tommasoc80/eventstoryline

Event StoryLine Corpus - annotated data, baselines and evaluation scripts, evaluation data.

eventstoryline's Issues

the parameter settings of the main function

I'm sorry to disturb you. Could you tell me the parameter settings of the main function. I / O files, etc

Causal and Corefer relationship appear at the same time

I found some trigger pairs have causal and corefer relationship at the same time，how to understand this phenomenon?
e.g.
evaluation_format\full_corpus\v1.0\coref_chain\3\3_6ecbplus.tab
"56 38_39_40 90 175_176 95_96 56 38_39_40 90 175_176 95_96 PRECONDITION"

Missing plot_link relation relType

There are some plot_links which were not fully annotated.

See the list below:
13_4ecbplus.xml.xml r_id= 99542
19_9ecbplus.xml.xml r_id= 96260
3_5ecbplus.xml.xml r_id= 96640
3_6ecbplus.xml.xml r_id= 96651
30_12ecbplus.xml.xml r_id= 213574
32_3ecbplus.xml.xml r_id= 208407
33_7ecbplus.xml.xml r_id= 213471
35_4ecbplus.xml.xml r_id= 215469
37_11ecbplus.xml.xml r_id= 90985
37_13ecbplus.xml.xml r_id= 91002
37_14ecbplus.xml.xml r_id= 91013
37_1ecbplus.xml.xml r_id= 91031
37_2ecbplus.xml.xml r_id= 91056
41_6ecbplus.xml.xml r_id= 104582

NAF directory

Hi there,

Thanks for sharing this amazing resource.

I was wondering what is the NAF file used in baseline_PPMI*.py. I can't find it in this repository.

Best,
I-Ta

Empty folder

I found this folder(evaluation_format/full_corpus/v1.0/event_mentions_extended/5) is empty, is there any problem?

How to use CAT tool?

The CAT tool has not been maintained and I can't create an account.

Is there any other methods to visualize this dataset?

Potential conflict when expanding causal relation with coreference chain

EventStoryLine/create_gold_document.py

Line 141 in f3c9aad

mention_elem[k[1]] = v

Hi, I found the processing script here does not consider conflicted relation inference. It is possible for one pair of mentions to have "PRECONDITION" and "FALLING_ACTION" relations inferred at the same time. The conflicted pairs would better be deleted.

How get CAT？

According to https://dh.fbk.eu/resources/cat-content-annotation-tool, I send a e-mail to: [email protected], but no reply.

Which tokenizer do you use?

Or you just use the split() function?

.

The number of PLOT LINK relation does not correspond to the number that paper illustrated

In ./evaluation_format/full_corpus/v0.9/event_mentions_extended/, there exists 5,602 PLOT LINK relations, which should be 5,519 that illustrated in The Event StoryLine Corpus: A New Benchmark for Causal and Temporal Relation Extraction.

In ./evaluation_format/full_corpus/v0.9/coref_chain/, there exists 2,243 PLOT LINK relations, which should be 2,265 illustrated in the same paper.

Event definition missing in v1.0

In EventStoryLine/annotated_data/v1.0/1/1_11ecbplus.xml.xml,

token t_id="36" sentence="1" number="4">rehab

should be an event.

However, it is missing. In v1.0/1/1_11ecbplus.xml.xml, t_id=36 is not labeled as an event.

But in v0.9, there is:
<ACTION_OCCURRENCE m_id="35" climaxEvent="TRUE" comment="">
<token_anchor t_id="36"/>