Hellisotherpeople commited on
Commit
23cd046
1 Parent(s): a544bb7

Mirror from github

Browse files
Files changed (1) hide show
  1. README.md +58 -0
README.md ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # DebateSum
2
+ Corresponding code repo for the upcoming paper at ARGMIN 2020: "DebateSum: A large-scale argument mining and summarization dataset"
3
+
4
+ Arxiv pre-print available here: https://arxiv.org/abs/2011.07251
5
+
6
+ Check out the presentation date and time here: https://argmining2020.i3s.unice.fr/node/9
7
+
8
+ Full paper as presented by the ACL is here: https://www.aclweb.org/anthology/2020.argmining-1.1/
9
+
10
+ Video of presentation at COLING 2020: https://underline.io/lecture/6461-debatesum-a-large-scale-argument-mining-and-summarization-dataset
11
+
12
+ The dataset is distributed as csv files.
13
+
14
+ A search engine over DebateSum (as well as some additional evidence not included in DebateSum) is available as [debate.cards](http://debate.cards/). It's very good quality and allows for the evidence to be viewed in the format that debaters use.
15
+
16
+ # Data
17
+
18
+ DebateSum consists of **187328** debate documents, arguements (also can be thought of as abstractive summaries, or queries), word-level extractive summaries, citations, and associated metadata organized by topic-year. This data is ready for analysis by NLP systems.
19
+
20
+ ## Download
21
+ All data is accesable in a parsed format organized by topic year [here](https://mega.nz/folder/ZdQGmK6b#-0hoBWc5fLYuxQuH25feXg)
22
+
23
+ Addtionally, the trained word-vectors for [debate2vec](https://github.com/Hellisotherpeople/debate2vec) are also found in that folder.
24
+
25
+ ## Regenerating it yourself
26
+
27
+ This is useful as the debaters who produce the evidence release their work every year. Soon enough I will update to include the 2020-2021 topic.
28
+
29
+ *Step 1: Download all open evidence files from [Open Evidence](https://openev.debatecoaches.org/) and unzip them into a directory. The links are as follows:*
30
+
31
+ * [2019](https://s3.amazonaws.com/openev/2019OpenEv.zip) - Resolved: The United States federal government should substantially reduce Direct Commercial Sales and/or Foreign Military Sales of arms from the United States.
32
+ * [2018](https://s3.amazonaws.com/openev/2018OpenEv.zip) - Resolved: The United States federal government should substantially reduce its restrictions on legal immigration to the United States.
33
+ * [2017](https://s3.amazonaws.com/openev/2017OpenEv.zip) - Resolved: The United States federal government should substantially increase its funding and/or regulation of elementary and/or secondary education in the United States.
34
+ * [2016](https://s3.amazonaws.com/openev/2016OpenEv.zip) - Resolved: The United States federal government should substantially increase its economic and/or diplomatic engagement with the People’s Republic of China.
35
+ * [2015](https://s3.amazonaws.com/openev/2015OpenEv.zip) - Resolved: The United States federal government should substantially curtail its domestic surveil-lance.
36
+ * [2014](https://s3.amazonaws.com/openev/2014OpenEv.zip) - Resolved: The United States federal government should substantially increase its non-military exploration and/or development of the Earth’s oceans.
37
+ * [2013](https://s3.amazonaws.com/openev/2013OpenEv.zip) - Resolved: The United States federal government should substantially increase its economic en-gagement toward Cuba, Mexico or Venezuela.
38
+
39
+
40
+ *Step 2: Convert all evidence from docx files to html5 files using [pandoc](https://pandoc.org/) with this command:*
41
+ ```
42
+ for f in *.docx; do pandoc "$f" -s -o "${f%.docx}.html5"; done
43
+ ```
44
+
45
+ *Step 3: install the dependencies for make_debate_dataset.py.*
46
+
47
+ ```
48
+ pip install -r requirements.txt
49
+ ```
50
+
51
+ *Step 4: Modify the folder and file locations as needed for your system, and run make_debate_dataset.py*
52
+
53
+ ```
54
+ python3 make_debate_dataset.py
55
+ ```
56
+
57
+ # Credits
58
+ Huge thanks to [Arvind Balaji](https://github.com/arvind-balaji) for making debate.cards and being second author on this paper!