Recreated dataset using pyarrow to avoid polars type conversion problems

README.md

@@ -14,7 +14,9 @@ Unofficial port of https://github.com/uberspot/OpenTriviaQA
 Code to reproduce the dataset:
 ```
 import os
-import polars as pl
+import pandas as pd
+import pyarrow as pa
+import pyarrow.parquet as pq
 
 source_dir = "OpenTriviaQA/categories"
 topics = sorted(os.listdir(source_dir))
@@ -50,8 +52,8 @@ for topic in topics:
     questions = [process_question(x, topic, i) for (i, x) in enumerate(questions)]
     questions = [x for x in questions if x]
 
-    data = pl.DataFrame(questions)
+    data = pa.Table.from_pandas(pd.DataFrame(questions))
 
     os.makedirs(os.path.join(target_dir, topic), exist_ok=True) 
-    data.write_parquet(os.path.join(target_dir, topic, "data.parquet"))
+    pq.write_table(data, os.path.join(target_dir, topic, "data.parquet"))
 ```