richardbrinkman/plagiarism

Plagiarism detection tool for Surpass, Testvision and Blackboard tests

plagiarism's Issues

exclude invalid rows from csv

Rows with Grade == "Ongeldig" should be excluded.

webapp still uses old plagiarism.py

The error was introduced after the refactoring to support TestVision.

Use multiprocessing pool

Calculation takes a lot of time because currently only a single CPU is used. It's therefore preferable to use a multiprocessing pool

Heroku times out

Use a periodic keep alive message to prevent Heroku to timeout.
You can send each 30 seconds a comment that starts with a colon.

Add web interface

Add a web interface. Because generating the Excel takes a lot of time, some kind of progress should be shown on the website. server sent events can be used for that.

Blackboard support

Blackboard does have it's own plagiarism detection. However, it supports only Microsoft Office documents and pdf.
It would be nice to have a universal plagiarism detection tool that can detect plagiarism for any filetype that can be converted to text.

Testvision question of type "MeervoudigInvul"

In the Excel file multiple rows correspond to the same question (of type MeervoudigInvul). This results in the following error:

pandas.errors.InvalidIndexError: Reindexing only valid with uniquely valued Index objects

extend web interface with Blackboard plagiarism detection

depends on issue #9

empty file is downloaded

Sometimes an empty file is downloaded. Although the correct plagiarism.xlsx file is placed on the server, if downloaded with http://localhost:8080/static/md5/plagiarism.xlsx it is truncated to 0 bytes.

Add plagiarism checks for Testvision

duplicate worksheet names

Catch xlsxwriter.exceptions.DuplicateWorksheetName exception and rename worksheet using counter.

add tab with average of all other tabs

add tab with student names and numbers