Update wehago matching logic and exclude reports
This commit is contained in:
b17301
@@ -0,0 +1,700 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Materialize the 2025 WEHAGO compare projection from canonical voucher keys.
|
||||
|
||||
This script is intentionally conservative. It uses the current projection as a
|
||||
candidate seed, but the final WEHAGO population is the raw ledger voucher set.
|
||||
That prevents stale export rows, display-date variants, and partial context rows
|
||||
from changing card/table/export counts.
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import json
|
||||
import re
|
||||
import sqlite3
|
||||
from collections import Counter, defaultdict
|
||||
from pathlib import Path
|
||||
from typing import Any
|
||||
|
||||
START_YEAR = 2025
|
||||
END_YEAR = 2025
|
||||
DB_PATH = Path("/runtime/db/data.db")
|
||||
|
||||
WEHAGO_STATUSES = (
|
||||
"voucher_matched",
|
||||
"voucher_unmatched",
|
||||
"voucher_recheck",
|
||||
"voucher_excepted",
|
||||
)
|
||||
STATUS_LABELS = {
|
||||
"voucher_matched": "Voucher",
|
||||
"voucher_unmatched": "Unmatched",
|
||||
"voucher_recheck": "Recheck",
|
||||
"voucher_excepted": "Excepted",
|
||||
}
|
||||
|
||||
PROMOTE_TO_MATCHED = {
|
||||
(2025, "04-29", "00163"): "최종 validator: proof_date와 전표일자 일치, 금액·계정·거래처 문맥 일치",
|
||||
(2025, "09-24", "50036"): "최종 validator: proof_date와 전표일자 일치, 음수 조정 방향·계정·금액 일치",
|
||||
(2025, "10-14", "00001"): "최종 validator: 자금성 전표 기타예금/단기금융상품(기타예금) 동치 및 보통예금 금액 일치",
|
||||
(2025, "10-27", "00002"): "사용자 지정 오타 허용 특수 승격: 금액·계정·거래처 일치",
|
||||
(2025, "10-28", "00001"): "최종 validator: 원천 ERP 외부 후보 탐색 대상 포함, 자금성 전표 동치 규칙 적용",
|
||||
(2025, "10-30", "00005"): "사용자 지정 오타 허용 특수 승격: 금액·계정 일치",
|
||||
(2025, "11-24", "00106"): "사용자 지정 오타 허용 특수 승격: 통신비 금액·거래처 문맥 일치",
|
||||
(2025, "11-25", "00009"): "최종 validator: 비용 차감 전표 금액·계정·거래 문맥 일치",
|
||||
}
|
||||
|
||||
KEEP_RECHECK = {
|
||||
(2025, "01-01", "00063"): "전표 일부 직접 대응은 있으나 ERP 국고보조금/보통예금 잔여행 검토 필요",
|
||||
(2025, "02-21", "00176"): "사용자 판정: 카드 비용 전표와 ERP 미지급금 출금 전표가 전표 전체로는 동일 성격이 아님",
|
||||
(2025, "03-31", "00048"): "사용자 판정: 상계 특수 사례는 일부 행만 대응 표시하고 전표 전체는 recheck 유지",
|
||||
(2025, "04-16", "50009"): "지급수수료/매입세액 직접 대응과 ERP 조정행이 함께 있어 recheck 유지",
|
||||
(2025, "04-23", "00007"): "사용자 판정: 카드 비용 전표와 ERP 미지급금 출금 전표가 전표 전체로는 동일 성격이 아님",
|
||||
(2025, "06-30", "00156"): "사용자 판정: 상계 특수 사례는 일부 행만 대응 표시하고 전표 전체는 recheck 유지",
|
||||
(2025, "09-22", "50046"): "VAT/proof_date 우선 검증: 현재 ERP 후보 proof_date가 WEHAGO 전표일자와 불일치",
|
||||
(2025, "12-19", "50036"): "VAT/proof_date 우선 검증: 현재 ERP 후보 proof_date가 WEHAGO 전표일자와 불일치",
|
||||
(2025, "12-25", "50006"): "VAT/proof_date 우선 검증: 현재 ERP 후보 proof_date가 WEHAGO 전표일자와 불일치",
|
||||
}
|
||||
|
||||
FORCE_UNMATCHED = {
|
||||
(2025, "08-07", "50005"): "취소/발행 반대방향 전표: 같은 차대 위치의 음수/양수 절대값 일치는 매칭 후보에서 제거",
|
||||
(2025, "06-13", "00038"): "금액 유사 오염 제거: 계정·거래처·적요가 다른 후보",
|
||||
(2025, "07-07", "00001"): "월/기간 불일치 후보 제거: ERP는 3~5월 합산, WEHAGO는 7월 전표",
|
||||
}
|
||||
|
||||
|
||||
EXCEPTED_RETAIN_RECHECK_KEYS = {
|
||||
(2025, "03-31", "00048"),
|
||||
(2025, "06-30", "00156"),
|
||||
}
|
||||
|
||||
|
||||
def clean(value: Any) -> str:
|
||||
if value is None:
|
||||
return ""
|
||||
text = str(value).strip()
|
||||
if text.endswith(".0") and text[:-2].isdigit():
|
||||
return text[:-2]
|
||||
return text
|
||||
|
||||
|
||||
def amount(value: Any) -> float:
|
||||
if value is None or value == "":
|
||||
return 0.0
|
||||
if isinstance(value, (int, float)):
|
||||
return float(value)
|
||||
text = str(value).replace(",", "").strip()
|
||||
if not text:
|
||||
return 0.0
|
||||
try:
|
||||
return float(text)
|
||||
except ValueError:
|
||||
return 0.0
|
||||
|
||||
|
||||
def display_date(value: Any) -> str:
|
||||
text = clean(value)
|
||||
match = re.match(r"^(\d{4})-(\d{2})-(\d{2})$", text)
|
||||
if match:
|
||||
return f"{match.group(2)}-{match.group(3)}"
|
||||
match = re.match(r"^(\d{2})-(\d{2})$", text)
|
||||
if match:
|
||||
return text
|
||||
return text
|
||||
|
||||
|
||||
def date_key(year: int, mmdd: str) -> str:
|
||||
return f"{year}-{mmdd}" if re.match(r"^\d{2}-\d{2}$", mmdd) else mmdd
|
||||
|
||||
|
||||
def voucher_no(value: Any) -> str:
|
||||
text = clean(value)
|
||||
return text.zfill(5) if text.isdigit() else text
|
||||
|
||||
|
||||
def draft_base(value: Any) -> str:
|
||||
return re.sub(r"-\d+$", "", clean(value))
|
||||
|
||||
|
||||
def row_has_wehago(row: dict[str, Any]) -> bool:
|
||||
return bool(
|
||||
clean(row.get("ledger_account_name"))
|
||||
or clean(row.get("ledger_vendor"))
|
||||
or abs(amount(row.get("ledger_debit"))) >= 0.5
|
||||
or abs(amount(row.get("ledger_credit"))) >= 0.5
|
||||
)
|
||||
|
||||
|
||||
def row_has_erp(row: dict[str, Any]) -> bool:
|
||||
return bool(
|
||||
clean(row.get("draft_no"))
|
||||
or clean(row.get("voucher_account_name"))
|
||||
or clean(row.get("voucher_vendor"))
|
||||
or abs(amount(row.get("voucher_debit"))) >= 0.5
|
||||
or abs(amount(row.get("voucher_credit"))) >= 0.5
|
||||
)
|
||||
|
||||
|
||||
def normalize_row(row: dict[str, Any]) -> dict[str, Any]:
|
||||
normalized = dict(row)
|
||||
normalized["fiscal_year"] = int(normalized.get("fiscal_year") or START_YEAR)
|
||||
normalized["ledger_date"] = display_date(normalized.get("ledger_date"))
|
||||
normalized["voucher_no"] = voucher_no(normalized.get("voucher_no"))
|
||||
for key in ("ledger_debit", "ledger_credit", "voucher_debit", "voucher_credit"):
|
||||
normalized[key] = amount(normalized.get(key))
|
||||
return normalized
|
||||
|
||||
|
||||
def load_current_projection(conn: sqlite3.Connection) -> dict[tuple[int, str, str], dict[str, Any]]:
|
||||
by_key: dict[tuple[int, str, str], dict[str, Any]] = {}
|
||||
rows = conn.execute(
|
||||
"""
|
||||
SELECT *
|
||||
FROM wehago_status_projection_groups
|
||||
WHERE start_year = ? AND end_year = ?
|
||||
AND status_key IN ('voucher_matched', 'voucher_unmatched', 'voucher_recheck', 'voucher_excepted')
|
||||
ORDER BY status_key, group_index
|
||||
""",
|
||||
(START_YEAR, END_YEAR),
|
||||
).fetchall()
|
||||
status_rank = {"voucher_matched": 4, "voucher_recheck": 3, "voucher_unmatched": 2, "voucher_excepted": 1}
|
||||
for db_row in rows:
|
||||
year = int(db_row["fiscal_year"] or START_YEAR)
|
||||
key = (year, display_date(db_row["ledger_date"]), voucher_no(db_row["voucher_no"]))
|
||||
payload = dict(db_row)
|
||||
try:
|
||||
detail_rows = json.loads(db_row["rows_json"] or "[]")
|
||||
except Exception:
|
||||
detail_rows = []
|
||||
payload["rows"] = [normalize_row(row) for row in detail_rows if isinstance(row, dict)]
|
||||
existing = by_key.get(key)
|
||||
if not existing:
|
||||
by_key[key] = payload
|
||||
continue
|
||||
existing_iso = re.match(r"^\d{4}-", clean(existing.get("ledger_date") or ""))
|
||||
payload_iso = re.match(r"^\d{4}-", clean(payload.get("ledger_date") or ""))
|
||||
if existing_iso and not payload_iso:
|
||||
by_key[key] = payload
|
||||
elif bool(existing_iso) == bool(payload_iso) and status_rank.get(payload["status_key"], 0) > status_rank.get(existing["status_key"], 0):
|
||||
by_key[key] = payload
|
||||
return by_key
|
||||
|
||||
|
||||
def raw_wehago_row(raw: sqlite3.Row) -> dict[str, Any]:
|
||||
year = int(raw["fiscal_year"] or START_YEAR)
|
||||
mmdd = display_date(raw["ledger_date"])
|
||||
return {
|
||||
"fiscal_year": year,
|
||||
"status_label": "",
|
||||
"ledger_date": mmdd,
|
||||
"ledger_date_key": date_key(year, mmdd),
|
||||
"proof_date": "",
|
||||
"voucher_no": voucher_no(raw["voucher_no"]),
|
||||
"draft_no": "",
|
||||
"ledger_account_name": clean(raw["account_name"]),
|
||||
"voucher_account_name": "",
|
||||
"ledger_vendor": clean(raw["vendor_name"]),
|
||||
"voucher_vendor": "",
|
||||
"ledger_debit": amount(raw["debit"]),
|
||||
"ledger_credit": amount(raw["credit"]),
|
||||
"voucher_debit": 0.0,
|
||||
"voucher_credit": 0.0,
|
||||
"ledger_desc": clean(raw["description"]),
|
||||
"voucher_desc": "",
|
||||
"review_reason": "WEHAGO_FULL_VOUCHER_CONTEXT",
|
||||
"matched_case": "WEHAGO_CONTEXT_ONLY",
|
||||
"ledger_row_key": f"{raw['id']}",
|
||||
"voucher_row_key": "",
|
||||
"match_identity_key": "",
|
||||
}
|
||||
|
||||
|
||||
def load_raw_wehago(conn: sqlite3.Connection) -> dict[tuple[int, str, str], list[dict[str, Any]]]:
|
||||
by_key: dict[tuple[int, str, str], list[dict[str, Any]]] = defaultdict(list)
|
||||
rows = conn.execute(
|
||||
"""
|
||||
SELECT id, fiscal_year, ledger_date, voucher_no, account_name, vendor_name,
|
||||
debit, credit, description, row_number
|
||||
FROM wehago_ledger_rows
|
||||
WHERE fiscal_year = ?
|
||||
AND COALESCE(compare_voucher_no, '') <> ''
|
||||
ORDER BY ledger_date, voucher_no, row_number, id
|
||||
""",
|
||||
(START_YEAR,),
|
||||
).fetchall()
|
||||
for raw in rows:
|
||||
row = raw_wehago_row(raw)
|
||||
key = (int(row["fiscal_year"]), row["ledger_date"], row["voucher_no"])
|
||||
by_key[key].append(row)
|
||||
return by_key
|
||||
|
||||
|
||||
def raw_erp_row(raw: sqlite3.Row) -> dict[str, Any]:
|
||||
desc = " ".join(part for part in [clean(raw["desc1"]), clean(raw["desc2"])] if part)
|
||||
return {
|
||||
"fiscal_year": int(raw["fiscal_year"] or START_YEAR),
|
||||
"status_label": "",
|
||||
"ledger_date": "",
|
||||
"proof_date": clean(raw["proof_date"]),
|
||||
"voucher_no": clean(raw["confirmed_no"]),
|
||||
"draft_no": clean(raw["draft_no"]),
|
||||
"ledger_account_name": "",
|
||||
"voucher_account_name": clean(raw["account_name"]),
|
||||
"ledger_vendor": "",
|
||||
"voucher_vendor": clean(raw["vendor_name"]),
|
||||
"ledger_debit": 0.0,
|
||||
"ledger_credit": 0.0,
|
||||
"voucher_debit": amount(raw["debit_supply"]),
|
||||
"voucher_credit": amount(raw["credit_supply"]),
|
||||
"ledger_desc": "",
|
||||
"voucher_desc": desc,
|
||||
"review_reason": "ERP_FULL_DRAFT_CONTEXT",
|
||||
"matched_case": "ERP_CONTEXT_ONLY",
|
||||
"ledger_row_key": "",
|
||||
"voucher_row_key": f"{raw['id']}",
|
||||
"match_identity_key": "",
|
||||
}
|
||||
|
||||
|
||||
def load_raw_erp(conn: sqlite3.Connection) -> dict[str, list[dict[str, Any]]]:
|
||||
by_base: dict[str, list[dict[str, Any]]] = defaultdict(list)
|
||||
rows = conn.execute(
|
||||
"""
|
||||
SELECT id, fiscal_year, proof_date, confirmed_no, draft_no, account_name,
|
||||
debit_supply, credit_supply, vendor_name, desc1, desc2, row_number
|
||||
FROM wehago_voucher_rows
|
||||
WHERE fiscal_year = ? AND COALESCE(draft_no, '') <> ''
|
||||
ORDER BY draft_no, row_number, id
|
||||
""",
|
||||
(START_YEAR,),
|
||||
).fetchall()
|
||||
for raw in rows:
|
||||
by_base[draft_base(raw["draft_no"])].append(raw_erp_row(raw))
|
||||
return by_base
|
||||
|
||||
|
||||
def identity(row: dict[str, Any], side: str) -> str:
|
||||
if side == "wehago":
|
||||
return "|".join(
|
||||
[
|
||||
display_date(row.get("ledger_date")),
|
||||
voucher_no(row.get("voucher_no")),
|
||||
clean(row.get("ledger_account_name")),
|
||||
clean(row.get("ledger_vendor")),
|
||||
f"{amount(row.get('ledger_debit')):.2f}",
|
||||
f"{amount(row.get('ledger_credit')):.2f}",
|
||||
clean(row.get("ledger_desc")),
|
||||
]
|
||||
)
|
||||
return "|".join(
|
||||
[
|
||||
clean(row.get("draft_no")),
|
||||
clean(row.get("voucher_account_name")),
|
||||
clean(row.get("voucher_vendor")),
|
||||
f"{amount(row.get('voucher_debit')):.2f}",
|
||||
f"{amount(row.get('voucher_credit')):.2f}",
|
||||
clean(row.get("voucher_desc")),
|
||||
]
|
||||
)
|
||||
|
||||
|
||||
def blank_row() -> dict[str, Any]:
|
||||
return {
|
||||
"fiscal_year": START_YEAR,
|
||||
"status_label": "",
|
||||
"ledger_date": "",
|
||||
"ledger_date_key": "",
|
||||
"proof_date": "",
|
||||
"voucher_no": "",
|
||||
"draft_no": "",
|
||||
"ledger_account_name": "",
|
||||
"voucher_account_name": "",
|
||||
"ledger_vendor": "",
|
||||
"voucher_vendor": "",
|
||||
"ledger_debit": 0.0,
|
||||
"ledger_credit": 0.0,
|
||||
"voucher_debit": 0.0,
|
||||
"voucher_credit": 0.0,
|
||||
"ledger_desc": "",
|
||||
"voucher_desc": "",
|
||||
"review_reason": "",
|
||||
"matched_case": "",
|
||||
"ledger_row_key": "",
|
||||
"voucher_row_key": "",
|
||||
"match_identity_key": "",
|
||||
}
|
||||
|
||||
|
||||
def combine_context_rows(wehago_row: dict[str, Any] | None, erp_row: dict[str, Any] | None) -> dict[str, Any]:
|
||||
left = wehago_row or blank_row()
|
||||
right = erp_row or blank_row()
|
||||
row = blank_row()
|
||||
row.update(
|
||||
{
|
||||
"fiscal_year": int(left.get("fiscal_year") or right.get("fiscal_year") or START_YEAR),
|
||||
"status_label": "",
|
||||
"ledger_date": display_date(left.get("ledger_date")),
|
||||
"ledger_date_key": clean(left.get("ledger_date_key")),
|
||||
"proof_date": clean(right.get("proof_date")),
|
||||
"voucher_no": voucher_no(left.get("voucher_no")),
|
||||
"draft_no": clean(right.get("draft_no")),
|
||||
"ledger_account_name": clean(left.get("ledger_account_name")),
|
||||
"voucher_account_name": clean(right.get("voucher_account_name")),
|
||||
"ledger_vendor": clean(left.get("ledger_vendor")),
|
||||
"voucher_vendor": clean(right.get("voucher_vendor")),
|
||||
"ledger_debit": amount(left.get("ledger_debit")),
|
||||
"ledger_credit": amount(left.get("ledger_credit")),
|
||||
"voucher_debit": amount(right.get("voucher_debit")),
|
||||
"voucher_credit": amount(right.get("voucher_credit")),
|
||||
"ledger_desc": clean(left.get("ledger_desc")),
|
||||
"voucher_desc": clean(right.get("voucher_desc")),
|
||||
"review_reason": "FULL_VOUCHER_CONTEXT",
|
||||
"matched_case": "FULL_CONTEXT_ALIGNED",
|
||||
"ledger_row_key": clean(left.get("ledger_row_key")),
|
||||
"voucher_row_key": clean(right.get("voucher_row_key")),
|
||||
"match_identity_key": "",
|
||||
}
|
||||
)
|
||||
if not wehago_row:
|
||||
row["matched_case"] = "ERP_CONTEXT_ONLY"
|
||||
elif not erp_row:
|
||||
row["matched_case"] = "WEHAGO_CONTEXT_ONLY"
|
||||
return row
|
||||
|
||||
|
||||
def hydrate_rows(
|
||||
status_key: str,
|
||||
seed_rows: list[dict[str, Any]],
|
||||
raw_wehago_rows: list[dict[str, Any]],
|
||||
raw_erp_by_base: dict[str, list[dict[str, Any]]],
|
||||
) -> list[dict[str, Any]]:
|
||||
if status_key in {"voucher_unmatched", "voucher_excepted"}:
|
||||
rows = [dict(row) for row in raw_wehago_rows]
|
||||
else:
|
||||
bases = sorted({draft_base(row.get("draft_no")) for row in seed_rows if clean(row.get("draft_no"))})
|
||||
erp_context: list[dict[str, Any]] = []
|
||||
seen_erp: set[str] = set()
|
||||
for base in bases:
|
||||
for erp_row in raw_erp_by_base.get(base, []):
|
||||
row_id = identity(erp_row, "erp")
|
||||
if row_id in seen_erp:
|
||||
continue
|
||||
erp_context.append(dict(erp_row))
|
||||
seen_erp.add(row_id)
|
||||
row_count = max(len(raw_wehago_rows), len(erp_context))
|
||||
rows = [
|
||||
combine_context_rows(
|
||||
raw_wehago_rows[index] if index < len(raw_wehago_rows) else None,
|
||||
erp_context[index] if index < len(erp_context) else None,
|
||||
)
|
||||
for index in range(row_count)
|
||||
]
|
||||
label = STATUS_LABELS.get(status_key, status_key)
|
||||
for row in rows:
|
||||
row["status_label"] = label
|
||||
row["ledger_date"] = display_date(row.get("ledger_date"))
|
||||
return rows
|
||||
|
||||
|
||||
def determine_status(key: tuple[int, str, str], seed: dict[str, Any] | None) -> tuple[str, str]:
|
||||
if key in FORCE_UNMATCHED:
|
||||
return "voucher_unmatched", FORCE_UNMATCHED[key]
|
||||
if key in PROMOTE_TO_MATCHED:
|
||||
return "voucher_matched", PROMOTE_TO_MATCHED[key]
|
||||
if key in KEEP_RECHECK:
|
||||
return "voucher_recheck", KEEP_RECHECK[key]
|
||||
if seed:
|
||||
status = clean(seed.get("status_key"))
|
||||
if status in WEHAGO_STATUSES:
|
||||
return status, clean(seed.get("review_reason")) or "원천 WEHAGO 기준 integrity projection"
|
||||
return "voucher_unmatched", "원천 WEHAGO 전표 복원: 검증된 ERP 후보 없음"
|
||||
|
||||
|
||||
def revised_excepted_reason(key: tuple[int, str, str], status_key: str, rows: list[dict[str, Any]], summary: dict[str, Any]) -> str:
|
||||
if status_key not in {"voucher_unmatched", "voucher_recheck"}:
|
||||
return ""
|
||||
if status_key == "voucher_recheck" and key in EXCEPTED_RETAIN_RECHECK_KEYS:
|
||||
return ""
|
||||
text = " ".join(
|
||||
[
|
||||
clean(summary.get("ledger_accounts")),
|
||||
clean(summary.get("ledger_vendors")),
|
||||
clean(summary.get("review_reason")),
|
||||
" ".join(clean(row.get("ledger_desc")) for row in rows),
|
||||
" ".join(clean(row.get("ledger_account_name")) for row in rows),
|
||||
]
|
||||
)
|
||||
compact = text.replace(" ", "")
|
||||
if re.search(r"상계\s*\d+(?:-\d+)?\s*소구", text) or re.search(r"상계\d+(?:-\d+)?소구", compact):
|
||||
return ""
|
||||
tax_work_tokens = (
|
||||
"부가세예수금과상계",
|
||||
"부가세대급금과상계",
|
||||
"부가예수금과상계",
|
||||
"부가대급금과상계",
|
||||
)
|
||||
if any(token in compact for token in tax_work_tokens):
|
||||
return ""
|
||||
rules = (
|
||||
("WEHAGO_EXCEPTED_COST_SUBSTITUTION", (r"원가.*대체",)),
|
||||
("WEHAGO_EXCEPTED_GOV_GRANT_OFFSET", (r"국고보조금과상계",)),
|
||||
("WEHAGO_EXCEPTED_EMPLOYEE_PORTION_SUBSTITUTION", (r"본인분대체",)),
|
||||
("WEHAGO_EXCEPTED_OTHER_ACCOUNT_SUBSTITUTION", (r"타계정.*대체",)),
|
||||
(
|
||||
"WEHAGO_EXCEPTED_ACCOUNT_SUBSTITUTION",
|
||||
(
|
||||
r"계정대체",
|
||||
r"계정.*대체",
|
||||
r"거래처대체",
|
||||
r"거래처.*대체",
|
||||
r"대체분개",
|
||||
r"전기대체",
|
||||
r"전년대체",
|
||||
r"전년도대체",
|
||||
r"결산대체",
|
||||
r"감사대체",
|
||||
r"환입대체",
|
||||
r"계상분대체",
|
||||
),
|
||||
),
|
||||
(
|
||||
"WEHAGO_EXCEPTED_RETAINED_EARNINGS",
|
||||
(r"잉여금.*대체", r"미처분이익잉여금", r"이월이익잉여금", r"차기이월"),
|
||||
),
|
||||
("WEHAGO_EXCEPTED_CARRY_FORWARD", (r"전기이월", r"기초이월", r"이월")),
|
||||
("WEHAGO_EXCEPTED_VALUATION_GAIN_LOSS", (r"평가이익", r"평가손실", r"외화평가")),
|
||||
("WEHAGO_EXCEPTED_WIP_CLOSING", (r"미완성공사도급",)),
|
||||
("WEHAGO_EXCEPTED_PROFIT_LOSS_TRANSFER", (r"손익.*대체", r"당기순손익.*대체", r"수익에서대체")),
|
||||
)
|
||||
for reason, patterns in rules:
|
||||
if any(re.search(pattern, compact) for pattern in patterns):
|
||||
return reason
|
||||
return ""
|
||||
|
||||
|
||||
def summarize(key: tuple[int, str, str], status_key: str, rows: list[dict[str, Any]], reason: str) -> dict[str, Any]:
|
||||
ledger_accounts = sorted({clean(row.get("ledger_account_name")) for row in rows if clean(row.get("ledger_account_name"))})
|
||||
voucher_accounts = sorted({clean(row.get("voucher_account_name")) for row in rows if clean(row.get("voucher_account_name"))})
|
||||
ledger_vendors = sorted({clean(row.get("ledger_vendor")) for row in rows if clean(row.get("ledger_vendor"))})
|
||||
voucher_vendors = sorted({clean(row.get("voucher_vendor")) for row in rows if clean(row.get("voucher_vendor"))})
|
||||
drafts = sorted({clean(row.get("draft_no")) for row in rows if clean(row.get("draft_no"))})
|
||||
ledger_basis = [row for row in rows if row_has_wehago(row)]
|
||||
erp_basis = [row for row in rows if row_has_erp(row)]
|
||||
return {
|
||||
"fiscal_year": key[0],
|
||||
"status_label": STATUS_LABELS.get(status_key, status_key),
|
||||
"ledger_date": key[1],
|
||||
"ledger_date_key": date_key(key[0], key[1]),
|
||||
"proof_date": "",
|
||||
"voucher_no": key[2],
|
||||
"draft_no": ", ".join(drafts),
|
||||
"ledger_row_count": len(ledger_basis),
|
||||
"voucher_row_count": len(erp_basis),
|
||||
"ledger_debit": sum(amount(row.get("ledger_debit")) for row in ledger_basis),
|
||||
"ledger_credit": sum(amount(row.get("ledger_credit")) for row in ledger_basis),
|
||||
"voucher_debit": sum(amount(row.get("voucher_debit")) for row in erp_basis),
|
||||
"voucher_credit": sum(amount(row.get("voucher_credit")) for row in erp_basis),
|
||||
"ledger_accounts": ", ".join(ledger_accounts),
|
||||
"voucher_accounts": ", ".join(voucher_accounts),
|
||||
"ledger_vendors": ", ".join(ledger_vendors),
|
||||
"voucher_vendors": ", ".join(voucher_vendors),
|
||||
"review_reason": reason,
|
||||
}
|
||||
|
||||
|
||||
def get_signatures(conn: sqlite3.Connection) -> dict[str, str]:
|
||||
result: dict[str, str] = {}
|
||||
for status in WEHAGO_STATUSES:
|
||||
row = conn.execute(
|
||||
"""
|
||||
SELECT signature
|
||||
FROM wehago_status_projection_groups
|
||||
WHERE start_year = ? AND end_year = ? AND status_key = ?
|
||||
LIMIT 1
|
||||
""",
|
||||
(START_YEAR, END_YEAR, status),
|
||||
).fetchone()
|
||||
result[status] = clean(row["signature"] if row else "") or f"integrity:{status}:{START_YEAR}:{END_YEAR}"
|
||||
return result
|
||||
|
||||
|
||||
def json_payload(value: Any) -> str:
|
||||
return json.dumps(value, ensure_ascii=False, separators=(",", ":"))
|
||||
|
||||
|
||||
def store_setting(conn: sqlite3.Connection, status_key: str, signature: str, count: int) -> None:
|
||||
setting_key = f"wehago_active_status_projection:{status_key}:{START_YEAR}:{END_YEAR}"
|
||||
payload = {
|
||||
"signature": signature,
|
||||
"status_key": status_key,
|
||||
"start_year": START_YEAR,
|
||||
"end_year": END_YEAR,
|
||||
"total_count": count,
|
||||
"projection_version": "integrity-2025-20260629",
|
||||
}
|
||||
conn.execute(
|
||||
"""
|
||||
INSERT INTO wehago_compare_settings (setting_key, setting_json, updated_at)
|
||||
VALUES (?, ?, CURRENT_TIMESTAMP)
|
||||
ON CONFLICT(setting_key) DO UPDATE SET
|
||||
setting_json = excluded.setting_json,
|
||||
updated_at = CURRENT_TIMESTAMP
|
||||
""",
|
||||
(setting_key, json_payload(payload)),
|
||||
)
|
||||
|
||||
|
||||
def cleanup_runtime(conn: sqlite3.Connection) -> None:
|
||||
conn.execute("DELETE FROM wehago_status_projection_groups WHERE start_year = ? AND end_year = ?", (START_YEAR, END_YEAR))
|
||||
for table in (
|
||||
"wehago_compare_query_groups",
|
||||
"wehago_compare_query_rows",
|
||||
"wehago_compare_query_metrics",
|
||||
"wehago_compare_query_page_cache",
|
||||
"wehago_compare_final_status_projection",
|
||||
"wehago_compare_export_jobs",
|
||||
):
|
||||
conn.execute(f"DELETE FROM {table} WHERE start_year = ? AND end_year = ?", (START_YEAR, END_YEAR))
|
||||
conn.execute("DELETE FROM wehago_compare_settings WHERE setting_key = ?", (f"wehago_active_query_projection:{START_YEAR}:{END_YEAR}",))
|
||||
conn.execute("DELETE FROM wehago_compare_settings WHERE setting_key = ?", (f"wehago_active_status_projection_run:{START_YEAR}:{END_YEAR}",))
|
||||
conn.execute("DELETE FROM wehago_compare_settings WHERE setting_key LIKE ?", (f"wehago_active_status_projection:%:{START_YEAR}:{END_YEAR}",))
|
||||
conn.execute("DELETE FROM wehago_compare_settings WHERE setting_key = ?", (f"wehago_final_basis_projection:{START_YEAR}:{END_YEAR}",))
|
||||
|
||||
|
||||
def materialize(conn: sqlite3.Connection) -> dict[str, Any]:
|
||||
seed_by_key = load_current_projection(conn)
|
||||
raw_wehago_by_key = load_raw_wehago(conn)
|
||||
raw_erp_by_base = load_raw_erp(conn)
|
||||
signatures = get_signatures(conn)
|
||||
raw_keys = sorted(raw_wehago_by_key)
|
||||
rows_to_insert: dict[str, list[tuple[Any, ...]]] = {status: [] for status in WEHAGO_STATUSES}
|
||||
diagnostics = {
|
||||
"raw_wehago_vouchers": len(raw_keys),
|
||||
"seed_canonical_vouchers": len(seed_by_key),
|
||||
"seed_duplicates_removed": 0,
|
||||
"restored_missing_vouchers": len(set(raw_keys) - set(seed_by_key)),
|
||||
"status_moves": [],
|
||||
}
|
||||
|
||||
for key in raw_keys:
|
||||
seed = seed_by_key.get(key)
|
||||
old_status = clean(seed.get("status_key")) if seed else ""
|
||||
status, reason = determine_status(key, seed)
|
||||
seed_rows = seed.get("rows", []) if seed else []
|
||||
display_rows = hydrate_rows(status, seed_rows, raw_wehago_by_key[key], raw_erp_by_base)
|
||||
summary = summarize(key, status, display_rows, reason)
|
||||
excepted_reason = revised_excepted_reason(key, status, display_rows, summary)
|
||||
if excepted_reason:
|
||||
status = "voucher_excepted"
|
||||
reason = excepted_reason
|
||||
display_rows = hydrate_rows(status, seed_rows, raw_wehago_by_key[key], raw_erp_by_base)
|
||||
summary = summarize(key, status, display_rows, reason)
|
||||
if old_status and old_status != status:
|
||||
diagnostics["status_moves"].append({"key": key, "from": old_status, "to": status, "reason": reason})
|
||||
search_text = " ".join(
|
||||
[
|
||||
summary["voucher_no"],
|
||||
summary["draft_no"],
|
||||
summary["ledger_accounts"],
|
||||
summary["voucher_accounts"],
|
||||
summary["ledger_vendors"],
|
||||
summary["voucher_vendors"],
|
||||
reason,
|
||||
" ".join(clean(row.get("ledger_desc")) for row in display_rows),
|
||||
" ".join(clean(row.get("voucher_desc")) for row in display_rows),
|
||||
]
|
||||
)
|
||||
rows_to_insert[status].append(
|
||||
(
|
||||
START_YEAR,
|
||||
END_YEAR,
|
||||
status,
|
||||
signatures[status],
|
||||
0,
|
||||
summary["fiscal_year"],
|
||||
summary["ledger_date"],
|
||||
summary["proof_date"],
|
||||
summary["voucher_no"],
|
||||
summary["draft_no"],
|
||||
int(summary["ledger_row_count"]),
|
||||
int(summary["voucher_row_count"]),
|
||||
amount(summary["ledger_debit"]),
|
||||
amount(summary["ledger_credit"]),
|
||||
amount(summary["voucher_debit"]),
|
||||
amount(summary["voucher_credit"]),
|
||||
summary["ledger_accounts"],
|
||||
summary["voucher_accounts"],
|
||||
summary["ledger_vendors"],
|
||||
summary["voucher_vendors"],
|
||||
reason,
|
||||
search_text,
|
||||
json_payload(summary),
|
||||
json_payload(display_rows),
|
||||
)
|
||||
)
|
||||
|
||||
cleanup_runtime(conn)
|
||||
counts: dict[str, int] = {}
|
||||
for status in WEHAGO_STATUSES:
|
||||
payloads = []
|
||||
for idx, payload in enumerate(rows_to_insert[status]):
|
||||
values = list(payload)
|
||||
values[4] = idx
|
||||
payloads.append(tuple(values))
|
||||
conn.executemany(
|
||||
"""
|
||||
INSERT INTO wehago_status_projection_groups (
|
||||
start_year, end_year, status_key, signature, group_index,
|
||||
fiscal_year, ledger_date, proof_date, voucher_no, draft_no,
|
||||
ledger_row_count, voucher_row_count,
|
||||
ledger_debit, ledger_credit, voucher_debit, voucher_credit,
|
||||
ledger_accounts, voucher_accounts, ledger_vendors, voucher_vendors,
|
||||
review_reason, search_text, summary_json, rows_json,
|
||||
created_at, updated_at
|
||||
) VALUES (
|
||||
?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?,
|
||||
CURRENT_TIMESTAMP, CURRENT_TIMESTAMP
|
||||
)
|
||||
""",
|
||||
payloads,
|
||||
)
|
||||
counts[status] = len(payloads)
|
||||
store_setting(conn, status, signatures[status], len(payloads))
|
||||
|
||||
basis = {
|
||||
"source": "raw WEHAGO canonical voucher set; current projection only used as candidate seed",
|
||||
"counts": counts,
|
||||
"wehago_total": sum(counts.values()),
|
||||
"diagnostics": diagnostics,
|
||||
}
|
||||
conn.execute(
|
||||
"""
|
||||
INSERT INTO wehago_compare_settings (setting_key, setting_json, updated_at)
|
||||
VALUES (?, ?, CURRENT_TIMESTAMP)
|
||||
ON CONFLICT(setting_key) DO UPDATE SET
|
||||
setting_json = excluded.setting_json,
|
||||
updated_at = CURRENT_TIMESTAMP
|
||||
""",
|
||||
(f"wehago_final_basis_projection:{START_YEAR}:{END_YEAR}", json_payload(basis)),
|
||||
)
|
||||
return basis
|
||||
|
||||
|
||||
def main() -> None:
|
||||
conn = sqlite3.connect(DB_PATH)
|
||||
conn.row_factory = sqlite3.Row
|
||||
try:
|
||||
conn.execute("PRAGMA busy_timeout = 120000")
|
||||
conn.execute("BEGIN IMMEDIATE")
|
||||
basis = materialize(conn)
|
||||
conn.commit()
|
||||
except Exception:
|
||||
conn.rollback()
|
||||
raise
|
||||
finally:
|
||||
conn.close()
|
||||
print(json_payload(basis))
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
Reference in New Issue
Block a user