a

corpusparser/Parser.py

@@ -3,24 +3,21 @@ import re
 import json
 from lxml import etree
 import logging
+import time
 
-logging.basicConfig(level=logging.INFO)
+# logging.basicConfig(level=logging.INFO)
 
 # Read input file(.xml, .json; kres or ssj500k).  
 # Create an iterator that outputs resulting sentences (python dict format).  
 class Parser():
 
-    def __init__(self, corpus, infiles, logger=None):
-
-        if corpus == "kres":
-            self.kres_folder = Path(infiles[0])
-            self.kres_srl_folder = Path(infiles[1])
-        elif corpus == "ssj":
-            self.ssj_file = Path(infiles[0])
-        else:
-            raise ValueError("Argument corpus should be 'ssj' or 'kres'.")
+    def __init__(self, corpus, kres_srl_folder=None, logger=None):
 
         self.corpus = corpus
+
+        if self.corpus == "kres":
+            self.kres_srl_folder = kres_srl_folder
+
         self.W_TAGS = ['w']
         self.C_TAGS = ['c']
         self.S_TAGS = ['S', 'pc']
@@ -30,6 +27,10 @@ class Parser():
             "missing_srl": []
         }
 
+        # for logging output
+        self.n_kres_files = -1
+        self.nth_kres_file = -1
+
     def parse_jos_links(self, sent_el):
         if self.corpus == "kres":
             return self.parse_jos_links_kres(sent_el)
@@ -60,7 +61,10 @@ class Parser():
 
     def parse_any_links_ssj(self, sent_el, links_type):
         lgrps = sent_el.findall(".//linkGrp")
-        links = [x for x in lgrps if x.get("type") == links_type][0]
+        try:
+            links = [x for x in lgrps if x.get("type") == links_type][0]
+        except:
+            return []
         res_links = []
         for link in links:
             tar = self.parse_ssj_target_arg(link.get("target"))
@@ -90,14 +94,34 @@ class Parser():
 
     def sentence_generator(self):
         # Using generators so we don't copy a whole corpu around in memory.
+        # Use parse_xml_file() instead for pre-file processing (parallelism?)
         if self.corpus == "kres":
+
+            # some logging output
+            if self.n_kres_files == -1:
+                self.n_kres_files = len(list(Path(self.kres_folder).glob('*')))
+
             for xml_file in self.kres_folder.iterdir():
-                # self.parse_xml_file(xml_file)
-                yield from self.parse_xml_file(xml_file)
+                self.nth_kres_file += 1
+                self.logger.info("{} ({}/{})".format(
+                    xml_file, self.nth_kres_file, self.n_kres_files))
+                yield from self.xml_file_to_generator(xml_file)
         else:
-            yield from self.parse_xml_file(self.ssj_file)
+            yield from self.xml_file_to_generator(self.ssj_file)
 
     def parse_xml_file(self, xml_file):
+        # tstart = time.time()
+        file_data = []
+        for tpl in self.xml_file_to_generator(xml_file):
+            file_data += [tpl[1]]
+        tend = time.time()
+        # self.logger.info("Parsed {} in {:.4f} s".format(xml_file, tend - tstart))
+        return file_data
+
+    def xml_file_to_generator(self, xml_file):
+        # for separate srl links, it will guess the srl file based on 
+        # self.kres_srl_folder
+
         srl_from_json = {}
         if self.corpus == "kres":  
             # in case of kres, read the SRL links form a separate json file
@@ -190,7 +214,7 @@ class Parser():
                         "text": sentence_text,
                         "tokens": sentence_tokens,
                         "jos_links": jos_links,
-                        "srl_links": srl_links_parsed
+                        "srl_links": srl_links_parsed,
                     }
                     self.stats["parsed_count"] += 1
                     yield (xml_file, sentence_entry)

corpusparser/__init__.py

@@ -1 +1,2 @@
-from corpusparser.Parser import Parser
+from corpusparser.Parser import Parser
+from corpusparser.main import enriched_lemma

corpusparser/main.py

@@ -4,121 +4,188 @@ import argparse
 import logging
 import json
 from pymongo import MongoClient
+import pymongo
 import sys
+from multiprocessing import Pool
+import time
 
-logging.basicConfig(level=logging.INFO)
+CORPORA = ["kres", "ssj"]
+
+# logging.basicConfig(filename=Path("/var/tmp/corpusparser.log"), filemode='a', level=logging.INFO)
 logger = logging.getLogger(__name__)
 
-## Main handles command line arguments and writing to files / DB.  
+# lfh = logging.FileHandler("/project/logs/fill-database.log")
+lfh = logging.StreamHandler(sys.stdout)
+formatter = logging.Formatter("%(asctime)s %(levelname)s %(message)s")
+lfh.setFormatter(formatter)
+logger.addHandler(lfh)
+logger.setLevel(logging.INFO)
 
-def ssj_to_json_file(sentence_generator, outfolder):
-    # this funciton is based on the fact that files are parsed sequentially
-    outfolder = Path(outfolder)
-    outfolder.mkdir(parents=True, exist_ok=True)
-    outfile = outfolder / "ssj500k.json"
+n_chunks = -1
 
-    data_buffer = []
-    for s in sentence_generator:
-        sdata = s[1]
-        data_buffer += [sdata]
 
-    # outfile = Path(outfile)
-    with outfile.open("w") as fp:
-        logger.info("Writing to {}".format(outfile))
-        json.dump(data_buffer, fp)
+def enriched_lemma(token):
+    return (token["lemma"] if token["msd"][0] == "G" else token["lemma"] + "_")
 
-def kres_to_json_files(sentence_generator, outfolder):
-    outfolder = Path(outfolder) / "kres_json"
-    outfolder.mkdir(parents=True, exist_ok=True)
 
-    def write_buffer_to_file(outfile, outfile_buffer):
-        logger.info("Writing file: {}".format(outfile))
-        with outfile.open("w") as fp:
-            json.dump(outfile_buffer, fp)
+def _helper_tid_to_token(tid, tokens):
+    for t in tokens:
+        if t["tid"] == tid:
+            return t
+    return None
 
-    outfile_buffer = None
-    current_outfile = None
-    for s in sentence_generator:
-        infile = s[0]
-        outfile = outfolder / Path(infile.name.split(".")[0]).with_suffix(".json")
 
-        # parser sequentially parses files; when we're done with a file, write it out
-        if current_outfile is None:
-            current_outfile = outfile
-            outfile_buffer = [] 
-        elif outfile != current_outfile:
-            write_buffer_to_file(current_outfile, outfile_buffer)
-            current_outfile = outfile
-            outfile_buffer = []
+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 = [enriched_lemma(t) for t in hw_tokens]
+        e["headwords"] = headwords
 
-        # update buffer
-        sdata = s[1]
-        outfile_buffer += [sdata]
-    write_buffer_to_file(current_outfile, outfile_buffer)
+        functors = list(set([x["afun"] for x in e["srl_links"]]))
+        e["functors"] = functors
+    return e
 
-def data_to_valdb(sentence_generator, dbaddr, username, password, collection_name):
-    logger.info("Connecting to: {}".format(dbaddr))
-    client = MongoClient(
-        "mongodb://{}".format(dbaddr),
-        username=username,
-        password=password,
+
+# handler for concurrency
+def _handle_kres_file_chunk(kres_file_chunk):
+    tstart = time.time()
+    kres_chunk_idx = kres_file_chunk[0]
+    kres_chunk = kres_file_chunk[1]
+
+    dbclient = None
+    db_payload = []
+    if 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'
+        )
+        # dbclient.valdb["kres"]
+    for kres_file in kres_chunk:
+        try:
+            kres_data = kres_parser.parse_xml_file(kres_file)
+        except:
+            logger.error("Failed to parse file: {}".format(kres_file))
+            continue
+        if args.output == "file":
+            kres_outdir = outdir / "kres_json"
+            kres_outdir.mkdir(parents=True, exist_ok=True)
+            kres_outfile = kres_outdir / Path(kres_file.name.split(".")[0]).with_suffix(".json")
+            with kres_outfile.open("w") as fp:
+                json.dump(kres_data, fp)
+
+        elif args.output == "db":
+            """
+            if dbclient.valdb["kres"].find({"sid": kres_data[0]["sid"]}).count() > 0:
+                logger.info("File {} already in DB, closing chunk ({}/{})".format(
+                    kres_file, kres_chunk_idx, n_chunks))
+                dbclient.close()
+                return
+            """
+
+            kres_data_1 = [_db_preprocess(x) for x in kres_data]
+            db_payload += kres_data_1
+
+    try:
+        dbclient.valdb["kres"].insert_many(db_payload, ordered=False)  # much much better (just make sure sid has a unique index)
+    except:
+        logger.error("Failed inserting kres files chunk ({}/{}) in {:.2f} s".format(
+            kres_chunk_idx, n_chunks, time.time() - tstart))
+    logger.info("Db insert: chunks ({}/{}) in {:.2f} s".format(
+        kres_chunk_idx, n_chunks, time.time() - tstart))
+    dbclient.close()
+
+def _get_dbclient(args):
+    dbclient = MongoClient(
+        "mongodb://{}".format(args.dbaddr),
+        username=args.dbuser,
+        password=args.dbpass,
         authSource="valdb",
-        authMechanism='SCRAM-SHA-256'
+        authMechanism='SCRAM-SHA-1'
     )
-    valdb = client.valdb
-    logger.info("Writing data to {}.".format(collection_name))
-    col = valdb[collection_name]
-    for s in sentence_generator:
-        sdata = s[1]
-        col.insert_one(sdata)
+    return dbclient
 
 
 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('--corpus', required=True)
+    parser.add_argument('--kres-folder', required=False)
+    parser.add_argument('--kres-srl-folder', required=False)
+    parser.add_argument('--ssj-file', required=False)
     parser.add_argument('--output', required=False, default=None)
     parser.add_argument('--outdir', required=False, default=None)
     parser.add_argument('--dbaddr', required=False, default=None)
     parser.add_argument('--dbuser', required=False, default=None)
     parser.add_argument('--dbpass', required=False, default=None)
+    parser.add_argument('--cores', required=False, default=1)
+    parser.add_argument('--chunk-size', required=False, default=3)
     args = parser.parse_args()
 
 
-    # parse ssj
-    logger.info("Parsing ssj500k: {}".format(args.ssj_file))
-    ssj_parser = Parser(
-        corpus="ssj",
-        infiles=[args.ssj_file],
-    )
-    # ssj to json
+    corpus = args.corpus
+    assert (corpus in CORPORA), "Wrong corpus name."
+
+
+    outdir = None
     if args.output == "file":
-        ssj_to_json_file(ssj_parser.sentence_generator(), args.outdir)
+        outdir = Path(args.outdir)
+        outdir.mkdir(parents=True, exist_ok=True)
     elif args.output == "db":
-        data_to_valdb(
-            ssj_parser.sentence_generator(),
-            args.dbaddr,
-            args.dbuser,
-            args.dbpass,
-            collection_name="ssj"
+        dbclient = _get_dbclient(args)
+        dbclient.valdb[corpus].ensure_index([("sid", pymongo.ASCENDING)])
+        dbclient.valdb[corpus].ensure_index([("headwords", pymongo.ASCENDING)])
+        dbclient.valdb[corpus].ensure_index([("functors", pymongo.ASCENDING)])
+        dbclient.close()
+
+
+    if corpus == "ssj":
+        logger.info("Parsing Ssj: {}".format(args.ssj_file))
+        ssj_parser = Parser(logger=logger, corpus="ssj")
+        ssj_data = ssj_parser.parse_xml_file(Path(args.ssj_file))
+        if args.output == "file":
+            ssj_outfile = outdir / "ssj500k.json"
+            with ssj_outfile.open("w") as fp:
+                json.dump(ssj_data, fp)
+        elif args.output == "db":
+            dbclient = _get_dbclient(args)
+            valdb = dbclient.valdb
+            ssj_col = valdb["ssj"]
+            for sentence in ssj_data:
+                sentence = _db_preprocess(sentence)
+                ssj_col.update({"sid": sentence["sid"]}, sentence, upsert=True)
+            dbclient.close()
+
+
+    if corpus == "kres":
+        # Kres
+        logger.info("Parsing Kres: {}".format(args.kres_folder))
+        kres_parser = Parser(
+            logger=logger,
+            corpus="kres",
+            kres_srl_folder=args.kres_srl_folder
         )
 
-    sys.exit()
-    # parse kres
-    logger.info("Parsing Kres: {}".format(args.ssj_file))
-    kres_parser = Parser(
-        corpus="kres",
-        infiles=[args.kres_folder, args.kres_srl_folder],
-    )
-    # kres to json
-    if args.output == "file":
-        kres_to_json_files(kres_parser.sentence_generator(), args.outdir)
-    elif args.output == "db":
-        data_to_valdb(
-            ssj_parser.sentence_generator(),
-            args.dbaddr,
-            args.dbuser,
-            args.dbpass,
-            collection_name="kres"
-        )
+        kres_files = [x for x in Path(args.kres_folder).iterdir()]
+        kres_files = sorted(kres_files, key=lambda x: x.name)
+        kres_files_chunks = []
+        i = 0
+        while i < len(kres_files):
+            # kres_files_chunks += kres_files[i:(i+args.chunk_size)]
+            new_i = i + int(args.chunk_size)
+            kres_files_chunks += [kres_files[i:new_i]]
+            i = new_i
+        kres_files_chunks = [x for x in enumerate(kres_files_chunks)]
+        n_chunks = len(kres_files_chunks)
+
+        p = Pool(int(args.cores))
+        p.map(_handle_kres_file_chunk, kres_files_chunks)
+
+    
+    logger.info("Finished parsing.")