def parse_profiler()

in bench_cluster/report.py [0:0]

• 43 lines of code
• 20 McCabe index (conditional complexity)

def parse_profiler(inp_dir):
    # Search for files ending in .json in the inp_dir and its subdirectories
    file_paths = glob.glob(os.path.join(inp_dir, "**", "*.json"), recursive=True)
    if not file_paths:
        raise ValueError(f"No .json file found in {inp_dir}")
    
    all_forward_durations = []
    all_backward_durations = []
    
    def _format_duration(duration):
        ms = duration // 1000
        us = duration % 1000
        return f"{ms}ms {us}μs"
        
    for file_path in file_paths:
        print(f"Processing file: {file_path}")
        
        with open(file_path, 'r') as f:
            trace_data = json.load(f)
        
        forward_durations = []
        backward_durations = []
        
        for event in trace_data['traceEvents']:
            if 'name' in event and 'dur' in event:
                if "forward" in event['name'].lower():
                    forward_durations.append(event['dur'])
                elif "backward" in event['name'].lower():
                    backward_durations.append(event['dur'])
        
        if forward_durations:
            all_forward_durations.extend(forward_durations)
        if backward_durations:
            all_backward_durations.extend(backward_durations)
        
        # Write the mean forward and backward durations to a csv file
        pattern = re.compile(r'dp-\d+_tp-\d+_pp-\d+_mbz-\d+')
        matching_index = next((i for i, part in enumerate(file_path.split("/")) if pattern.match(part)), None)
        
        if matching_index is None:
            raise ValueError(f"Could not find the specified pattern in {file_paths[0]}")

        assert matching_index < len(file_path.split("/")) - 1, "Matching index is out of bounds"
        output_file = "/".join(file_path.split("/")[:matching_index + 1]) + "/profiler.csv"
        
        if all_forward_durations or all_backward_durations:
            with open(output_file, 'w', newline='') as f:
                writer = csv.writer(f)
                writer.writerow(["forward", "backward"])
                writer.writerow([
                    _format_duration(int(mean(all_forward_durations))) if all_forward_durations else "N/A",
                    _format_duration(int(mean(all_backward_durations))) if all_backward_durations else "N/A"
                ])
            print(f"Results written to {output_file}")
        else:
            print("No forward or backward durations found in any file.")