forked from kristjan/cjvt-srl-tagging
62 lines
1.5 KiB
Python
62 lines
1.5 KiB
Python
from parser.parser import Parser
|
|
import os
|
|
from os.path import join, dirname
|
|
from pathlib import Path
|
|
import re
|
|
import sys
|
|
import cProfile
|
|
import configparser
|
|
import logging
|
|
from multiprocessing import Pool
|
|
|
|
SSJ500K_2_1 = 27829 # number of sentences
|
|
par = Parser()
|
|
|
|
# path to data
|
|
config = configparser.ConfigParser()
|
|
config.read("tools.cfg")
|
|
INDIR = Path(config["tools"]["kres_orig"])
|
|
OUTDIR = Path(config["tools"]["kres_tsv"])
|
|
|
|
LOGFILE = Path(config["tools"]["logfile"]).absolute()
|
|
LOGFILE.touch(exist_ok=True)
|
|
LOGFILE.resolve()
|
|
|
|
logging.basicConfig(filename=str(LOGFILE), level=logging.INFO)
|
|
|
|
"""
|
|
print("parsing ssj")
|
|
ssj_file = "../data/ssj500k-sl.sample.xml"
|
|
ssj_dict = par.parse_tei(ssj_file)
|
|
# assert (len(ssj_dict) == 27829), "Parsed wrong number of sentences."
|
|
print("end parsing ssj")
|
|
"""
|
|
|
|
logging.info("parsing kres")
|
|
# kres_file = "../data/kres_example/F0019343.xml.parsed.xml"
|
|
OUTDIR.mkdir(exist_ok=True)
|
|
|
|
infiles = list(enumerate([x for x in INDIR.iterdir() if x.is_file()]))
|
|
|
|
def handle_file(infile):
|
|
i = infile[0]
|
|
kres_file = infile[1]
|
|
|
|
logging.info("Processing file ({}/{}): {}".format(i+1, len(infiles), str(kres_file)))
|
|
res_dict = par.parse_tei(kres_file)
|
|
kres_out_str = ""
|
|
|
|
for _, sentence in res_dict.items():
|
|
kres_out_str += par.to_conll_2009_SRL(sentence)
|
|
|
|
with (OUTDIR / kres_file.name).with_suffix(".tsv").open("wb+") as fp:
|
|
fp.write(kres_out_str.encode("utf-8"))
|
|
|
|
with Pool(3) as p:
|
|
p.map(handle_file, infiles)
|
|
|
|
|
|
logging.info("end parsing kres")
|
|
|
|
|