"""

    AUTHOR:

        Khushal P Soonderji

    DATE:

        Tuesday, 26th Nov., 2024

    OBJECTIVE:

        To parse raw mail bodies and give a structure that is suitable for storing in No-SQL databases like MongoDB. The
        raw mail's text is expected to be compliant with standard defined in RFC 5322, RFC 2045, and maybe a few more.

    REFERENCES:

        1. GitHub: https://github.com/SpamScope/mail-parser
        2. RFC 5322: https://datatracker.ietf.org/doc/html/rfc5322
        3. RFC 2045: https://datatracker.ietf.org/doc/html/rfc2045

    DOWNLOADS:

        N/A

"""


#   *****************************************************************************************************************
#  *****                                                                                                          ****
# ***                                                   IMPORT                                                      ***
#  *****                                                                                                          ****
#   *****************************************************************************************************************


# To make sibling directories accessible for imports:
import sys
sys.path.append(".")
sys.path.append("..")

# System-level activities:
import io

# My utils:
from utils_v2.string import json
from utils_v2.string import regex
from utils_v2.date_time import date_time

# To work with mails:
import mailparser

# To work with datatypes:
from typing import Any, Dict


#   *****************************************************************************************************************
#  *****                                                                                                          ****
# ***                                           MACROS / ONE-TIME INIT                                              ***
#  *****                                                                                                          ****
#   *****************************************************************************************************************


# --- Nothing Yet


#   *****************************************************************************************************************
#  *****                                                                                                          ****
# ***                                                VARIABLES                                                      ***
#  *****                                                                                                          ****
#   *****************************************************************************************************************


# --- Nothing Yet


#   *****************************************************************************************************************
#  *****                                                                                                          ****
# ***                                                  FUNCTIONS                                                    ***
#  *****                                                                                                          ****
#   *****************************************************************************************************************


def parse(raw_mail: str | bytes) -> Dict[str, Any]:

    """
    To parse the raw mail text to a usable JSON that can even be stored on a No-SQL database like MongoDB.
    DOCUMENTATION:
        1. GitHub: https://github.com/SpamScope/mail-parser
        2. RFC 5322: https://datatracker.ietf.org/doc/html/rfc5322
        3. RFC 2045: https://datatracker.ietf.org/doc/html/rfc2045
    :param raw_mail: The raw mail body that adheres to RFC 5322 and RFC 2045 (among others).
    :return: The parsed JSON format (dict) of the mail.
    """

    # Parse the raw format:
    if isinstance(raw_mail, str): parsed_mail = mailparser.parse_from_string(raw_mail)
    else: parsed_mail = mailparser.parse_from_bytes(raw_mail)

    # Format the attachments:
    message_attachments = [
        {
            "filename": attachment["filename"],
            "type": attachment["mail_content_type"],
            "cid": regex.find_first(text = attachment["content-id"], pattern = r"(?<=<).*(?=>)"),
            "rawCid": attachment["content-id"],
            "contentDisposition": (cd := attachment["content-disposition"]),
            "isInline": True if cd.lower().find("inline") >= 0 else False,
            "charset": attachment["charset"],
            "contentTransferEncoding": attachment["content_transfer_encoding"],
            "payload": attachment["payload"]
        } for attachment in parsed_mail.attachments
    ]

    # Put everything together:
    return {
        "ts": date_time.to_timezone(parsed_mail.date, timezone = date_time.TIMEZONE_UTC),
        "headers": parsed_mail.headers,
        "from": [{"name": _[0] or _[1], "email": _[1]} for _ in parsed_mail.headers["From"]],
        "to": [{"name": _[0] or _[1], "email": _[1]} for _ in parsed_mail.headers["To"]],
        "cc": [{"name": _[0] or _[1], "email": _[1]} for _ in parsed_mail.headers.get("Cc", [])],
        "bcc": [{"name": _[0] or _[1], "email": _[1]} for _ in parsed_mail.headers.get("Bcc", [])],
        "text": parsed_mail.text_plain,
        "html": parsed_mail.text_html,
        "attachments": message_attachments,
    }


#   *****************************************************************************************************************
#  *****                                                                                                          ****
# ***                                               MAIN PROGRAM                                                    ***
#  *****                                                                                                          ****
#   *****************************************************************************************************************


if __name__ == "__main__":

    pass