view_bash_profile_output.py

"""
Requires python3 and pandas.

Add this to a file:

    PS4='+ $EPOCHREALTIME\011 '
    exec 3>&2 2>/tmp/bashstart.$$.log
    set -x
    ...
    set +x
    exec 2>&3 3>&-

Then profile the output with this script, e.g.:
    python3 view_bash_profile_output.py /tmp/bashstart.35464.log 0.0221729

                                                   cmds  time_diff
1                             ++SHELL_SESSION_HISTORY=0   0.088930
2                    ++HISTCONTROL=ignoreboth:erasedups   0.022173
4                                 ++shopt -s histappend   0.026941
5                                    ++shopt -s cmdhist   0.026941
6     ++test -e /Users/danielnordberg/.iterm2_shell_...   0.044107
...                                                 ...        ...
7187                  +++export PYENV_VIRTUALENV_INIT=1   0.032187
7189  +++[[ trap '__bp_preexec_invoke_exec "$_"' DEB...   0.093222
7190  +++PROMPT_COMMAND='_pyenv_virtualenv_hook; tra...   0.063896
7191  ++alias 'git-pull-all=find . -maxdepth 3 -name...   0.051975
7193                                   ++'[' 1 == 1 ']'   0.028133

"""
import pandas as pd
import sys


pd.set_option('display.max_rows', None)


def clean(text):
    return text.replace('\n', '').replace('\t', '')


def main():
    if len(sys.argv) < 2:
        print(
            "Usage: python3 view_bash_profile_output.py {file} {microseconds_time_diff_greater_than:optional}")
        sys.exit()
    logfile = sys.argv[1]
    time_diff_greater_than = None
    if len(sys.argv) > 2:
        time_diff_greater_than = float(sys.argv[2])
    profile_info = []
    with open(logfile) as fp:
        counter = 0
        last_time = None
        for line in fp:
            if line.startswith('+'):
                counter += 1
                level, time, *cmds = line.split()
                # get time diff in milliseconds
                profile_info.append({
                    'cmds': level + " ".join([
                        clean(cmd)
                        for cmd in cmds]),
                    'time_diff': last_time and (
                        (float(time) - float(last_time)) * 10 ** 3) or 0
                })
                last_time = time
            else:
                # append to last command
                profile_info[counter - 1]['cmds'] += clean(line)

    df = pd.DataFrame(profile_info)
    if time_diff_greater_than:
        print(df[df.time_diff > time_diff_greater_than])
    else:
        print(df)


if __name__ == "__main__":
    main()