valhallasw / plagiabot

plagiabot

Plagiabot is a copyright violation detection bot.

Repository for Turntin-based plagiarism detection for Wikipedia. See https://en.wikipedia.org/wiki/Wikipedia:Turnitin for details.

Running the bot

The bot support standard pywikibot page generators - for most of them it check the latest revision. The bot also supports special generators to check specific edit based on the diff:

• recentchanges (DB based)
• recentchanges_api (api based)
• live - recent changes using streaming or IRC

See command line help for more details

valhallasw@lisilwen:~/src/plagiabot$ python -i plagiabot.py
Logging in...
Finding folder to upload into, with name 'Wikipedia'...
Upload test text to iThenticate...
Polling iThenticate until document has been processed... . . .
Part #14558041 has a 62% match. Getting details...
Details are available on https://api.ithenticate.com/report/14557806/similarity
Sources found were:
 * I  62% 42 words at http://lrd.yahooapis.com/_ylc=X3oDMTVnb2
 * I  62% 42 words at http://lrd.yahooapis.com/_ylc=X3oDMTVncn
 * I  62% 42 words at http://lrd.yahooapis.com/_ylc=X3oDMTU4aD
 * I  62% 42 words at http://www.games2.about2006.com/aboutsit
 * I  62% 42 words at http://medlibrary.org/medwiki/All_your_b
 * I  62% 42 words at http://plumbot.com/All_your_base_are_bel
 * I  62% 42 words at http://lembolies.com/Your
 * I  62% 42 words at http://dvdradix.com/capture-flash-video-
 * I  62% 42 words at http://www.dvdradix.com/capture-flash-vi
 * I  62% 42 words at http://www.reachinformation.com/define/A
 * I  60% 41 words at http://www.reference.com/browse/wiki/se/
 * I  60% 41 words at http://www.buellersdownunder.com/archive
 * I  38% 26 words at http://lrd.yahooapis.com/_ylc=X3oDMTVnbm

API

You can query suspected diffs using the API available in: http://tools.wmflabs.org/eranbot/plagiabot/api.py

Examples:

• http://tools.wmflabs.org/eranbot/plagiabot/api.py?action=suspected_diffs
• http://tools.wmflabs.org/eranbot/plagiabot/api.py?action=suspected_diffs&format=csv
• http://tools.wmflabs.org/eranbot/plagiabot/api.py?action=suspected_diffs&page_title=Rajesh_Khanna&report=1

i18n

The bot supports English, French, Portuguese and Hebrew.

For running the bot on new languages:

1. Make sure ithenticate backend index pages in the desired language: http://www.ithenticate.com/products/faqs
2. Add relevant messages to help the bot skip reverted edits

UI

The bot can either generate simple wiki report pages, or write to a database to be used by other tools.

See also: https://github.com/wikimedia/CopyPatrol

Useful links:

• http://www.ithenticate.com/hs-fs/hub/92785/file-1383985272-pdf/iTh_documentation/iThenticate_API_Manual.pdf

plagiabot's Issues

suspected_diffs query should return more items

The current limit is 50. It would be much easier to consume if that number got increased to something that is well above the number of new issues that typically get added each day. That way, another system (such as dashboard.wikiedu.org, which has started consuming this data) could just do a single suspected_diffs query and get all the new diffs.

Can we try increasing it to like 500?

suspected_diffs action should provide report_id, matching urls, % match, and number of words

Right now, if I query for suspected diffs of an article (like this) I don't get enough information to investigate the suspected plagiarism.

I can only query suspected_diffs one article or revision ID at a time

It'd be nice if it could handle batches, for example, of 50 revision ids ... or better yet, article ids, at a time.

Error while handling API request, no other debugging info

URL: https://tools.wmflabs.org/eranbot/plagiabot/api.py?action=suspected_diffs&page_title=User:EarwigBot/Copyvios/Tests/2&lang=en&report=1

Output:

Error encountered while handling the request. Please report a bug: https://github.com/valhallasw/plagiabot/issues
[]

suspected_diffs API returning error

It looks like the suspected_diffs API has been broken for around the last month: https://tools.wmflabs.org/eranbot/plagiabot/api.py?action=suspected_diffs

February 27 is the last time dashboard.wikiedu.org registered any new suspected diffs from student editors, so I'm guessing it's been throwing that error since soon after that date.

SyntaxError: invalid syntax on Earwigs copyvio tool

On https://tools.wmflabs.org/copyvios/?lang=en&project=wikipedia&title=Kevin+Byard&oldid=&action=search&use_engine=0&use_links=0&turnitin=1 I get the following error

Traceback (most recent call last):
  File "/data/project/copyvios/www/python/src/app.py", line 38, in inner
    return func(*args, **kwargs)
  File "/data/project/copyvios/www/python/src/app.py", line 103, in index
    query = do_check()
  File "./copyvios/checker.py", line 38, in do_check
    _get_results(query, follow=not _coerce_bool(query.noredirect))
  File "./copyvios/checker.py", line 74, in _get_results
    query.turnitin_result = search_turnitin(page.title, query.lang)
  File "./copyvios/turnitin.py", line 22, in search_turnitin
    return TurnitinResult(_make_api_request(page_title, lang))
  File "./copyvios/turnitin.py", line 35, in _make_api_request
    parsed_api_result = literal_eval(result.text)
  File "/usr/lib/python2.7/ast.py", line 49, in literal_eval
    node_or_string = parse(node_or_string, mode='eval')
  File "/usr/lib/python2.7/ast.py", line 37, in parse
    return compile(source, filename, mode, PyCF_ONLY_AST)
  File "", line 1
    Error encountered while handling the request. Please report a bug: https://github.com/valhallasw/plagiabot/issues
                    ^
SyntaxError: invalid syntax

iThenticate login error does not include error details

I'm not sure whether the error printing code is intentionally commented out or not: https://github.com/valhallasw/plagiabot/blob/master/webservice/api.py#L75

But in any case, ;-( is not a useful error message.

api cannot handle unicode

For example: http://tools.wmflabs.org/eranbot/plagiabot/api.py?action=suspected_diffs&page_title=%27Ch%C3%A2teau%20Laroque%27

Transfer repository to separate github organisation

Would it make sense to move this repository under its own organisation, e.g. https://github.com/plagiabot/plagiabot? Apart from the initial implementation many years ago I haven't really been involved.

I'm not entirely sure what this would do in terms of redirect (e.g. if an existing git checkout will still be able to pull), but this should be relatively easy to fix.

Another option is moving everything to Wikimedia's new Gitlab instance -> https://gitlab.wikimedia.org

API does not handle errors

The API should handle common errors with an error code/message or at least a blank result, rather than serving error pages with a HTML stack trace, like this one: http://tools.wmflabs.org/eranbot/plagiabot/api.py?action=suspected_diffs&page_title=%D0%9D%D0%B8%D0%BA%D0%BE%D0%BB%D1%8C%D1%81%D0%BA%D0%B8%D0%B9%20%D0%9D%D0%B8%D0%BA%D0%BE%D0%BB%D0%B0%D0%B9%D0%98%D0%B2%D0%B0%D0%BD%D0%BE%D0%B2%D0%B8%D1%87&report=1