arg parse

analyze_startup.py

@@ -4,18 +4,7 @@ import re
 import os
 from datetime import datetime
 import json
-
-# Optional variable to define the directory of logs
-logs_dir = None  # Example: "./logs"
-
-# Explicit list of files and labels
-files_and_labels = [
-    # ("./logs/startup_lr62-pre-02.log", "lr62-pre-02"),
-    # ("./logs/startup_lr62-pre-01.log", "lr62-pre-01"),
-    # ("./logs/startup_lr62-dev.log", "lr62-dev"),
-    ("./logs/startup_lr62-prod-02.log", "lr62-prod-02"),
-    ("./logs/startup_lr62-prod-01.log", "lr62-prod-01"),
-]
+import argparse
 
 # Function to load data from files
 def load_data(file_path):
@@ -43,9 +32,45 @@ def find_files_and_labels(logs_dir):
             files_and_labels.append((file_path, label))
     return files_and_labels
 
-# If logs_dir is defined, populate files_and_labels from the directory
-if logs_dir:
-    files_and_labels = find_files_and_labels(logs_dir)
+# Parsing command-line arguments
+parser = argparse.ArgumentParser(
+    description='Process server startup logs.',
+    epilog='''
+Examples of usage:
+
+1. Use log directory and results directory with a specific date range:
+   python analyze_startup.py --log-dir ./logs --results-dir ./results --min-date 2023-07-01 --max-date 2023-12-31
+
+2. Use specific log files with automatic labels:
+   python analyze_startup.py -f ./logs/startup_lr62-prod-01.log -f ./logs/startup_lr62-prod-02.log
+
+3. Use specific log files with custom labels:
+   python analyze_startup.py -f ./logs/startup_lr62-prod-01.log -l prod01 -f ./logs/startup_lr62-prod-02.log -l prod02
+
+4. Combine log directory and specific files:
+   python analyze_startup.py --log-dir ./logs -f ./logs/startup_lr62-dev.log -l dev --results-dir ./results --min-date 2023-07-01
+    ''',
+    formatter_class=argparse.RawTextHelpFormatter
+)
+parser.add_argument('--log-dir', type=str, default=None, help='Directory containing log files')
+parser.add_argument('--results-dir', type=str, default='./results', help='Directory to save results')
+parser.add_argument('--max-date', type=str, default=None, help='Maximum date for filtering logs (YYYY-MM-DD)')
+parser.add_argument('--min-date', type=str, default='2023-07-01', help='Minimum date for filtering logs (YYYY-MM-DD)')
+parser.add_argument('-f', '--file', type=str, action='append', help='Explicit log file')
+parser.add_argument('-l', '--label', type=str, action='append', help='Label for explicit log file')
+
+args = parser.parse_args()
+
+# Handling explicit files and labels
+files_and_labels = []
+if args.file:
+    for i, file in enumerate(args.file):
+        label = args.label[i] if args.label and len(args.label) > i else re.search(r'startup_(.+)\.log', file).group(1)
+        files_and_labels.append((file, label))
+
+# If log_dir is defined, populate files_and_labels from the directory
+if args.log_dir:
+    files_and_labels += find_files_and_labels(args.log_dir)
 
 # Print the list of calculated files and labels in a readable format
 print("files_and_labels = [")
@@ -53,20 +78,16 @@ for file_path, label in files_and_labels:
     print(f'    ("{file_path}", "{label}"),')
 print("]")
 
-min_date_str = "2023-07-01"  # Default start date
-max_date_str = None  # Default end date
-
 # Convert date strings to datetime objects
-min_date = pd.to_datetime(min_date_str) if min_date_str else None
-max_date = pd.to_datetime(max_date_str) if max_date_str else None
+min_date = pd.to_datetime(args.min_date) if args.min_date else None
+max_date = pd.to_datetime(args.max_date) if args.max_date else None
 
 # Use the current date if max_date_str is None
 current_date_str = datetime.now().strftime("%Y-%m-%d")
-date_range_str = f"{min_date_str}_to_{max_date_str if max_date_str else current_date_str}"
+date_range_str = f"{args.min_date}_to_{args.max_date if args.max_date else current_date_str}"
 
 # Ensure the results directory exists
-results_dir = "./results"
-os.makedirs(results_dir, exist_ok=True)
+os.makedirs(args.results_dir, exist_ok=True)
 
 # Plot setup
 plt.figure(figsize=(14, 7))
@@ -94,7 +115,7 @@ for file_path, label in files_and_labels:
 
 # Create the filename base using the date range and labels
 labels_str = "_".join([label for _, label in files_and_labels])
-filename_base = f"{results_dir}/server_startup_{labels_str}_{date_range_str}"
+filename_base = f"{args.results_dir}/server_startup_{labels_str}_{date_range_str}"
 png_filename = f"{filename_base}.png"
 json_filename = f"{filename_base}.json"