Datasets:
Tasks:
Question Answering
Modalities:
Text
Languages:
English
Size:
10K - 100K
ArXiv:
Tags:
chained-qa
License:
Update files from the datasets library (from 1.16.0)
Browse filesRelease notes: https://github.com/huggingface/datasets/releases/tag/1.16.0
- README.md +1 -0
- babi_qa.py +71 -67
README.md
CHANGED
@@ -1,4 +1,5 @@
|
|
1 |
---
|
|
|
2 |
annotations_creators:
|
3 |
- machine-generated
|
4 |
language_creators:
|
|
|
1 |
---
|
2 |
+
pretty_name: BabiQa
|
3 |
annotations_creators:
|
4 |
- machine-generated
|
5 |
language_creators:
|
babi_qa.py
CHANGED
@@ -15,8 +15,6 @@
|
|
15 |
"""The bAbI tasks dataset."""
|
16 |
|
17 |
|
18 |
-
import os
|
19 |
-
|
20 |
import datasets
|
21 |
|
22 |
|
@@ -851,20 +849,22 @@ class BabiQa(datasets.GeneratorBasedBuilder):
|
|
851 |
def _split_generators(self, dl_manager):
|
852 |
"""Returns SplitGenerators."""
|
853 |
my_urls = ZIP_URL
|
854 |
-
|
855 |
splits = [
|
856 |
datasets.SplitGenerator(
|
857 |
name=datasets.Split.TRAIN,
|
858 |
# These kwargs will be passed to _generate_examples
|
859 |
gen_kwargs={
|
860 |
-
"filepath":
|
|
|
861 |
},
|
862 |
),
|
863 |
datasets.SplitGenerator(
|
864 |
name=datasets.Split.TEST,
|
865 |
# These kwargs will be passed to _generate_examples
|
866 |
gen_kwargs={
|
867 |
-
"filepath":
|
|
|
868 |
},
|
869 |
),
|
870 |
]
|
@@ -874,73 +874,77 @@ class BabiQa(datasets.GeneratorBasedBuilder):
|
|
874 |
name=datasets.Split.VALIDATION,
|
875 |
# These kwargs will be passed to _generate_examples
|
876 |
gen_kwargs={
|
877 |
-
"filepath":
|
|
|
878 |
},
|
879 |
),
|
880 |
]
|
881 |
return splits
|
882 |
|
883 |
-
def _generate_examples(self, filepath):
|
884 |
|
885 |
-
|
886 |
-
|
887 |
-
|
888 |
-
|
889 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
890 |
if story != []:
|
891 |
yield example_idx, {"story": story}
|
892 |
-
|
893 |
-
story = []
|
894 |
-
elif line.strip().split()[0] == "1": # New story
|
895 |
-
if story != []: # Already some story, flush it out
|
896 |
-
yield example_idx, {"story": story}
|
897 |
-
example_idx += 1
|
898 |
-
story = []
|
899 |
-
line_no = line.split()[0]
|
900 |
-
line_split = line[len(line_no) :].strip().split("\t")
|
901 |
-
if len(line_split) > 1:
|
902 |
-
story.append(
|
903 |
-
{
|
904 |
-
"id": line_no,
|
905 |
-
"type": 1, # question
|
906 |
-
"supporting_ids": line_split[-1].split(" "),
|
907 |
-
"text": line_split[0].strip(),
|
908 |
-
"answer": line_split[1].strip(),
|
909 |
-
}
|
910 |
-
)
|
911 |
-
else:
|
912 |
-
story.append(
|
913 |
-
{
|
914 |
-
"id": line_no,
|
915 |
-
"type": 0, # context
|
916 |
-
"supporting_ids": [],
|
917 |
-
"text": line_split[0].strip(),
|
918 |
-
"answer": "",
|
919 |
-
}
|
920 |
-
)
|
921 |
-
else:
|
922 |
-
line_no = line.split()[0]
|
923 |
-
line_split = line[len(line_no) :].strip().split("\t")
|
924 |
-
if len(line_split) > 1:
|
925 |
-
story.append(
|
926 |
-
{
|
927 |
-
"id": line_no,
|
928 |
-
"type": 1, # question
|
929 |
-
"supporting_ids": line_split[-1].split(" "),
|
930 |
-
"text": line_split[0].strip(),
|
931 |
-
"answer": line_split[1].strip(),
|
932 |
-
}
|
933 |
-
)
|
934 |
-
else:
|
935 |
-
story.append(
|
936 |
-
{
|
937 |
-
"id": line_no,
|
938 |
-
"type": 0, # context
|
939 |
-
"supporting_ids": [],
|
940 |
-
"text": line_split[0].strip(),
|
941 |
-
"answer": "",
|
942 |
-
}
|
943 |
-
)
|
944 |
-
else: # After last line
|
945 |
-
if story != []:
|
946 |
-
yield example_idx, {"story": story}
|
|
|
15 |
"""The bAbI tasks dataset."""
|
16 |
|
17 |
|
|
|
|
|
18 |
import datasets
|
19 |
|
20 |
|
|
|
849 |
def _split_generators(self, dl_manager):
|
850 |
"""Returns SplitGenerators."""
|
851 |
my_urls = ZIP_URL
|
852 |
+
archive = dl_manager.download(my_urls)
|
853 |
splits = [
|
854 |
datasets.SplitGenerator(
|
855 |
name=datasets.Split.TRAIN,
|
856 |
# These kwargs will be passed to _generate_examples
|
857 |
gen_kwargs={
|
858 |
+
"filepath": paths[self.config.type][self.config.task_no]["train"],
|
859 |
+
"files": dl_manager.iter_archive(archive),
|
860 |
},
|
861 |
),
|
862 |
datasets.SplitGenerator(
|
863 |
name=datasets.Split.TEST,
|
864 |
# These kwargs will be passed to _generate_examples
|
865 |
gen_kwargs={
|
866 |
+
"filepath": paths[self.config.type][self.config.task_no]["test"],
|
867 |
+
"files": dl_manager.iter_archive(archive),
|
868 |
},
|
869 |
),
|
870 |
]
|
|
|
874 |
name=datasets.Split.VALIDATION,
|
875 |
# These kwargs will be passed to _generate_examples
|
876 |
gen_kwargs={
|
877 |
+
"filepath": paths[self.config.type][self.config.task_no]["valid"],
|
878 |
+
"files": dl_manager.iter_archive(archive),
|
879 |
},
|
880 |
),
|
881 |
]
|
882 |
return splits
|
883 |
|
884 |
+
def _generate_examples(self, filepath, files):
|
885 |
|
886 |
+
for path, f in files:
|
887 |
+
if path == filepath:
|
888 |
+
story = []
|
889 |
+
example_idx = 0
|
890 |
+
for idx, line in enumerate(f):
|
891 |
+
line = line.decode("utf-8")
|
892 |
+
if line.strip() == "":
|
893 |
+
if story != []:
|
894 |
+
yield example_idx, {"story": story}
|
895 |
+
example_idx += 1
|
896 |
+
story = []
|
897 |
+
elif line.strip().split()[0] == "1": # New story
|
898 |
+
if story != []: # Already some story, flush it out
|
899 |
+
yield example_idx, {"story": story}
|
900 |
+
example_idx += 1
|
901 |
+
story = []
|
902 |
+
line_no = line.split()[0]
|
903 |
+
line_split = line[len(line_no) :].strip().split("\t")
|
904 |
+
if len(line_split) > 1:
|
905 |
+
story.append(
|
906 |
+
{
|
907 |
+
"id": line_no,
|
908 |
+
"type": 1, # question
|
909 |
+
"supporting_ids": line_split[-1].split(" "),
|
910 |
+
"text": line_split[0].strip(),
|
911 |
+
"answer": line_split[1].strip(),
|
912 |
+
}
|
913 |
+
)
|
914 |
+
else:
|
915 |
+
story.append(
|
916 |
+
{
|
917 |
+
"id": line_no,
|
918 |
+
"type": 0, # context
|
919 |
+
"supporting_ids": [],
|
920 |
+
"text": line_split[0].strip(),
|
921 |
+
"answer": "",
|
922 |
+
}
|
923 |
+
)
|
924 |
+
else:
|
925 |
+
line_no = line.split()[0]
|
926 |
+
line_split = line[len(line_no) :].strip().split("\t")
|
927 |
+
if len(line_split) > 1:
|
928 |
+
story.append(
|
929 |
+
{
|
930 |
+
"id": line_no,
|
931 |
+
"type": 1, # question
|
932 |
+
"supporting_ids": line_split[-1].split(" "),
|
933 |
+
"text": line_split[0].strip(),
|
934 |
+
"answer": line_split[1].strip(),
|
935 |
+
}
|
936 |
+
)
|
937 |
+
else:
|
938 |
+
story.append(
|
939 |
+
{
|
940 |
+
"id": line_no,
|
941 |
+
"type": 0, # context
|
942 |
+
"supporting_ids": [],
|
943 |
+
"text": line_split[0].strip(),
|
944 |
+
"answer": "",
|
945 |
+
}
|
946 |
+
)
|
947 |
+
else: # After last line
|
948 |
if story != []:
|
949 |
yield example_idx, {"story": story}
|
950 |
+
break
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|