Visualize Construct in Python

visistruct.py

"""
MIT License

Copyright 2022 Dan Davis

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
"""

import itertools
import re
from dataclasses import dataclass
from typing import List, Optional

import construct as c
from rich import print
from rich.text import Text


FORMAT_CHAR = {
    ">B": "Int8ub",
    ">H": "Int16ub",
    ">L": "Int32ul",
    ">Q": "Int32ub",
    ">b": "Int8sb",
    ">h": "Int16sb",
    ">l": "Int32sb",
    ">q": "Int64sb",
    "<B": "Int8ul",
    "<H": "Int16ul",
    "<L": "Int32ul",
    "<Q": "Int64ul",
    "<b": "Int8ub",
    "<h": "Int16ub",
    "<l": "Int32ub",
    "<q": "Int64ub",
    "=B": "Int8un",
    "=H": "Int16un",
    "=L": "Int32un",
    "=Q": "Int32un",
    "=b": "Int8sn",
    "=h": "Int16sn",
    "=l": "Int32sn",
    "=q": "Int64sn",
    ">e": "Float16b",
    "<e": "Float16l",
    "=e": "Float16n",
    ">f": "Float32b",
    "<f": "Float32l",
    "=f": "Float32n",
    ">d": "Float64b",
    "<d": "Float64l",
    "=d": "Float64n",
}


def tokens(text):
    """Hacky regex based parser."""
    TOKEN_RX = r"""(?xm)
        (?P<nonbuild>   \+nonbuild        )|
        (?P<docs>       \+docs            )|
        (?P<type>       [^<\s]\w+[^>\s]   )|
        (?P<name>       \s\w+\s           )|
        (?P<open>       [<]               )|
        (?P<close>      [>]               )|
        (               \#.$              )|
        (               \s+               )
    """
    for match in re.finditer(TOKEN_RX, text):
        if match.lastgroup:
            if match.lastgroup == "name":
                yield (match.lastgroup, match[0].strip())
                continue
            yield (match.lastgroup, match[0])


@dataclass
class Field:
    """Used instead of a dict, its just a container."""
    name: str
    type: str
    length: int
    offset: int = 0  # has to be set later
    color: str = ""


class VisiStruct:
    """
    Visualize a Construct

    Given a Construct and either some bytes that can build it or an already
    parsed construct, try to work out the type, size, and offsets of every
    subconstruct.

    Notes:
        - incredibly hacky
        - would want flag for color output
    """

    def __init__(
        self,
        format: c.Construct,
        raw: Optional[bytes] = None,
        parsed: Optional[c.Container] = None,
    ):
        self.format = format
        self._raw = raw
        self._parsed = parsed
        self.fields: List[Field]

    @property
    def raw(self):
        if not self._raw:
            if self._parsed:
                self._raw = self.format.build(self._parsed)
        return self._raw

    @property
    def parsed(self):
        if not self._parsed:
            if self._raw:
                self._parsed = self.format.parse(self.raw)
        return self._parsed

    def create_fields(self, subcon=None) -> List[Field]:
        """Called recursively until all subcons parsed."""
        fields = []
        subcons = subcon if subcon else self.format.subcons
        for sub in subcons:
            types = []
            name = ""
            for kind, text in tokens(str(sub)):
                if kind == "type" and text != "Renamed":
                    types.append(text)
                elif kind == "name":
                    name = text
            # print(sub)
            # print(f" name: {name}")
            # print(f" types: {types}")
            # print()

            if "Const" in types:
                fields.append(Field(name, type=" ".join(types), length=sub.sizeof()))

            elif "FormatField" in types:
                if "Enum" in types:
                    type = sub.subcon.subcon.fmtstr
                    length = sub.subcon.subcon.length
                else:
                    type = sub.subcon.fmtstr
                    length = sub.subcon.length
                fields.append(Field(name, FORMAT_CHAR[type], length))

            # Only handles CString rn
            elif "StringEncoded" in types:
                enc = sub.subcon.encoding
                length = len(self.parsed[name].encode(enc))
                if "NullTerminated" in types:
                    type = "CString"
                    length += 1  # NULL byte
                fields.append(Field(name, type, length))

            elif types == ["Struct"]:
                fields.extend(self.create_fields(sub.subcon.subcons))

        color_wheel = itertools.cycle(
            ["cyan", "grey", "green", "yellow", "purple", "orange"]
        )

        # Set offsets and colors
        offset = 0
        for f in fields:
            offset += f.length
            f.offset = offset
            f.color = next(color_wheel)
        self.fields = fields
        return fields

    def chunk_bytes(self, chunk_size: int) -> list:
        if not self._raw:
            raise Exception("raw is None")

        fields = iter(self.fields)
        field = next(fields)

        as_hex = self._raw.hex()
        as_hex += ".." * (len(as_hex) // 8 + 1)
        out = []
        n = 0
        while n < (len(as_hex) - 2):
            if n // 2 > field.offset - 1 and n > 0:
                try:
                    field = next(fields)
                except StopIteration:
                    pass
            text = as_hex[n : n + 2]
            if text == "..":
                encoded = Text(f" {text} ")
            else:
                encoded = Text(f" {text} ", style=f"bold {field.color}")
            out.append(encoded)
            n += 2
        return [out[i : i + chunk_size] for i in range(0, len(out), chunk_size)]

    def print(self, width: int = 8):
        # width sets how many bytes to print per line
        fields = con.create_fields()
        print("Name         Type              Sz        Offset")
        print("-----------------------------------------------")
        for field in fields:
            print(
                Text(
                    f"{field.name:12} {field.type:16} {field.length:4} {field.offset:12}",
                    style=f"{field.color}",
                )
            )
        print()
        for chunk in self.chunk_bytes(width):
            text = Text()
            [text.append(c) for c in chunk]
            print(text)


if __name__ == "__main__":
    inner = c.Struct(
        "my_id" / c.Int16ul,
        "my_value" / c.Enum(c.Int32ul, HOT=1, COLD=2, JUST_RIGHT=3),
    )

    format = c.Struct(
        "my_header" / c.Const(b"FAKE"),
        "my_int" / c.Int32ul,
        "my_string" / c.CString("ascii"),
        "my_enum" / c.Enum(c.Int8ul, ONE=1, TWO=2, THREE=3),
        "my_inner" / inner,
    )

    args = dict(
        my_int=17,
        my_string="helloworld",
        my_enum="ONE",
        my_inner=dict(my_id=3, my_value="HOT"),
    )

    raw = format.build(args)
    con = VisiStruct(format, raw)
    con.print(8)