add yago-facts.ttl to facts.tar.gz (adds ~30+ million more facts)

README.md

@@ -28,9 +28,9 @@ dataset_info:
   config_name: default
   splits:
     - name: train
-      num_bytes: 36585902753
-      num_examples: 210248472
-  dataset_size: 36585902753
+      num_bytes: 42709902295
+      num_examples: 249675587
+  dataset_size: 42709902295
 viewer: false
 ---
 

facts.tar.gz

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:671a7b0411a60077cc60aa96e4124b6d956717188ccf8976f292d60999f7f11b
-size 3273108327
+oid sha256:f451bfa0d6535ffaa223c3dea8c1a148be2f9ecd15e2b5efab7f989821a66331
+size 3856455011

yago45en.py

@@ -10,7 +10,7 @@ SCHEMA = Namespace('http://schema.org/')
 YAGO = Namespace('http://yago-knowledge.org/resource/')
 
 class YAGO45DatasetBuilder(GeneratorBasedBuilder):
-    VERSION = "1.0.0"
+    VERSION = "1.0.1"
 
     def _info(self):
         return DatasetInfo(
@@ -32,10 +32,10 @@ class YAGO45DatasetBuilder(GeneratorBasedBuilder):
         # Download and extract the dataset files
         facts, taxonomy = dl_manager.download_and_extract(["facts.tar.gz", "yago-taxonomy.ttl"])
 
-        facts = os.path.join(facts, "tmp/yago/")
+        facts = os.path.join(facts, "facts/")
 
         # Define splits for each chunk of your dataset.
-        chunk_paths = [os.path.join(facts, chunk) for chunk in os.listdir(facts) if chunk.endswith('.nt')]
+        chunk_paths = [os.path.join(facts, chunk) for chunk in os.listdir(facts) if chunk.endswith('.nt')] + [taxonomy]
         return [SplitGenerator(name=datasets.Split.TRAIN, 
                                gen_kwargs={'chunk_paths': chunk_paths})]