Analyze runtime of SuperCollider test CI

analyze_log.py

import re
import sys

import pandas as pd

# please insert a *recent* "View raw logs" URL of a macOS/linux test run here
# as the url is only valid within a timeframe, the URL needs to have been generated recently
#
# paste the URL into here or pass it as an argument
LOG_URL = ""

if not LOG_URL:
    LOG_URL = sys.argv[1]

print(f"Try to parse {LOG_URL}")

# load log file and split datetime and message (this is only approx but works 99.9%)
raw_df = pd.read_fwf(LOG_URL, colspecs=[(0, 29), (29, -1)], header=None)

# set column names for convenience
raw_df.columns = ['time', 'message']

# force str cast
raw_df['message'] = raw_df['message'].astype(str)

# remove color formatting escape chars
ascii_escape_chars = re.compile(r'\x1b\[[0-9;]*[a-zA-Z]')
raw_df['message'] = raw_df['message'].apply(lambda x: ascii_escape_chars.sub('', x))

# extract info via regex
regex = r"\[\d*/\d*\]\s*(?P<file_name>.*):(?P<test_name>\S*)\s*\((?P<time>\S*)s\)"
df = raw_df['message'].str.extract(regex)

# remove un-matched rows
df = df.dropna()

# force cast of time
df['time'] = df['time'].astype(float)

# save values to html table
df_top50 = df.sort_values('time', ascending=False).head(50)
df_top50.to_html("top_50.html")

top30_files_df = df.groupby("file_name")['time',].sum().sort_values('time', ascending=False).head(30)
top30_files_df.to_html("top_30_files.html")

print("Finished")