def main(args):
span_file = args.spans_file
article_file = args.article_file
print_line_numbers = bool(args.add_line_numbers)
fix_from_char_index = int(args.fix_from_char_index)
offset = int(args.offset)
propaganda_techniques_file = args.propaganda_techniques_file
annotations = aa.Articles_annotations()
an.Annotation.set_propaganda_technique_list_obj(pt.Propaganda_Techniques(filename=propaganda_techniques_file))
annotations.load_article_annotations_from_csv_file(span_file)
annotations.shift_spans(fix_from_char_index, offset)
with codecs.open(article_file, "r", encoding="utf8") as f:
article_content = f.read()
output_text, footnotes, legend = annotations.mark_text(article_content, print_line_numbers)
print(output_text)
print(legend)
print(footnotes)
if offset != 0:
annotations.save_annotations_to_file(span_file + ".fix")
print("Fixed annotations saved to file %s.fix"%(span_file))
def main(args):
span_file = args.spans_file
article_file = args.article_file
propaganda_techniques_list_file = args.propaganda_techniques_list_file
debug_on_std = bool(args.debug_on_std)
if not debug_on_std:
logging.getLogger("propaganda_scorer").setLevel(logging.ERROR)
propaganda_techniques = pt.Propaganda_Techniques(
propaganda_techniques_list_file)
annotations = aa.Articles_annotations()
aa.Articles_annotations.techniques = propaganda_techniques
annotations.load_article_annotations_from_csv_file(span_file)
with codecs.open(article_file, "r", encoding="utf8") as f:
article_content = f.read()
#print("\n".join([str(i)+") "+x for i,x in enumerate(str(aa.techniques).split("\n"))]))
output_text, footnotes = annotations.tag_text_with_annotations(
article_content) #add html tags
#output_text, footnotes, legend = annotations.mark_text(article_content) #mark annotations for terminal
print(output_text)
print(footnotes)
def add_annotation(self, annotation: an.Annotation, article_id: str):
"""
Add a single annotation to the article with id article_id.
If such article does not exists, the annotation is created.
"""
if not self.has_article(article_id):
self.annotations[article_id] = aa.Articles_annotations(
article_id=article_id)
self.annotations[article_id].add_annotation(annotation)
def main(args):
span_file = args.spans_file
write_technique_on_output = bool(args.write_technique_on_output)
prop_vs_non_propaganda = bool(args.fragments_only)
output_file = args.output_file
if not output_file:
output_file = span_file + ".merged"
article_annotations = an.Articles_annotations()
article_annotations.load_article_annotations_from_csv_file(span_file)
article_annotations.has_overlapping_spans(prop_vs_non_propaganda, True)
article_annotations.set_output_format(True, True,
write_technique_on_output)
article_annotations.save_annotations_to_file(output_file)
def main(args):
span_file = args.spans_file
article_file = args.article_file
print_line_numbers = bool(args.add_line_numbers)
an.techniques = pt.Propaganda_Techniques(filename="data/propaganda-techniques-names-semeval2020task11.txt")
annotations = an.Articles_annotations()
annotations.load_article_annotations_from_csv_file(span_file)
with codecs.open(article_file, "r", encoding="utf8") as f:
article_content = f.read()
output_text, footnotes, legend = annotations.mark_text(article_content, print_line_numbers)
print(output_text)
print(legend)
print(footnotes)
def main(args):
span_file = args.spans_file
article_file = args.article_file
print_line_numbers = bool(args.add_line_numbers)
annotations = an.Articles_annotations()
annotations.load_article_annotations_from_csv_file(span_file)
with codecs.open(article_file, "r", encoding="utf8") as f:
article_content = f.read()
output_text, footnotes, legend = annotations.mark_text(
article_content, print_line_numbers)
print(output_text)
print(legend)
print(footnotes)
def main(args):
span_file = args.spans_file
article_file = args.article_file
propaganda_techniques_list_file = args.propaganda_techniques_list_file
propaganda_techniques = pt.Propaganda_Techniques(
propaganda_techniques_list_file)
annotations = aa.Articles_annotations()
aa.Articles_annotations.techniques = propaganda_techniques
annotations.load_article_annotations_from_csv_file(span_file)
with codecs.open(article_file, "r", encoding="utf8") as f:
article_content = f.read()
# print("\n".join([str(i)+") "+x for i,x in enumerate(str(aa.techniques).split("\n"))]))
# output_text, footnotes = annotations.tag_text_with_annotations(article_content)
output_text, footnotes, legend = annotations.mark_text(article_content)
print(output_text)
print(footnotes)
def create_article_annotations_object(self, article_id: str) -> None:
self.annotations[article_id] = aa.Articles_annotations(
article_id=article_id)
def add_annotation(self, annotation, article_id):
if not self.has_article(article_id):
self.annotations[article_id] = an.Articles_annotations(
article_id=article_id)
self.annotations[article_id].add_annotation(annotation)
def main(args):
span_file = args.spans_file
article_file = args.article_file
original_text_file = args.original_text_file
gold_spans = args.gold_spans
debug = args.debug
if args.output_file_name is None:
output_file_name = span_file + ".aligned.txt"
else:
output_file_name = args.output_file_name
annotations = aa.Articles_annotations()
annotations.load_article_annotations_from_csv_file(
span_file, an.AnnotationWithOutLabel)
annotations.sort_spans()
annotations.has_overlapping_spans(True, True)
with codecs.open(article_file, "r", encoding="utf8") as f:
article_content = f.read()
with codecs.open(original_text_file, "r", encoding="utf8") as f:
original_text = f.read()
if debug == True:
gold_annotations = aa.Articles_annotations()
gold_annotations.load_article_annotations_from_csv_file(
gold_spans, an.AnnotationWithOutLabel)
gold_annotations.sort_spans()
gold_annotations_content = gold_annotations.get_spans_content(
original_text)
print(gold_annotations)
print("gold annotations content (no spaces):\n%s\n---" %
(gold_annotations_content))
before = annotations.get_spans_content(article_content)
print(annotations)
print("content of annotations of converted text:\n%s\n---" % (before))
original_annotations = copy.deepcopy(annotations)
annotations.align_annotation_to_new_text(original_text, article_content)
after = annotations.get_spans_content(original_text)
annotations.set_output_format(True, True, False)
annotations.save_annotations_to_file(output_file_name)
if gold_spans is not None:
gold_annotations = aa.Articles_annotations()
gold_annotations.load_article_annotations_from_csv_file(
gold_spans, an.AnnotationWithOutLabel)
gold_annotations.sort_spans()
gold_annotations.has_overlapping_spans(True, True)
annotations.sort_spans()
if not gold_annotations == annotations:
print("%s: The following annotations are different: " %
(article_file),
end=""),
for p in annotations - gold_annotations:
print("1: %s; 2: %s -- %s -- %s" %
(p[0], p[1], p[0].get_span_content(article_content),
p[1].get_span_content(original_text))),
print(
"%s: The following annotations are problematic:\n(USER ANNOTATION) %s\n(GOLD ANNOTATION) %s\n"
% (article_file, annotations, gold_annotations))
gold_annotations_content = gold_annotations.get_spans_content(
original_text)
print("gold annotations content (no spaces):\n%s\n---" %
(gold_annotations_content))
before = original_annotations.get_spans_content(article_content)
print(
"content of annotations of converted text before alignment:\n%s\n---"
% (before))
else:
print("OK: %s -> %s" % (article_file, output_file_name))
import sys
sys.path.append("../")
import src.propaganda_techniques as pt
import src.annotation as an
import src.article_annotations as aa
def test_remove_annotation(artannotations):
before = str(artannotations)
print("removing annotation: " + str(artannotations[0]))
artannotations.remove_annotation(artannotations[0])
after = str(artannotations)
assert after == before.replace("\n\t[0, 59] -> Exaggeration,Minimisation",
"", 1)
if __name__ == "__main__":
propaganda_techniques = pt.Propaganda_Techniques(
filename="../data/propaganda-techniques-names.txt")
an.Annotation.set_propaganda_technique_list_obj(propaganda_techniques)
artannotations = aa.Articles_annotations()
artannotations.load_article_annotations_from_csv_file(
"../data/article736757214.task-FLC.labels")
test_remove_annotation(artannotations)
