Assumptions

CART makes a few assumptions about the world.

Assumption #1: Abstracts exist

CART assumes both that a folder of "abstracts" exists, and that within this folder are csv files with the following headings: unique_id,review_count,url,title,abstract,user,vote,in_progress,time.

These columns adhere to the following properties:

column	object type	default value	comment
unique_id	int (in csv as string)	user supplied	keeps track of each paper
review_count	int (in csv as string)	0	count of number of votes cast on this paper
url	string	user supplied	hyperlink for paper
title	string	user supplied	paper title (HTML permissible)
abstract	string	user supplied	paper abstract (HTML permissible) required to end with "------endofabstract------xxxx" where the xxxx is the unique_id of the paper
user	string	none	team member acting on entry
vote	string	none	what team member's vote is
in_progress	string	no	flag for team member opening abstract
time	string	1707541131.0723252	helpful when viewing latest actions

For example, the paper titled 6583.csv would be a file that looks like this:

unique_id,review_count,url,title,abstract,user,vote,in_progress,time
6583,
0,
https://www.usenix.org/conference/usenixsecurity20/presentation/wei,
"What Twitter Knows...",
"Although targeted advertising...------endofabstract------6583",
none,
none,
no,
1691766446.157932

Getting Papers: A DEMO

If you would like to see how populating the abstracts folder could look, you have a few options. The first option is copy some of the example files into the abstracts folder.

The second option is to run the demo scraper get_data.py located in the sample_get_data folder. If you run this file, you will be scraping paper titles and abstracts from the Privacy Enhancing Technologies Symposium for the year 2023. This script will create a new folder /abstracts/-sample_from_scrape/SAMPLE_PETS_at_<TIME GOES HERE>/'which will house the .csv files (numbered) that CART expects to see. You can then drag and drop these files to the abstracts/ location and try them out. Researchers may choose to expand this scraping script to include more years and conferences when creating an initial batch of papers to review.

Assumption #2: Member-Driven-Server-Is-OK

The second assumption made by CART is that one team member keeps cart.py running on their local machine for as long as the culling process takes. Although it is true that a viable alternative would be to point a server at CART, this depends on resources that are not typically freely accessible. Therefore, CART makes every effort possible to provide teams with a "zero-step" culling process. One team member initiates CART with the free ngrok service, allowing everyone to view and grade papers at their leisure.

Assumption #3: Users are friendly

CART assumes (to use security terminology) that users are friendly, but curious. There is no strict login process because it is assumed that users are not going to be logging in with other usernames and making fraudulent votes. This is also helpful when an administrator needs to review votes cast by other members or debug (logging in as the other user helps reproduce the issue).