Simple Python Logger

Question

Here is a simple Python logger I would like some general feedback on.

Features:

• Dynamically creates log folder and files wherever user runs the program
• Abillity to create multiple logs from a single program
• Names logs from time stamp to the latest second preventing name conflicts

I understand there may be far better alternatives to this but I just wanted to tinker around with this concept a bit!

Thank you!

logger.py:

#!/usr/bin/python
# -*- coding: utf-8 -*-
import datetime
import time
import os
import sys
# get user's current directory
directory = os.path.dirname(os.path.realpath(__file__)) + '/logs'
def createNewLog():
 # ensure log folder exists
 if not os.path.exists(directory):
 os.mkdir(directory)
 # get timestamp for naming the log
 timeInSeconds = time.time()
 timestamp = \
 datetime.datetime.fromtimestamp(timeInSeconds).strftime('%Y-%m-%d %H:%M:%S'
 )
 logFile = directory + '/' + timestamp + '.txt'
 mostRecentLog = directory + '/mostRecentLog.txt'
 # create log file and file with the log file's name
 with open(logFile, 'w') as log:
 with open(mostRecentLog, 'w') as recentLog:
 recentLog.write(timestamp)
def findMostRecentLog():
 # ensure logger has been intiated
 try:
 with open(directory + '/mostRecentLog.txt', 'r') as logFile:
 logName = logFile.read()
 except FileNotFoundError:
 print("Must initiate logger first!")
 sys.exit(1)
 return directory + '/' + logName + '.txt'
def log(comment):
 comment = str(comment)
 # write to log file retriving most recent log from corresponding file
 with open(findMostRecentLog(), 'a') as log:
 log.write(comment + '\n')

Answer 1

In case you do not know it yet, I suggest you to use pylint to parse your code and have automatic feedbacks. In fact some of my notes come from pylint.

Styling notes:

• Constant names like var "directory" require uppercase style PEP8
• Function and variable names require snake_case style PEP8
• Functions require docstring, as well as the module itself PEP8

Inline notes

#NOTE: 'from datetime import datetime' allows a shorter line below
import datetime
import time
import os
import sys
# get user's current directory
# NOTE: use os.path.join for concat filesystem paths for cross compatibility (also below)
directory = os.path.dirname(os.path.realpath(__file__)) + '/logs'
def createNewLog():
 # ensure log folder exists
 if not os.path.exists(directory):
 # NOTE: what if user has no permission to create files/directory in current directory? (also below)
 # HINT: give a way to choose where to store logs
 os.mkdir(directory)
 # get timestamp for naming the log
 timeInSeconds = time.time()
 # NOTE: fromtimestamp may raise an exception which is not not managed here
 timestamp = \
 datetime.datetime.fromtimestamp(timeInSeconds).strftime('%Y-%m-%d %H:%M:%S'
 )
 logFile = directory + '/' + timestamp + '.txt'
 mostRecentLog = directory + '/mostRecentLog.txt'
 # create log file and file with the log file's name
 # NOTE: redefining name 'log' (it's also a function name)
 # NOTE: this 'log' is not used. What it's that for?
 with open(logFile, 'w') as log:
 with open(mostRecentLog, 'w') as recentLog:
 recentLog.write(timestamp)
def findMostRecentLog():
 # ensure logger has been intiated
 try:
 with open(directory + '/mostRecentLog.txt', 'r') as logFile:
 logName = logFile.read()
 except FileNotFoundError:
 print("Must initiate logger first!")
 # NOTE: kind of rugh, I suggest returning an error and managing at caller level
 sys.exit(1)
 return directory + '/' + logName + '.txt'
def log(comment):
 # NOTE: why forcing the type?
 comment = str(comment)
 # write to log file retriving most recent log from corresponding file
 # NOTE: this dependency will make hard testing this function alone.
 with open(findMostRecentLog(), 'a') as log:
 log.write(comment + '\n')

• Daniel: The fromtimestamp assignment won't raise an exception. The log variable is only altered in the local scope of createNewLog, so it's mostly harmless (except for hurting readability, perhaps).
• Carlo Lobrano: @Coal_ could you explain better the first part? I saw from here that datetime.fromtimestamp could raise a OverflowError. What am I missing? Second part, I agree it's harmless in this context, but I consider it something to avoid (for readability, for future changes, etc.) not an error.