Ugly text boxes

app.py

@@ -4,14 +4,53 @@ import numpy as np
 import gradio as gr
 
 
-# Sample stats
+# Sample test results with test names
 MODELS = {
-    "llama" : {"passed": 14, "failed": 1, "skipped": 6, "error": 0},
-    "gemma3" : {"passed": 42, "failed": 6, "skipped": 12, "error": 0},
-    "csm" : {"passed": 0, "failed": 0, "skipped": 0, "error": 1},
+    "llama": {
+        "amd": {
+            "passed": ["auth_login", "data_validation", "api_response", "file_upload", "cache_hit", "user_permissions", "db_query", "session_mgmt", "input_sanitize", "rate_limit", "error_handling", "memory_alloc", "thread_safety", "backup_restore"],
+            "failed": ["network_timeout"],
+            "skipped": ["gpu_accel", "cuda_ops", "ml_inference", "tensor_ops", "distributed", "multi_gpu"],
+            "error": []
+        },
+        "nvidia": {
+            "passed": ["auth_login", "data_validation", "api_response", "file_upload", "cache_hit", "user_permissions", "db_query", "session_mgmt", "input_sanitize", "rate_limit", "error_handling", "memory_alloc", "thread_safety", "backup_restore", "gpu_accel", "cuda_ops", "ml_inference", "tensor_ops"],
+            "failed": ["network_timeout", "distributed"],
+            "skipped": ["multi_gpu"],
+            "error": []
+        }
+    },
+    "gemma3": {
+        "amd": {
+            "passed": ["auth_login", "data_validation", "api_response", "file_upload", "cache_hit", "user_permissions", "db_query", "session_mgmt", "input_sanitize", "rate_limit", "error_handling", "memory_alloc", "thread_safety", "backup_restore", "config_load", "log_rotation", "health_check", "metrics", "alerts", "monitoring", "security_scan", "password_hash", "jwt_token", "oauth_flow", "csrf_protect", "xss_filter", "sql_injection", "rate_limiter", "load_balance", "circuit_break", "retry_logic", "timeout_handle", "graceful_shutdown", "hot_reload", "config_watch", "env_vars", "secrets_mgmt", "tls_cert", "encryption", "compression", "serialization", "deserialization", "validation"],
+            "failed": ["gpu_accel", "cuda_ops", "ml_inference", "tensor_ops", "distributed", "multi_gpu"],
+            "skipped": ["perf_test", "stress_test", "load_test", "endurance", "benchmark", "profiling", "memory_leak", "cpu_usage", "disk_io", "network_bw", "latency", "throughput"],
+            "error": []
+        },
+        "nvidia": {
+            "passed": ["auth_login", "data_validation", "api_response", "file_upload", "cache_hit", "user_permissions", "db_query", "session_mgmt", "input_sanitize", "rate_limit", "error_handling", "memory_alloc", "thread_safety", "backup_restore", "config_load", "log_rotation", "health_check", "metrics", "alerts", "monitoring", "security_scan", "password_hash", "jwt_token", "oauth_flow", "csrf_protect", "xss_filter", "sql_injection", "rate_limiter", "load_balance", "circuit_break", "retry_logic", "timeout_handle", "graceful_shutdown", "hot_reload", "config_watch", "env_vars", "secrets_mgmt", "tls_cert", "encryption", "compression", "serialization", "deserialization", "validation", "gpu_accel", "cuda_ops", "ml_inference", "tensor_ops"],
+            "failed": ["distributed", "multi_gpu"],
+            "skipped": ["perf_test", "stress_test", "load_test", "endurance", "benchmark", "profiling", "memory_leak", "cpu_usage", "disk_io", "network_bw"],
+            "error": []
+        }
+    },
+    "csm": {
+        "amd": {
+            "passed": [],
+            "failed": [],
+            "skipped": [],
+            "error": ["system_crash"]
+        },
+        "nvidia": {
+            "passed": [],
+            "failed": [],
+            "skipped": [],
+            "error": ["system_crash"]
+        }
+    }
 }
 
-def plot_model_stats(model_name: str) -> plt.Figure:
+def plot_model_stats(model_name: str) -> tuple[plt.Figure, str, str]:
     """Draws a pie chart of model's passed, failed, skipped, and error stats."""
     model_stats = MODELS[model_name]
     
@@ -23,10 +62,15 @@ def plot_model_stats(model_name: str) -> plt.Figure:
         'error': '#8B0000'      # Dark red
     }
     
+    # Convert test lists to counts for chart display
+    amd_stats = {k: len(v) for k, v in model_stats['amd'].items()}
+    nvidia_stats = {k: len(v) for k, v in model_stats['nvidia'].items()}
+    
     # Filter out categories with 0 values for cleaner visualization
-    filtered_stats = {k: v for k, v in model_stats.items() if v > 0}
+    amd_filtered = {k: v for k, v in amd_stats.items() if v > 0}
+    nvidia_filtered = {k: v for k, v in nvidia_stats.items() if v > 0}
     
-    if not filtered_stats:
+    if not amd_filtered and not nvidia_filtered:
         # Handle case where all values are 0 - minimal empty state
         fig, ax = plt.subplots(figsize=(10, 8), facecolor='#000000')
         ax.set_facecolor('#000000')
@@ -37,17 +81,27 @@ def plot_model_stats(model_name: str) -> plt.Figure:
         ax.set_xlim(0, 1)
         ax.set_ylim(0, 1)
         ax.axis('off')
-        return fig
+        return fig, "", ""
     
     # Create figure with two subplots side by side with padding
     fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(18, 9), facecolor='#000000')
     ax1.set_facecolor('#000000')
     ax2.set_facecolor('#000000')
     
-    # Get colors for filtered categories
-    chart_colors = [colors[category] for category in filtered_stats.keys()]
-    
-    def create_pie_chart(ax, device_label):
+    def create_pie_chart(ax, device_label, filtered_stats):
+        if not filtered_stats:
+            ax.text(0.5, 0.5, 'No test results', 
+                   horizontalalignment='center', verticalalignment='center',
+                   transform=ax.transAxes, fontsize=14, color='#888888',
+                   fontfamily='monospace', weight='normal')
+            ax.set_title(device_label, 
+                        fontsize=28, weight='bold', pad=2, color='#FFFFFF', 
+                        fontfamily='monospace')
+            ax.axis('off')
+            return
+            
+        chart_colors = [colors[category] for category in filtered_stats.keys()]
+        
         # Create minimal pie chart - full pie, no donut effect
         wedges, texts, autotexts = ax.pie(
             filtered_stats.values(), 
@@ -81,18 +135,19 @@ def plot_model_stats(model_name: str) -> plt.Figure:
                     fontfamily='monospace')
     
     # Create both pie charts with device labels
-    create_pie_chart(ax1, "AMD")
-    create_pie_chart(ax2, "Nvidia")
+    create_pie_chart(ax1, "amd", amd_filtered)
+    create_pie_chart(ax2, "nvidia", nvidia_filtered)
     
-    # Add subtle separation line between charts
+    # Add subtle separation line between charts - longer to match layout
     line_x = 0.5
-    fig.add_artist(plt.Line2D([line_x, line_x], [0.15, 0.8], 
+    fig.add_artist(plt.Line2D([line_x, line_x], [0.05, 0.9], 
                               color='#333333', linewidth=1, alpha=0.5,
                               transform=fig.transFigure))
     
     # Add central shared title for model name and test count
-    total_tests = sum(model_stats.values())
-    fig.suptitle(f'{model_name.lower()} • {total_tests} tests', 
+    total_amd = sum(amd_stats.values())
+    total_nvidia = sum(nvidia_stats.values())
+    fig.suptitle(f'{model_name.lower()} • amd: {total_amd} tests, nvidia: {total_nvidia} tests', 
                 fontsize=18, weight='normal', color='#CCCCCC', 
                 fontfamily='monospace', y=0.95)
     
@@ -100,14 +155,26 @@ def plot_model_stats(model_name: str) -> plt.Figure:
     plt.tight_layout()
     plt.subplots_adjust(top=0.85, wspace=0.4)  # Added wspace for padding between charts
     
-    return fig
+    # Generate separate failed tests info for AMD and NVIDIA
+    amd_failed = model_stats['amd']['failed']
+    nvidia_failed = model_stats['nvidia']['failed']
+    
+    amd_failed_info = "\n".join(amd_failed) if amd_failed else "None"
+    nvidia_failed_info = "\n".join(nvidia_failed) if nvidia_failed else "None"
+    
+    return fig, amd_failed_info, nvidia_failed_info
 
 def get_model_stats_summary(model_name: str) -> tuple:
     """Get summary stats for a model (total tests, success rate, status indicator)."""
     stats = MODELS[model_name]
-    total = sum(stats.values())
-    passed = stats['passed']
-    success_rate = (passed / total * 100) if total > 0 else 0
+    # Combine AMD and NVIDIA results
+    total_passed = len(stats['amd']['passed']) + len(stats['nvidia']['passed'])
+    total_failed = len(stats['amd']['failed']) + len(stats['nvidia']['failed'])
+    total_skipped = len(stats['amd']['skipped']) + len(stats['nvidia']['skipped'])
+    total_error = len(stats['amd']['error']) + len(stats['nvidia']['error'])
+    
+    total = total_passed + total_failed + total_skipped + total_error
+    success_rate = (total_passed / total * 100) if total > 0 else 0
     
     # Determine status indicator color
     if success_rate >= 80:
@@ -281,6 +348,21 @@ h1, h2, h3, p, .markdown {
     padding: 20px !important;
     margin-left: 300px !important;
 }
+
+/* Failed tests display */
+.failed-tests {
+    background-color: #1a1a1a !important;
+    color: #CCCCCC !important;
+    font-family: monospace !important;
+    font-size: 12px !important;
+    padding: 15px !important;
+    margin-top: 20px !important;
+    border-radius: 5px !important;
+    border: 1px solid #333333 !important;
+    max-height: 200px !important;
+    overflow-y: auto !important;
+    white-space: pre-line !important;
+}
 """
 
 # Create the Gradio interface with sidebar and dark theme
@@ -313,18 +395,39 @@ with gr.Blocks(title="Model Test Results Dashboard", css=dark_theme_css) as demo
                 format="png",
                 elem_classes=["plot-container"]
             )
+            
+            # Create two separate failed tests displays in a row layout
+            with gr.Row():
+                with gr.Column(scale=1):
+                    amd_failed_tests_output = gr.Textbox(
+                        label="AMD Failed Tests",
+                        value="",
+                        lines=8,
+                        max_lines=8,
+                        interactive=False,
+                        elem_classes=["failed-tests"]
+                    )
+                with gr.Column(scale=1):
+                    nvidia_failed_tests_output = gr.Textbox(
+                        label="NVIDIA Failed Tests",
+                        value="",
+                        lines=8,
+                        max_lines=8,
+                        interactive=False,
+                        elem_classes=["failed-tests"]
+                    )
     
     # Set up click handlers for each button
     for i, (model_name, button) in enumerate(zip(MODELS.keys(), model_buttons)):
         button.click(
             fn=lambda name=model_name: plot_model_stats(name),
-            outputs=plot_output
+            outputs=[plot_output, amd_failed_tests_output, nvidia_failed_tests_output]
         )
     
     # Initialize with the first model
     demo.load(
         fn=lambda: plot_model_stats(list(MODELS.keys())[0]),
-        outputs=plot_output
+        outputs=[plot_output, amd_failed_tests_output, nvidia_failed_tests_output]
     )
 
 if __name__ == "__main__":