def index(mapper, bulk_size=10000):
"""Indexing"""
pres_prof_mapping = gen_mappings()
file = open(WP_ST_F, "r")
index_name = WP_ST_INDEX_ID
mappings = {
"content": Elastic.analyzed_field(),
"professions": Elastic.notanalyzed_field()
}
elastic = Elastic(index_name)
elastic.create_index(mappings, force=True)
doc_id = 0
docs = {}
for line in file:
doc_id += 1
profs = []
while ("[" in line): # replace [A|B] with A
matchObj = re.search('\[(.*?)\]', line)
entity = matchObj.group(1).split("|")[0]
name = entity.replace("_", " ")
entity_id = mapper.get_id_from_person(name)
prof_list = pres_prof_mapping[name]
prof_list = [mapper.get_id_from_prof(prof) for prof in prof_list]
profs += prof_list
line = line.replace("[" + matchObj.group(1) + "]", entity_id)
docs[doc_id] = {"content": line, "professions": list(set(profs))}
if len(docs) == bulk_size: # bulk add 10000 sentences into elastic
elastic.add_docs_bulk(docs)
docs = {}
print(doc_id / 1000, "K documents indexed.")
# if len(docs) < 10000: # index the last butch of sentences
elastic.add_docs_bulk(docs)
def build(self, callback_get_doc_content, bulk_size=1000):
"""Builds the DBpedia index from the mongo collection.
To speedup indexing, we index documents as a bulk.
There is an optimum value for the bulk size; try to figure it out.
:param callback_get_doc_content: a function that get a documet from mongo and return the content for indexing
:param bulk_size: Number of documents to be added to the index as a bulk
"""
PLOGGER.info("Building " + self.__index_name + " ...")
elastic = Elastic(self.__index_name)
elastic.create_index(self.__mappings, model=self.__model, force=True)
i = 0
docs = dict()
for mdoc in self.__mongo.find_all(no_timeout=True):
docid = Mongo.unescape(mdoc[Mongo.ID_FIELD])
# get back document from mongo with keys and _id field unescaped
doc = callback_get_doc_content(Mongo.unescape_doc(mdoc))
if doc is None:
continue
docs[docid] = doc
i += 1
if i % bulk_size == 0:
elastic.add_docs_bulk(docs)
docs = dict()
PLOGGER.info(str(i / 1000) + "K documents indexed")
# indexing the last bulk of documents
elastic.add_docs_bulk(docs)
PLOGGER.info("Finished indexing (" + str(i) + " documents in total)")
def main():
index_name = "toy_index"
mappings = {
# "id": Elastic.notanalyzed_field(),
"title": Elastic.analyzed_field(),
"content": Elastic.analyzed_field(),
}
docs = {
1: {
"title":
"Rap God",
"content":
"gonna, gonna, Look, I was gonna go easy on you and not to hurt your feelings"
},
2: {
"title":
"Lose Yourself",
"content":
"Yo, if you could just, for one minute Or one split second in time, forget everything Everything that bothers you, or your problems Everything, and follow me"
},
3: {
"title":
"Love The Way You Lie",
"content":
"Just gonna stand there and watch me burn But that's alright, because I like the way it hurts"
},
4: {
"title":
"The Monster",
"content": [
"gonna gonna I'm friends with the monster",
"That's under my bed Get along with the voices inside of my head"
]
},
5: {
"title":
"Beautiful",
"content":
"Lately I've been hard to reach I've been too long on my own Everybody has a private world Where they can be alone"
}
}
elastic = Elastic(index_name)
elastic.create_index(mappings, force=True)
elastic.add_docs_bulk(docs)
print("index has been built")
class IndexerDBpediaTypes(object):
__DOC_TYPE = "doc" # we don't make use of types
__MAPPINGS = {
"id": Elastic.notanalyzed_field(),
"content": Elastic.analyzed_field(),
}
def __init__(self, config):
self.__elastic = None
self.__config = config
self.__index_name = config["index_name"]
self.__dbpedia_path = config["dbpedia_files_path"]
# For triple parsing
self.__prefix = URIPrefix()
self.__triple = Triple()
self.__ntparser = NTriplesParser(self.__triple)
# Entity abstract and type assignments kept in memory
self.__entity_abstracts = {}
self.__load_entity_abstracts()
self.__types_entities = defaultdict(list)
self.__load_entity_types()
@property
def name(self):
return self.__index_name
def __parse_line(self, line):
"""Parses a line from a ttl file and returns subject and object pair.
It is used for parsing DBpedia abstracts and entity types.
The subject is always prefixed.
For object URIs, it is returned prefixed if from DBpedia otherwise
None (i.e., types); literal objects are always returned (i.e.,
abstracts).
"""
line = line.decode("utf-8") if isinstance(line, bytes) else line
try:
self.__ntparser.parsestring(line)
except ParseError: # skip lines that couldn't be parsed
return None, None
if self.__triple.subject() is None: # only if parsed as a triple
return None, None
subj = self.__prefix.get_prefixed(self.__triple.subject())
obj = None
if type(self.__triple.object()) is URIRef:
if self.__triple.object().startswith(
"http://dbpedia.org/ontology"):
obj = self.__prefix.get_prefixed(self.__triple.object())
else:
obj = self.__triple.object().encode("utf-8")
return subj, obj
def __load_entity_abstracts(self):
num_lines = 0
filename = os.sep.join([self.__dbpedia_path, ENTITY_ABSTRACTS_FILE])
PLOGGER.info("Loading entity abstracts from {}".format(filename))
for line in FileUtils.read_file_as_list(filename):
entity, abstract = self.__parse_line(line)
if abstract and len(abstract) > 0: # skip empty objects
self.__entity_abstracts[entity] = abstract
num_lines += 1
if num_lines % 10000 == 0:
PLOGGER.info(" {}K lines processed".format(num_lines // 1000))
PLOGGER.info(" Done.")
def __load_entity_types(self):
num_lines = 0
for types_file in ENTITY_TYPES_FILES:
filename = os.sep.join([self.__dbpedia_path, types_file])
PLOGGER.info("Loading entity types from {}".format(filename))
for line in FileUtils.read_file_as_list(filename):
entity, entity_type = self.__parse_line(line)
if type(entity_type) != str: # Likely result of parsing error
continue
if not entity_type.startswith("<dbo:"):
PLOGGER.info(" Non-DBpedia type: {}".format(entity_type))
continue
if not entity.startswith("<dbpedia:"):
PLOGGER.info(" Invalid entity: {}".format(entity))
continue
self.__types_entities[entity_type].append(entity)
num_lines += 1
if num_lines % 10000 == 0:
PLOGGER.info(" {}K lines processed".format(num_lines //
1000))
PLOGGER.info(" Done.")
def __make_type_doc(self, type_name):
"""Gets the document representation of a type to be indexed, from its
entity short abstracts."""
content = "\n".join([
self.__entity_abstracts.get(e, b"").decode("utf-8")
for e in self.__types_entities[type_name]
])
if len(content) > MAX_BULKING_DOC_SIZE:
PLOGGER.info("Type {} has content larger than allowed: {}.".format(
type_name, len(content)))
# we randomly sample a subset of Y entity abstracts, s.t.
# Y * AVG_SHORT_ABSTRACT_LEN <= MAX_BULKING_DOC_SIZE
num_entities = len(self.__types_entities[type_name])
amount_abstracts_to_sample = min(
floor(MAX_BULKING_DOC_SIZE / AVG_SHORT_ABSTRACT_LEN),
num_entities)
entities_sample = [
self.__types_entities[type_name][i] for i in sample(
range(num_entities), amount_abstracts_to_sample)
]
content = "" # reset content
for entity in entities_sample:
new_content_candidate = "\n".join([
content,
self.__entity_abstracts.get(entity, b"").decode("utf-8")
])
# we add an abstract only if by doing so it will not exceed
# MAX_BULKING_DOC_SIZE
if len(new_content_candidate) > MAX_BULKING_DOC_SIZE:
break
content = new_content_candidate
return {"content": content}
def build_index(self, force=False):
"""Builds the index.
Note: since DBpedia only has a few hundred types, no bulk indexing is
needed.
:param force: True iff it is required to overwrite the index (i.e. by
creating it by force); False by default.
:type force: bool
:return:
"""
PLOGGER.info("Building type index {}".format(self.__index_name))
self.__elastic = Elastic(self.__index_name)
self.__elastic.create_index(mappings=self.__MAPPINGS, force=force)
for type_name in self.__types_entities:
PLOGGER.info(" Adding {} ...".format(type_name))
contents = self.__make_type_doc(type_name)
self.__elastic.add_doc(type_name, contents)
PLOGGER.info(" Done.")
class IndexerDBpediaTypes(object):
__DOC_TYPE = "doc" # we don't make use of types
__MAPPINGS = {
# ID_KEY: Elastic.notanalyzed_field(),
CONTENT_KEY: Elastic.analyzed_field(),
}
def __init__(self, config):
self.__elastic = None
self.__config = config
self.__model = config.get("model", Elastic.BM25)
self.__index_name = config["index_name"]
self.__type2entity_file = config["type2entity_file"]
self.__entity_abstracts = {}
self.__load_entity_abstracts(config["entity_abstracts_file"])
@property
def name(self):
return self.__index_name
def __load_entity_abstracts(self, filename):
prefix = URIPrefix()
t = Triple()
p = NTriplesParser(t)
lines_counter = 0
PLOGGER.info("Loading entity abstracts from {}".format(filename))
for line in FileUtils.read_file_as_list(filename):
# basic line parsing
line = line.decode("utf-8") if isinstance(line, bytes) else line
try:
p.parsestring(line)
except ParseError: # skip lines that couldn't be parsed
continue
if t.subject() is None: # only if parsed as a triple
continue
# Subject and object identification
subj = prefix.get_prefixed(t.subject())
obj = ""
if type(t.object()) is URIRef:
# PLOGGER.error("Error: it is URIRef the parsed obj")
pass
else:
obj = t.object().encode("utf-8")
if len(obj) == 0:
continue # skip empty objects
self.__entity_abstracts[subj] = obj
lines_counter += 1
if lines_counter % 10000 == 0:
PLOGGER.info("\t{}K lines processed".format(lines_counter // 1000))
pass
PLOGGER.info("\n### Loading entity abstracts... Done.")
def __make_type_doc(self, entities, last_type):
"""Gets the document representation of a type to be indexed, from its entity short abstracts."""
content = ABSTRACTS_SEPARATOR.join([self.__entity_abstracts.get(e, b"").decode("utf-8")
for e in entities])
if len(content) > MAX_BULKING_DOC_SIZE:
PLOGGER.info("Type {} has content larger than allowed: {}.".format(last_type, len(content)))
# we randomly sample a subset of Y entity abstracts, s.t. Y * AVG_SHORT_ABSTRACT_LEN <= MAX_BULKING_DOC_SIZE
amount_abstracts_to_sample = min(floor(MAX_BULKING_DOC_SIZE / AVG_SHORT_ABSTRACT_LEN), len(entities))
entities_sample = [entities[i] for i in sample(range(len(entities)), amount_abstracts_to_sample)]
content = "" # reset content
for entity in entities_sample:
new_content_candidate = (content + ABSTRACTS_SEPARATOR +
self.__entity_abstracts.get(entity, b"").decode("utf-8"))
# we add an abstract only if by doing so it will not exceed MAX_BULKING_DOC_SIZE
if len(new_content_candidate) <= MAX_BULKING_DOC_SIZE:
content = new_content_candidate
else:
break
return {CONTENT_KEY: content}
def build_index(self, force=False):
"""Builds the index.
:param force: True iff it is required to overwrite the index (i.e. by creating it by force); False by default.
:type force: bool
:return:
"""
self.__elastic = Elastic(self.__index_name)
self.__elastic.create_index(mappings=self.__MAPPINGS, force=force)
prefix = URIPrefix()
# For indexing types in bulk
types_bulk = {} # dict from type id to type(=doc)
# process type2entity file
last_type = None
entities = []
lines_counter = 0
types_counter = 0
with FileUtils.open_file_by_type(self.__type2entity_file) as f:
for line in f:
line = line.decode() # o.w. line is made of bytes
if not line.startswith("<"): # bad-formed lines in dataset
continue
subj, obj = line.rstrip().split()
type = prefix.get_prefixed(subj) # subject prefixed
entity = prefix.get_prefixed(obj)
# use only DBpedia Ontology native types (no bibo, foaf, schema, etc.)
if not type.startswith(DBO_PREFIX):
continue
if last_type is not None and type != last_type:
# moving to new type, so:
# create a doc for this type, with all the abstracts for its entities, and store it in a bulk
types_counter += 1
# PLOGGER.info("\n\tFound {}-th type: {}\t\t with # of entities: {}".format(types_counter,
# last_type,
# len(entities)))
types_bulk[last_type] = self.__make_type_doc(entities, last_type)
entities = [] # important to reset it
if types_counter % BULK_LEN == 0: # index the bulk of BULK_LEN docs
self.__elastic.add_docs_bulk(types_bulk)
types_bulk.clear() # NOTE: important to reset it
PLOGGER.info("\tIndexing a bulk of {} docs (types)... OK. "
"{} types already indexed.".format(BULK_LEN, types_counter))
last_type = type
entities.append(entity)
lines_counter += 1
if lines_counter % 10000 == 0:
# PLOGGER.info("\t{}K lines processed".format(lines_counter // 1000))
pass
pass
# index the last type
types_counter += 1
PLOGGER.info("\n\tFound {}-th (last) type: {}\t\t with # of entities: {}".format(types_counter, last_type,
len(entities)))
types_bulk[last_type] = self.__make_type_doc(entities, last_type)
self.__elastic.add_docs_bulk(types_bulk) # a tiny bulk :)
# no need to reset neither entities nor types_bulk :P
# PLOGGER.info("Indexing a bulk of {} docs (types)... OK.".format(BULK_LEN))
PLOGGER.info("\n### Indexing all {} found docs (types)... Done.".format(types_counter))