You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
cjvt-valency/src/pkg/corpusparser/corpusparser/main.py

59 lines
1.8 KiB

from corpusparser import Parser
import argparse
import logging
import json
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
## Main handles command line arguments and writing to files / DB.
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Parsing corpora kres and ssj500k.")
parser.add_argument('--kres-folder', required=True)
parser.add_argument('--kres-srl-folder', required=True)
parser.add_argument('--ssj-file', required=True)
parser.add_argument('--output', required=False, default=None)
parser.add_argument('--outdir', required=False, default=None)
parser.add_argument('--dbaddr', required=False, default=None)
args = parser.parse_args()
# parse ssj
logger.info("Parsing ssj500k: {}".format(args.ssj_file))
ssj_parser = Parser(
corpus="ssj",
infiles=[args.ssj_file],
output=args.output,
outdir=args.outdir,
)
res = [x[1]["sid"] for x in ssj_parser.sentence_generator()]
logger.info("Parsed {} sentences (ssj500k)".format(len(res)))
# parse kres
logger.info("Parsing Kres: {}".format(args.ssj_file))
kres_parser = Parser(
corpus="kres",
infiles=[args.kres_folder, args.kres_srl_folder],
output=args.output,
outdir=args.outdir,
)
res = [x[1]["sid"] for x in kres_parser.sentence_generator()]
logger.info("Parsed {} sentences (kres)".format(len(res)))
## Handling output is situational --- implement it outside of Parser.
## Parser returns tuples (orig_file, element)
# 1. parse per-file and output to file (JSON)
# 2. parse and save to DB
# TODO
def handle_output(self, sent_ent, xml_file):
if self.output is None:
pass
if self.output == "file":
outfile = Path(self.outdir) / Path(xml_file.name.split(".")[0]).with_suffix(".json")
with outfile.open("a") as fp:
print(sent_ent)
json.dumps(sent_ent, fp)