sort run column in generate_plot.py

Author: Smaran-Teja

alarm.png

@@ -2,33 +2,78 @@
 import json
 import pandas as pd
 import matplotlib.pyplot as plt
+from datetime import datetime
 
 base_dir = "results"
 records = []
 
-for folder in os.listdir(base_dir):
+def extract_timestamp(folder_name):
+    """
+    Extract timestamp from folder name like 'run_2025-09-07_18-27-28'
+    Returns a datetime object for sorting.
+    """
+    try:
+        # Extract the timestamp part after 'run_'
+        timestamp_str = folder_name.split('run_')[1]
+        # Parse the timestamp (assuming format: YYYY-MM-DD_HH-MM-SS)
+        return datetime.strptime(timestamp_str, '%Y-%m-%d_%H-%M-%S')
+    except (IndexError, ValueError):
+        # If parsing fails, return the folder name as-is for fallback sorting
+        return datetime.min
+
+# Get and sort folders by timestamp
+folders = [f for f in os.listdir(base_dir) if os.path.isdir(os.path.join(base_dir, f))]
+folders.sort(key=extract_timestamp)
+
+# Process folders in sorted order
+for folder in folders:
     folder_path = os.path.join(base_dir, folder)
-    if os.path.isdir(folder_path):
-        for fname in os.listdir(folder_path):
-            if fname.endswith(".json"):
-                with open(os.path.join(folder_path, fname)) as f:
-                    data = json.load(f)
-                    data["run"] = folder   # add folder as run label
-                    data["filename"] = fname
-                    records.append(data)
+    for fname in os.listdir(folder_path):
+        if fname.endswith(".json"):
+            with open(os.path.join(folder_path, fname)) as f:
+                data = json.load(f)
+                data["run"] = folder  # add folder as run label
+                data["filename"] = fname
+                records.append(data)
 
 df = pd.DataFrame(records)
 print(df.head())
 
+# Convert run column to categorical with the sorted order to maintain it in plots
+df['run'] = pd.Categorical(df['run'], categories=folders, ordered=True)
+
+# Sort dataframe by run to ensure correct order
+df = df.sort_values('run')
+
 for filename, group in df.groupby("filename"):
-    plt.figure()
-    plt.plot(group["run"], group["real_time_ms"], marker="o", label="real_time_ms")
-    plt.plot(group["run"], group["cpu_time_ms"], marker="o", label="cpu_time_ms")
-    plt.plot(group["run"], group["gc_time_ms"], marker="o", label="gc_time_ms")
+    plt.figure(figsize=(10, 6))  # Increased figure size for better readability
+    
+    # Sort group by run to ensure correct plotting order
+    group = group.sort_values('run')
+    
+    # Use index positions for x-axis to ensure proper spacing
+    x_positions = range(len(group))
+    
+    plt.plot(x_positions, group["real_time_ms"], marker="o", label="real_time_ms")
+    plt.plot(x_positions, group["cpu_time_ms"], marker="o", label="cpu_time_ms")
+    plt.plot(x_positions, group["gc_time_ms"], marker="o", label="gc_time_ms")
+    
     plt.title(f"Timing for {filename}")
     plt.xlabel("Run")
     plt.ylabel("Time (ms)")
     plt.legend()
-    plt.xticks(rotation=45)
+    
+    # Set x-tick labels to the run names, but only show every nth label if there are many
+    run_labels = group["run"].tolist()
+    if len(run_labels) > 10:
+        # Show every 2nd or 3rd label if there are many runs
+        step = max(1, len(run_labels) // 10)
+        tick_positions = list(range(0, len(run_labels), step))
+        tick_labels = [run_labels[i] for i in tick_positions]
+        plt.xticks(tick_positions, tick_labels, rotation=45, ha='right')
+    else:
+        plt.xticks(x_positions, run_labels, rotation=45, ha='right')
+    
+    plt.grid(True, alpha=0.3)
     plt.tight_layout()
-    plt.savefig(f"{filename[:-5]}.png")
+    plt.savefig(f"{filename[:-5]}.png", dpi=100, bbox_inches='tight')