From e97c777b789d11046143ee156a70b8d9481b2246 Mon Sep 17 00:00:00 2001
From: vvye <ekaiser.hellwege@gmail.com>
Date: Fri, 24 Sep 2021 14:31:38 +0200
Subject: [PATCH] Remove debug logic
---
dataset.py | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/dataset.py b/dataset.py
index d53ca44..a61d63d 100644
--- a/dataset.py
+++ b/dataset.py
@@ -128,8 +128,6 @@ def get_crisis_dataset():
date_path = topic_path / 'public' / 'content' / pub_date
for article_filename in util.files(date_path, extension='.cont'):
article_file_path = date_path / article_filename
- if '1093.htm.cont' not in str(article_file_path):
- continue
print(article_file_path)
article = {'pub_date': pub_date, 'sentences': []}
@@ -144,8 +142,8 @@ def get_crisis_dataset():
# get date mentions using HeidelTime
# and add them to the sentence data
mentioned_dates_by_sentence = heideltime_util.mentioned_dates_by_sentence(article_file_path, pub_date)
- mentioned_dates_by_sentence = mentioned_dates_by_sentence[1:]
- assert len(mentioned_dates_by_sentence) == len(sentences_in_article) # skip first line (headline)
+ mentioned_dates_by_sentence = mentioned_dates_by_sentence[1:] # skip first line (headline)
+ assert len(mentioned_dates_by_sentence) == len(sentences_in_article)
for i in range(len(sentences_in_article)):
sentence = sentences_in_article[i]
sentence['mentioned_dates'] = mentioned_dates_by_sentence[i]
--
GitLab