analyse_performance.py

analyse_performance.py

import json
import sys
import pandas as pd
import glob

def analyse(files):
    '''
    - latency (responseStart - requestStart)
    - time to interactive (domInteractive - navigationStart)
    - load time (loadEventStart - navigationStart)
    - first meaningful paint
    '''

    N_bins = 4

    print("Files in analysis: {}".format(len(files)))

    analysis_results = []

    for _file in files:
        data = json.load(open(_file))

        res = []
        for row in data:
            res.append({
                'number_of_images': row['number_of_images'],
                'latency': row['requestStart'] - row['requestStart'],
                'render': row['domComplete'] - row['domLoading'],
                'time_to_interactive': row['domInteractive'] - row['navigationStart'],
                'load_time': row['loadEventStart'] - row['navigationStart'],
                'first_meaningful_paint': row['firstMeaningfulPaint'],
                'url': row['url']
            })
        df = pd.DataFrame(res)
        df['number_of_images_bin'] = pd.cut(df['number_of_images'], N_bins)

        per_json_analysis = df.groupby(df['number_of_images_bin'])\
                              .agg(['mean'])\
                              .reset_index()
        per_json_analysis['file'] = _file

        analysis_results.append(per_json_analysis)

    analysis_df = pd.concat(analysis_results)
    print(analysis_df.groupby(analysis_df['number_of_images_bin']).agg(['mean', 'std']))

if __name__ == '__main__':
    analyse(sys.argv[1:])