Datasets:

Modalities:
Text
Formats:
parquet
Languages:
Norwegian
Libraries:
Datasets
pandas
SushantGautam commited on
Commit
ce96a8b
·
verified ·
1 Parent(s): e741f1b

Update sync_from_json.py

Browse files
Files changed (1) hide show
  1. sync_from_json.py +20 -9
sync_from_json.py CHANGED
@@ -1,14 +1,25 @@
1
- from datasets import Dataset, DatasetDict, load_dataset
2
  from huggingface_hub import login
3
-
4
- # Load the JSON file
5
  import json
6
- dataset = load_dataset("json", data_files="norwegian_medical_qa_v2.json", split="train")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7
 
8
- # Optional: Add metadata or manipulate dataset
9
- dataset = dataset.rename_column("question_text", "question")
10
- dataset = dataset.rename_column("answer_text", "correct_answer")
11
- dataset = dataset.rename_column("wrong_answers_text", "wrong_answers")
12
 
13
- # Push to HF Hub
14
  dataset.push_to_hub("SimulaMet/NorMedQA")
 
1
+ from datasets import Dataset
2
  from huggingface_hub import login
 
 
3
  import json
4
+ # Load and process JSON data
5
+ with open("norwegian_medical_qa_v2.json") as f:
6
+ raw_data = json.load(f)
7
+
8
+ # Transform the data
9
+ processed_data = []
10
+ for item in raw_data:
11
+ wrong_answers = item["wrong_answers_text"]
12
+ # wrong_answers= [ans.strip() for ans in wrong_answers_raw.split(";")]
13
+ processed_data.append({
14
+ "question": item["question_text"],
15
+ "answer": item["answer_text"],
16
+ "wrong_answers": wrong_answers,
17
+ "source_file": item["source_file"],
18
+ "question_number": item["question_number"]
19
+ })
20
 
21
+ # Create Hugging Face Dataset
22
+ dataset = Dataset.from_list(processed_data)
 
 
23
 
24
+ # Push to the Hub
25
  dataset.push_to_hub("SimulaMet/NorMedQA")