使用带有DateTimeLine的pygal值配置

Question

I'm plotting some CPU time statistics and would like to be able to annotate certain values. I wanted to use pygal's value configuration for this, but this does not seem to work in combination with the DateTimeLine chart I am using.

def generate_cpu_time_plot(csv_file_path, output_file):
    user = []
    system = []
    with open(csv_file_path, encoding="utf-8") as csv_file:
        reader = csv.DictReader(csv_file)
        for row in reader:
            time = datetime.fromtimestamp(int(row['time_millis']) / 1000)
            user.append((time, {
                'value': float(row['cpu_time_user'])
            }))
            system.append((time, {
                'value': float(row['cpu_time_system'])
            }))

    chart = pygal.DateTimeLine(x_label_rotation=35,
                               x_value_formatter=lambda dt: dt.strftime(
                                   '%d/%m/%Y %H:%M'), x_title='Time',
                               y_title='CPU time',
                               title=os.path.basename(csv_file_path))
    chart.add("User", user)
    chart.add("System", system)
    chart.render_to_file(output_file)

This gives me a TypeError: TypeError: '<' not supported between instances of 'dict' and 'dict'

Is there a way I can make this combination work? If I use the floats directly, without the dict, this works fine.

Answer 1

The tuple containing the x and y values you provide to the XY chart is the value. When you use the dict format for providing values you need to set the value property to this tuple.

At the moment your code tries to put a dict for the y value within the tuple. Changing the lines where the values are appended to the user and system lists to the following should fix it:

user.append({'value': (time, float(row['cpu_time_user']))})
system.append({'value': (time, float(row['cpu_time_system']))})