Update app.py

app.py

@@ -1,512 +1,512 @@
-import gradio as gr
-import mtdna_backend
-import json
-from iterate3 import data_preprocess, model, pipeline
-import os
-import hashlib
-import threading
-# Gradio UI
-#stop_flag = gr.State(value=False)
-class StopFlag:
-    def __init__(self):
-        self.value = False
-global_stop_flag = StopFlag()  # Shared between run + stop
-
-with gr.Blocks() as interface:
-    gr.Markdown("# 🧬 mtDNA Location Classifier (MVP)")
-
-    #inputMode = gr.Radio(choices=["Single Accession", "Batch Input"], value="Single Accession", label="Choose Input Mode")
-    user_email = gr.Textbox(label="📧 Your email (used to track free quota)")
-    usage_display = gr.Markdown("", visible=False)
-
-    # with gr.Group() as single_input_group:
-    #     single_accession = gr.Textbox(label="Enter Single Accession (e.g., KU131308)")
-
-    # with gr.Group(visible=False) as batch_input_group:
-    #     raw_text = gr.Textbox(label="🧬 Paste Accession Numbers (e.g., MF362736.1,MF362738.1,KU131308,MW291678)")
-    #     resume_file = gr.File(label="🗃️ Previously saved Excel output (optional)", file_types=[".xlsx"], interactive=True)
-    #     gr.HTML("""<a href="https://drive.google.com/file/d/1t-TFeIsGVu5Jh3CUZS-VE9jQWzNFCs_c/view?usp=sharing" download target="_blank">Download Example CSV Format</a>""")
-    #     gr.HTML("""<a href="https://docs.google.com/spreadsheets/d/1lKqPp17EfHsshJGZRWEpcNOZlGo3F5qU/edit?usp=sharing&ouid=112390323314156876153&rtpof=true&sd=true" download target="_blank">Download Example Excel Format</a>""")
-    #     file_upload = gr.File(label="📁 Or Upload CSV/Excel File", file_types=[".csv", ".xlsx"], interactive=True, elem_id="file-upload-box")
-    raw_text = gr.Textbox(label="🧚 Input Accession Number(s) (single (KU131308) or comma-separated (e.g., MF362736.1,MF362738.1,KU131308,MW291678))")
-    #resume_file = gr.File(label="🗃️ Previously saved Excel output (optional)", file_types=[".xlsx"], interactive=True)
-    gr.HTML("""<a href="https://docs.google.com/spreadsheets/d/1lKqPp17EfHsshJGZRWEpcNOZlGo3F5qU/edit?usp=sharing" download target="_blank">Download Example Excel Format</a>""")
-    file_upload = gr.File(label="📁 Or Upload CSV/Excel File", file_types=[".csv", ".xlsx"], interactive=True)
-
-    with gr.Row():
-        run_button = gr.Button("🔍 Submit and Classify")
-        stop_button = gr.Button("❌ Stop Batch", visible=True)
-        reset_button = gr.Button("🔄 Reset")
-
-    status = gr.Markdown(visible=False)
-    
-    with gr.Group(visible=False) as results_group:
-      # with gr.Accordion("Open to See the Result", open=False) as results:  
-      #     with gr.Row():
-      #         output_summary = gr.Markdown(elem_id="output-summary")
-      #         output_flag = gr.Markdown(elem_id="output-flag")
-          
-      #     gr.Markdown("---")
-
-      with gr.Accordion("Open to See the Output Table", open=False) as table_accordion:    
-          output_table = gr.HTML(render=True)
-
-      with gr.Row():
-          output_type = gr.Dropdown(choices=["Excel", "JSON", "TXT"], label="Select Output Format", value="Excel")
-          download_button = gr.Button("⬇️ Download Output")
-      #download_file = gr.File(label="Download File Here",visible=False)
-      download_file = gr.File(label="Download File Here", visible=False, interactive=True)
-      progress_box = gr.Textbox(label="Live Processing Log", lines=20, interactive=False)
-
-      gr.Markdown("---")
-
-      gr.Markdown("### 💬 Feedback (required)")
-      q1 = gr.Textbox(label="1️⃣ Was the inferred location accurate or helpful? Please explain.")
-      q2 = gr.Textbox(label="2️⃣ What would improve your experience with this tool?")
-      contact = gr.Textbox(label="📧 Your email or institution (optional)")
-      submit_feedback = gr.Button("✅ Submit Feedback")
-      feedback_status = gr.Markdown()
-
-    # Functions
-    # def toggle_input_mode(mode):
-    #     if mode == "Single Accession":
-    #         return gr.update(visible=True), gr.update(visible=False)
-    #     else:
-    #         return gr.update(visible=False), gr.update(visible=True)
-
-    def classify_with_loading():
-        return gr.update(value="⏳ Please wait... processing...",visible=True)  # Show processing message
-         
-    # def classify_dynamic(single_accession, file, text, resume, email, mode):
-    #     if mode == "Single Accession":
-    #         return classify_main(single_accession)  + (gr.update(visible=False),)
-    #     else:
-    #         #return summarize_batch(file, text) + (gr.update(visible=False),)  # Hide processing message
-    #         return classify_mulAcc(file, text, resume) + (gr.update(visible=False),)  # Hide processing message
-    # Logging helpers defined early to avoid NameError
-    
-
-    # def classify_dynamic(single_accession, file, text, resume, email, mode):
-    #   if mode == "Single Accession":
-    #       return classify_main(single_accession) + (gr.update(value="", visible=False),)
-    #   else:
-    #       return classify_mulAcc(file, text, resume, email, log_callback=real_time_logger, log_collector=log_collector)
-
-    # for single accession
-    # def classify_main(accession):
-    #     #table, summary, labelAncient_Modern, explain_label = mtdna_backend.summarize_results(accession)
-    #     table = mtdna_backend.summarize_results(accession)
-    #     #flag_output = f"### 🏺 Ancient/Modern Flag\n**{labelAncient_Modern}**\n\n_Explanation:_ {explain_label}"
-    #     return (
-    #         #table,
-    #         make_html_table(table),
-    #         # summary,
-    #         # flag_output,
-    #         gr.update(visible=True),
-    #         gr.update(visible=False),
-    #         gr.update(visible=False)
-    #     )
-    
-    #stop_flag = gr.State(value=False)
-    #stop_flag = StopFlag()
-
-    # def stop_batch(stop_flag):
-    #   stop_flag.value = True
-    #   return gr.update(value="❌ Stopping...", visible=True), stop_flag
-    def stop_batch():
-      global_stop_flag.value = True
-      return gr.update(value="❌ Stopping...", visible=True)
-
-    # def threaded_batch_runner(file, text, email):
-    #   global_stop_flag.value = False
-    #   log_lines = []
-
-    #   def update_log(line):
-    #       log_lines.append(line)
-    #       yield (
-    #           gr.update(visible=False),  # output_table (not yet)
-    #           gr.update(visible=False),  # results_group
-    #           gr.update(visible=False),  # download_file
-    #           gr.update(visible=False),  # usage_display
-    #           gr.update(value="⏳ Still processing...", visible=True),  # status
-    #           gr.update(value="\n".join(log_lines))  # progress_box
-    #       )
-
-    #   # Start a dummy update to say "Starting..."
-    #   yield from update_log("🚀 Starting batch processing...")
-
-    #   rows, file_path, count, final_log, warning = mtdna_backend.summarize_batch(
-    #       file=file,
-    #       raw_text=text,
-    #       resume_file=None,
-    #       user_email=email,
-    #       stop_flag=global_stop_flag,
-    #       yield_callback=lambda line: (yield from update_log(line))
-    #   )
-
-    #   html = make_html_table(rows)
-    #   file_update = gr.update(value=file_path, visible=True) if os.path.exists(file_path) else gr.update(visible=False)
-    #   usage_or_warning_text = f"**{count}** samples used by this email." if email.strip() else warning
-
-    #   yield (
-    #       html,
-    #       gr.update(visible=True),        # results_group
-    #       file_update,                    # download_file
-    #       gr.update(value=usage_or_warning_text, visible=True),
-    #       gr.update(value="✅ Done", visible=True),
-    #       gr.update(value=final_log)
-    #   )
-    
-    def threaded_batch_runner(file=None, text="", email=""):
-      print("📧 EMAIL RECEIVED:", email)
-      import tempfile
-      from mtdna_backend import (
-          extract_accessions_from_input,
-          summarize_results,
-          save_to_excel,
-          hash_user_id,
-          increment_usage,
-      )
-      import os
-
-      global_stop_flag.value = False  # reset stop flag
-
-      tmp_dir = tempfile.mkdtemp()
-      output_file_path = os.path.join(tmp_dir, "batch_output_live.xlsx")
-      limited_acc = 50 + (10 if email.strip() else 0)
-
-      # Step 1: Parse input
-      accessions, error = extract_accessions_from_input(file, text)
-      if error:
-          yield (
-              "",                          # output_table
-              gr.update(visible=False),   # results_group
-              gr.update(visible=False),   # download_file
-              "",                          # usage_display
-              "❌ Error",                  # status
-              str(error)                        # progress_box
-          )
-          return
-
-      total = len(accessions)
-      if total > limited_acc:
-          accessions = accessions[:limited_acc]
-          warning = f"⚠️ Only processing first {limited_acc} accessions."
-      else:
-          warning = f"✅ All {total} accessions will be processed."
-
-      all_rows = []
-      log_lines = []
-
-      # Step 2: Loop through accessions
-      for i, acc in enumerate(accessions):
-          if global_stop_flag.value:
-              log_lines.append(f"🛑 Stopped at {acc} ({i+1}/{total})")
-              usage_text = ""
-              if email.strip():
-                  # user_hash = hash_user_id(email)
-                  # usage_count = increment_usage(user_hash, len(all_rows))
-                  usage_count = increment_usage(email, len(all_rows))
-                  usage_text = f"**{usage_count}** samples used by this email. Ten more samples are added first (you now have 60 limited accessions), then wait we will contact you via this email."
-              else:
-                  usage_text = f"The limited accession is 50. The user has used {len(all_rows)}, and only {50-len(all_rows)} left."    
-              yield (
-                  make_html_table(all_rows),
-                  gr.update(visible=True),
-                  gr.update(value=output_file_path, visible=True),
-                  gr.update(value=usage_text, visible=True),
-                  "🛑 Stopped",
-                  "\n".join(log_lines)
-              )
-              return
-
-          log_lines.append(f"[{i+1}/{total}] Processing {acc}")
-          yield (
-              make_html_table(all_rows),
-              gr.update(visible=True),
-              gr.update(visible=False),
-              "",
-              "⏳ Processing...",
-              "\n".join(log_lines)
-          )
-
-          try:
-              rows = summarize_results(acc)
-              all_rows.extend(rows)
-              save_to_excel(all_rows, "", "", output_file_path, is_resume=False)
-              log_lines.append(f"✅ Processed {acc} ({i+1}/{total})")
-          except Exception as e:
-              log_lines.append(f"❌ Failed to process {acc}: {e}")
-
-          yield (
-              make_html_table(all_rows),
-              gr.update(visible=True),
-              gr.update(visible=False),
-              "",
-              "⏳ Processing...",
-              "\n".join(log_lines)
-          )
-
-      # Final update
-      usage_text = ""
-      if email.strip():
-          # user_hash = hash_user_id(email)
-          # usage_count = increment_usage(user_hash, len(all_rows))
-          usage_count = increment_usage(email, len(all_rows))
-          usage_text = f"**{usage_count}** samples used by this email. Ten more samples are added first (you now have 60 limited accessions), then wait we will contact you via this email."
-      else:
-          usage_text = f"The limited accession is 50. The user has used {len(all_rows)}, and only {50-len(all_rows)} left."
-      yield (
-          make_html_table(all_rows),
-          gr.update(visible=True),
-          gr.update(value=output_file_path, visible=True),
-          gr.update(value=usage_text, visible=True),
-          "✅ Done",
-          "\n".join(log_lines)
-      )
-
-    # def threaded_batch_runner(file=None, text="", email=""):
-    #   global_stop_flag.value = False
-
-    #   # Dummy test output that matches expected schema
-    #   return (
-    #       "<div>✅ Dummy output table</div>",   # HTML string
-    #       gr.update(visible=True),             # Group visibility
-    #       gr.update(visible=False),            # Download file
-    #       "**0** samples used.",               # Markdown
-    #       "✅ Done",                           # Status string
-    #       "Processing finished."               # Progress string
-    #   )
-
-    
-    # def classify_mulAcc(file, text, resume, email, log_callback=None, log_collector=None):
-    #     stop_flag.value = False
-    #     return threaded_batch_runner(file, text, resume, email, status, stop_flag, log_callback=log_callback, log_collector=log_collector)
-    
-
-    def make_html_table(rows):
-      html = """
-      <div style='overflow-x: auto; padding: 10px;'>
-          <div style='max-height: 400px; overflow-y: auto; border: 1px solid #444; border-radius: 8px;'>
-              <table style='width:100%; border-collapse: collapse; table-layout: auto; font-size: 14px; color: #f1f1f1; background-color: #1e1e1e;'>
-                  <thead style='position: sticky; top: 0; background-color: #2c2c2c; z-index: 1;'>
-                      <tr>
-      """
-      headers = ["Sample ID", "Predicted Country", "Country Explanation", "Predicted Sample Type", "Sample Type Explanation", "Sources", "Time cost"]
-      html += "".join(
-          f"<th style='padding: 10px; border: 1px solid #555; text-align: left; white-space: nowrap;'>{h}</th>"
-          for h in headers
-      )
-      html += "</tr></thead><tbody>"
-
-      for row in rows:
-          html += "<tr>"
-          for i, col in enumerate(row):
-              header = headers[i]
-              style = "padding: 10px; border: 1px solid #555; vertical-align: top;"
-
-              # For specific columns like Haplogroup, force nowrap
-              if header in ["Country Explanation", "Sample Type Explanation"]:
-                style += " max-width: 400px; word-wrap: break-word; white-space: normal;"
-              elif header in ["Sample ID", "Predicted Country", "Predicted Sample Type", "Time cost"]:
-                  style += " white-space: nowrap; text-overflow: ellipsis; max-width: 200px; overflow: hidden;"
-
-              # if header == "Sources" and isinstance(col, str) and col.strip().lower().startswith("http"):
-              #     col = f"<a href='{col}' target='_blank' style='color: #4ea1f3; text-decoration: underline;'>{col}</a>"
-
-              #html += f"<td style='{style}'>{col}</td>"
-              if header == "Sources" and isinstance(col, str):
-                  links = [f"<a href='{url.strip()}' target='_blank' style='color: #4ea1f3; text-decoration: underline;'>{url.strip()}</a>" for url in col.strip().split("\n") if url.strip()]
-                  col = "- "+"<br>- ".join(links)
-              elif isinstance(col, str):
-                  # lines = []
-                  # for line in col.split("\n"):
-                  #     line = line.strip()
-                  #     if not line:
-                  #         continue
-                  #     if line.lower().startswith("rag_llm-"):
-                  #         content = line[len("rag_llm-"):].strip()
-                  #         line = f"{content} (Method: RAG_LLM)"
-                  #     lines.append(f"- {line}")
-                  col = col.replace("\n", "<br>")
-                  #col = col.replace("\t", "&nbsp;&nbsp;&nbsp;&nbsp;")
-                  #col = "<br>".join(lines)
-
-              html += f"<td style='{style}'>{col}</td>"
-          html += "</tr>"
-
-      html += "</tbody></table></div></div>"
-      return html
-  
-
-    # def reset_fields():
-    #     global_stop_flag.value = False  # 💡 Add this to reset the flag
-    #     return (
-    #         #gr.update(value=""),  # single_accession
-    #         gr.update(value=""),  # raw_text
-    #         gr.update(value=None), # file_upload
-    #         #gr.update(value=None),  # resume_file
-    #         #gr.update(value="Single Accession"), # inputMode
-    #         gr.update(value=[], visible=True), # output_table
-    #         # gr.update(value="", visible=True), # output_summary
-    #         # gr.update(value="", visible=True), # output_flag
-    #         gr.update(visible=False), # status
-    #         gr.update(visible=False),  # results_group
-    #         gr.update(value="", visible=False),  # usage_display
-    #         gr.update(value="", visible=False),  # progress_box
-    #     )
-    def reset_fields():
-      global_stop_flag.value = False  # Reset the stop flag
-
-      return (
-          gr.update(value=""),          # raw_text
-          gr.update(value=None),        # file_upload
-          gr.update(value=[], visible=True),   # output_table
-          gr.update(value="", visible=True),   # status — reset and make visible again
-          gr.update(visible=False),     # results_group
-          gr.update(value="", visible=True),   # usage_display — reset and make visible again
-          gr.update(value="", visible=True),   # progress_box — reset AND visible!
-      )
-    #inputMode.change(fn=toggle_input_mode, inputs=inputMode, outputs=[single_input_group, batch_input_group])
-    #run_button.click(fn=classify_with_loading, inputs=[], outputs=[status])
-    # run_button.click(
-    #     fn=classify_dynamic,
-    #     inputs=[single_accession, file_upload, raw_text, resume_file,user_email,inputMode],
-    #     outputs=[output_table, 
-    #     #output_summary, output_flag, 
-    #     results_group, download_file, usage_display,status, progress_box]
-    # )
-    
-    # run_button.click(
-    #     fn=threaded_batch_runner,
-    #     #inputs=[file_upload, raw_text, resume_file, user_email],
-    #     inputs=[file_upload, raw_text, user_email],
-    #     outputs=[output_table, results_group, download_file, usage_display, status, progress_box]
-    # )
-#     run_button.click(
-#     fn=threaded_batch_runner,
-#     inputs=[file_upload, raw_text, user_email],
-#     outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
-#     every=0.5  # <-- this tells Gradio to expect streaming
-# )
-    # output_table = gr.HTML()
-    # results_group = gr.Group(visible=False)
-    # download_file = gr.File(visible=False)
-    # usage_display = gr.Markdown(visible=False)
-    # status = gr.Markdown(visible=False)
-    # progress_box = gr.Textbox(visible=False)
-
-#     run_button.click(
-#     fn=threaded_batch_runner,
-#     inputs=[file_upload, raw_text, user_email],
-#     outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
-#     every=0.5,  # streaming enabled
-#     show_progress="full"
-# )
-    print("🎯 DEBUG COMPONENT TYPES")
-    print(type(output_table))
-    print(type(results_group))
-    print(type(download_file))
-    print(type(usage_display))
-    print(type(status))
-    print(type(progress_box))
-    
-
-    # interface.stream(
-    #     fn=threaded_batch_runner,
-    #     inputs=[file_upload, raw_text, user_email],
-    #     outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
-    #     trigger=run_button,
-    #     every=0.5,
-    #     show_progress="full",
-    # )
-    interface.queue()  # No arguments here!
-
-    run_button.click(
-        fn=threaded_batch_runner,
-        inputs=[file_upload, raw_text, user_email],
-        outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
-        concurrency_limit=1,  # ✅ correct in Gradio 5.x
-        queue=True,            # ✅ ensure the queue is used
-        #every=0.5
-    )
-    
-
-
-
-    stop_button.click(fn=stop_batch, inputs=[], outputs=[status])
-
-    # reset_button.click(
-    #     #fn=reset_fields,
-    #     fn=lambda: (
-    #         gr.update(value=""), gr.update(value=""), gr.update(value=None), gr.update(value=None), gr.update(value="Single Accession"),
-    #         gr.update(value=[], visible=True), gr.update(visible=False), gr.update(visible=False), gr.update(value="", visible=False), gr.update(value="", visible=False)
-    #     ),
-    #     inputs=[],
-    #     outputs=[
-    #         single_accession, raw_text, file_upload, resume_file,inputMode,
-    #         output_table,# output_summary, output_flag,
-    #         status, results_group, usage_display, progress_box
-    #     ]
-    # )
-    #stop_button.click(fn=lambda sf: (gr.update(value="❌ Stopping...", visible=True), setattr(sf, "value", True) or sf), inputs=[gr.State(stop_flag)], outputs=[status, gr.State(stop_flag)])
-
-    reset_button.click(
-        fn=reset_fields,
-        inputs=[],
-        #outputs=[raw_text, file_upload, resume_file, output_table, status, results_group, usage_display, progress_box]
-        outputs=[raw_text, file_upload, output_table, status, results_group, usage_display, progress_box]
-    )  
-
-    download_button.click(
-      fn=mtdna_backend.save_batch_output, 
-      #inputs=[output_table, output_summary, output_flag, output_type], 
-      inputs=[output_table, output_type], 
-      outputs=[download_file])
-
-    # submit_feedback.click(
-    #     fn=mtdna_backend.store_feedback_to_google_sheets, 
-    #     inputs=[single_accession, q1, q2, contact], outputs=feedback_status
-    # )
-    submit_feedback.click(
-        fn=mtdna_backend.store_feedback_to_google_sheets,
-        inputs=[raw_text, q1, q2, contact],
-        outputs=[feedback_status]
-    )
-    #     # Custom CSS styles
-    # gr.HTML("""
-    # <style>
-    #   /* Ensures both sections are equally spaced with the same background size */
-    #   #output-summary, #output-flag {
-    #       background-color: #f0f4f8; /* Light Grey for both */
-    #       padding: 20px;
-    #       border-radius: 10px;
-    #       margin-top: 10px;
-    #       width: 100%; /* Ensure full width */
-    #       min-height: 150px; /* Ensures both have a minimum height */
-    #       box-sizing: border-box; /* Prevents padding from increasing size */
-    #       display: flex;
-    #       flex-direction: column;
-    #       justify-content: space-between;
-    #   }
-      
-    #   /* Specific background colors */
-    #   #output-summary {
-    #       background-color: #434a4b; 
-    #   }
-
-    #   #output-flag {
-    #       background-color: #141616; 
-    #   }
-
-    #   /* Ensuring they are in a row and evenly spaced */
-    #   .gradio-row {
-    #       display: flex;
-    #       justify-content: space-between;
-    #       width: 100%;
-    #   }
-    # </style>
-    # """)
-
-
+import gradio as gr
+import mtdna_backend
+import json
+import data_preprocess, model, pipeline
+import os
+import hashlib
+import threading
+# Gradio UI
+#stop_flag = gr.State(value=False)
+class StopFlag:
+    def __init__(self):
+        self.value = False
+global_stop_flag = StopFlag()  # Shared between run + stop
+
+with gr.Blocks() as interface:
+    gr.Markdown("# 🧬 mtDNA Location Classifier (MVP)")
+
+    #inputMode = gr.Radio(choices=["Single Accession", "Batch Input"], value="Single Accession", label="Choose Input Mode")
+    user_email = gr.Textbox(label="📧 Your email (used to track free quota)")
+    usage_display = gr.Markdown("", visible=False)
+
+    # with gr.Group() as single_input_group:
+    #     single_accession = gr.Textbox(label="Enter Single Accession (e.g., KU131308)")
+
+    # with gr.Group(visible=False) as batch_input_group:
+    #     raw_text = gr.Textbox(label="🧬 Paste Accession Numbers (e.g., MF362736.1,MF362738.1,KU131308,MW291678)")
+    #     resume_file = gr.File(label="🗃️ Previously saved Excel output (optional)", file_types=[".xlsx"], interactive=True)
+    #     gr.HTML("""<a href="https://drive.google.com/file/d/1t-TFeIsGVu5Jh3CUZS-VE9jQWzNFCs_c/view?usp=sharing" download target="_blank">Download Example CSV Format</a>""")
+    #     gr.HTML("""<a href="https://docs.google.com/spreadsheets/d/1lKqPp17EfHsshJGZRWEpcNOZlGo3F5qU/edit?usp=sharing&ouid=112390323314156876153&rtpof=true&sd=true" download target="_blank">Download Example Excel Format</a>""")
+    #     file_upload = gr.File(label="📁 Or Upload CSV/Excel File", file_types=[".csv", ".xlsx"], interactive=True, elem_id="file-upload-box")
+    raw_text = gr.Textbox(label="🧚 Input Accession Number(s) (single (KU131308) or comma-separated (e.g., MF362736.1,MF362738.1,KU131308,MW291678))")
+    #resume_file = gr.File(label="🗃️ Previously saved Excel output (optional)", file_types=[".xlsx"], interactive=True)
+    gr.HTML("""<a href="https://docs.google.com/spreadsheets/d/1lKqPp17EfHsshJGZRWEpcNOZlGo3F5qU/edit?usp=sharing" download target="_blank">Download Example Excel Format</a>""")
+    file_upload = gr.File(label="📁 Or Upload CSV/Excel File", file_types=[".csv", ".xlsx"], interactive=True)
+
+    with gr.Row():
+        run_button = gr.Button("🔍 Submit and Classify")
+        stop_button = gr.Button("❌ Stop Batch", visible=True)
+        reset_button = gr.Button("🔄 Reset")
+
+    status = gr.Markdown(visible=False)
+    
+    with gr.Group(visible=False) as results_group:
+      # with gr.Accordion("Open to See the Result", open=False) as results:  
+      #     with gr.Row():
+      #         output_summary = gr.Markdown(elem_id="output-summary")
+      #         output_flag = gr.Markdown(elem_id="output-flag")
+          
+      #     gr.Markdown("---")
+
+      with gr.Accordion("Open to See the Output Table", open=False) as table_accordion:    
+          output_table = gr.HTML(render=True)
+
+      with gr.Row():
+          output_type = gr.Dropdown(choices=["Excel", "JSON", "TXT"], label="Select Output Format", value="Excel")
+          download_button = gr.Button("⬇️ Download Output")
+      #download_file = gr.File(label="Download File Here",visible=False)
+      download_file = gr.File(label="Download File Here", visible=False, interactive=True)
+      progress_box = gr.Textbox(label="Live Processing Log", lines=20, interactive=False)
+
+      gr.Markdown("---")
+
+      gr.Markdown("### 💬 Feedback (required)")
+      q1 = gr.Textbox(label="1️⃣ Was the inferred location accurate or helpful? Please explain.")
+      q2 = gr.Textbox(label="2️⃣ What would improve your experience with this tool?")
+      contact = gr.Textbox(label="📧 Your email or institution (optional)")
+      submit_feedback = gr.Button("✅ Submit Feedback")
+      feedback_status = gr.Markdown()
+
+    # Functions
+    # def toggle_input_mode(mode):
+    #     if mode == "Single Accession":
+    #         return gr.update(visible=True), gr.update(visible=False)
+    #     else:
+    #         return gr.update(visible=False), gr.update(visible=True)
+
+    def classify_with_loading():
+        return gr.update(value="⏳ Please wait... processing...",visible=True)  # Show processing message
+         
+    # def classify_dynamic(single_accession, file, text, resume, email, mode):
+    #     if mode == "Single Accession":
+    #         return classify_main(single_accession)  + (gr.update(visible=False),)
+    #     else:
+    #         #return summarize_batch(file, text) + (gr.update(visible=False),)  # Hide processing message
+    #         return classify_mulAcc(file, text, resume) + (gr.update(visible=False),)  # Hide processing message
+    # Logging helpers defined early to avoid NameError
+    
+
+    # def classify_dynamic(single_accession, file, text, resume, email, mode):
+    #   if mode == "Single Accession":
+    #       return classify_main(single_accession) + (gr.update(value="", visible=False),)
+    #   else:
+    #       return classify_mulAcc(file, text, resume, email, log_callback=real_time_logger, log_collector=log_collector)
+
+    # for single accession
+    # def classify_main(accession):
+    #     #table, summary, labelAncient_Modern, explain_label = mtdna_backend.summarize_results(accession)
+    #     table = mtdna_backend.summarize_results(accession)
+    #     #flag_output = f"### 🏺 Ancient/Modern Flag\n**{labelAncient_Modern}**\n\n_Explanation:_ {explain_label}"
+    #     return (
+    #         #table,
+    #         make_html_table(table),
+    #         # summary,
+    #         # flag_output,
+    #         gr.update(visible=True),
+    #         gr.update(visible=False),
+    #         gr.update(visible=False)
+    #     )
+    
+    #stop_flag = gr.State(value=False)
+    #stop_flag = StopFlag()
+
+    # def stop_batch(stop_flag):
+    #   stop_flag.value = True
+    #   return gr.update(value="❌ Stopping...", visible=True), stop_flag
+    def stop_batch():
+      global_stop_flag.value = True
+      return gr.update(value="❌ Stopping...", visible=True)
+
+    # def threaded_batch_runner(file, text, email):
+    #   global_stop_flag.value = False
+    #   log_lines = []
+
+    #   def update_log(line):
+    #       log_lines.append(line)
+    #       yield (
+    #           gr.update(visible=False),  # output_table (not yet)
+    #           gr.update(visible=False),  # results_group
+    #           gr.update(visible=False),  # download_file
+    #           gr.update(visible=False),  # usage_display
+    #           gr.update(value="⏳ Still processing...", visible=True),  # status
+    #           gr.update(value="\n".join(log_lines))  # progress_box
+    #       )
+
+    #   # Start a dummy update to say "Starting..."
+    #   yield from update_log("🚀 Starting batch processing...")
+
+    #   rows, file_path, count, final_log, warning = mtdna_backend.summarize_batch(
+    #       file=file,
+    #       raw_text=text,
+    #       resume_file=None,
+    #       user_email=email,
+    #       stop_flag=global_stop_flag,
+    #       yield_callback=lambda line: (yield from update_log(line))
+    #   )
+
+    #   html = make_html_table(rows)
+    #   file_update = gr.update(value=file_path, visible=True) if os.path.exists(file_path) else gr.update(visible=False)
+    #   usage_or_warning_text = f"**{count}** samples used by this email." if email.strip() else warning
+
+    #   yield (
+    #       html,
+    #       gr.update(visible=True),        # results_group
+    #       file_update,                    # download_file
+    #       gr.update(value=usage_or_warning_text, visible=True),
+    #       gr.update(value="✅ Done", visible=True),
+    #       gr.update(value=final_log)
+    #   )
+    
+    def threaded_batch_runner(file=None, text="", email=""):
+      print("📧 EMAIL RECEIVED:", email)
+      import tempfile
+      from mtdna_backend import (
+          extract_accessions_from_input,
+          summarize_results,
+          save_to_excel,
+          hash_user_id,
+          increment_usage,
+      )
+      import os
+
+      global_stop_flag.value = False  # reset stop flag
+
+      tmp_dir = tempfile.mkdtemp()
+      output_file_path = os.path.join(tmp_dir, "batch_output_live.xlsx")
+      limited_acc = 50 + (10 if email.strip() else 0)
+
+      # Step 1: Parse input
+      accessions, error = extract_accessions_from_input(file, text)
+      if error:
+          yield (
+              "",                          # output_table
+              gr.update(visible=False),   # results_group
+              gr.update(visible=False),   # download_file
+              "",                          # usage_display
+              "❌ Error",                  # status
+              str(error)                        # progress_box
+          )
+          return
+
+      total = len(accessions)
+      if total > limited_acc:
+          accessions = accessions[:limited_acc]
+          warning = f"⚠️ Only processing first {limited_acc} accessions."
+      else:
+          warning = f"✅ All {total} accessions will be processed."
+
+      all_rows = []
+      log_lines = []
+
+      # Step 2: Loop through accessions
+      for i, acc in enumerate(accessions):
+          if global_stop_flag.value:
+              log_lines.append(f"🛑 Stopped at {acc} ({i+1}/{total})")
+              usage_text = ""
+              if email.strip():
+                  # user_hash = hash_user_id(email)
+                  # usage_count = increment_usage(user_hash, len(all_rows))
+                  usage_count = increment_usage(email, len(all_rows))
+                  usage_text = f"**{usage_count}** samples used by this email. Ten more samples are added first (you now have 60 limited accessions), then wait we will contact you via this email."
+              else:
+                  usage_text = f"The limited accession is 50. The user has used {len(all_rows)}, and only {50-len(all_rows)} left."    
+              yield (
+                  make_html_table(all_rows),
+                  gr.update(visible=True),
+                  gr.update(value=output_file_path, visible=True),
+                  gr.update(value=usage_text, visible=True),
+                  "🛑 Stopped",
+                  "\n".join(log_lines)
+              )
+              return
+
+          log_lines.append(f"[{i+1}/{total}] Processing {acc}")
+          yield (
+              make_html_table(all_rows),
+              gr.update(visible=True),
+              gr.update(visible=False),
+              "",
+              "⏳ Processing...",
+              "\n".join(log_lines)
+          )
+
+          try:
+              rows = summarize_results(acc)
+              all_rows.extend(rows)
+              save_to_excel(all_rows, "", "", output_file_path, is_resume=False)
+              log_lines.append(f"✅ Processed {acc} ({i+1}/{total})")
+          except Exception as e:
+              log_lines.append(f"❌ Failed to process {acc}: {e}")
+
+          yield (
+              make_html_table(all_rows),
+              gr.update(visible=True),
+              gr.update(visible=False),
+              "",
+              "⏳ Processing...",
+              "\n".join(log_lines)
+          )
+
+      # Final update
+      usage_text = ""
+      if email.strip():
+          # user_hash = hash_user_id(email)
+          # usage_count = increment_usage(user_hash, len(all_rows))
+          usage_count = increment_usage(email, len(all_rows))
+          usage_text = f"**{usage_count}** samples used by this email. Ten more samples are added first (you now have 60 limited accessions), then wait we will contact you via this email."
+      else:
+          usage_text = f"The limited accession is 50. The user has used {len(all_rows)}, and only {50-len(all_rows)} left."
+      yield (
+          make_html_table(all_rows),
+          gr.update(visible=True),
+          gr.update(value=output_file_path, visible=True),
+          gr.update(value=usage_text, visible=True),
+          "✅ Done",
+          "\n".join(log_lines)
+      )
+
+    # def threaded_batch_runner(file=None, text="", email=""):
+    #   global_stop_flag.value = False
+
+    #   # Dummy test output that matches expected schema
+    #   return (
+    #       "<div>✅ Dummy output table</div>",   # HTML string
+    #       gr.update(visible=True),             # Group visibility
+    #       gr.update(visible=False),            # Download file
+    #       "**0** samples used.",               # Markdown
+    #       "✅ Done",                           # Status string
+    #       "Processing finished."               # Progress string
+    #   )
+
+    
+    # def classify_mulAcc(file, text, resume, email, log_callback=None, log_collector=None):
+    #     stop_flag.value = False
+    #     return threaded_batch_runner(file, text, resume, email, status, stop_flag, log_callback=log_callback, log_collector=log_collector)
+    
+
+    def make_html_table(rows):
+      html = """
+      <div style='overflow-x: auto; padding: 10px;'>
+          <div style='max-height: 400px; overflow-y: auto; border: 1px solid #444; border-radius: 8px;'>
+              <table style='width:100%; border-collapse: collapse; table-layout: auto; font-size: 14px; color: #f1f1f1; background-color: #1e1e1e;'>
+                  <thead style='position: sticky; top: 0; background-color: #2c2c2c; z-index: 1;'>
+                      <tr>
+      """
+      headers = ["Sample ID", "Predicted Country", "Country Explanation", "Predicted Sample Type", "Sample Type Explanation", "Sources", "Time cost"]
+      html += "".join(
+          f"<th style='padding: 10px; border: 1px solid #555; text-align: left; white-space: nowrap;'>{h}</th>"
+          for h in headers
+      )
+      html += "</tr></thead><tbody>"
+
+      for row in rows:
+          html += "<tr>"
+          for i, col in enumerate(row):
+              header = headers[i]
+              style = "padding: 10px; border: 1px solid #555; vertical-align: top;"
+
+              # For specific columns like Haplogroup, force nowrap
+              if header in ["Country Explanation", "Sample Type Explanation"]:
+                style += " max-width: 400px; word-wrap: break-word; white-space: normal;"
+              elif header in ["Sample ID", "Predicted Country", "Predicted Sample Type", "Time cost"]:
+                  style += " white-space: nowrap; text-overflow: ellipsis; max-width: 200px; overflow: hidden;"
+
+              # if header == "Sources" and isinstance(col, str) and col.strip().lower().startswith("http"):
+              #     col = f"<a href='{col}' target='_blank' style='color: #4ea1f3; text-decoration: underline;'>{col}</a>"
+
+              #html += f"<td style='{style}'>{col}</td>"
+              if header == "Sources" and isinstance(col, str):
+                  links = [f"<a href='{url.strip()}' target='_blank' style='color: #4ea1f3; text-decoration: underline;'>{url.strip()}</a>" for url in col.strip().split("\n") if url.strip()]
+                  col = "- "+"<br>- ".join(links)
+              elif isinstance(col, str):
+                  # lines = []
+                  # for line in col.split("\n"):
+                  #     line = line.strip()
+                  #     if not line:
+                  #         continue
+                  #     if line.lower().startswith("rag_llm-"):
+                  #         content = line[len("rag_llm-"):].strip()
+                  #         line = f"{content} (Method: RAG_LLM)"
+                  #     lines.append(f"- {line}")
+                  col = col.replace("\n", "<br>")
+                  #col = col.replace("\t", "&nbsp;&nbsp;&nbsp;&nbsp;")
+                  #col = "<br>".join(lines)
+
+              html += f"<td style='{style}'>{col}</td>"
+          html += "</tr>"
+
+      html += "</tbody></table></div></div>"
+      return html
+  
+
+    # def reset_fields():
+    #     global_stop_flag.value = False  # 💡 Add this to reset the flag
+    #     return (
+    #         #gr.update(value=""),  # single_accession
+    #         gr.update(value=""),  # raw_text
+    #         gr.update(value=None), # file_upload
+    #         #gr.update(value=None),  # resume_file
+    #         #gr.update(value="Single Accession"), # inputMode
+    #         gr.update(value=[], visible=True), # output_table
+    #         # gr.update(value="", visible=True), # output_summary
+    #         # gr.update(value="", visible=True), # output_flag
+    #         gr.update(visible=False), # status
+    #         gr.update(visible=False),  # results_group
+    #         gr.update(value="", visible=False),  # usage_display
+    #         gr.update(value="", visible=False),  # progress_box
+    #     )
+    def reset_fields():
+      global_stop_flag.value = False  # Reset the stop flag
+
+      return (
+          gr.update(value=""),          # raw_text
+          gr.update(value=None),        # file_upload
+          gr.update(value=[], visible=True),   # output_table
+          gr.update(value="", visible=True),   # status — reset and make visible again
+          gr.update(visible=False),     # results_group
+          gr.update(value="", visible=True),   # usage_display — reset and make visible again
+          gr.update(value="", visible=True),   # progress_box — reset AND visible!
+      )
+    #inputMode.change(fn=toggle_input_mode, inputs=inputMode, outputs=[single_input_group, batch_input_group])
+    #run_button.click(fn=classify_with_loading, inputs=[], outputs=[status])
+    # run_button.click(
+    #     fn=classify_dynamic,
+    #     inputs=[single_accession, file_upload, raw_text, resume_file,user_email,inputMode],
+    #     outputs=[output_table, 
+    #     #output_summary, output_flag, 
+    #     results_group, download_file, usage_display,status, progress_box]
+    # )
+    
+    # run_button.click(
+    #     fn=threaded_batch_runner,
+    #     #inputs=[file_upload, raw_text, resume_file, user_email],
+    #     inputs=[file_upload, raw_text, user_email],
+    #     outputs=[output_table, results_group, download_file, usage_display, status, progress_box]
+    # )
+#     run_button.click(
+#     fn=threaded_batch_runner,
+#     inputs=[file_upload, raw_text, user_email],
+#     outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
+#     every=0.5  # <-- this tells Gradio to expect streaming
+# )
+    # output_table = gr.HTML()
+    # results_group = gr.Group(visible=False)
+    # download_file = gr.File(visible=False)
+    # usage_display = gr.Markdown(visible=False)
+    # status = gr.Markdown(visible=False)
+    # progress_box = gr.Textbox(visible=False)
+
+#     run_button.click(
+#     fn=threaded_batch_runner,
+#     inputs=[file_upload, raw_text, user_email],
+#     outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
+#     every=0.5,  # streaming enabled
+#     show_progress="full"
+# )
+    print("🎯 DEBUG COMPONENT TYPES")
+    print(type(output_table))
+    print(type(results_group))
+    print(type(download_file))
+    print(type(usage_display))
+    print(type(status))
+    print(type(progress_box))
+    
+
+    # interface.stream(
+    #     fn=threaded_batch_runner,
+    #     inputs=[file_upload, raw_text, user_email],
+    #     outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
+    #     trigger=run_button,
+    #     every=0.5,
+    #     show_progress="full",
+    # )
+    interface.queue()  # No arguments here!
+
+    run_button.click(
+        fn=threaded_batch_runner,
+        inputs=[file_upload, raw_text, user_email],
+        outputs=[output_table, results_group, download_file, usage_display, status, progress_box],
+        concurrency_limit=1,  # ✅ correct in Gradio 5.x
+        queue=True,            # ✅ ensure the queue is used
+        #every=0.5
+    )
+    
+
+
+
+    stop_button.click(fn=stop_batch, inputs=[], outputs=[status])
+
+    # reset_button.click(
+    #     #fn=reset_fields,
+    #     fn=lambda: (
+    #         gr.update(value=""), gr.update(value=""), gr.update(value=None), gr.update(value=None), gr.update(value="Single Accession"),
+    #         gr.update(value=[], visible=True), gr.update(visible=False), gr.update(visible=False), gr.update(value="", visible=False), gr.update(value="", visible=False)
+    #     ),
+    #     inputs=[],
+    #     outputs=[
+    #         single_accession, raw_text, file_upload, resume_file,inputMode,
+    #         output_table,# output_summary, output_flag,
+    #         status, results_group, usage_display, progress_box
+    #     ]
+    # )
+    #stop_button.click(fn=lambda sf: (gr.update(value="❌ Stopping...", visible=True), setattr(sf, "value", True) or sf), inputs=[gr.State(stop_flag)], outputs=[status, gr.State(stop_flag)])
+
+    reset_button.click(
+        fn=reset_fields,
+        inputs=[],
+        #outputs=[raw_text, file_upload, resume_file, output_table, status, results_group, usage_display, progress_box]
+        outputs=[raw_text, file_upload, output_table, status, results_group, usage_display, progress_box]
+    )  
+
+    download_button.click(
+      fn=mtdna_backend.save_batch_output, 
+      #inputs=[output_table, output_summary, output_flag, output_type], 
+      inputs=[output_table, output_type], 
+      outputs=[download_file])
+
+    # submit_feedback.click(
+    #     fn=mtdna_backend.store_feedback_to_google_sheets, 
+    #     inputs=[single_accession, q1, q2, contact], outputs=feedback_status
+    # )
+    submit_feedback.click(
+        fn=mtdna_backend.store_feedback_to_google_sheets,
+        inputs=[raw_text, q1, q2, contact],
+        outputs=[feedback_status]
+    )
+    #     # Custom CSS styles
+    # gr.HTML("""
+    # <style>
+    #   /* Ensures both sections are equally spaced with the same background size */
+    #   #output-summary, #output-flag {
+    #       background-color: #f0f4f8; /* Light Grey for both */
+    #       padding: 20px;
+    #       border-radius: 10px;
+    #       margin-top: 10px;
+    #       width: 100%; /* Ensure full width */
+    #       min-height: 150px; /* Ensures both have a minimum height */
+    #       box-sizing: border-box; /* Prevents padding from increasing size */
+    #       display: flex;
+    #       flex-direction: column;
+    #       justify-content: space-between;
+    #   }
+      
+    #   /* Specific background colors */
+    #   #output-summary {
+    #       background-color: #434a4b; 
+    #   }
+
+    #   #output-flag {
+    #       background-color: #141616; 
+    #   }
+
+    #   /* Ensuring they are in a row and evenly spaced */
+    #   .gradio-row {
+    #       display: flex;
+    #       justify-content: space-between;
+    #       width: 100%;
+    #   }
+    # </style>
+    # """)
+
+
 interface.launch(share=True,debug=True)