provit 0.2

A a light, dezentralized provenance tracking framework using the W3C PROV-O vocabulary

PROVIT is a light, dezentralized provenance tracking framework. It allows the user to track workflows and modifications of data and files. A small subset of the W3C PROV-O vocabulary is implemented. Its aim is to provided an easy to use interface for users who have never worked with provenance tracking before. It you feel limited by PROVIT you should have a look at the more extensive implementation prov.

Requirements

This Software was tested with Python 3.5 and 3.6.

Installation

Installation via pip is recommended for end users. We strongly encourage end users to make use of a virtualenv.

pip

Clone the repository and create a virtualenv.

$ git clone https://github.com/diggr/pit
$ mkvirtualenv provit

Install it with pip

$ pip install PATH_TO_PROVIT_REPOSITORY

git / Development

Clone the repository and create a virtualenv.

$ git clone https://github.com/diggr/pit
$ mkvirtualenv provit

Install it with pip in editable mode

$ pip install -e PATH_TO_PROVI_REPOSITORY

Usage

Provenance Integration Tools provide a command line client which can be used out of the box to enrich any file based data with provenance information. Furthermore the provenance class and vocabulary shipped with PIT can be used within other applications.

Command Line Client

Usage:

$ pit [OPTIONS] FILEPATH

Options:

--add

Add provenance information layer to file

-a, --agent=TEXT

Provenance information: agent

--activity=TEXT

Provenance information: activity

-d, --desc=TEXT

Provenance information: Description of the data manipulation process

-o, --origin=TEXT

Provenance information: Data origin

-s, --sources=TEXT

Provenance information: Source files

-b, --browser

Provenance browser

-n, --namespace=TEXT

Provenance Namespace, default: http://provit.diggr.link/

--help

Show this message and exit.

Provenance Class

from pit.prov import Provenance

# load prov data for a file, or create new prov for file
prov = Provenance(<filepath>)

# add provenance metadata
prov.add(agent="agent", activity="activity", description="...")
prov.add_primary_source("primary_source", url="http://...", comment="...")
prov.add_sources(["filepath1", "filepath2"])

# return provenance as json tree
prov_dict = prov.tree()

# save provenance metadata into "<filename>.prov" file
prov.save()

Roadmap

General Roadmap containing features we’d like to realize in the project

• Add Persons to Agent, to allow more granular activity tracking

Feature Wishlist

A more detailed list of specific (smaller) features and functionality.

Notify user if source referenced file changes

Provenance files contain the version of a file if referenced, if an older version (i.e. not the current version) of a file is referenced, a warning should be displayed.

File Browser

A file browser showing e.g. all files with missing provenance.

Reference Clustering

Inspect files in your research folder, and display all references, to identify clusters. This could help structuring a messy research directory without breaking scripts, or at least knowing, which scripts possibly need to be updated.

FAQ / Paradigms

Can I add multiple agents to an activity?

No. The reason is: If you can distinguish the activities or impact of the agent, then you have multiple agents with multiple activities. E.g. if you let three students help you proofreading a file and you get back 1 revised version, then the three students are 1 agent as you cannot distinguish between their results. If you get back 3 versions you have 3 agents and 3 activities.

Overview

Authors:

P. Mühleder [e-mail], F. Rämisch [e-mail]

License:

MIT

Copyright:

2018, Peter Mühleder and Universitätsbibliothek Leipzig