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 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):
user_submission_file = args.submission
gold_folder = args.gold
output_log_file = args.log_file
prop_vs_non_propaganda = bool(args.fragments_only)
merge_user_annotations = bool(args.merge_user_annotations)
per_article_evaluation = bool(args.per_article_evaluation)
if args.debug_on_std:
ch.setLevel(logging.DEBUG)
if output_log_file is not None:
logger.info("Logging execution to file " + output_log_file)
fileLogger = logging.FileHandler(output_log_file)
fileLogger.setLevel(logging.DEBUG)
fileLogger.setFormatter(formatter)
logger.addHandler(fileLogger)
techniques = pt.Propaganda_Techniques()
#techniques_names = pt.load_technique_names_from_file(args.techniques_file) # load technique names
submission_annotations = an.Annotations()
submission_annotations.load_annotation_list_from_file(user_submission_file)
if gold_folder is None:
# no gold file provided, perform only some checks on the submission files
logger.info('Checking format of user submitted file %s' % (user_submission_file))
for article_id, annotations in submission_annotations.items():
check_article_annotations_format(annotations, article_id, techniques_names)
logger.warning("The format of the submitted file is ok. However, more checks, requiring the gold file, are needed "
"for the submission to be correct: the number of article and their ids must correspond to the "
"ones of the gold file, etc")
else:
logger.info('Checking user submitted file %s against gold folder %s' % (user_submission_file, gold_folder))
gold_annotations = an.Annotations()
gold_annotations.load_annotation_list_from_folder(gold_folder)
#gold_annotations = load_annotation_list_from_folder(gold_folder, techniques_names)
check_data_file_lists(submission_annotations, gold_annotations)
if prop_vs_non_propaganda:
if not check_annotation_spans(submission_annotations, merge_user_annotations):
logger.info("Error in file %s" %(user_submission_file))
sys.exit()
check_annotation_spans(gold_annotations, True)
else:
if not check_annotation_spans_with_category_matching(submission_annotations, merge_user_annotations):
logger.info("Error in file %s" % (user_submission_file))
sys.exit()
check_annotation_spans_with_category_matching(gold_annotations, True)
logger.info('Scoring user submitted file %s against gold files %s/*.labels' % (user_submission_file, gold_folder))
#logger.info("Scoring the submission with max")
#score_max = compute_score_max(submission_annotations, gold_annotations, techniques_names, prop_vs_non_propaganda)
#logger.info("Scoring the submission with min")
#score_min = compute_score_min(submission_annotations, gold_annotations, techniques_names, prop_vs_non_propaganda)
logger.info("Scoring the submission with precision and recall")
score_pr = compute_score_pr(submission_annotations, gold_annotations, techniques_names,
prop_vs_non_propaganda, per_article_evaluation)
return score_pr
def main(args):
user_submission_file = args.submission
gold_file = args.gold
output_log_file = args.log_file
propaganda_techniques_list_file = args.propaganda_techniques_list_file
output_for_script = bool(args.output_for_script)
if not output_for_script:
logger.addHandler(ch)
if args.debug_on_std:
ch.setLevel(logging.DEBUG)
if output_log_file is not None:
logger.info("Logging execution to file " + output_log_file)
fileLogger = logging.FileHandler(output_log_file)
fileLogger.setLevel(logging.DEBUG)
fileLogger.setFormatter(formatter)
logger.addHandler(fileLogger)
propaganda_techniques = pt.Propaganda_Techniques(
propaganda_techniques_list_file)
an.Annotation.set_propaganda_technique_list_obj(propaganda_techniques)
user_annotations = ans.Annotations()
user_annotations.load_annotation_list_from_file(user_submission_file)
for article in user_annotations.get_article_id_list():
user_annotations.get_article_annotations_obj(article).sort_spans()
gold_annotations = ans.Annotations()
gold_annotations.load_annotation_list_from_file(gold_file)
for article in gold_annotations.get_article_id_list():
gold_annotations.get_article_annotations_obj(article).sort_spans()
logger.info("Checking format: User Predictions -- Gold Annotations")
if not user_annotations.compare_annotations_identical_article_lists(
gold_annotations
) or not user_annotations.compare_annotations_identical(gold_annotations):
logger.error("wrong format, no scoring will be performed")
sys.exit()
logger.info("OK: submission file format appears to be correct")
res_for_output, res_for_script = user_annotations.TC_score_to_string(
gold_annotations, output_for_script)
logger.info("Scoring submission" + res_for_output)
if output_for_script:
print(res_for_script)
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
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)
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)