Spaces:
Sleeping
Sleeping
convert df to datasetdict
Browse files- src/utils.py +14 -0
src/utils.py
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from datasets import DatasetDict, Dataset
|
2 |
+
import pandas as pd
|
3 |
+
|
4 |
+
|
5 |
+
def convert_df_to_dataset_dict(df_train:pd.DataFrame, df_test:pd.DataFrame)-> DatasetDict:
|
6 |
+
mapper = {"#2_tweet": "tweet", "#3_country_label": "label"}
|
7 |
+
df_train = df_train.rename(columns=mapper)
|
8 |
+
df_test = df_test.rename(columns=mapper)
|
9 |
+
print(df_train.columns)
|
10 |
+
columns_to_keep = ["tweet", "label"]
|
11 |
+
train_dataset = Dataset.from_pandas(df_train[columns_to_keep])
|
12 |
+
test_dataset = Dataset.from_pandas(df_test[columns_to_keep])
|
13 |
+
|
14 |
+
return DatasetDict({'train': train_dataset, 'test': test_dataset})
|