User:EranBot

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search


EranBot
This is a bot account owned by User:ערן (aka Eran).
Robot icon.svg
Copy & Paste detection

This is a copy and paste detection bot based on the multi year efforts here WP:Turnitin. The bot runs as the background service for CopyPatrol.

CopyPatrol
Patrol copyright violations

How it works[edit]

  1. All recent edits to En Wikipedia over a certain size (+500 [after removing wikicode]) are scanned (that wasn't there in the previous and previous-previous revisions). The text is sent to plagiarism detection service iThenticate.
  2. Edits with similar text to external sources are consider as possible copyright violations, and are reported in the CopyPatrol tool.
    1. If the external source is mirror of Wikipedia, it is either removed by iThenticate itself or afterwards by the bot (based on blacklist)
    2. If the source is broken link the bot removes it
  3. Each entry in the report page have the following fields: Title of the edited page, Diff with link to the relevant edit diff and page history, Editor, Source - link to report page in iThenticate (titled "report") and links to possible sources of the edit (titled "compare"), Status - Should be filled manually with TP/FP. The bot adds hints for possible good edits:
    • citation - the added text mention the source. For short text it is OK (in copyright sense) and for long text it is a violation (see also Wikipedia:Close paraphrasing).
    • Mirror? - the added text comes from a source that may be a possible mirror site of Wikipedia. E.g the source seems to be unknown mirror (that doesn't appear in our blacklist, but have attribution to Wikipedia). Editors can add such sites to blacklist, so they don't appear in future.
    • (CC) - the added text comes from a source that probably have creative commons license.

Current state[edit]

Currently it only on all en/es/fr/cs Wikipedia ( There may be potential for it to expand to other languages - just ask!). It has been a great help for medical articles. Efforts to make it more functional are ongoing. The results are being placed at CopyPatrol tool and the bot runs 24/7.

There is NO plan for this bot to make edits to mainspace. The concept has been discussed with the WMF legal team who are happy with it.

Who runs the bot[edit]

It is run by the Hebrew Wikipedia User:Eran. Doc James and User:Ocaasi have been guiding its development.

Source[edit]

The bot is based on pywikibot and you can find its source code in github. It is possible to run the bot in other Wikipedia languages, but to run the bot you have to request account for iThenticate.


Feel free to edit this page