Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -14,20 +14,16 @@ def convert_parquet_to_jsonl(parquet_file=None, parquet_url=None):
|
|
14 |
else:
|
15 |
raise ValueError("Either parquet_file or parquet_url must be provided")
|
16 |
|
17 |
-
#
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
return val
|
23 |
|
24 |
-
#
|
25 |
-
df = df.applymap(clean_string)
|
26 |
-
|
27 |
-
# Now safely convert to JSON Lines
|
28 |
jsonl_data = df.to_json(orient="records", lines=True)
|
29 |
-
|
30 |
-
# Write
|
31 |
output_file_path = "output.jsonl"
|
32 |
with open(output_file_path, "w", encoding="utf-8") as f:
|
33 |
f.write(jsonl_data)
|
|
|
14 |
else:
|
15 |
raise ValueError("Either parquet_file or parquet_url must be provided")
|
16 |
|
17 |
+
# Clean string columns to replace invalid UTF-8 sequences
|
18 |
+
for col in df.select_dtypes(include=["object"]).columns:
|
19 |
+
df[col] = df[col].apply(
|
20 |
+
lambda x: x.encode("utf-8", errors="replace").decode("utf-8", errors="replace") if isinstance(x, str) else x
|
21 |
+
)
|
|
|
22 |
|
23 |
+
# Convert to JSON Lines
|
|
|
|
|
|
|
24 |
jsonl_data = df.to_json(orient="records", lines=True)
|
25 |
+
|
26 |
+
# Write the output to a file using UTF-8 encoding explicitly
|
27 |
output_file_path = "output.jsonl"
|
28 |
with open(output_file_path, "w", encoding="utf-8") as f:
|
29 |
f.write(jsonl_data)
|