Update app.py

app.py

@@ -1,6 +1,7 @@
 import streamlit as st
 from datasets import load_dataset, concatenate_datasets
 import json
+import os
 
 def load_and_combine_datasets():
     python_codes_dataset = load_dataset('flytech/python-codes-25k', split='train')
@@ -18,7 +19,7 @@ def save_combined_dataset_as_jsonl(combined_dataset, file_path):
             f.write('\n')
 
 def main():
-    st.title("Combined Dataset Viewer")
+    st.title("Combined Dataset Viewer and Downloader")
 
     # Load and combine datasets
     combined_dataset = load_and_combine_datasets()
@@ -26,18 +27,25 @@ def main():
     # Display a subset of the combined dataset
     st.write("Subset of Combined Dataset:", combined_dataset[:10])
 
+    # Take input for output dataset name
+    output_dataset_name = st.text_input("Enter output dataset name (without extension):", "combined_dataset")
+
     # Add option to save the combined dataset as JSONL
     if st.button("Save Combined Dataset (JSONL)"):
-        file_path = "combined_dataset.jsonl"
+        file_path = os.path.join(os.getcwd(), f"{output_dataset_name}.jsonl")
         save_combined_dataset_as_jsonl(combined_dataset, file_path)
         st.write(f"Combined dataset saved as JSONL file: {file_path}")
 
     # Add option to download the JSONL file
     if st.button("Download Combined Dataset (JSONL)"):
-        file_path = "combined_dataset.jsonl"
+        file_path = os.path.join(os.getcwd(), f"{output_dataset_name}.jsonl")
         save_combined_dataset_as_jsonl(combined_dataset, file_path)
         st.write("Download the combined dataset as JSONL file:")
-        st.download_button(label="Download", data=open(file_path, "rb"), file_name="combined_dataset.jsonl")
+        with open(file_path, "rb") as f:
+            bytes_data = f.read()
+        b64 = base64.b64encode(bytes_data).decode()
+        href = f'<a href="data:file/jsonl;base64,{b64}" download="{output_dataset_name}.jsonl">Download JSONL File</a>'
+        st.markdown(href, unsafe_allow_html=True)
 
 if __name__ == "__main__":
     main()