Update quac_expanded.py
Browse files- quac_expanded.py +3 -3
quac_expanded.py
CHANGED
|
@@ -130,17 +130,17 @@ class Quac(datasets.GeneratorBasedBuilder):
|
|
| 130 |
section_title = section.get("section_title", "").strip()
|
| 131 |
|
| 132 |
for dialogue in section["paragraphs"]:
|
| 133 |
-
context = dialogue["context"].strip()
|
| 134 |
dialogue_id = dialogue["id"]
|
| 135 |
|
| 136 |
for turn in dialogue["qas"]:
|
| 137 |
answer = {
|
| 138 |
-
"texts": [t["text"].strip() for t in turn["answers"]],
|
| 139 |
"answer_starts": [t["answer_start"] for t in turn["answers"]],
|
| 140 |
}
|
| 141 |
|
| 142 |
orig_answer = {
|
| 143 |
-
"text": turn["orig_answer"]["text"],
|
| 144 |
"answer_start": turn["orig_answer"]["answer_start"],
|
| 145 |
}
|
| 146 |
|
|
|
|
| 130 |
section_title = section.get("section_title", "").strip()
|
| 131 |
|
| 132 |
for dialogue in section["paragraphs"]:
|
| 133 |
+
context = dialogue["context"].replace("CANNOTANSWER","").strip()
|
| 134 |
dialogue_id = dialogue["id"]
|
| 135 |
|
| 136 |
for turn in dialogue["qas"]:
|
| 137 |
answer = {
|
| 138 |
+
"texts": [t["text"].replace("CANNOTANSWER","unknown").strip() for t in turn["answers"]],
|
| 139 |
"answer_starts": [t["answer_start"] for t in turn["answers"]],
|
| 140 |
}
|
| 141 |
|
| 142 |
orig_answer = {
|
| 143 |
+
"text": turn["orig_answer"]["text"].replace("CANNOTANSWER","unknown"),
|
| 144 |
"answer_start": turn["orig_answer"]["answer_start"],
|
| 145 |
}
|
| 146 |
|