Refund rows: separate the selected payment's statuses before totaling

Four invented refund rows total $50.00, matching the invented original charge. That does not make the selected payment fully refunded: its supplied rows contain $12.50 marked succeeded, $7.50 pending and $10.00 failed. Another payment contributes the remaining $20.00. Actual receipt and whether another refund is needed remain unknown.

Prepare the selected snapshot

This fictional U.S. direct-merchant snapshot has one account and a fixed USD payment. The existing credit-note guide [cited revision] handles allocation and funding; this calculation leaves that guidance intact.

Stripe's Refund object, Attributes describes integer amounts in the smallest currency unit, the refunded charge relationship and five status labels: succeeded, pending, failed, canceled and requires_action. Its currency guide, Specify amounts uses 1000 for 10 USD; this fixture handles USD only.

Authored row	Relation to selected payment	Fixture amount/status
refund-success	Selected charge A	$12.50 / succeeded
refund-pending	Selected charge A	$7.50 / pending
refund-failed	Selected charge A	$10.00 / failed
refund-other	Other charge B	$20.00 / succeeded, excluded

1. Bind account, selected payment and currency before totaling.
2. Use one current row per refund ID. Identical duplicates and same-ID histories stop; resolve them before rerunning.
3. Keep matching statuses separate and list other-charge exclusions. Missing rows yield zero only in this inventory, without proving no refund exists.
4. Hand back stages and unknowns; infer no amount owed, entitlement or permission to refund again.

Positive amounts, nonnull statuses, unexpanded IDs and fixed payment are own subset boundaries, not Stripe API schema. Other currencies/accounts or new fields/evidence need separate evaluation. Stripe's refund guidance, state transitions and reversal/trace sections illustrates why recorded success is not a substitute for receipt evidence: states can change and a reversal need not show a separate credit.

Reproduce the exact calculation

Save the five named fences as UTF-8 files with one final LF each in an empty directory. On CPython 3.12.3 run:

python sum_refunds.py fixture.json 363df99c1d548fd02e022fa61c6286c6ffbf7691f29dc0c701c937138c876b89
python check.py

The first output equals expected.json: 642 bytes, SHA-256 e95ba243ec164accf114b75b8ffcc73f1eb896481f04906e2fce71a3b2af84e6. The checker compares all 10 top-level fields, nested subtotals and full bytes: four positive cases, one wrong-total demonstration and 18 stops, recorded individually. Python JSON documentation, loads and object_pairs_hook supports the standard decoder; no replacement parser or general refund validator is used.

Evidence: Sources read through 2026-10-05T19:48:31.254905Z; no update dates established. Python reference 3.12.15, execution 3.12.3 at 2026-10-05T19:52:06.831189Z, exit 0; expectations authored beforehand. A distinct same-controller checker read the sources, replayed all 23 public cases and checked ten additional executable boundaries; no independent formal coverage. No account/API/payment/refund/bank/contact action. Recheck sources and changed inputs before use and by 2026-11-05T19:52:06.831189Z. Feedback should name a failed public case without private transaction details. Billing standard [current article].

Complete public files and measured results

fixture.json

{
  "origin": "fictional-authored",
  "fixture_account": "fictional-us-direct-merchant",
  "selected_payment": {
    "charge": "fictional-charge-A",
    "amount": 5000,
    "currency": "usd"
  },
  "refunds": [
    {
      "fixture_account": "fictional-us-direct-merchant",
      "id": "fictional-refund-success",
      "charge": "fictional-charge-A",
      "amount": 1250,
      "currency": "usd",
      "status": "succeeded"
    },
    {
      "fixture_account": "fictional-us-direct-merchant",
      "id": "fictional-refund-pending",
      "charge": "fictional-charge-A",
      "amount": 750,
      "currency": "usd",
      "status": "pending"
    },
    {
      "fixture_account": "fictional-us-direct-merchant",
      "id": "fictional-refund-failed",
      "charge": "fictional-charge-A",
      "amount": 1000,
      "currency": "usd",
      "status": "failed"
    },
    {
      "fixture_account": "fictional-us-direct-merchant",
      "id": "fictional-refund-other",
      "charge": "fictional-charge-B",
      "amount": 2000,
      "currency": "usd",
      "status": "succeeded"
    }
  ]
}

sum_refunds.py

"""Own fictional USD snapshot arithmetic; no API validation or financial action."""
import hashlib
import json
import sys

ACCOUNT = "fictional-us-direct-merchant"
PAYMENT = {"charge": "fictional-charge-A", "amount": 5000, "currency": "usd"}
STATES = ("succeeded", "pending", "failed", "canceled", "requires_action")

def require(condition, message):
    if not condition:
        raise ValueError(message)

def keys(value, names):
    require(type(value) is dict and set(value) == set(names.split()), "missing or unsupported fields")

def unique_members(pairs):
    result = {}
    for name, value in pairs:
        require(name not in result, "duplicate JSON member")
        result[name] = value
    return result

def decode(raw):
    return json.loads(raw, object_pairs_hook=unique_members)

def subtotal(data):
    keys(data, "origin fixture_account selected_payment refunds")
    require(data["origin"] == "fictional-authored" and data["fixture_account"] == ACCOUNT, "outside authored account context")
    payment = data["selected_payment"]
    keys(payment, "charge amount currency")
    require(type(payment["amount"]) is int and payment == PAYMENT, "changed selected payment")
    require(type(data["refunds"]) is list, "refund snapshot list required")
    totals = dict.fromkeys(STATES, 0)
    seen, excluded = set(), []
    for row in data["refunds"]:
        keys(row, "fixture_account id charge amount currency status")
        require(row["fixture_account"] == ACCOUNT, "mixed account context")
        require(type(row["id"]) is str and bool(row["id"]) and type(row["charge"]) is str and bool(row["charge"]), "missing relationship or identifier")
        require(row["id"] not in seen, "duplicate refund snapshot; select one current record first")
        seen.add(row["id"])
        require(type(row["amount"]) is int and row["amount"] > 0, "positive integer minor-unit amount required")
        require(row["currency"] == "usd", "outside USD-only fixture")
        require(row["status"] in STATES, "unknown or unset fixture status")
        if row["charge"] != payment["charge"]:
            excluded.append(row["id"])
        else:
            totals[row["status"]] += row["amount"]
    return {
        "fixture_account": ACCOUNT,
        "selected_charge": payment["charge"],
        "currency": "usd",
        "fixture_original_charge_minor_units": payment["amount"],
        "fixture_status_subtotals_minor_units": totals,
        "excluded_other_charge_ids": sorted(excluded),
        "actual_bank_receipt": None,
        "inventory_complete": None,
        "additional_refund_needed": None,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
    }

def main():
    raw = open(sys.argv[1], "rb").read()
    require(hashlib.sha256(raw).hexdigest() == sys.argv[2], "fixture bytes changed")
    print(json.dumps(subtotal(decode(raw)), indent=2))

if __name__ == "__main__":
    try:
        main()
    except (ValueError, TypeError, KeyError, IndexError) as error:
        raise SystemExit("STOP: " + str(error))

expected.json

{
  "fixture_account": "fictional-us-direct-merchant",
  "selected_charge": "fictional-charge-A",
  "currency": "usd",
  "fixture_original_charge_minor_units": 5000,
  "fixture_status_subtotals_minor_units": {
    "succeeded": 1250,
    "pending": 750,
    "failed": 1000,
    "canceled": 0,
    "requires_action": 0
  },
  "excluded_other_charge_ids": [
    "fictional-refund-other"
  ],
  "actual_bank_receipt": null,
  "inventory_complete": null,
  "additional_refund_needed": null,
  "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
}

check.py

"""Replay own declared USD refund-snapshot outputs and local stop boundaries."""
import copy
import datetime
import hashlib
import importlib.util
import json
from pathlib import Path
import platform
import subprocess
import sys

P = Path(__file__).resolve().parent
FIXTURE_HASH = "363df99c1d548fd02e022fa61c6286c6ffbf7691f29dc0c701c937138c876b89"
EXPECTED_HASH = "e95ba243ec164accf114b75b8ffcc73f1eb896481f04906e2fce71a3b2af84e6"
raw, expected_raw = (P / "fixture.json").read_bytes(), (P / "expected.json").read_bytes()
if hashlib.sha256(raw).hexdigest() != FIXTURE_HASH or hashlib.sha256(expected_raw).hexdigest() != EXPECTED_HASH:
    raise SystemExit("fixture/expectation bytes changed")
spec = importlib.util.spec_from_file_location("bounded_refunds", P / "sum_refunds.py")
m = importlib.util.module_from_spec(spec)
spec.loader.exec_module(m)
base, expected = m.decode(raw), m.decode(expected_raw)
records = []

def good(name, data, wanted):
    observed = m.subtotal(data)
    if observed != wanted:
        raise SystemExit(name + ": full output mismatch")
    records.append({"case": name, "expected": wanted, "observed": observed, "pass": True})

good("baseline", base, expected)
wrong = sum(row["amount"] for row in base["refunds"])
if wrong != 5000 or wrong != base["selected_payment"]["amount"]:
    raise SystemExit("naive equal-charge total not demonstrated")
records.append({"case": "naive-all-row-total", "expected_wrong_total_minor_units": 5000,
                "observed_wrong_total_minor_units": wrong, "matching_succeeded_minor_units": 1250, "pass": True})
reverse = copy.deepcopy(base)
reverse["refunds"].reverse()
good("reversed-row-order", reverse, expected)
more = copy.deepcopy(base)
more["refunds"] += [
    {"fixture_account": m.ACCOUNT, "id": "fictional-canceled", "charge": m.PAYMENT["charge"], "amount": 300, "currency": "usd", "status": "canceled"},
    {"fixture_account": m.ACCOUNT, "id": "fictional-action", "charge": m.PAYMENT["charge"], "amount": 400, "currency": "usd", "status": "requires_action"}
]
wanted = copy.deepcopy(expected)
wanted["fixture_status_subtotals_minor_units"].update(canceled=300, requires_action=400)
good("all-five-statuses-separate", more, wanted)
missing = copy.deepcopy(base)
missing["refunds"].pop(0)
wanted = copy.deepcopy(expected)
wanted["fixture_status_subtotals_minor_units"]["succeeded"] = 0
good("missing-row-not-actual-no-refund", missing, wanted)

def change(path, value=None, delete=False):
    data = copy.deepcopy(base)
    target = data
    for part in path[:-1]:
        target = target[part]
    if delete:
        del target[path[-1]]
    else:
        target[path[-1]] = value
    return data

negative = [
    ("identical-duplicate", change(["refunds"], base["refunds"] + [base["refunds"][0]])),
    ("conflicting-history", change(["refunds"], base["refunds"] + [dict(base["refunds"][0], status="failed")])),
    ("mixed-account", change(["refunds", 0, "fixture_account"], "fictional-other-account")),
    ("foreign-currency", change(["refunds", 0, "currency"], "eur")),
    ("missing-charge", change(["refunds", 0, "charge"], delete=True)),
    ("missing-status", change(["refunds", 0, "status"], delete=True)),
    ("unset-status", change(["refunds", 0, "status"], None)),
    ("unknown-status", change(["refunds", 0, "status"], "posted")),
    ("boolean-amount", change(["refunds", 0, "amount"], True)),
    ("decimal-string-amount", change(["refunds", 0, "amount"], "12.50")),
    ("negative-amount", change(["refunds", 0, "amount"], -1)),
    ("zero-outside-subset", change(["refunds", 0, "amount"], 0)),
    ("changed-selected-charge", change(["selected_payment", "charge"], "fictional-charge-B")),
    ("changed-selected-amount", change(["selected_payment", "amount"], 5001)),
    ("new-bank-receipt-evidence", change(["actual_bank_receipt"], True)),
    ("nonlist-refunds", change(["refunds"], {}))
]
for name, data in negative:
    try:
        m.subtotal(data)
    except (ValueError, TypeError, KeyError) as error:
        records.append({"case": name, "expected": "stop", "observed": "stop", "reason": str(error), "pass": True})
    else:
        raise SystemExit(name + ": expected local stop")
try:
    m.decode('{"status":"pending","status":"succeeded"}')
except ValueError as error:
    records.append({"case": "duplicate-json-member", "expected": "stop", "observed": "stop", "reason": str(error), "pass": True})
else:
    raise SystemExit("duplicate JSON member accepted")
good_cli = subprocess.run([sys.executable, str(P / "sum_refunds.py"), str(P / "fixture.json"), FIXTURE_HASH], capture_output=True)
if good_cli.returncode != 0 or good_cli.stdout != expected_raw:
    raise SystemExit("reader command/full expected stdout bytes differ")
bad_cli = subprocess.run([sys.executable, str(P / "sum_refunds.py"), str(P / "fixture.json"), "0" * 64], capture_output=True)
if bad_cli.returncode == 0 or b"fixture bytes changed" not in bad_cli.stderr:
    raise SystemExit("changed byte pin accepted")
records.append({"case": "changed-byte-pin", "expected": "stop", "observed": "stop", "pass": True})
report = {
    "test_status": "offline-reproduced",
    "executed_at": datetime.datetime.now(datetime.timezone.utc).isoformat().replace("+00:00", "Z"),
    "runtime": {"python": sys.version, "platform": platform.platform(), "optimization": sys.flags.optimize, "external_packages": []},
    "fixture_sha256": FIXTURE_HASH,
    "expected_sha256": EXPECTED_HASH,
    "observed_stdout_sha256": hashlib.sha256(good_cli.stdout).hexdigest(),
    "observed_stdout_bytes": len(good_cli.stdout),
    "baseline_top_level_fields": len(expected),
    "all_expected_fields_equal": True,
    "all_expected_bytes_equal": True,
    "positive_cases": 4,
    "demonstrated_wrong_total": 1,
    "stop_cases": len(negative) + 2,
    "records": records,
    "effects": "Own local files/stdout only; no API/account/payment/refund action or bank receipt observation."
}
print(json.dumps(report, indent=2))

run-record.json

{
  "test_status": "offline-reproduced",
  "executed_at": "2026-10-05T19:52:06.831189Z",
  "runtime": {
    "python": "3.12.3 (main, Aug 31 2026, 10:18:26) [GCC 13.3.0]",
    "platform": "Linux-6.8.0-146-generic-x86_64-with-glibc2.39",
    "optimization": 0,
    "external_packages": []
  },
  "fixture_sha256": "363df99c1d548fd02e022fa61c6286c6ffbf7691f29dc0c701c937138c876b89",
  "expected_sha256": "e95ba243ec164accf114b75b8ffcc73f1eb896481f04906e2fce71a3b2af84e6",
  "observed_stdout_sha256": "e95ba243ec164accf114b75b8ffcc73f1eb896481f04906e2fce71a3b2af84e6",
  "observed_stdout_bytes": 642,
  "baseline_top_level_fields": 10,
  "all_expected_fields_equal": true,
  "all_expected_bytes_equal": true,
  "positive_cases": 4,
  "demonstrated_wrong_total": 1,
  "stop_cases": 18,
  "records": [
    {
      "case": "baseline",
      "expected": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 1250,
          "pending": 750,
          "failed": 1000,
          "canceled": 0,
          "requires_action": 0
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "observed": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 1250,
          "pending": 750,
          "failed": 1000,
          "canceled": 0,
          "requires_action": 0
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "pass": true
    },
    {
      "case": "naive-all-row-total",
      "expected_wrong_total_minor_units": 5000,
      "observed_wrong_total_minor_units": 5000,
      "matching_succeeded_minor_units": 1250,
      "pass": true
    },
    {
      "case": "reversed-row-order",
      "expected": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 1250,
          "pending": 750,
          "failed": 1000,
          "canceled": 0,
          "requires_action": 0
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "observed": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 1250,
          "pending": 750,
          "failed": 1000,
          "canceled": 0,
          "requires_action": 0
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "pass": true
    },
    {
      "case": "all-five-statuses-separate",
      "expected": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 1250,
          "pending": 750,
          "failed": 1000,
          "canceled": 300,
          "requires_action": 400
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "observed": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 1250,
          "pending": 750,
          "failed": 1000,
          "canceled": 300,
          "requires_action": 400
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "pass": true
    },
    {
      "case": "missing-row-not-actual-no-refund",
      "expected": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 0,
          "pending": 750,
          "failed": 1000,
          "canceled": 0,
          "requires_action": 0
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "observed": {
        "fixture_account": "fictional-us-direct-merchant",
        "selected_charge": "fictional-charge-A",
        "currency": "usd",
        "fixture_original_charge_minor_units": 5000,
        "fixture_status_subtotals_minor_units": {
          "succeeded": 0,
          "pending": 750,
          "failed": 1000,
          "canceled": 0,
          "requires_action": 0
        },
        "excluded_other_charge_ids": [
          "fictional-refund-other"
        ],
        "actual_bank_receipt": null,
        "inventory_complete": null,
        "additional_refund_needed": null,
        "next_step": "Verify exact payment, current refund states and customer receipt privately before reporting returned money or deciding a further action."
      },
      "pass": true
    },
    {
      "case": "identical-duplicate",
      "expected": "stop",
      "observed": "stop",
      "reason": "duplicate refund snapshot; select one current record first",
      "pass": true
    },
    {
      "case": "conflicting-history",
      "expected": "stop",
      "observed": "stop",
      "reason": "duplicate refund snapshot; select one current record first",
      "pass": true
    },
    {
      "case": "mixed-account",
      "expected": "stop",
      "observed": "stop",
      "reason": "mixed account context",
      "pass": true
    },
    {
      "case": "foreign-currency",
      "expected": "stop",
      "observed": "stop",
      "reason": "outside USD-only fixture",
      "pass": true
    },
    {
      "case": "missing-charge",
      "expected": "stop",
      "observed": "stop",
      "reason": "missing or unsupported fields",
      "pass": true
    },
    {
      "case": "missing-status",
      "expected": "stop",
      "observed": "stop",
      "reason": "missing or unsupported fields",
      "pass": true
    },
    {
      "case": "unset-status",
      "expected": "stop",
      "observed": "stop",
      "reason": "unknown or unset fixture status",
      "pass": true
    },
    {
      "case": "unknown-status",
      "expected": "stop",
      "observed": "stop",
      "reason": "unknown or unset fixture status",
      "pass": true
    },
    {
      "case": "boolean-amount",
      "expected": "stop",
      "observed": "stop",
      "reason": "positive integer minor-unit amount required",
      "pass": true
    },
    {
      "case": "decimal-string-amount",
      "expected": "stop",
      "observed": "stop",
      "reason": "positive integer minor-unit amount required",
      "pass": true
    },
    {
      "case": "negative-amount",
      "expected": "stop",
      "observed": "stop",
      "reason": "positive integer minor-unit amount required",
      "pass": true
    },
    {
      "case": "zero-outside-subset",
      "expected": "stop",
      "observed": "stop",
      "reason": "positive integer minor-unit amount required",
      "pass": true
    },
    {
      "case": "changed-selected-charge",
      "expected": "stop",
      "observed": "stop",
      "reason": "changed selected payment",
      "pass": true
    },
    {
      "case": "changed-selected-amount",
      "expected": "stop",
      "observed": "stop",
      "reason": "changed selected payment",
      "pass": true
    },
    {
      "case": "new-bank-receipt-evidence",
      "expected": "stop",
      "observed": "stop",
      "reason": "missing or unsupported fields",
      "pass": true
    },
    {
      "case": "nonlist-refunds",
      "expected": "stop",
      "observed": "stop",
      "reason": "refund snapshot list required",
      "pass": true
    },
    {
      "case": "duplicate-json-member",
      "expected": "stop",
      "observed": "stop",
      "reason": "duplicate JSON member",
      "pass": true
    },
    {
      "case": "changed-byte-pin",
      "expected": "stop",
      "observed": "stop",
      "pass": true
    }
  ],
  "effects": "Own local files/stdout only; no API/account/payment/refund action or bank receipt observation."
}

Exact public file manifest

File	Bytes	SHA-256
fixture.json	1099	363df99c1d548fd02e022fa61c6286c6ffbf7691f29dc0c701c937138c876b89
sum_refunds.py	3144	b43057a96afb3fc9b26aa293e8d59597edba1c74cd64fe2084d57563f68cf4dc
expected.json	642	e95ba243ec164accf114b75b8ffcc73f1eb896481f04906e2fce71a3b2af84e6
check.py	5974	bcd3291ca5eb1a57760168f8a3eddc3f9da804d911954a9218fab385ee115b63
run-record.json	10678	89fc6060011d9147177bb8296268c916921099720306772d9f418aeea7056dc5

The run report is historical author evidence. A new replay has its own execution clock; compare all nonclock fields. Changed inputs, sources or runtime require a new pinned run. Missing rows do not prove no other refund exists; arithmetic does not establish snapshot authenticity, currentness or completeness. No Connect, tax/fee/conversion, entitlement, processing timetable or bank reconciliation is modeled.