Datasets:
maybe this work?
Browse files- repobench-p.py +26 -4
repobench-p.py
CHANGED
@@ -125,8 +125,30 @@ class RepoBenchP(datasets.GeneratorBasedBuilder):
|
|
125 |
with gzip.open(data_dir[split], "rb") as f:
|
126 |
data = pickle.load(f)
|
127 |
|
128 |
-
|
129 |
-
|
130 |
-
|
131 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
132 |
|
|
|
125 |
with gzip.open(data_dir[split], "rb") as f:
|
126 |
data = pickle.load(f)
|
127 |
|
128 |
+
if split == "cff" or split == "cfr":
|
129 |
+
for i, example in enumerate(data):
|
130 |
+
|
131 |
+
yield i, {
|
132 |
+
"repo_name": example["repo_name"],
|
133 |
+
"file_path": example["file_path"],
|
134 |
+
"context": example["context"],
|
135 |
+
"import_statement": example["import_statement"],
|
136 |
+
"code": example["code"],
|
137 |
+
"next_line": example["next_line"],
|
138 |
+
"gold_snippet_index": example["gold_snippet_index"]
|
139 |
+
}
|
140 |
+
|
141 |
+
elif split == "if":
|
142 |
+
for i, example in enumerate(data):
|
143 |
+
|
144 |
+
yield i, {
|
145 |
+
"repo_name": example["repo_name"],
|
146 |
+
"file_path": example["file_path"],
|
147 |
+
"context": example["context"],
|
148 |
+
"import_statement": example["import_statement"],
|
149 |
+
"code": example["code"],
|
150 |
+
"next_line": example["next_line"],
|
151 |
+
# in-file data has no gold_snippet_index
|
152 |
+
"gold_snippet_index": -1
|
153 |
+
}
|
154 |
|