Update app.py

app.py

@@ -13,19 +13,20 @@ import numpy as np
 import torch
 import tempfile
 
-# Set up logging
+# Configure logging to track application events and errors
 logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
 logger = logging.getLogger(__name__)
 
-# Log dependency versions
+# Log versions of key dependencies for debugging
 logger.info(f"NumPy version: {np.__version__}")
 logger.info(f"Pandas version: {pd.__version__}")
 logger.info(f"PyTorch version: {torch.__version__}")
 logger.info(f"Gradio version: {gr.__version__}")
 
-# Store logs for UI display
+# Initialize list to store log messages for UI display
 log_messages = []
 
+# Custom logging handler to capture logs for Gradio UI
 class UILogHandler(logging.Handler):
     def emit(self, record):
         log_entry = self.format(record)
@@ -35,12 +36,12 @@ ui_handler = UILogHandler()
 ui_handler.setFormatter(logging.Formatter('%(asctime)s - %(levelname)s - %(message)s'))
 logger.addHandler(ui_handler)
 
-# Initialize Hugging Face summarization pipeline
+# Initialize Hugging Face summarization model (BART) for generating summaries
 try:
     summarizer = pipeline(
         "summarization",
         model="facebook/bart-large-cnn",
-        device=0 if torch.cuda.is_available() else -1,
+        device=0 if torch.cuda.is_available() else -1,  # Use GPU if available
         framework="pt"
     )
     logger.info("Hugging Face summarization pipeline initialized")
@@ -48,29 +49,32 @@ except Exception as e:
     logger.error(f"Failed to initialize Hugging Face pipeline: {e}")
     summarizer = None
 
-# Global variable to store logs data
+# Global DataFrame to store logs data
 logs = pd.DataFrame()
 
-# Load logs from uploaded CSV file
+# Load and validate logs from uploaded CSV file
 def load_logs(logs_file):
     global logs
     try:
         if logs_file is None:
             logger.error("No logs CSV file uploaded")
             return pd.DataFrame()
-        # Read CSV from uploaded file
+        # Write uploaded file to a temporary CSV
         with tempfile.NamedTemporaryFile(delete=False, suffix='.csv') as tmp:
             tmp.write(logs_file.read())
             tmp_path = tmp.name
         logs = pd.read_csv(tmp_path)
-        os.unlink(tmp_path)  # Clean up temporary file
-
-        # Validate required columns
+        os.unlink(tmp_path)  # Remove temporary file
+        # Define required columns; 'comments' is optional
         required_columns = ['device_id', 'lab_id', 'timestamp', 'status', 'usage_count', 'type']
         if not all(col in logs.columns for col in required_columns):
             logger.error(f"Missing required columns in logs: {required_columns}")
             return pd.DataFrame()
-
+        # Add 'comments' column if missing, filling with 'No comment'
+        if 'comments' not in logs.columns:
+            logs['comments'] = 'No comment'
+            logger.info("Added missing 'comments' column with default value 'No comment'")
+        # Convert timestamp to datetime and check for invalid entries
         logs['timestamp'] = pd.to_datetime(logs['timestamp'], errors='coerce')
         if logs['timestamp'].isna().any():
             logger.error("Invalid timestamps detected in logs")
@@ -81,11 +85,10 @@ def load_logs(logs_file):
         logger.error(f"Error loading logs: {e}")
         return pd.DataFrame()
 
-# Update dropdowns and validate uploads
+# Update dropdown menus based on uploaded CSV data
 def update_dropdowns(logs_file):
     global logs
     logs = load_logs(logs_file)
-    
     if logs.empty:
         return (
             gr.Dropdown(choices=[], value=None, label="Select Lab"),
@@ -93,11 +96,10 @@ def update_dropdowns(logs_file):
             "<p style='color: red;'>Please upload a valid logs CSV file to proceed.</p>",
             False  # Disable Generate Dashboard button
         )
-    
-    lab_choices = sorted(logs['lab_id'].unique().tolist())  # Sort for better UX
-    equipment_choices = sorted(logs['type'].unique().tolist())  # Sort for better UX
+    # Extract unique labs and equipment types, sorted for better UX
+    lab_choices = sorted(logs['lab_id'].unique().tolist())
+    equipment_choices = sorted(logs['type'].unique().tolist())
     upload_status = "<p style='color: green;'>Logs loaded successfully.</p>"
-    
     return (
         gr.Dropdown(choices=lab_choices, value=lab_choices[0] if lab_choices else None, label="Select Lab"),
         gr.Dropdown(choices=equipment_choices, value=equipment_choices[0] if equipment_choices else None, label="Select Equipment Type"),
@@ -105,7 +107,7 @@ def update_dropdowns(logs_file):
         True  # Enable Generate Dashboard button
     )
 
-# Anomaly detection
+# Detect anomalies using Isolation Forest
 def detect_anomalies(logs):
     try:
         if logs.empty:
@@ -120,14 +122,15 @@ def detect_anomalies(logs):
         logger.error(f"Error in anomaly detection: {e}")
         return logs
 
-# Generate text summary using Hugging Face
+# Generate text summary including comments
 def generate_text_summary(logs):
     if summarizer is None or logs.empty:
         return "No summary available due to missing data or model initialization."
     try:
         log_text = "\n".join(
             f"Device {row['device_id']} in lab {row['lab_id']} on {row['timestamp'].strftime('%Y-%m-%d %H:%M:%S')} "
-            f"was {row['status']} with usage count {row['usage_count']} (Type: {row['type']})."
+            f"was {row['status']} with usage count {row['usage_count']} (Type: {row['type']}). "
+            f"Comment: {row['comments']}"
             for _, row in logs.iterrows()
         )
         summary = summarizer(log_text, max_length=150, min_length=40, do_sample=False)[0]['summary_text']
@@ -137,28 +140,28 @@ def generate_text_summary(logs):
         logger.error(f"Error generating text summary: {e}")
         return f"Error generating summary: {str(e)}"
 
-# Generate executive insights
+# Generate executive insights including comments
 def generate_executive_insights(logs, anomalies):
     if summarizer is None or logs.empty:
         return "No executive insights available due to missing data or model initialization."
     try:
         log_text = "\n".join(
             f"Device {row['device_id']} in lab {row['lab_id']} on {row['timestamp'].strftime('%Y-%m-%d %H:%M:%S')} "
-            f"was {row['status']} with usage count {row['usage_count']} (Type: {row['type']})."
+            f"was {row['status']} with usage count {row['usage_count']} (Type: {row['type']}). "
+            f"Comment: {row['comments']}"
             for _, row in logs.iterrows()
         )
         anomaly_text = "\n".join(
-            f"Device {row['device_id']} showed anomalous usage count {row['usage_count']} on {row['timestamp'].strftime('%Y-%m-%d %H:%M:%S')}."
+            f"Device {row['device_id']} showed anomalous usage count {row['usage_count']} on "
+            f"{row['timestamp'].strftime('%Y-%m-%d %H:%M:%S')}. Comment: {row['comments']}"
             for _, row in anomalies.iterrows()
         ) if not anomalies.empty else "No anomalies detected."
-        
         prompt = (
             f"Summarize the following lab operations data into concise executive insights:\n\n"
             f"Logs:\n{log_text}\n\n"
             f"Anomalies:\n{anomaly_text}\n\n"
-            f"Provide high-level insights for lab managers, focusing on operational status and issues."
+            f"Provide high-level insights for lab managers, focusing on operational status, issues, and comments."
         )
-        
         insights = summarizer(prompt, max_length=200, min_length=50, do_sample=False)[0]['summary_text']
         logger.info("Executive insights generated successfully")
         return insights
@@ -166,7 +169,7 @@ def generate_executive_insights(logs, anomalies):
         logger.error(f"Error generating executive insights: {e}")
         return f"Error generating insights: {str(e)}"
 
-# Generate PDF report with summary and insights
+# Generate PDF report with summary, insights, and comments
 def generate_pdf(lab, equipment_type, filtered_logs, summary, insights):
     try:
         pdf_file = f"labops_report_{lab}_{equipment_type}_{datetime.datetime.now().strftime('%Y%m%d_%H%M%S')}.pdf"
@@ -200,13 +203,14 @@ def generate_pdf(lab, equipment_type, filtered_logs, summary, insights):
                 c.showPage()
                 y = 750
         
-        # Add device status
+        # Add device status with comments
         c.setFont("Helvetica-Bold", 14)
         c.drawString(100, y, "Device Status Summary")
         c.setFont("Helvetica", 12)
         y -= 20
         for _, row in filtered_logs.iterrows():
-            c.drawString(100, y, f"Device: {row['device_id']}, Status: {row['status']}, Usage: {row['usage_count']}")
+            c.drawString(100, y, f"Device: {row['device_id']}, Status: {row['status']}, "
+                                f"Usage: {row['usage_count']}, Comment: {row['comments'][:30]}")
             y -= 20
             if y < 100:
                 c.showPage()
@@ -222,7 +226,7 @@ def generate_pdf(lab, equipment_type, filtered_logs, summary, insights):
         logger.error(f"Error generating PDF: {e}")
         return None
 
-# Validate date format
+# Validate date format (YYYY-MM-DD)
 def validate_date(date_str):
     pattern = r'^\d{4}-\d{2}-\d{2}$'
     if not isinstance(date_str, str) or not re.match(pattern, date_str):
@@ -233,11 +237,11 @@ def validate_date(date_str):
     except ValueError:
         return False
 
-# Dashboard rendering
+# Render the dashboard with filtered data
 def render_dashboard(lab, equipment_type, start_date, end_date):
     try:
-        logger.info(f"Rendering dashboard with filters: lab={lab}, equipment_type={equipment_type}, start_date={start_date}, end_date={end_date}")
-
+        logger.info(f"Rendering dashboard with filters: lab={lab}, equipment_type={equipment_type}, "
+                    f"start_date={start_date}, end_date={end_date}")
         # Validate inputs
         if not lab or not equipment_type:
             return (
@@ -246,29 +250,24 @@ def render_dashboard(lab, equipment_type, start_date, end_date):
                 None, "No summary available.", "No insights available.",
                 "\n".join(log_messages[-10:]) or "No logs available."
             )
-
-        # Validate dates
+        # Validate and adjust dates
         if not validate_date(start_date):
             logger.warning(f"Invalid start_date format: {start_date}. Using default 2025-05-01.")
             start_date = "2025-05-01"
         if not validate_date(end_date):
             logger.warning(f"Invalid end_date format: {end_date}. Using default 2025-05-30.")
             end_date = "2025-05-30"
-
         start_dt = pd.to_datetime(start_date, format='%Y-%m-%d')
         end_dt = pd.to_datetime(end_date, format='%Y-%m-%d')
-
         if start_dt > end_dt:
             logger.warning("start_date is after end_date. Swapping dates.")
             start_dt, end_dt = end_dt, start_dt
-
-        # Filter logs
+        # Filter logs by lab, equipment type, and date range
         filtered_logs = logs[(logs['lab_id'] == lab) & (logs['type'] == equipment_type)]
         filtered_logs = filtered_logs[
             (filtered_logs['timestamp'] >= start_dt) &
             (filtered_logs['timestamp'] <= end_dt)
         ]
-
         if filtered_logs.empty:
             logger.warning("No data available for the selected filters")
             return (
@@ -277,11 +276,9 @@ def render_dashboard(lab, equipment_type, start_date, end_date):
                 None, "No summary available.", "No insights available.",
                 "\n".join(log_messages[-10:]) or "No logs available."
             )
-
         # Apply anomaly detection
         filtered_logs = detect_anomalies(filtered_logs)
-
-        # Device Cards
+        # Generate device cards for display
         device_cards = "<div style='display: flex; flex-wrap: wrap; gap: 20px;'>"
         for _, row in filtered_logs.iterrows():
             status_color = "green" if row['status'] == "active" else "red"
@@ -290,12 +287,12 @@ def render_dashboard(lab, equipment_type, start_date, end_date):
                     <h3 style='margin: 0; font-size: 18px;'>Device: {row['device_id']}</h3>
                     <p style='color: {status_color};'>Status: {row['status'].capitalize()}</p>
                     <p>Usage Count: {row['usage_count']}</p>
+                    <p>Comment: {row['comments'][:30]}</p>
                     <p>Last Log: {row['timestamp'].strftime('%Y-%m-%d %H:%M:%S')}</p>
                 </div>
             """
         device_cards += "</div>"
-
-        # Usage Trend Chart
+        # Generate usage trend chart
         fig_usage = px.line(
             filtered_logs,
             x='timestamp',
@@ -308,8 +305,7 @@ def render_dashboard(lab, equipment_type, start_date, end_date):
             yaxis_title="Usage Count",
             margin=dict(l=20, r=20, t=40, b=20)
         )
-
-        # Uptime Chart
+        # Generate uptime chart
         uptime = filtered_logs.groupby('device_id')['status'].apply(lambda x: (x == 'active').mean() * 100)
         fig_uptime = px.bar(
             uptime,
@@ -322,25 +318,20 @@ def render_dashboard(lab, equipment_type, start_date, end_date):
             yaxis_title="Uptime %",
             margin=dict(l=20, r=20, t=40, b=20)
         )
-
-        # Anomaly Alerts
+        # Generate anomaly table
         anomalies = filtered_logs[filtered_logs['anomaly'] == -1] if 'anomaly' in filtered_logs.columns else pd.DataFrame()
-        anomaly_table = anomalies[['device_id', 'timestamp', 'usage_count']].to_html(
+        anomaly_table = anomalies[['device_id', 'timestamp', 'usage_count', 'comments']].to_html(
             index=False,
             classes="table table-striped",
             border=0
         ) if not anomalies.empty else "<p>No anomalies detected.</p>"
-
         # Generate summary and insights
         summary = generate_text_summary(filtered_logs)
         insights = generate_executive_insights(filtered_logs, anomalies)
-
-        # PDF Download
+        # Generate PDF report
         pdf_data = generate_pdf(lab, equipment_type, filtered_logs, summary, insights)
-
-        # Logs for display
+        # Display recent logs
         logs_output = "\n".join(log_messages[-10:]) or "No logs available."
-
         logger.info("Dashboard rendered successfully")
         return (
             device_cards,
@@ -366,7 +357,7 @@ def render_dashboard(lab, equipment_type, start_date, end_date):
             logs_output
         )
 
-# Gradio Interface
+# Define Gradio interface
 with gr.Blocks(
     css="""
         .gradio-container {max-width: 1200px; margin: auto; padding: 20px;}
@@ -379,7 +370,7 @@ with gr.Blocks(
 ) as demo:
     gr.Markdown("## LabOps Central Dashboard")
     
-    # File Upload
+    # File Upload Section
     with gr.Group():
         gr.Markdown("### Upload Data File")
         gr.Markdown("**Note**: Please upload a logs.csv file to proceed.")
@@ -387,7 +378,7 @@ with gr.Blocks(
         upload_btn = gr.Button("Load Data", variant="primary", elem_classes="submit-btn")
         upload_status = gr.HTML(label="Upload Status")
 
-    # Filters
+    # Filter Options
     with gr.Group():
         gr.Markdown("### Filter Options")
         with gr.Row():
@@ -418,7 +409,7 @@ with gr.Blocks(
             )
         submit_btn = gr.Button("Generate Dashboard", variant="primary", elem_classes="submit-btn", interactive=False)
 
-    # Outputs
+    # Output Section
     with gr.Group():
         gr.Markdown("### Dashboard Results")
         with gr.Tabs():
@@ -437,14 +428,13 @@ with gr.Blocks(
         pdf_download = gr.File(label="Download PDF Report")
         log_display = gr.Textbox(label="Debug Logs", lines=10, interactive=False)
     
-    # Update dropdowns on file upload
+    # Connect upload button to dropdown update function
     upload_btn.click(
         fn=update_dropdowns,
         inputs=[logs_file],
         outputs=[lab, equipment_type, upload_status, submit_btn]
     )
-
-    # Update dashboard on submit
+    # Connect submit button to dashboard rendering
     inputs = [lab, equipment_type, start_date, end_date]
     outputs = [device_cards, usage_chart, uptime_chart, anomaly_table, pdf_download, summary_output, insights_output, log_display]
     submit_btn.click(render_dashboard, inputs=inputs, outputs=outputs)