"""
parse commandline output, reconstruct a fileststem and determine the sum of all folders < 100kB
"""

text = """$ cd /
$ ls
dir a
14848514 b.txt
8504156 c.dat
dir d
$ cd a
$ ls
dir e
29116 f
2557 g
62596 h.lst
$ cd e
$ ls
584 i
$ cd ..
$ cd ..
$ cd d
$ ls
4060174 j
8033020 d.log
5626152 d.ext
7214296 k
"""

with open('2022/7/stdout.txt') as filein:
    text = filein.read()

def parse_stdin(stdin:str):
    outputs = stdin.split('$')
    outputs = [[l.strip() for l in o.splitlines()] for o in outputs if o]
    return recursive_parse(outputs)

def recursive_parse(outputs:list):
    structure = {}
    while True:
        try:
            current = outputs.pop(0)
        except IndexError:
            return structure
        command = current.pop(0)

        if command.startswith('ls'):
            content = [c.split() for c in current]
            for value, name in content:
                structure[name] = int(value) if value.isnumeric() else value

        elif command.startswith('cd'):
            dest = command.split()[1]
            if dest == '..':
                return structure
            else:
                structure[dest] = recursive_parse(outputs)

def recursive_sum(d:dict, folders:list=None, initial=True):
    s=0
    folders = [] if folders==None else folders

    for name, value in d.items():
        if isinstance(value, dict):
            zs = recursive_sum(value, folders, False)
            folders.append((name,zs))
        else:
            zs = value
        s+=zs

    if initial:
        return folders
    else:
        return s

def folders_up_to_size(folders:list, size:int):
    return  [(k,v) for k,v in folders if v<=size]

def recursive_print(d:dict, indent = 0, increment=2):
    for k,v in d.items():
        if isinstance( v, dict):
            print(' '*indent, 'dir', k)
            recursive_print(v, indent+increment)
        else:
            print(' '*indent, k,v)


system = parse_stdin(text)
# recursive_print(system)
sums = recursive_sum(system)
l = folders_up_to_size(sums,100000)
print(l)
print(sum([v for _,v in l]))

"""
now find the smallest folser that can be deleted to free up the space needed to keep the total size under 40MB
"""


def folders_over_size(folders:list, size:int):
    return  [(k,v) for k,v in folders if v>size]

total = sums[-1][1]
over = total-40000000
print('filesystem overfull by:', over)

l = folders_over_size(sums,over)
l.sort(key = lambda x: x[1])
print(l[0])