gistfile1.txt

# (venv) ➜  argspec_test python go.py
# Estimated file length is 3121
# Loading python file took 0.1354 seconds
# Loading yaml(python parser) file took 4.1489 seconds
# Loading yaml(C parser) file took 0.4825 seconds
# (venv) ➜  argspec_test 



import contextlib
import csv
import json
import pprint
import time
import urllib.request
import yaml

URL = "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_daily_reports/04-23-2020.csv"


def build_sources():
    with contextlib.closing(urllib.request.urlopen(URL)) as fp:
        raw_bytes = fp.read()

    data_str = raw_bytes.decode("utf8")
    print("Estimated file length is %s" % len(data_str.splitlines()))
    dump = {'data': [dict(row) for row in csv.DictReader(data_str.splitlines())]}

    with open('data.yaml', 'w') as file:
        yaml.dump(dump, file)

    dump = "DATA = " + pprint.pformat(dump, indent=4)
    with open("data.py", "w") as file1: 
        file1.writelines(dump) 

def time_python():
    tic = time.perf_counter()
    from data import DATA as laoded
    toc = time.perf_counter()
    py_elapsed = toc - tic
    print(f"Loading python file took {py_elapsed:0.4f} seconds")

def time_yamlp():
    tic = time.perf_counter()
    with open('data.yaml') as f:
        data = yaml.load(f, Loader=yaml.SafeLoader)
    toc = time.perf_counter()
    yaml_elasped = toc - tic
    print(f"Loading yaml(python parser) file took {yaml_elasped:0.4f} seconds")

def time_yamlc():
    tic = time.perf_counter()
    with open('data.yaml') as f:
        data = yaml.load(f, Loader=yaml.CLoader)
    toc = time.perf_counter()
    yaml_elasped = toc - tic
    print(f"Loading yaml(C parser) file took {yaml_elasped:0.4f} seconds")

if __name__ == "__main__": 
    build_sources()
    time_python()
    time_yamlp()
    time_yamlc()