30 v5 v7
user2 2019-07-05 12:00:45 v7 v9
user2 2019-07-05 12:01:00 v9 v6
user3 2019-07-05 18:00:00 v10 v5
user3 2019-07-05 18:01:15 v10 v7
user3 2019-07-05 18:03:35 v10 v9
user3 2019-07-05 18:06:00 v7 v4
user3 2019-07-05 18:07:22 v5 v2
item class1 class2 class3
v1 x \alpha h
v2 y \beta h
v3 y \beta f
v4 x \beta h
v5 z \gammma f
v6 y \alpha h
v7 z \alpha f
v8 x \gammma f
v9 y \alpha f
v10 z \gammma h

If you want to run example 1, you need to be in the directory datasets/example1, and run:

$ python3 example_1.py

Example 2

This figure is the Gephi graph of dataset 2, where each color correspond to a different media.

The file describing requests has the same structure than the one in example 1.

The file describing pages is more concrete than the one in example 1:

item media continent
item0 Politics Europe
item1 Health Asia
item2 Politics North America

If you want to run example 2, you need to be in the directory datasets/example2, and run:

$ python3 example_2.py

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for logdiv, version 0.0.2.2
Filename, size File type Python version Upload date Hashes
Filename, size logdiv-0.0.2.2-py3-none-any.whl (38.5 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size logdiv-0.0.2.2.tar.gz (24.3 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page