Update app.py

app.py

@@ -71,17 +71,21 @@ audit_state = AuditState()
 def run_audit(file, pasted_urls, domain, batch_size, timeout, delay, workers):
     """Main audit generator — yields progress updates."""
     if sb is None:
-        yield "❌ Supabase not connected. Set SUPABASE_URL and SUPABASE_KEY in Space secrets.", "", gr.update(), gr.update()
+        yield ("❌ Supabase not connected. Set SUPABASE_URL and SUPABASE_KEY in Space secrets.",
+               "",
+               gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+               gr.Button(value="🚀 Run Audit", variant="primary", interactive=True))
         return
 
     # Parse URLs
     urls = []
     if file is not None:
         try:
-            if file.name.endswith('.csv'):
-                df = pd.read_csv(file.name)
+            fpath = file.name if hasattr(file, 'name') else file
+            if str(fpath).endswith('.csv'):
+                df = pd.read_csv(fpath)
             else:
-                df = pd.read_excel(file.name)
+                df = pd.read_excel(fpath)
             url_col = None
             for col in df.columns:
                 sample = str(df[col].iloc[0]).strip().lower()
@@ -92,13 +96,17 @@ def run_audit(file, pasted_urls, domain, batch_size, timeout, delay, workers):
                 url_col = df.columns[0]
             urls = [u for u in df[url_col].dropna().astype(str).str.strip().tolist() if u.startswith('http')]
         except Exception as e:
-            yield f"❌ File error: {e}", "", gr.update(), gr.update()
+            yield (f"❌ File error: {e}", "",
+                   gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+                   gr.Button(value="🚀 Run Audit", variant="primary", interactive=True))
             return
     elif pasted_urls and pasted_urls.strip():
         urls = [u.strip() for u in pasted_urls.strip().split('\n') if u.strip().startswith('http')]
 
     if not urls:
-        yield "⚠ No valid URLs found. Upload a file or paste URLs.", "", gr.update(), gr.update()
+        yield ("⚠ No valid URLs found. Upload a file or paste URLs.", "",
+               gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+               gr.Button(value="🚀 Run Audit", variant="primary", interactive=True))
         return
 
     # Deduplicate preserving order
@@ -117,21 +125,27 @@ def run_audit(file, pasted_urls, domain, batch_size, timeout, delay, workers):
     audit_state.resume()  # Reset pause flag
 
     total = len(urls)
+    batch_size = int(batch_size)
+    timeout = int(timeout)
+    workers = int(workers)
     start_time = time.time()
     batch_num = 0
     log_lines = []
 
-    yield f"🚀 Started: {run_name}\n📦 {total} URLs · Batch size: {batch_size}", "", gr.update(interactive=True), gr.update(interactive=False)
+    yield (f"🚀 Started: {run_name}\n📦 {total} URLs · Batch size: {batch_size}", "",
+           gr.Button(value="⏸️ Pause", variant="stop", interactive=True),
+           gr.Button(value="🚀 Running...", variant="primary", interactive=False))
 
     try:
         for batch_start in range(0, total, batch_size):
-            # Check for pause
             if audit_state.is_paused():
                 completed = get_completed_count(sb, run_id)
                 update_run_status(sb, run_id, "paused", completed)
                 log_lines.append(f"⏸️ PAUSED at {completed}/{total} — resume from Past Runs")
                 audit_state.set_running(False)
-                yield "\n".join(log_lines[-40:]), "", gr.update(interactive=False), gr.update(interactive=False)
+                yield ("\n".join(log_lines[-40:]), f"⏸️ Paused at {completed}/{total}",
+                       gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+                       gr.Button(value="🚀 Run Audit", variant="primary", interactive=True))
                 return
 
             batch_end = min(batch_start + batch_size, total)
@@ -140,16 +154,16 @@ def run_audit(file, pasted_urls, domain, batch_size, timeout, delay, workers):
             batch_results = []
 
             for j, url in enumerate(batch_urls):
-                # Check pause between each URL
                 if audit_state.is_paused():
-                    # Save partial batch
                     if batch_results:
                         save_batch_results(sb, run_id, batch_results)
                     completed = get_completed_count(sb, run_id)
                     update_run_status(sb, run_id, "paused", completed)
                     log_lines.append(f"⏸️ PAUSED at {completed}/{total}")
                     audit_state.set_running(False)
-                    yield "\n".join(log_lines[-40:]), "", gr.update(interactive=False), gr.update(interactive=False)
+                    yield ("\n".join(log_lines[-40:]), f"⏸️ Paused at {completed}/{total}",
+                           gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+                           gr.Button(value="🚀 Run Audit", variant="primary", interactive=True))
                     return
 
                 global_idx = batch_start + j + 1
@@ -179,13 +193,15 @@ def run_audit(file, pasted_urls, domain, batch_size, timeout, delay, workers):
                     flag_str = " · ".join(flags) if flags else "✅"
                     log_lines.append(f"[{global_idx}/{total}] {short} — Int:{result['int_count']} Ext:{result['ext_count']} · {flag_str}")
 
-                progress_text = f"📊 Progress: {global_idx}/{total} ({global_idx*100//total}%) · Batch {batch_num} · ETA: {eta_str}"
-                yield "\n".join(log_lines[-40:]), progress_text, gr.update(interactive=True), gr.update(interactive=False)
+                progress_msg = f"📊 Progress: {global_idx}/{total} ({global_idx*100//total}%) · Batch {batch_num} · ETA: {eta_str}"
+                yield ("\n".join(log_lines[-40:]), progress_msg,
+                       gr.Button(value="⏸️ Pause", variant="stop", interactive=True),
+                       gr.Button(value="🚀 Running...", variant="primary", interactive=False))
 
                 if j < len(batch_urls) - 1:
                     time.sleep(delay)
 
-            # Save batch to Supabase
+            # Save batch
             if batch_results:
                 try:
                     save_batch_results(sb, run_id, batch_results)
@@ -195,12 +211,16 @@ def run_audit(file, pasted_urls, domain, batch_size, timeout, delay, workers):
                 except Exception as e:
                     log_lines.append(f"⚠ Batch save error: {str(e)[:60]}")
 
-                yield "\n".join(log_lines[-40:]), f"📊 Progress: {min(batch_end, total)}/{total} · Saved batch {batch_num}", gr.update(interactive=True), gr.update(interactive=False)
+                yield ("\n".join(log_lines[-40:]), f"📊 Saved batch {batch_num} — {min(batch_end, total)}/{total}",
+                       gr.Button(value="⏸️ Pause", variant="stop", interactive=True),
+                       gr.Button(value="🚀 Running...", variant="primary", interactive=False))
                 del batch_results
 
         # ── ALL DONE — Orphan analysis ──
         log_lines.append("🔍 Running orphan page analysis...")
-        yield "\n".join(log_lines[-40:]), f"📊 Orphan analysis...", gr.update(interactive=False), gr.update(interactive=False)
+        yield ("\n".join(log_lines[-40:]), "📊 Orphan analysis...",
+               gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+               gr.Button(value="🚀 Running...", variant="primary", interactive=False))
 
         all_pages = get_all_page_results(sb, run_id)
         all_results = [p['result'] for p in all_pages]
@@ -233,15 +253,22 @@ def run_audit(file, pasted_urls, domain, batch_size, timeout, delay, workers):
         log_lines.append("→ Go to Past Runs tab to generate report")
 
         audit_state.set_running(False)
-        yield "\n".join(log_lines[-40:]), f"✅ Complete — {len(all_results)} pages", gr.update(interactive=False), gr.update(interactive=False)
+        yield ("\n".join(log_lines[-40:]), f"✅ Complete — {len(all_results)} pages",
+               gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+               gr.Button(value="🚀 Run Audit", variant="primary", interactive=True))
 
     except Exception as e:
         log_lines.append(f"❌ Error: {str(e)}")
         audit_state.set_running(False)
         if run_id:
-            completed = get_completed_count(sb, run_id)
-            update_run_status(sb, run_id, "paused", completed)
-        yield "\n".join(log_lines[-40:]), f"❌ Error — saved progress to Supabase", gr.update(interactive=False), gr.update(interactive=False)
+            try:
+                completed = get_completed_count(sb, run_id)
+                update_run_status(sb, run_id, "paused", completed)
+            except:
+                pass
+        yield ("\n".join(log_lines[-40:]), f"❌ Error — saved progress to Supabase",
+               gr.Button(value="⏸️ Pause", variant="stop", interactive=False),
+               gr.Button(value="🚀 Run Audit", variant="primary", interactive=True))
 
 
 def pause_audit():
@@ -259,11 +286,15 @@ def pause_audit():
 def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
     """Resume a paused/interrupted run."""
     if sb is None:
-        yield "❌ Supabase not connected.", "", gr.update(), gr.update()
+        yield ("❌ Supabase not connected.", "",
+               gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+               gr.Button(value="▶️ Resume Audit", variant="primary", interactive=True))
         return
 
     if not run_id:
-        yield "⚠ No run selected.", "", gr.update(), gr.update()
+        yield ("⚠ No run selected.", "",
+               gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+               gr.Button(value="▶️ Resume Audit", variant="primary", interactive=True))
         return
 
     all_urls_for_run = get_pending_urls(sb, run_id)
@@ -272,19 +303,35 @@ def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
 
     if not remaining:
         update_run_status(sb, run_id, "completed", len(done_urls))
-        yield "✅ All pages already audited!", "", gr.update(), gr.update()
+        yield ("✅ All pages already audited!", "",
+               gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+               gr.Button(value="▶️ Resume Audit", variant="primary", interactive=True))
         return
 
+    # Get domain from run
+    try:
+        runs = get_all_runs(sb)
+        run_data = next((r for r in runs if r['id'] == run_id), None)
+        if run_data:
+            domain = run_data.get('domain', domain)
+    except:
+        pass
+
     audit_state.set_running(True, run_id)
     audit_state.resume()
     update_run_status(sb, run_id, "running")
 
     total = len(all_urls_for_run)
+    batch_size = int(batch_size)
+    timeout = int(timeout)
+    workers = int(workers)
     start_time = time.time()
     batch_num = 0
     log_lines = [f"▶️ Resuming — {len(remaining)} pages remaining ({len(done_urls)} already done)"]
 
-    yield "\n".join(log_lines), f"📊 Resuming: {len(done_urls)}/{total}", gr.update(interactive=True), gr.update(interactive=False)
+    yield ("\n".join(log_lines), f"📊 Resuming: {len(done_urls)}/{total}",
+           gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=True),
+           gr.Button(value="▶️ Resuming...", variant="primary", interactive=False))
 
     try:
         for batch_start in range(0, len(remaining), batch_size):
@@ -293,7 +340,9 @@ def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
                 update_run_status(sb, run_id, "paused", completed)
                 log_lines.append(f"⏸️ PAUSED at {completed}/{total}")
                 audit_state.set_running(False)
-                yield "\n".join(log_lines[-40:]), "", gr.update(interactive=False), gr.update(interactive=False)
+                yield ("\n".join(log_lines[-40:]), f"⏸️ Paused at {completed}/{total}",
+                       gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+                       gr.Button(value="▶️ Resume Audit", variant="primary", interactive=True))
                 return
 
             batch_end = min(batch_start + batch_size, len(remaining))
@@ -309,7 +358,9 @@ def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
                     update_run_status(sb, run_id, "paused", completed)
                     log_lines.append(f"⏸️ PAUSED at {completed}/{total}")
                     audit_state.set_running(False)
-                    yield "\n".join(log_lines[-40:]), "", gr.update(interactive=False), gr.update(interactive=False)
+                    yield ("\n".join(log_lines[-40:]), f"⏸️ Paused at {completed}/{total}",
+                           gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+                           gr.Button(value="▶️ Resume Audit", variant="primary", interactive=True))
                     return
 
                 global_idx = len(done_urls) + batch_start + j + 1
@@ -334,7 +385,10 @@ def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
                     flag_str = f"🔴 {b} broken" if b else "✅"
                     log_lines.append(f"[{global_idx}/{total}] {short} · {flag_str}")
 
-                yield "\n".join(log_lines[-40:]), f"📊 Progress: {global_idx}/{total} ({global_idx*100//total}%) · ETA: {eta_str}", gr.update(interactive=True), gr.update(interactive=False)
+                yield ("\n".join(log_lines[-40:]),
+                       f"📊 Progress: {global_idx}/{total} ({global_idx*100//total}%) · ETA: {eta_str}",
+                       gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=True),
+                       gr.Button(value="▶️ Resuming...", variant="primary", interactive=False))
                 if j < len(batch_urls) - 1:
                     time.sleep(delay)
 
@@ -347,7 +401,9 @@ def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
 
         # Orphan analysis
         log_lines.append("🔍 Orphan analysis...")
-        yield "\n".join(log_lines[-40:]), "📊 Orphan analysis...", gr.update(interactive=False), gr.update(interactive=False)
+        yield ("\n".join(log_lines[-40:]), "📊 Orphan analysis...",
+               gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+               gr.Button(value="▶️ Resuming...", variant="primary", interactive=False))
 
         all_pages = get_all_page_results(sb, run_id)
         all_results = [p['result'] for p in all_pages]
@@ -377,14 +433,21 @@ def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
         total_time = time.time() - start_time
         log_lines.append(f"✅ COMPLETE! {len(all_results)} pages in {total_time:.0f}s · {len(orphans)} orphans")
         audit_state.set_running(False)
-        yield "\n".join(log_lines[-40:]), f"✅ Complete — {len(all_results)} pages", gr.update(interactive=False), gr.update(interactive=False)
+        yield ("\n".join(log_lines[-40:]), f"✅ Complete — {len(all_results)} pages",
+               gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+               gr.Button(value="▶️ Resume Audit", variant="primary", interactive=True))
 
     except Exception as e:
         log_lines.append(f"❌ Error: {str(e)}")
         audit_state.set_running(False)
-        completed = get_completed_count(sb, run_id)
-        update_run_status(sb, run_id, "paused", completed)
-        yield "\n".join(log_lines[-40:]), "❌ Error", gr.update(interactive=False), gr.update(interactive=False)
+        try:
+            completed = get_completed_count(sb, run_id)
+            update_run_status(sb, run_id, "paused", completed)
+        except:
+            pass
+        yield ("\n".join(log_lines[-40:]), "❌ Error",
+               gr.Button(value="⏸️ Pause Resume", variant="stop", interactive=False),
+               gr.Button(value="▶️ Resume Audit", variant="primary", interactive=True))
 
 
 # ═══════════════════════════════════════════════════
@@ -393,13 +456,12 @@ def resume_audit(run_id, domain, batch_size, timeout, delay, workers):
 
 def load_past_runs():
     if sb is None:
-        return "<p>❌ Supabase not connected</p>", gr.update(choices=[], value=None)
+        return "<p>❌ Supabase not connected</p>", gr.Dropdown(choices=[], value=None)
 
     runs = get_all_runs(sb)
     if not runs:
-        return "<p>No saved runs yet.</p>", gr.update(choices=[], value=None)
+        return "<p>No saved runs yet.</p>", gr.Dropdown(choices=[], value=None)
 
-    # Build choices for dropdown
     choices = []
     for r in runs:
         status = r.get('status', 'unknown')
@@ -408,7 +470,6 @@ def load_past_runs():
         label = f"{r.get('name', 'Untitled')} [{status.upper()}] ({completed}/{total})"
         choices.append((label, r['id']))
 
-    # Build HTML table
     html = '<div style="max-height:400px;overflow-y:auto;">'
     html += '<table style="width:100%;border-collapse:collapse;font-size:13px;">'
     html += '<tr style="background:#f1f5f9;"><th style="padding:10px;text-align:left;">Run Name</th><th style="padding:10px;text-align:center;">Status</th><th style="padding:10px;text-align:center;">Pages</th><th style="padding:10px;text-align:center;">Broken</th><th style="padding:10px;text-align:center;">Flags</th><th style="padding:10px;text-align:center;">Dups</th><th style="padding:10px;text-align:center;">Orphans</th></tr>'
@@ -430,7 +491,7 @@ def load_past_runs():
         </tr>'''
 
     html += '</table></div>'
-    return html, gr.update(choices=choices, value=choices[0][1] if choices else None)
+    return html, gr.Dropdown(choices=choices, value=choices[0][1] if choices else None)
 
 
 def generate_report_for_run(run_id, domain):
@@ -456,7 +517,6 @@ def generate_report_for_run(run_id, domain):
 
         report_html = generate_report(results, orphan_urls, report_domain)
 
-        # Save to temp file
         tmp = tempfile.NamedTemporaryFile(delete=False, suffix='.html', prefix='Link_Audit_')
         tmp.write(report_html.encode('utf-8'))
         tmp.close()
@@ -516,21 +576,18 @@ css = """
 .main-header h1 { margin: 0 0 4px 0; font-size: 24px; }
 .main-header p { margin: 0; opacity: 0.8; font-size: 13px; }
 .status-bar { background: #f1f5f9; border: 1px solid #e2e8f0; border-radius: 8px; padding: 10px 16px; font-family: monospace; font-size: 13px; font-weight: 600; }
-.log-area textarea { font-family: 'JetBrains Mono', monospace !important; font-size: 12px !important; line-height: 1.6 !important; }
 """
 
 with gr.Blocks(css=css, title="🔗 Link Audit Tool", theme=gr.themes.Soft()) as app:
 
-    # Header
     gr.HTML("""
     <div class="main-header">
-        <p style="font-size:10px;font-weight:700;letter-spacing:1.5px;text-transform:uppercase;color:#93c5fd;margin-bottom:8px;">SEO Link Audit Tool</p>
+        <p style="font-size:10px;font-weight:700;letter-spacing:1.5px;text-transform:uppercase;color:#93c5fd;margin-bottom:8px;">SEO LINK AUDIT TOOL</p>
         <h1>🔗 Bulk Link Audit</h1>
         <p>Upload URLs → batch crawl with auto-save → pause/resume anytime → generate interactive report</p>
     </div>
     """)
 
-    # Connection status
     conn_status = "✅ Supabase Connected" if sb else "❌ Supabase Not Connected — add SUPABASE_URL and SUPABASE_KEY to Space secrets"
     gr.HTML(f'<div class="status-bar">🗄️ {conn_status}</div>')
 
@@ -544,27 +601,24 @@ with gr.Blocks(css=css, title="🔗 Link Audit Tool", theme=gr.themes.Soft()) as
                     pasted_urls = gr.Textbox(label="Or paste URLs (one per line)", lines=5, placeholder="https://www.example.com/blog/page1\nhttps://www.example.com/blog/page2")
 
                 with gr.Column(scale=1):
-                    domain = gr.Textbox(label="Your Domain", value="edstellar.com")
-                    batch_size = gr.Slider(5, 50, value=25, step=5, label="Batch Size")
-                    timeout = gr.Slider(5, 60, value=15, step=5, label="Timeout (s)")
-                    delay = gr.Slider(0, 5, value=1.0, step=0.5, label="Delay between pages (s)")
-                    workers = gr.Slider(1, 10, value=5, step=1, label="Parallel link checks")
+                    domain_input = gr.Textbox(label="Your Domain", value="edstellar.com")
+                    batch_size_input = gr.Slider(5, 50, value=25, step=5, label="Batch Size")
+                    timeout_input = gr.Slider(5, 60, value=15, step=5, label="Timeout (s)")
+                    delay_input = gr.Slider(0, 5, value=1.0, step=0.5, label="Delay between pages (s)")
+                    workers_input = gr.Slider(1, 10, value=5, step=1, label="Parallel link checks")
 
             with gr.Row():
                 run_btn = gr.Button("🚀 Run Audit", variant="primary", scale=2)
                 pause_btn = gr.Button("⏸️ Pause", variant="stop", scale=1, interactive=False)
 
-            progress_text = gr.Textbox(label="Status", interactive=False, elem_classes=["status-bar"])
-            log_output = gr.Textbox(label="Audit Log", lines=20, interactive=False, elem_classes=["log-area"])
+            progress_text = gr.Textbox(label="Status", interactive=False)
+            log_output = gr.Textbox(label="Audit Log", lines=20, interactive=False)
 
-            # Wire up run button (generator for streaming)
             run_btn.click(
                 fn=run_audit,
-                inputs=[file_input, pasted_urls, domain, batch_size, timeout, delay, workers],
+                inputs=[file_input, pasted_urls, domain_input, batch_size_input, timeout_input, delay_input, workers_input],
                 outputs=[log_output, progress_text, pause_btn, run_btn],
             )
-
-            # Wire up pause button
             pause_btn.click(fn=pause_audit, outputs=[progress_text])
 
         # ═══ TAB 2: PAST RUNS ═══
@@ -583,33 +637,20 @@ with gr.Blocks(css=css, title="🔗 Link Audit Tool", theme=gr.themes.Soft()) as
             report_file = gr.File(label="Download Report", interactive=False)
             csv_file = gr.File(label="Download CSV", interactive=False)
 
-            # Resume log & progress (shared with new audit display format)
-            resume_progress = gr.Textbox(label="Resume Status", interactive=False, elem_classes=["status-bar"])
-            resume_log = gr.Textbox(label="Resume Log", lines=15, interactive=False, elem_classes=["log-area"])
+            resume_progress = gr.Textbox(label="Resume Status", interactive=False)
+            resume_log = gr.Textbox(label="Resume Log", lines=15, interactive=False)
             resume_pause_btn = gr.Button("⏸️ Pause Resume", variant="stop", interactive=False)
 
-            # Refresh
             refresh_btn.click(fn=load_past_runs, outputs=[runs_html, run_dropdown])
 
-            # Generate report
-            def gen_report_wrapper(run_id, domain_val):
-                filepath, msg = generate_report_for_run(run_id, domain_val)
-                return filepath, msg
-
             report_btn.click(
-                fn=gen_report_wrapper,
-                inputs=[run_dropdown, domain],
+                fn=generate_report_for_run,
+                inputs=[run_dropdown, domain_input],
                 outputs=[report_file, action_status],
             )
 
-            # CSV
-            def csv_wrapper(run_id):
-                filepath, msg = generate_csv_for_run(run_id)
-                return filepath, msg
-
-            csv_btn.click(fn=csv_wrapper, inputs=[run_dropdown], outputs=[csv_file, action_status])
+            csv_btn.click(fn=generate_csv_for_run, inputs=[run_dropdown], outputs=[csv_file, action_status])
 
-            # Delete
             def delete_wrapper(run_id):
                 msg = delete_selected_run(run_id)
                 html, dropdown = load_past_runs()
@@ -617,15 +658,13 @@ with gr.Blocks(css=css, title="🔗 Link Audit Tool", theme=gr.themes.Soft()) as
 
             delete_btn.click(fn=delete_wrapper, inputs=[run_dropdown], outputs=[action_status, runs_html, run_dropdown])
 
-            # Resume
             resume_btn.click(
                 fn=resume_audit,
-                inputs=[run_dropdown, domain, batch_size, timeout, delay, workers],
+                inputs=[run_dropdown, domain_input, batch_size_input, timeout_input, delay_input, workers_input],
                 outputs=[resume_log, resume_progress, resume_pause_btn, resume_btn],
             )
             resume_pause_btn.click(fn=pause_audit, outputs=[resume_progress])
 
-    # Auto-load runs on startup
     app.load(fn=load_past_runs, outputs=[runs_html, run_dropdown])