def get_parses(self, sentences):
""" Gets a dependency semantic parse as conll from a sentence str. """
# TODO: server version of Malt?
with deleting(mkdtemp(suffix='XXXXXX', prefix='semafor.')) as temp_dir:
input_filename = os.path.join(temp_dir, "sentence")
output_filename = os.path.join(temp_dir, "conll")
with codecs.open(input_filename, 'w', encoding="utf8") as input_file:
input_file.write(u'\n'.join(sentences))
os.system("cd %s && ./bin/runMalt.sh %s %s" %
(SEMAFOR_HOME, input_filename, temp_dir))
with codecs.open(output_filename, encoding="utf8") as output_file:
output = output_file.read()
return output
def get_parses(self, sentences):
""" Gets a dependency semantic parse as conll from a sentence str. """
# TODO: server version of Malt?
with deleting(mkdtemp(suffix='XXXXXX', prefix='semafor.')) as temp_dir:
input_filename = os.path.join(temp_dir, "sentence")
output_filename = os.path.join(temp_dir, "conll")
with codecs.open(input_filename, 'w', encoding="utf8") as input_file:
input_file.write(u'\n'.join(sentences))
os.system("cd %s && ./bin/runMalt.sh %s %s" %
(SEMAFOR_HOME, input_filename, temp_dir))
with codecs.open(output_filename, encoding="utf8") as output_file:
output = output_file.read()
return output
def tag_sentences(self, sentences):
""" Runs tokenization and part-of-speech tagging a sentence str. """
#TODO: could probably just use nltk
with deleting(mkdtemp(suffix='XXXXXX', prefix='semafor.')) as temp_dir:
input_filename = os.path.join(temp_dir, "sentence")
output_filename = os.path.join(temp_dir, "pos.tagged")
with codecs.open(input_filename, 'w', encoding="utf8") as input_file:
input_file.write(u'\n'.join(sentences))
os.system("cd %s && ./tokenize_and_postag.sh %s %s" %
(TAGGING_SCRIPT_HOME, input_filename, output_filename))
with codecs.open(output_filename, encoding="utf8") as output_file:
output = output_file.read()
return output
def tag_sentences(self, sentences):
""" Runs tokenization and part-of-speech tagging a sentence str. """
#TODO: could probably just use nltk
with deleting(mkdtemp(suffix='XXXXXX', prefix='semafor.')) as temp_dir:
input_filename = os.path.join(temp_dir, "sentence")
output_filename = os.path.join(temp_dir, "pos.tagged")
with codecs.open(input_filename, 'w', encoding="utf8") as input_file:
input_file.write(u'\n'.join(sentences))
os.system("cd %s && ./tokenize_and_postag.sh %s %s" %
(TAGGING_SCRIPT_HOME, input_filename, output_filename))
with codecs.open(output_filename, encoding="utf8") as output_file:
output = output_file.read()
return output
