def save_document_triples(data, dataset=FLAGS.dataset):
path = io.get_document_triples_path(dataset)
io.write_pickle(path, data)
csv_path = path.replace(".pickle.bz2", ".csv")
io.write_csv(csv_path,
data,
sep=",",
header=["doc1", "doc2", "relations", "detail"])
def save_ordered_document_triples_metrics(data,
edge_type,
dataset=FLAGS.dataset):
path = io.get_ordered_document_triples_metrics_path(edge_type, dataset)
io.write_csv(path,
data,
sep=",",
header=[
"doc1", "doc2", "count", "idf", "idf_wiki", "count_norm",
"count_norm_pmi", "idf_norm", "idf_wiki_norm",
"idf_norm_pmi", "idf_wiki_norm_pmi"
])
def save_document_triples_metrics(data, dataset=FLAGS.dataset):
path = io.get_document_triples_metrics_path(dataset)
io.write_pickle(path, data)
csv_path = path.replace(".pickle.bz2", ".csv")
io.write_csv(csv_path,
data,
sep=",",
header=[
"doc1", "doc2", "count", "idf", "idf_wiki", "count_norm",
"count_norm_pmi", "idf_norm", "idf_wiki_norm",
"idf_norm_pmi", "idf_wiki_norm_pmi"
])
def save_eval_logs(data, dataset=FLAGS.dataset, version=FLAGS.version):
io.write_csv(io.get_eval_log_path(dataset, version), data, sep=';')
def save_document_embeddings(data, layer):
io.write_csv(io.get_document_embeddings_path(layer), data.numpy(), sep=",")
def save_filtered_triples(data, dataset=FLAGS.dataset):
io.write_csv(io.get_filtered_word_triples_path(dataset),
data,
sep=",",
header=["entity1", "relations", "entity2"])
def save_all_relations(data):
io.write_csv(
io.get_all_wiki_relations_path(),
data,
sep="+",
header=["ID", "label", "description", "aliases", "type", "count"])
def save_filtered_relations(data):
io.write_csv(io.get_filtered_wiki_relations_path(),
data,
sep="+",
header=["ID"])
def save_doc2idf(data, dataset=FLAGS.dataset):
io.write_csv(io.get_doc2idf_path(dataset),
data,
sep=",",
header=["doc", "relation", "idf", "idf_wiki"])
def save_doc2id(data, dataset=FLAGS.dataset):
io.write_csv(io.get_doc2id_path(dataset),
data,
sep=",",
header=["doc", "wikiID"])
def save_entity2id(data, dataset=FLAGS.dataset):
io.write_csv(io.get_entity2id_path(dataset),
data,
sep=",",
header=["word", "wikiID"])
