update 3 tabs

app.py

@@ -1,13 +1,25 @@
 import os
 import gradio as gr
 import hashlib
+import pandas as pd
+import zipfile
+import tempfile
 from datetime import datetime
-from huggingface_hub import HfApi
+from huggingface_hub import HfApi, hf_hub_download,CommitOperationDelete
 from pathlib import Path
 
+
+label_codes = {
+    "1":"Engine",
+    "2":"Environmental",
+    "3":"Mechanical"
+}
+label_decoder = {v: k for k, v in label_codes.items()}
+
 # --- CONFIGURATION ---
 DATASET_REPO_ID = "MeysamSh/ENSIMSoundDataCollection"
 HF_TOKEN = os.environ.get("HF_TOKEN")
+COUPON_SALT = os.environ.get("COUPON_SALT")#"ENSIM_2026_SECRET"
 
 # Admin Credentials
 ADMIN_USERNAME = "admin"
@@ -22,6 +34,11 @@ api = HfApi()
 
 # --- LOGIC FUNCTIONS ---
 
+def generate_coupon(filename):
+    """Creates a unique string for the student to save."""
+    return hashlib.sha1(f"{filename}{COUPON_SALT}".encode()).hexdigest()[:10].upper()
+
+
 def verify_user(email):
     if not email: return gr.update(visible=False), "⚠️ Enter email."
     clean_email = email.strip().lower()
@@ -31,42 +48,116 @@ def verify_user(email):
     return gr.update(visible=False), "🚫 Not authorized."
 
 def upload_data(email, label, audio_path):
-    if not audio_path: return "⚠️ No audio recorded."
+    if not audio_path: return "⚠️ No audio.", ""
     try:
         clean_email = email.strip().lower()
         email_index = AUTHORIZED_USERS.index(clean_email) if clean_email in AUTHORIZED_USERS else "unknown"
+        
+        # Determine if this is file #1, #2, etc. for this user to assign Training/Test
+        all_files = api.list_repo_files(repo_id=DATASET_REPO_ID, repo_type="dataset")
+        user_submissions = [f for f in all_files if f.startswith(f"data/{email_index}_")]
+        submission_number = len(user_submissions) + 1
+        
         timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
         unique_filename = f"{email_index}_{timestamp}.wav"
         
         # Upload Audio
         api.upload_file(path_or_fileobj=audio_path, path_in_repo=f"data/{unique_filename}",
                         repo_id=DATASET_REPO_ID, repo_type="dataset", token=HF_TOKEN)
+        coupon = generate_coupon(unique_filename)
         # Upload Metadata
-        meta_content = f"user_id,label,file_name,timestamp\n{clean_email},{label},{unique_filename},{timestamp}"
+        meta_content = f"user_id,label,file_name,time,order\n{clean_email},{label},{unique_filename},{timestamp},{submission_number}"
         api.upload_file(path_or_fileobj=meta_content.encode(), path_in_repo=f"metadata/meta_{email_index}_{timestamp}.csv",
                         repo_id=DATASET_REPO_ID, repo_type="dataset", token=HF_TOKEN)
-        return f"🎉 Success! Audio saved as {label}."
-    except Exception as e: return f"❌ Error: {str(e)}"
+        
+        
+        return f"🎉 Success! Submission #{submission_number} saved.", coupon
+    except Exception as e: return f"❌ Error: {str(e)}", ""
 
 # --- ADMIN LOGIC ---
 
+def delete_all_files(confirm):
+    if not confirm:
+        return "⚠️ You must check the 'Confirm' box to delete everything.", gr.update()
+    
+    try:
+        # 1. Get all files in the repo
+        all_files = api.list_repo_files(repo_id=DATASET_REPO_ID, repo_type="dataset")
+        
+        # 2. Filter for files in our managed folders
+        files_to_delete = [f for f in all_files if f.startswith("data/") or f.startswith("metadata/")]
+        
+        if not files_to_delete:
+            return "ℹ️ The dataset is already empty.", gr.update(choices=[])
+
+        # 3. Use bulk deletion to avoid hundreds of individual API calls
+        # This is much faster for "Delete All"
+        
+        operations = [CommitOperationDelete(path_in_repo=f) for f in files_to_delete]
+        
+        api.create_commit(
+            repo_id=DATASET_REPO_ID,
+            repo_type="dataset",
+            operations=operations,
+            commit_message=f"Admin: Bulk delete of {len(files_to_delete)} files",
+            token=HF_TOKEN
+        )
+
+        return f"💥 Success! Deleted {len(files_to_delete)} files. Dataset is now clean.", gr.update(choices=[], value=None)
+    
+    except Exception as e:
+        return f"❌ Bulk delete failed: {str(e)}", gr.update()
+
 def get_stats():
-    """Helper to calculate stats from repository files"""
+    """Helper to calculate stats and label distribution from repository"""
     try:
-        files = api.list_repo_files(repo_id=DATASET_REPO_ID, repo_type="dataset")
-        audio_files = [f for f in files if f.startswith("data/") and f.endswith(".wav")]
+        # List all files once to avoid multiple API calls
+        all_files = api.list_repo_files(repo_id=DATASET_REPO_ID, repo_type="dataset")
         
-        # Extract user indices from filenames like 'data/0_20260511.wav'
+        audio_files = [f for f in all_files if f.startswith("data/") and f.endswith(".wav")]
+        metadata_files = [f for f in all_files if f.startswith("metadata/") and f.endswith(".csv")]
+        print(f"Found {len(audio_files)} audio files and {len(metadata_files)} metadata files in the repository.")
+        
+        # 1. Count Unique Contributors
         user_indices = set()
         for f in audio_files:
-            filename = f.split("/")[-1] # get '0_20260511.wav'
+            filename = f.split("/")[-1]
             user_id = filename.split("_")[0]
             user_indices.add(user_id)
             
-        stats_text = f"📊 **Stats:** {len(audio_files)} recordings from {len(user_indices)} unique contributors."
-        return audio_files, stats_text
-    except:
-        return [], "⚠️ Could not retrieve stats."
+        # 2. Count Files per Category (Label)
+        category_counts = {label_codes["1"]: 0, label_codes["2"]: 0, label_codes["3"]: 0}
+        
+        for m_file in metadata_files:
+            try:
+                # Download and read the small metadata file
+                file_path = hf_hub_download(repo_id=DATASET_REPO_ID, filename=m_file, repo_type="dataset", token=HF_TOKEN)
+                with open(file_path, 'r') as f:
+                    content = f.readlines()
+                    if len(content) > 1:
+                        # The label is the second column in: user_id,label,file_name,timestamp
+                        label = content[1].split(",")[1].strip()
+                        if label in category_counts:
+                            category_counts[label] += 1
+                        else:
+                            # Handle cases where label might not match exactly
+                            category_counts[label] = category_counts.get(label, 0) + 1
+            except Exception:
+                print(f"⚠️ Failed to process metadata file: {m_file}")
+                continue # Skip files that fail to download or parse
+
+        # 3. Format the stats string
+        stats_md = f"### 📊 Dataset Statistics\n"
+        stats_md += f"**Total Recordings:** {len(audio_files)}  \n"
+        stats_md += f"**Unique Contributors:** {len(user_indices)}  \n\n"
+        stats_md += "**Category Breakdown:**\n"
+        for cat, count in category_counts.items():
+            stats_md += f"- **{cat}:** {count} files\n"
+            
+        return audio_files, stats_md
+    except Exception as e:
+        return [], f"⚠️ Error retrieving stats: {str(e)}"
+    
 
 def admin_login(user, pwd):
     pwd_hash = hashlib.sha256(pwd.encode()).hexdigest()
@@ -88,6 +179,79 @@ def delete_selected_file(file_path):
         return f"🗑️ Deleted {file_path}. {stats_text}", gr.update(choices=audio_files, value=None)
     except Exception as e: return f"❌ Error: {str(e)}", gr.update()
 
+def access_dataset_zip(email, coupons_str):
+    """Verifies coupons and creates a ZIP of accessible data."""
+    if not email or not coupons_str:
+        return None, "⚠️ Please provide your email and coupons."
+    
+    # Clean coupons input
+    coupons_list = [c.strip().upper() for c in coupons_str.split(",") if c.strip()]
+    num_coupons = len(coupons_list)
+    
+    if num_coupons == 0:
+        return None, "⚠️ No valid coupons provided."
+
+    # try:
+    all_files = api.list_repo_files(repo_id=DATASET_REPO_ID, repo_type="dataset")
+    meta_files = [f for f in all_files if f.startswith("metadata/")]
+    
+    # Create a temporary directory for the ZIP
+    tmp_dir = tempfile.mkdtemp()
+    zip_path = os.path.join(tmp_dir, f"ENSIM_Data_{num_coupons}_coupons.zip")
+    print(f"Creating ZIP at: {zip_path}")
+
+    training_metadata = []
+
+    with zipfile.ZipFile(zip_path, 'w') as zipf:
+        for m_file in meta_files:
+            # Get metadata for each entry
+            print(f"Processing metadata file: {m_file}")
+            local_meta = hf_hub_download(repo_id=DATASET_REPO_ID, filename=m_file, repo_type="dataset", token=HF_TOKEN)
+            print(f"Downloaded metadata to: {local_meta}")
+            df = pd.read_csv(local_meta)
+            row = df.iloc[0]
+            
+            order = int(row['order'])
+            audio_filename = row['file_name']
+            audio_repo_path = f"data/{audio_filename}"
+            print(f"Processing {audio_filename} with order {order} for coupon access...")
+            
+            # ACCESS RULES:
+            # 1. Training Set (Odd): Only if student has enough coupons
+            # 2. Test Set (Even): Always included if they have at least 1 coupon
+            
+            is_training = order % 2 != 0
+            
+            if is_training and order <= num_coupons:
+                # Download and add to Training folder
+                audio_local = hf_hub_download(repo_id=DATASET_REPO_ID, filename=audio_repo_path, repo_type="dataset", token=HF_TOKEN)
+                print(f"Adding to ZIP: {audio_filename} (Label: {row['label']})")
+                zipf.write(audio_local, arcname=f"training_set/{audio_filename}")
+                print(f"Added {audio_filename} to ZIP under training_set/")
+                training_metadata.append({"file": audio_filename, "label": row['label']})
+            
+            elif not is_training:
+                # Download and add to Test folder (No labels added to zip)
+                audio_local = hf_hub_download(repo_id=DATASET_REPO_ID, filename=audio_repo_path, repo_type="dataset", token=HF_TOKEN)
+                print(f"Adding to ZIP: {audio_filename} (Test Set)")
+                zipf.write(audio_local, arcname=f"test_set/{audio_filename}")
+                print(f"Added {audio_filename} to ZIP under test_set/")
+
+        # Add Training Metadata CSV to the ZIP if data exists
+        if training_metadata:
+            print(f"Adding {len(training_metadata)} training metadata entries to ZIP...")
+            train_df = pd.DataFrame(training_metadata)
+            csv_path = os.path.join(tmp_dir, "training_labels.csv")
+            train_df.to_csv(csv_path, index=False)
+            zipf.write(csv_path, arcname="training_set/labels.csv")
+
+    return zip_path, f"✅ Successfully bundled data for {num_coupons} coupons."
+
+    # except Exception as e:
+    #     return None, f"❌ Error creating ZIP: {str(e)}"
+    
+    # except Exception as e: return f"❌ Error: {str(e)}"
+
 # --- UI ---
 
 with gr.Blocks() as demo:
@@ -102,11 +266,26 @@ with gr.Blocks() as demo:
             login_status = gr.Markdown("Waiting for login...")
 
             with gr.Column(visible=False) as recording_zone:
-                label_input = gr.Radio(choices=["Engine", "Environmental", "Mechanical"], label="Category")
+                label_input = gr.Radio(choices=[label_codes["1"], label_codes["2"], label_codes["3"]], label="Category")
                 audio_input = gr.Audio(label="Record (40s)", sources=["microphone"], type="filepath")
                 submit_btn = gr.Button("🚀 Submit", variant="primary")
-                final_status = gr.Textbox(label="Status", interactive=False)
+                res_msg = gr.Textbox(label="Status", interactive=False)
+                coupon_display = gr.Textbox(label="🎟️ YOUR COUPON (Save this!)", interactive=False)
 
+        # 2. DATASET ACCESS TAB
+        with gr.TabItem("2. Dataset Access"):
+            gr.Markdown("""
+            ### 🔓 Unlock Your Data Partition
+            - **Training Data:** You receive Training samples (Audio + Label) proportional to your coupons.
+            - **Test Data:** You receive the full global Test set (Audio Only) to evaluate your models.
+            """)
+            acc_email = gr.Textbox(label="Email")
+            coupons_input = gr.Textbox(label="Coupons List (comma separated)", placeholder="C1, C2, C3...")
+            download_btn = gr.Button("📦 Generate Data ZIP", variant="primary")
+            
+            status_out = gr.Textbox(label="Status")
+            file_out = gr.File(label="Download Your Data")
+        
         # ADMIN TAB
         with gr.TabItem("Administration"):
             with gr.Row():
@@ -119,14 +298,20 @@ with gr.Blocks() as demo:
             admin_stats_display = gr.Markdown("") 
             
             with gr.Column(visible=False) as admin_panel:
-                gr.Separator()
                 file_dropdown = gr.Dropdown(label="Select File to Remove", choices=[])
                 delete_btn = gr.Button("🗑️ Delete Selected File", variant="stop")
                 delete_status = gr.Textbox(label="Delete Progress")
 
+                gr.Markdown("### 🧨 Danger Zone")
+                confirm_check = gr.Checkbox(label="I understand this will permanently delete ALL recordings and metadata.")
+                delete_all_btn = gr.Button("🔥 DELETE ALL DATASET FILES", variant="stop")
+            
+                delete_status = gr.Textbox(label="Status Log")
+
+
     # --- EVENT HANDLERS ---
     login_btn.click(verify_user, [email_input], [recording_zone, login_status])
-    submit_btn.click(upload_data, [email_input, label_input, audio_input], final_status)
+    submit_btn.click(upload_data, [email_input, label_input, audio_input], [res_msg, coupon_display])
     
     admin_login_btn.click(
         admin_login, 
@@ -140,5 +325,17 @@ with gr.Blocks() as demo:
         [delete_status, file_dropdown]
     )
 
+    download_btn.click(
+        fn=access_dataset_zip,
+        inputs=[acc_email, coupons_input],
+        outputs=[file_out, status_out]
+    )
+
+    delete_all_btn.click(
+        fn=delete_all_files,
+        inputs=[confirm_check],
+        outputs=[delete_status, file_dropdown]
+    )
+
 if __name__ == "__main__":
     demo.launch(theme=gr.themes.Soft())