"""python-fixed-width-test-data-cobol.py

Generate a deterministic fixed-width transaction file for the COBOL reader in
python-fixed-width.cob, then parse the file back and print the
same control totals the COBOL program prints.

Record layout (64 bytes per line, LF line ending, matches the copybook
code/python-fixed-width-test-data-cobol.cpy):

  cols  1- 8  Txn-Id       PIC 9(8)      zero padded
  cols  9-14  Txn-Account  PIC 9(6)      zero padded
  cols 15-16  Txn-Type     PIC X(2)      "DR" or "CR"
  cols 17-27  Txn-Amount   PIC 9(9)V99   11 digits, implied decimal point (cents)
  cols 28-35  Txn-Date     PIC 9(8)      YYYYMMDD
  cols 36-64  Txn-Memo     PIC X(29)     left justified, space padded

Inputs : none (random.seed(20260905) makes the file reproducible)
Outputs: datasets/python-fixed-width-test-data-cobol.dat and control totals on stdout
Run    : python code/python-fixed-width-test-data-cobol.py   (from the site folder)
Standard library only.
"""
from __future__ import annotations

import os
import random
import struct
import sys

ROWS = 200
RECORD_LEN = 64
OUT = os.path.join("datasets", "python-fixed-width-test-data-cobol.dat")
ACCOUNTS = [100234, 100235, 100240, 100251, 100262, 100270, 100281, 100299]
MEMOS = ["Card purchase", "Standing order", "Salary", "Refund", "Direct debit",
         "Cash withdrawal", "Transfer in", "Transfer out", "Interest", "Fee"]

# struct format: 8s 6s 2s 11s 8s 29s = 64 bytes; "s" fields are raw bytes
LAYOUT = struct.Struct("8s6s2s11s8s29s")
assert LAYOUT.size == RECORD_LEN


def make_record(n: int, rng: random.Random) -> str:
    txn_id = f"{n:08d}"
    account = f"{rng.choice(ACCOUNTS):06d}"
    txn_type = rng.choice(["DR", "DR", "CR"])
    cents = rng.randint(1, 250_000_00)          # 0.01 .. 250,000.00
    amount = f"{cents:011d}"                    # PIC 9(9)V99: no point, cents included
    day = rng.randint(1, 28)
    month = rng.randint(1, 9)
    date = f"2026{month:02d}{day:02d}"
    memo = rng.choice(MEMOS)[:29].ljust(29)     # PIC X(29): space padded on the right
    rec = txn_id + account + txn_type + amount + date + memo
    assert len(rec) == RECORD_LEN, len(rec)
    return rec


def write_file(path: str) -> None:
    rng = random.Random(20260905)
    os.makedirs(os.path.dirname(path), exist_ok=True)
    with open(path, "w", encoding="ascii", newline="\n") as fh:   # LF only, like a Unix line sequential file
        for n in range(1, ROWS + 1):
            fh.write(make_record(n, rng) + "\n")


def parse_file(path: str) -> dict:
    totals = {"read": 0, "dr_count": 0, "cr_count": 0, "bad": 0, "dr_total": 0, "cr_total": 0}
    with open(path, "rb") as fh:
        for raw in fh:
            line = raw.rstrip(b"\n")
            if len(line) != RECORD_LEN:
                raise ValueError(f"record of {len(line)} bytes, expected {RECORD_LEN}")
            txn_id, account, txn_type, amount, date, memo = LAYOUT.unpack(line)
            totals["read"] += 1
            if not amount.isdigit():           # the COBOL program's IS NUMERIC test
                totals["bad"] += 1
                continue
            cents = int(amount)                # implied decimal point: value is cents
            if txn_type == b"DR":
                totals["dr_count"] += 1
                totals["dr_total"] += cents
            elif txn_type == b"CR":
                totals["cr_count"] += 1
                totals["cr_total"] += cents
            else:
                totals["bad"] += 1
    return totals


def money(cents: int, width: int = 16) -> str:
    """Render like COBOL PIC Z,ZZZ,ZZZ,ZZ9.99 (16 chars, leading zeros suppressed)."""
    return f"{cents / 100:,.2f}".rjust(width)


def signed_money(cents: int, width: int = 16) -> str:
    """Render like COBOL PIC -,ZZZ,ZZZ,ZZ9.99: one minus is a FIXED sign in the leftmost position
    (a space when not negative), then the zero-suppressed digits (16 chars in all)."""
    return ("-" if cents < 0 else " ") + f"{abs(cents) / 100:,.2f}".rjust(width - 1)


def main() -> int:
    write_file(OUT)
    t = parse_file(OUT)
    print(f"Wrote {ROWS} records of {RECORD_LEN} bytes to {OUT}")
    print("Expected COBOL output:")
    print(f"Records read    : {t['read']:>7,}")
    print(f"Debits          : {t['dr_count']:>7,}  {money(t['dr_total'])}")
    print(f"Credits         : {t['cr_count']:>7,}  {money(t['cr_total'])}")
    print(f"Rejected        : {t['bad']:>7,}")
    print(f"Net (CR - DR)   : {signed_money(t['cr_total'] - t['dr_total'])}")
    return 0


if __name__ == "__main__":
    sys.exit(main())