|
|
@ -4,20 +4,44 @@ import argparse
|
|
|
|
import logging
|
|
|
|
import logging
|
|
|
|
import json
|
|
|
|
import json
|
|
|
|
from pymongo import MongoClient
|
|
|
|
from pymongo import MongoClient
|
|
|
|
|
|
|
|
import pymongo
|
|
|
|
import sys
|
|
|
|
import sys
|
|
|
|
from multiprocessing import Pool
|
|
|
|
from multiprocessing import Pool
|
|
|
|
|
|
|
|
import time
|
|
|
|
|
|
|
|
|
|
|
|
logging.basicConfig(level=logging.INFO)
|
|
|
|
logging.basicConfig(level=logging.INFO)
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
|
|
|
|
|
|
n_kres_files = -1 # for logging
|
|
|
|
n_kres_files = -1 # for logging
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def _helper_tid_to_token(tid, tokens):
|
|
|
|
|
|
|
|
for t in tokens:
|
|
|
|
|
|
|
|
if t["tid"] == tid:
|
|
|
|
|
|
|
|
return t
|
|
|
|
|
|
|
|
return None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def _db_preprocess(e):
|
|
|
|
|
|
|
|
if e["srl_links"] is None:
|
|
|
|
|
|
|
|
e["headwords"] = []
|
|
|
|
|
|
|
|
e["functors"] = []
|
|
|
|
|
|
|
|
else:
|
|
|
|
|
|
|
|
hw_tids = list(set([x["from"] for x in e["srl_links"]]))
|
|
|
|
|
|
|
|
hw_tokens = [_helper_tid_to_token(tid, e["tokens"]) for tid in hw_tids]
|
|
|
|
|
|
|
|
headwords = [(t["lemma"] if t["msd"][0] == "G" else t["lemma"] + "_") for t in hw_tokens]
|
|
|
|
|
|
|
|
e["headwords"] = headwords
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
functors = list(set([x["afun"] for x in e["srl_links"]]))
|
|
|
|
|
|
|
|
e["functors"] = functors
|
|
|
|
|
|
|
|
return e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# handler for concurrency
|
|
|
|
# handler for concurrency
|
|
|
|
def _handle_kres_file_tpl(kres_file_tpl):
|
|
|
|
def _handle_kres_file_tpl(kres_file_tpl):
|
|
|
|
|
|
|
|
tstart = time.time()
|
|
|
|
kres_file_idx = kres_file_tpl[0]
|
|
|
|
kres_file_idx = kres_file_tpl[0]
|
|
|
|
kres_file = kres_file_tpl[1]
|
|
|
|
kres_file = kres_file_tpl[1]
|
|
|
|
logging.info("Handling {} ({}/{})".format(
|
|
|
|
|
|
|
|
kres_file, kres_file_idx, n_kres_files))
|
|
|
|
|
|
|
|
kres_data = kres_parser.parse_xml_file(kres_file)
|
|
|
|
kres_data = kres_parser.parse_xml_file(kres_file)
|
|
|
|
if args.output == "file":
|
|
|
|
if args.output == "file":
|
|
|
|
kres_outdir = outdir / "kres_json"
|
|
|
|
kres_outdir = outdir / "kres_json"
|
|
|
@ -26,9 +50,37 @@ def _handle_kres_file_tpl(kres_file_tpl):
|
|
|
|
with kres_outfile.open("w") as fp:
|
|
|
|
with kres_outfile.open("w") as fp:
|
|
|
|
json.dump(kres_data, fp)
|
|
|
|
json.dump(kres_data, fp)
|
|
|
|
elif args.output == "db":
|
|
|
|
elif args.output == "db":
|
|
|
|
|
|
|
|
# mongoclient needs to be created after forking
|
|
|
|
|
|
|
|
dbclient = MongoClient(
|
|
|
|
|
|
|
|
"mongodb://{}".format(args.dbaddr),
|
|
|
|
|
|
|
|
username=args.dbuser,
|
|
|
|
|
|
|
|
password=args.dbpass,
|
|
|
|
|
|
|
|
authSource="valdb",
|
|
|
|
|
|
|
|
authMechanism='SCRAM-SHA-1'
|
|
|
|
|
|
|
|
)
|
|
|
|
|
|
|
|
valdb = dbclient.valdb
|
|
|
|
kres_col = valdb["kres"]
|
|
|
|
kres_col = valdb["kres"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# HUUUUGE BOTTLENECK
|
|
|
|
|
|
|
|
"""
|
|
|
|
for sentence in kres_data:
|
|
|
|
for sentence in kres_data:
|
|
|
|
kres_col.update({"sid": sentence["sid"]}, sentence, upsert=True)
|
|
|
|
kres_col.update({"sid": sentence["sid"]}, sentence, upsert=True)
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
kres_data_1 = [_db_preprocess(x) for x in kres_data]
|
|
|
|
|
|
|
|
kres_col.insert_many(kres_data_1) # much much better (just make sure sid has a unique index)
|
|
|
|
|
|
|
|
logging.info("Handled {} ({}/{}) in {:.2f} s".format(
|
|
|
|
|
|
|
|
kres_file, kres_file_idx, n_kres_files, time.time() - tstart))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def _get_dbclient(args):
|
|
|
|
|
|
|
|
dbclient = MongoClient(
|
|
|
|
|
|
|
|
"mongodb://{}".format(args.dbaddr),
|
|
|
|
|
|
|
|
username=args.dbuser,
|
|
|
|
|
|
|
|
password=args.dbpass,
|
|
|
|
|
|
|
|
authSource="valdb",
|
|
|
|
|
|
|
|
authMechanism='SCRAM-SHA-1'
|
|
|
|
|
|
|
|
)
|
|
|
|
|
|
|
|
return dbclient
|
|
|
|
|
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
if __name__ == "__main__":
|
|
|
|
parser = argparse.ArgumentParser(description="Parsing corpora kres and ssj500k.")
|
|
|
|
parser = argparse.ArgumentParser(description="Parsing corpora kres and ssj500k.")
|
|
|
@ -44,19 +96,16 @@ if __name__ == "__main__":
|
|
|
|
args = parser.parse_args()
|
|
|
|
args = parser.parse_args()
|
|
|
|
|
|
|
|
|
|
|
|
outdir = None
|
|
|
|
outdir = None
|
|
|
|
valdb = None
|
|
|
|
|
|
|
|
if args.output == "file":
|
|
|
|
if args.output == "file":
|
|
|
|
outdir = Path(args.outdir)
|
|
|
|
outdir = Path(args.outdir)
|
|
|
|
outdir.mkdir(parents=True, exist_ok=True)
|
|
|
|
outdir.mkdir(parents=True, exist_ok=True)
|
|
|
|
elif args.output == "db":
|
|
|
|
elif args.output == "db":
|
|
|
|
dbclient = MongoClient(
|
|
|
|
# Force unique sid
|
|
|
|
"mongodb://{}".format(args.dbaddr),
|
|
|
|
dbclient = _get_dbclient(args)
|
|
|
|
username=args.dbuser,
|
|
|
|
for corpus in ["kres", "ssj"]:
|
|
|
|
password=args.dbpass,
|
|
|
|
dbclient.valdb[corpus].ensure_index([("sid", pymongo.ASCENDING)])
|
|
|
|
authSource="valdb",
|
|
|
|
dbclient.valdb[corpus].ensure_index([("headwords", pymongo.ASCENDING)])
|
|
|
|
authMechanism='SCRAM-SHA-1'
|
|
|
|
dbclient.valdb[corpus].ensure_index([("functors", pymongo.ASCENDING)])
|
|
|
|
)
|
|
|
|
|
|
|
|
valdb = dbclient.valdb
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# SSJ
|
|
|
|
# SSJ
|
|
|
|
logger.info("Parsing Ssj: {}".format(args.ssj_file))
|
|
|
|
logger.info("Parsing Ssj: {}".format(args.ssj_file))
|
|
|
@ -67,8 +116,11 @@ if __name__ == "__main__":
|
|
|
|
with ssj_outfile.open("w") as fp:
|
|
|
|
with ssj_outfile.open("w") as fp:
|
|
|
|
json.dump(ssj_data, fp)
|
|
|
|
json.dump(ssj_data, fp)
|
|
|
|
elif args.output == "db":
|
|
|
|
elif args.output == "db":
|
|
|
|
|
|
|
|
dbclient = _get_dbclient(args)
|
|
|
|
|
|
|
|
valdb = dbclient.valdb
|
|
|
|
ssj_col = valdb["ssj"]
|
|
|
|
ssj_col = valdb["ssj"]
|
|
|
|
for sentence in ssj_data:
|
|
|
|
for sentence in ssj_data:
|
|
|
|
|
|
|
|
sentence = _db_preprocess(sentence)
|
|
|
|
ssj_col.update({"sid": sentence["sid"]}, sentence, upsert=True)
|
|
|
|
ssj_col.update({"sid": sentence["sid"]}, sentence, upsert=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|