Datasets:
use GeneratorBasedBuilder as superclass
Browse files- yago45en.py +12 -11
yago45en.py
CHANGED
@@ -1,5 +1,6 @@
|
|
1 |
import os
|
2 |
-
|
|
|
3 |
from rdflib import Graph, URIRef, Literal, BNode
|
4 |
from rdflib.namespace import RDF, RDFS, OWL, XSD, Namespace, NamespaceManager
|
5 |
from datasets.features import Features, Value
|
@@ -8,13 +9,12 @@ SCHEMA = Namespace('http://schema.org/')
|
|
8 |
|
9 |
YAGO = Namespace('http://yago-knowledge.org/resource/')
|
10 |
|
11 |
-
class YAGO45DatasetBuilder(
|
12 |
VERSION = "1.0.0"
|
13 |
|
14 |
taxonomy = Graph(bind_namespaces="core")
|
15 |
|
16 |
def _info(self):
|
17 |
-
print("INFO")
|
18 |
return DatasetInfo(
|
19 |
description="A subset of the YAGO 4.5 dataset maintaining only English labels",
|
20 |
citation="@article{suchanek2023integrating,title={Integrating the Wikidata Taxonomy into YAGO},author={Suchanek, Fabian M and Alam, Mehwish and Bonald, Thomas and Paris, Pierre-Henri and Soria, Jules},journal={arXiv preprint arXiv:2308.11884},year={2023}}",
|
@@ -32,18 +32,21 @@ class YAGO45DatasetBuilder(DatasetBuilder):
|
|
32 |
# Define splits for each chunk of your dataset.
|
33 |
|
34 |
# Download and extract the dataset files
|
35 |
-
|
36 |
-
|
37 |
-
|
|
|
38 |
# Load yago-taxonomy.ttl file in every process
|
39 |
-
self.taxonomy.parse(
|
40 |
|
41 |
# Extract prefix mappings
|
42 |
prefix_mappings = {prefix: namespace for prefix, namespace in self.taxonomy.namespaces()}
|
43 |
|
44 |
# Define splits for each chunk
|
45 |
-
chunk_paths = [os.path.join(
|
46 |
-
return [SplitGenerator(name=
|
|
|
|
|
47 |
|
48 |
def _generate_examples(self, chunk_paths, prefix_mappings):
|
49 |
# Load the chunks into an rdflib graph
|
@@ -61,5 +64,3 @@ class YAGO45DatasetBuilder(DatasetBuilder):
|
|
61 |
'predicate': str(p),
|
62 |
'object': str(o)
|
63 |
}
|
64 |
-
|
65 |
-
|
|
|
1 |
import os
|
2 |
+
import datasets
|
3 |
+
from datasets import DatasetBuilder, SplitGenerator, DownloadConfig, load_dataset, DownloadManager, DatasetInfo, GeneratorBasedBuilder
|
4 |
from rdflib import Graph, URIRef, Literal, BNode
|
5 |
from rdflib.namespace import RDF, RDFS, OWL, XSD, Namespace, NamespaceManager
|
6 |
from datasets.features import Features, Value
|
|
|
9 |
|
10 |
YAGO = Namespace('http://yago-knowledge.org/resource/')
|
11 |
|
12 |
+
class YAGO45DatasetBuilder(GeneratorBasedBuilder):
|
13 |
VERSION = "1.0.0"
|
14 |
|
15 |
taxonomy = Graph(bind_namespaces="core")
|
16 |
|
17 |
def _info(self):
|
|
|
18 |
return DatasetInfo(
|
19 |
description="A subset of the YAGO 4.5 dataset maintaining only English labels",
|
20 |
citation="@article{suchanek2023integrating,title={Integrating the Wikidata Taxonomy into YAGO},author={Suchanek, Fabian M and Alam, Mehwish and Bonald, Thomas and Paris, Pierre-Henri and Soria, Jules},journal={arXiv preprint arXiv:2308.11884},year={2023}}",
|
|
|
32 |
# Define splits for each chunk of your dataset.
|
33 |
|
34 |
# Download and extract the dataset files
|
35 |
+
facts, taxonomy = dl_manager.download_and_extract(["facts.tar.gz", "yago-taxonomy.ttl"])
|
36 |
+
|
37 |
+
facts = os.path.join(facts, "tmp/yago/")
|
38 |
+
|
39 |
# Load yago-taxonomy.ttl file in every process
|
40 |
+
self.taxonomy.parse(taxonomy)
|
41 |
|
42 |
# Extract prefix mappings
|
43 |
prefix_mappings = {prefix: namespace for prefix, namespace in self.taxonomy.namespaces()}
|
44 |
|
45 |
# Define splits for each chunk
|
46 |
+
chunk_paths = [os.path.join(facts, chunk) for chunk in os.listdir(facts) if chunk.endswith('.nt')]
|
47 |
+
return [SplitGenerator(name=datasets.Split.TRAIN,
|
48 |
+
gen_kwargs={'chunk_paths': chunk_paths,
|
49 |
+
'prefix_mappings': prefix_mappings})]
|
50 |
|
51 |
def _generate_examples(self, chunk_paths, prefix_mappings):
|
52 |
# Load the chunks into an rdflib graph
|
|
|
64 |
'predicate': str(p),
|
65 |
'object': str(o)
|
66 |
}
|
|
|
|