analyze_command_history.py

#!env python
# -*- coding: utf-8 -*-
import os
import collections
import argparse
from pathlib import Path

class Node:
    def __init__(self):
        self.count, self.key = 0, ""
        self.children = collections.defaultdict(Node)

def collect(file, root):
    isZshHistory = True if "zsh" in file else False
    with open(file, 'rb') as f:
        for line in f:
            try:
                line = line.decode()
                if isZshHistory:
                    index = line.strip().find(";")
                    line = line[index+1:]
                l = line.split()
                node = root
                node.count += 1
                for i in range(min(4, len(l))):
                    node = node.children[l[i]]
                    node.key = " ".join(l[:i+1])
                    node.count += 1
            except Exception as e:
                continue

def printNode(node, root, maxl, level):
    if node == root:
        print(('{0: <' + str(maxl) + "}").format("TOTAL " + str(root.count))[:maxl], "="*100)

    sortedChildren = sorted(node.children.values(), key=lambda x:x.count, reverse = True)
    for child in sortedChildren:
        if child.count/root.count > 0.01:
            l = "  "*level +  " " + child.key + " " +  str(child.count)
            l = ('{0: <' + str(maxl) + "}").format(l)[:maxl]
            print(l, "="*int(child.count/root.count*100))
            printNode(child, root, maxl, level+1)


if __name__ == '__main__':
    # params
    parser = argparse.ArgumentParser(description='analyze command history.')
    parser.add_argument('--history-file', action="store", dest="historyFile", default=str(Path.home())+"/.zsh_history")
    args = parser.parse_args()

    root = Node()
    collect(args.historyFile, root)
    printNode(root, root, 40, 0)