Upload normalizers.py with huggingface_hub
Browse files- normalizers.py +26 -0
normalizers.py
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from .operator import StreamInstanceOperator
|
| 2 |
+
|
| 3 |
+
from typing import List, Dict, Any
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
class NormalizeListFields(StreamInstanceOperator):
|
| 7 |
+
fields: List[str]
|
| 8 |
+
key_prefix: str = ""
|
| 9 |
+
empty_value: str = ""
|
| 10 |
+
separator: str = ", "
|
| 11 |
+
|
| 12 |
+
def process(self, instance: Dict[str, Any], stream_name: str = None) -> Dict[str, Any]:
|
| 13 |
+
for field in self.fields:
|
| 14 |
+
assert field in instance, f"Field {field} not found in instance {instance}"
|
| 15 |
+
assert isinstance(instance[field], list), f"Field {field} should be a list, got {type(instance[field])}"
|
| 16 |
+
|
| 17 |
+
target_key = self.key_prefix + field
|
| 18 |
+
|
| 19 |
+
if len(instance[field]) == 0:
|
| 20 |
+
instance[target_key] = self.empty_value
|
| 21 |
+
elif len(instance[field]) == 1:
|
| 22 |
+
instance[target_key] = instance[field][0]
|
| 23 |
+
else:
|
| 24 |
+
instance[target_key] = self.separator.join(instance[field])
|
| 25 |
+
|
| 26 |
+
return instance
|