def parse(options, filename, annotated_sentences, tmp_dir):
tagged_conll_filename = output_filename(tmp_dir, filename, "tag.conll")
parsed_filename = output_filename(tmp_dir, filename, "conll")
log_filename = output_filename(tmp_dir, filename, "log")
# The parser command line is dependent on the input and
# output files, so we build that one for each data file
parser_cmdline = ["java", "-Xmx2000m",
"-jar", os.path.expanduser(options.malt),
"-m", "parse",
"-i", tagged_conll_filename,
"-o", parsed_filename,
"-w", tmp_dir,
"-c", os.path.basename(options.parsing_model)]
# Conversion from .tag file to tagged.conll (input format for the parser)
tagged_conll_file = open(tagged_conll_filename, "w", encoding="utf-8")
tagged_to_tagged_conll(annotated_sentences, tagged_conll_file)
tagged_conll_file.close()
# Run the parser
with open(log_filename, "w", encoding="utf-8") as log_file:
returncode = Popen(parser_cmdline, stdout=log_file, stderr=log_file).wait()
if returncode:
sys.exit("Parsing failed! Log file may contain more information: %s" % log_filename)
return parsed_filename
def test_verb(self):
annotated_sentences = [[
('Jag jag PRON|Case=Nom|Definite=Def|Gender=Com|Number=Sing PN|UTR|SIN|DEF|SUB'
).split(" "),
('har ha VERB|Mood=Ind|Tense=Pres|VerbForm=Fin|Voice=Act VB|PRS|AKT'
).split(" "),
('en en DET|Definite=Ind|Gender=Com|Number=Sing DT|UTR|SIN|IND'
).split(" "),
('dröm dröm NOUN|Case=Nom|Definite=Ind|Gender=Com|Number=Sing NN|UTR|SIN|IND|NOM'
).split(" "),
('. . PUNCT|_ MAD').split(" "),
]]
file_data = ""
with tempfile.TemporaryFile(mode="w+") as outfile:
conll.tagged_to_tagged_conll(annotated_sentences, outfile)
outfile.seek(0)
file_data = outfile.read()
self.assertEqual(
file_data.splitlines(),
textwrap.dedent("""
0 Jag jag PRON PN|UTR|SIN|DEF|SUB Case=Nom|Definite=Def|Gender=Com|Number=Sing
1 har ha VERB VB|PRS|AKT Mood=Ind|Tense=Pres|VerbForm=Fin|Voice=Act
2 en en DET DT|UTR|SIN|IND Definite=Ind|Gender=Com|Number=Sing
3 dröm dröm NOUN NN|UTR|SIN|IND|NOM Case=Nom|Definite=Ind|Gender=Com|Number=Sing
4 . . PUNCT MAD _
""").lstrip().splitlines())
def parse(options, filename, annotated_sentences, tmp_dir):
tagged_conll_filename = output_filename(tmp_dir, filename, "tag.conll")
parsed_filename = output_filename(tmp_dir, filename, "conll")
log_filename = output_filename(tmp_dir, filename, "log")
# The parser command line is dependent on the input and
# output files, so we build that one for each data file
parser_cmdline = [
"java", "-Xmx2000m", "-jar",
os.path.expanduser(options.malt), "-m", "parse", "-i",
tagged_conll_filename, "-o", parsed_filename, "-w", tmp_dir, "-c",
os.path.basename(options.parsing_model)
]
# Conversion from .tag file to tagged.conll (input format for the parser)
tagged_conll_file = open(tagged_conll_filename, "w", encoding="utf-8")
tagged_to_tagged_conll(annotated_sentences, tagged_conll_file)
tagged_conll_file.close()
# Run the parser
with open(log_filename, "w", encoding="utf-8") as log_file:
returncode = Popen(parser_cmdline, stdout=log_file,
stderr=log_file).wait()
if returncode:
sys.exit("Parsing failed! See log file: %s" % log_filename)
return parsed_filename
def test_verb(self):
annotated_sentences = [[
('Jag jag PRON|Case=Nom|Definite=Def|Gender=Com|Number=Sing PN|UTR|SIN|DEF|SUB').split(" "),
('har ha VERB|Mood=Ind|Tense=Pres|VerbForm=Fin|Voice=Act VB|PRS|AKT').split(" "),
('en en DET|Definite=Ind|Gender=Com|Number=Sing DT|UTR|SIN|IND').split(" "),
('dröm dröm NOUN|Case=Nom|Definite=Ind|Gender=Com|Number=Sing NN|UTR|SIN|IND|NOM').split(" "),
('. . PUNCT|_ MAD').split(" "),
]]
file_data = ""
with tempfile.TemporaryFile(mode="w+") as outfile:
conll.tagged_to_tagged_conll(annotated_sentences, outfile)
outfile.seek(0)
file_data = outfile.read()
self.assertEqual(file_data.splitlines(), textwrap.dedent("""
0 Jag jag PRON PN|UTR|SIN|DEF|SUB Case=Nom|Definite=Def|Gender=Com|Number=Sing
1 har ha VERB VB|PRS|AKT Mood=Ind|Tense=Pres|VerbForm=Fin|Voice=Act
2 en en DET DT|UTR|SIN|IND Definite=Ind|Gender=Com|Number=Sing
3 dröm dröm NOUN NN|UTR|SIN|IND|NOM Case=Nom|Definite=Ind|Gender=Com|Number=Sing
4 . . PUNCT MAD _
""").lstrip().splitlines())
