scripts: add script to visualise scheduling history

scripts/scheduler-log.py

@@ -0,0 +1,67 @@
+#!/usr/bin/env python3
+
+import pandas as pd
+import sys
+
+if len(sys.argv) != 2:
+    print("Usage: scheduler-log.py <log file>")
+    sys.exit(1)
+
+logfile = sys.argv[1]
+
+# event keywords to look for in the log
+KEYWORDS = ['switching to', 'leaving']
+
+
+with open(logfile, 'r') as f:
+    log_lines = f.read()
+
+log_lines = log_lines.splitlines()
+
+# Initialize a dictionary to hold the current thread for each CPU
+cpu_state: dict[int, str | None] = {}
+log_data = []
+
+for lineid, line in enumerate(log_lines):
+    if not any(keyword in line for keyword in KEYWORDS):
+        continue
+
+    line_split = line.split('|')[:4]
+    line_split = [x.strip() for x in line_split]
+
+    cpu_num = int(line_split[1].split()[1])
+
+    if KEYWORDS[0] in line:
+        thread = line.split(KEYWORDS[0] + ' ')[-1].split(',')[0]
+
+        # detect if a thread is running on multiple CPUs
+        if thread in cpu_state.values():
+            print(f"Thread {thread} is running on multiple CPUs:")
+            print()
+
+            # print relevant log lines
+            for i in range(lineid - 5, lineid + 5):
+                if i < 0 or i >= len(log_lines):
+                    continue
+                if i == lineid:
+                    print("\033[31m", log_lines[i], "\033[m", sep="")
+                else:
+                    print(log_lines[i], "\033[m", sep="")
+
+    elif KEYWORDS[1] in line:
+        thread = None
+
+    cpu_state[cpu_num] = thread
+    log_data.append(cpu_state.copy())
+
+# Convert the log data to a DataFrame for easier viewing
+df = pd.DataFrame(log_data)
+df.columns = [f"CPU {i}" for i in df.columns]  # we can do this because dict keys are sorted
+df['Notes'] = df.apply(lambda row: 'Multiple CPUs' if row.dropna().duplicated().any() else '', axis=1)
+
+# sort the columns
+df = df[[f"CPU {i}" for i in range(df.shape[1] - 1)] + ['Notes']]
+
+# Display the table
+print()
+print(df.to_string())