init
Browse files
kowiki.py
CHANGED
@@ -60,6 +60,8 @@ def replace_link(text):
|
|
60 |
|
61 |
|
62 |
def process_text(text: str):
|
|
|
|
|
63 |
text = text.strip()
|
64 |
text = re.sub("<[^>]+>", "", text)
|
65 |
text = re.sub("\[\[파일:[^\]]+\]\]", "", text)
|
@@ -143,19 +145,19 @@ class KoWiki(datasets.GeneratorBasedBuilder):
|
|
143 |
]
|
144 |
|
145 |
def _generate_examples(self, data_file):
|
146 |
-
n = 0
|
147 |
-
output = {}
|
148 |
if data_file.startswith(("https://", "http://")):
|
149 |
data_file = requests.get(data_file, stream=True).raw
|
150 |
|
|
|
151 |
for event, node in ET.iterparse(data_file):
|
152 |
-
|
|
|
153 |
output["title"] = node.text
|
154 |
-
|
155 |
-
|
156 |
-
text =
|
157 |
-
|
158 |
-
|
159 |
output["id"] = f"{n:08d}"
|
160 |
yield n, output
|
161 |
output = {}
|
|
|
60 |
|
61 |
|
62 |
def process_text(text: str):
|
63 |
+
if not isinstance(text, str):
|
64 |
+
return ""
|
65 |
text = text.strip()
|
66 |
text = re.sub("<[^>]+>", "", text)
|
67 |
text = re.sub("\[\[파일:[^\]]+\]\]", "", text)
|
|
|
145 |
]
|
146 |
|
147 |
def _generate_examples(self, data_file):
|
|
|
|
|
148 |
if data_file.startswith(("https://", "http://")):
|
149 |
data_file = requests.get(data_file, stream=True).raw
|
150 |
|
151 |
+
n, output = 0, {}
|
152 |
for event, node in ET.iterparse(data_file):
|
153 |
+
tag_name = node.tag
|
154 |
+
if tag_name.endswith("title"):
|
155 |
output["title"] = node.text
|
156 |
+
|
157 |
+
elif tag_name.endswith("text"):
|
158 |
+
output["text"] = process_text(node.text)
|
159 |
+
|
160 |
+
elif tag_name.endswith("page"):
|
161 |
output["id"] = f"{n:08d}"
|
162 |
yield n, output
|
163 |
output = {}
|