perf formatter

formatter.py

#!/usr/bin/env python                                                                                                                                                                                                                                                                                                  [31/1804]
#fileencoding: utf-8

from __future__ import print_function
import sys

MASK_32BIT = 0xFFFFFFFF

class BranchEntry:
    def __init__(self, line):
        symaddr, fname_with_offset = line.strip().split(" ")
        if "+" in fname_with_offset:
            fname, symoffset = fname_with_offset.strip().split("+")
            symoffset = int(symoffset, 16)
        else:
            fname = "[unknown]"
            symoffset = 0
        symaddr = int(symaddr, 16)
        self.fname = fname
        self.addr = symaddr - symoffset
        self.addr = self.addr & MASK_32BIT
        self.offset = symoffset & MASK_32BIT
        self.level = 0
        self.count = 1
        self.child = []

    def countup(self):
        self.count += 1

    def add_child(self, entry):
        self.child.append(entry)

    def fingerprint(self):
        return [self.fname, self.addr, self.offset]

    def set_level(self, level):
        self.level = level

    def get_child(self):
        return self.child

    def get_count(self):
        return self.count

    def print_all_leaf(self):
        print(self)
        for i in sorted(self.child, key=lambda x: x.get_count(), reverse=True):
            i.print_all_leaf()

    def calc_ratio_all_leaf(self, total_entry_count):
        self.ratio = float(self.count * 100) / total_entry_count
        for i in self.child:
            i.calc_ratio_all_leaf(total_entry_count)

    def __str__(self):
        return "{:6.2f}%, {}, 0x{:x}, {}, 0x{:x}".format(self.ratio, self.level, self.addr, self.fname, self.offset)


def is_same(a, b):
    for i, j in zip(a.fingerprint(), b.fingerprint()):
        if i != j:
            return False
    return True

def do_merge(dst, src):
    for d in dst.get_child():
        for s in src.get_child():
            if is_same(d, s):
                d.countup()
                do_merge(d, s)
                return
    for i in src.get_child():
        dst.add_child(i)

def merge_branch(root, branch_root):
    for r in root:
        if is_same(r, branch_root):
            r.countup()
            do_merge(r, branch_root)
            return
    root.append(branch_root)

def gen_branch(entries):
    branch_root = entries.pop()
    current = branch_root
    level = 0
    for e in entries[::-1]:
        level += 1
        e.set_level(level)
        current.add_child(e)
        current = e
    return branch_root

def gen_tree(filename):
    root = []
    entries = []
    total_entry_count = 0
    with open(filename, "r") as f:
        for line in f:
            if len(line) == 1:
                total_entry_count += 1
                branch_root = gen_branch(entries)
                merge_branch(root, branch_root)
                entries = []
            elif "\t" in line:
                entries.append(BranchEntry(line))
        for r in root:
            r.calc_ratio_all_leaf(total_entry_count)
        for r in sorted(root, key=lambda x: x.get_count(), reverse=True):
            r.print_all_leaf()

def main():
    if len(sys.argv) != 2:
        print("usage: formatter.py filename")
        sys.exit(1)
    gen_tree(sys.argv[1])

if __name__ == "__main__":
    main()