I have a list text file with the following

Aerospace
    Aerodynamics
        Aerospace Technician/Engineer
        Performance Engineer
    Flight Dynamics & Control
        Aircraft Loads Professional
        Ground Test Rig Architecture Designer
Biomedical
    Bioinstrumentation
        Biomedical Engineer
        Biomaterials Engineer
    Biomaterials
        Researcher
        Systems Development Engineer

and I want to convert it into a format that looks something like this

[
    {
        "label" : "Aerospace",
        "children" : [
            {
                "label" : "Aerodynamics",
                "children" : [
                     ...
                ]
            }
        ]
    },
    {
        "label" : "Biomedical",
        "children" : ...
    }
]

I tried my hand at recursion, but I wasn't able to create an algo to do this. I'm hoping someone smarter than I can figure this out / give me pointers as to where I went wrong.

This is what I tried

def recurse(remaining_lines, depth):
    out = []
    for i in range(len(remaining_lines)):
        line = remaining_lines[i]
        nspaces = len(line) - len(line.lstrip())
        if nspaces > depth:
            out.extend(recurse(remaining_lines[i 1:], nspaces))
        else:
            return [{
                "id" : str(uuid.uuid4()),
                "label" : line.strip(),
                "children" : out,
            }]
    return out


# lines is an an array created by splitting the text file by \n
out = recurse(lines, -1)

CodePudding user response：

You can maintain a stack that maintains the current path in the hierarchy of (children) lists, where each list is accompanied with its indent. Whenever a node is created at that indent, it will be added to the corresponding list:

def create_hierarchy(lines):
    stack = [(-1, [])]
    for line in lines:
        label = line.lstrip()
        indent = len(line) - len(label)
        while indent <= stack[-1][0]:
            stack.pop()
        node = { "label": label, "children": [] }
        stack[-1][1].append(node)
        stack.append((indent, node["children"]))
    return stack[0][1]

CodePudding user response：

This seems to work

def parse_line(line, indent_str=' '*4):
    ind = 0
    while line.startswith(indent_str):
        line = line[len(indent_str):]
        ind  = 1
    return line, ind

# In Python >= 3.9 you can use:
# def parse_line(line, indent_str=' '*4):
#     ind = 0
#     while line != (line := line.removeprefix(indent_str)):
#         ind  = 1
#     return line, ind

assert(parse_line("Aerospace") == ('Aerospace', 0))
assert(parse_line("        Researcher") == ('Researcher', 2))
assert(parse_line("---A", indent_str='-') == ('A', 3))


def parse_lines(lines, ind=0, indent_str=' '*4):
    lines = lines.copy()
    lines_out = []
    while len(lines) > 0:
        line, new_ind = parse_line(lines[0], indent_str)
        if new_ind < ind:
            break
        if new_ind == ind   1:
            sub_lines, lines, new_ind = parse_lines(lines, new_ind, indent_str)
            lines_out[-1] = {"label": lines_out[-1], "children": sub_lines}
        elif new_ind == ind:
            lines_out.append(line)
            lines.pop(0)
        elif new_ind > ind   1:
            raise ValueError("Invalid indenting")
    if ind > 0:
        return lines_out, lines, new_ind
    assert(len(lines) == 0)  # make sure everything was parsed
    return lines_out


assert(parse_lines([' a', ' b'], 1, indent_str=' ') == (['a', 'b'], [], 1))
assert(parse_lines(['a', 'b', ' c', ' d', '  e', 'f'], indent_str=' ') == 
       ['a', {'label': 'b', 'children': ['c', {'label': 'd', 'children': ['e']}]}, 'f'])```

Output:

In [10]: parse_lines(lines)
Out[10]: 
[{'label': 'Aerospace',
  'children': [{'label': 'Aerodynamics',
    'children': ['Aerospace Technician/Engineer', 'Performance Engineer']},
   {'label': 'Flight Dynamics & Control',
    'children': ['Aircraft Loads Professional',
     'Ground Test Rig Architecture Designer']}]},
 {'label': 'Biomedical',
  'children': [{'label': 'Bioinstrumentation',
    'children': ['Biomedical Engineer', 'Biomaterials Engineer']},
   {'label': 'Biomaterials',
    'children': ['Researcher', 'Systems Development Engineer']}]}]