Update intranet tools and voucher comparison

scripts/analyze_wehago_voucher_coverage.py

@@ -0,0 +1,427 @@
+from __future__ import annotations
+
+import json
+import re
+import sqlite3
+from collections import Counter, defaultdict
+from datetime import datetime
+from pathlib import Path
+from typing import Any
+
+from openpyxl import Workbook
+from openpyxl.styles import Font, PatternFill
+from openpyxl.utils import get_column_letter
+
+import sys
+
+ROOT = Path(__file__).resolve().parents[1]
+sys.path.insert(0, str(ROOT))
+
+from runtime_config import DB_PATH  # noqa: E402
+
+EXPORT_DIR = ROOT / "static" / "exports" / "wehago_compare"
+YEAR = 2025
+WEHAGO_STATUSES = ("voucher_matched", "voucher_unmatched", "voucher_recheck")
+
+
+def clean(value: Any) -> str:
+    return "" if value is None else str(value).strip()
+
+
+def key_from_display(year: int, ledger_date: str, voucher_no: str) -> str:
+    voucher = clean(voucher_no)
+    if re.fullmatch(r"\d{8}-\d{5}", voucher):
+        return voucher
+    match = re.fullmatch(r"(\d{4})-(\d{2})-(\d{5})", voucher)
+    if match:
+        return f"{match.group(1)}{match.group(2)}-{match.group(3)}"
+    date_text = clean(ledger_date)
+    date_match = re.search(r"(\d{1,2})[-./](\d{1,2})", date_text)
+    voucher_digits = re.sub(r"\D+", "", voucher)
+    if date_match and voucher_digits:
+        month = int(date_match.group(1))
+        day = int(date_match.group(2))
+        return f"{int(year):04d}{month:02d}{day:02d}-{int(voucher_digits):05d}"
+    return ""
+
+
+def latest_projection_signature(cur: sqlite3.Cursor) -> str:
+    row = cur.execute(
+        """
+        SELECT payload_json
+        FROM wehago_action_history
+        WHERE action_type = 'auto_recheck_promote'
+        ORDER BY id DESC
+        LIMIT 1
+        """
+    ).fetchone()
+    if row:
+        try:
+            payload = json.loads(row[0] or "{}")
+            signature = clean(payload.get("signature"))
+            if signature:
+                exists = cur.execute(
+                    """
+                    SELECT 1
+                    FROM wehago_compare_query_groups
+                    WHERE start_year = ? AND end_year = ? AND signature = ?
+                    LIMIT 1
+                    """,
+                    (YEAR, YEAR, signature),
+                ).fetchone()
+                if exists:
+                    return signature
+        except Exception:
+            pass
+    row = cur.execute(
+        """
+        SELECT signature, MAX(updated_at) AS max_updated_at
+        FROM wehago_compare_query_groups
+        WHERE start_year = ? AND end_year = ? AND signature LIKE 'compare-query-v5|%'
+        GROUP BY signature
+        ORDER BY max_updated_at DESC
+        LIMIT 1
+        """,
+        (YEAR, YEAR),
+    ).fetchone()
+    return clean(row[0]) if row else ""
+
+
+def append_rows(ws, rows: list[list[Any]]) -> None:
+    for row in rows:
+        ws.append(row)
+
+
+def style_sheet(ws) -> None:
+    header_fill = PatternFill("solid", fgColor="D9EAF7")
+    for cell in ws[1]:
+        cell.font = Font(bold=True)
+        cell.fill = header_fill
+    ws.freeze_panes = "A2"
+    ws.auto_filter.ref = ws.dimensions
+    for col_idx, column_cells in enumerate(ws.columns, start=1):
+        max_len = 10
+        for cell in column_cells:
+            max_len = max(max_len, min(len(clean(cell.value)), 80))
+        ws.column_dimensions[get_column_letter(col_idx)].width = max_len + 2
+
+
+def main() -> None:
+    conn = sqlite3.connect(DB_PATH)
+    conn.row_factory = sqlite3.Row
+    cur = conn.cursor()
+    signature = latest_projection_signature(cur)
+    if not signature:
+        raise RuntimeError("compare query projection signature not found")
+
+    db_rows = cur.execute(
+        """
+        SELECT fiscal_year, voucher_no, status, voucher_row_count, ledger_row_count,
+               voucher_debit, voucher_credit, ledger_debit, ledger_credit,
+               voucher_accounts, ledger_accounts, voucher_vendors, ledger_vendors, notes
+        FROM wehago_comparison_results
+        WHERE fiscal_year = ?
+          AND status <> 'voucher_only'
+        ORDER BY voucher_no
+        """,
+        (YEAR,),
+    ).fetchall()
+    db_by_key = {clean(row["voucher_no"]): dict(row) for row in db_rows if clean(row["voucher_no"])}
+    comparison_total = cur.execute(
+        """
+        SELECT COUNT(*) AS count_all,
+               SUM(CASE WHEN status = 'voucher_only' THEN 1 ELSE 0 END) AS voucher_only_count
+        FROM wehago_comparison_results
+        WHERE fiscal_year = ?
+        """,
+        (YEAR,),
+    ).fetchone()
+
+    ledger_source = cur.execute(
+        """
+        SELECT COUNT(*) AS row_count,
+               COUNT(DISTINCT fiscal_year || '|' || compare_voucher_no) AS voucher_count
+        FROM wehago_ledger_rows
+        WHERE fiscal_year = ? AND COALESCE(compare_voucher_no, '') <> ''
+        """,
+        (YEAR,),
+    ).fetchone()
+    voucher_source = cur.execute(
+        """
+        SELECT COUNT(*) AS row_count,
+               COUNT(DISTINCT fiscal_year || '|' || compare_voucher_no) AS voucher_count
+        FROM wehago_voucher_rows
+        WHERE fiscal_year = ? AND COALESCE(compare_voucher_no, '') <> ''
+        """,
+        (YEAR,),
+    ).fetchone()
+
+    ui_rows = cur.execute(
+        f"""
+        SELECT status_key, 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
+        FROM wehago_compare_query_groups
+        WHERE start_year = ? AND end_year = ? AND signature = ?
+          AND status_key IN ({','.join('?' for _ in WEHAGO_STATUSES)})
+        ORDER BY status_key, group_index
+        """,
+        (YEAR, YEAR, signature, *WEHAGO_STATUSES),
+    ).fetchall()
+
+    ui_group_records: list[dict[str, Any]] = []
+    for row in ui_rows:
+        record = dict(row)
+        record["compare_key"] = key_from_display(record["fiscal_year"], record["ledger_date"], record["voucher_no"])
+        ui_group_records.append(record)
+
+    section_counts = Counter(record["status_key"] for record in ui_group_records)
+    ui_by_key: dict[str, list[dict[str, Any]]] = defaultdict(list)
+    bad_key_records: list[dict[str, Any]] = []
+    for record in ui_group_records:
+        key = clean(record.get("compare_key"))
+        if key:
+            ui_by_key[key].append(record)
+        else:
+            bad_key_records.append(record)
+
+    db_keys = set(db_by_key)
+    ui_keys = set(ui_by_key)
+    duplicate_ui = {key: records for key, records in ui_by_key.items() if len(records) > 1}
+    ui_not_in_db = {key: records for key, records in ui_by_key.items() if key not in db_keys}
+    db_not_in_ui = {key: db_by_key[key] for key in sorted(db_keys - ui_keys)}
+
+    section_sum = sum(section_counts.get(status, 0) for status in WEHAGO_STATUSES)
+    db_count = len(db_by_key)
+    ui_unique_count = len(ui_keys)
+    duplicate_extra_count = section_sum - ui_unique_count
+
+    wb = Workbook()
+    ws = wb.active
+    ws.title = "요약"
+    append_rows(
+        ws,
+        [
+            ["항목", "값", "설명"],
+            ["분석 기준 연도", YEAR, ""],
+            ["사용 projection signature", signature, "화면 WEHAGO/Recheck 섹션에 사용된 최신 projection"],
+            ["DB WEHAGO 전표 수", db_count, "wehago_comparison_results의 fiscal_year+voucher_no 기준"],
+            ["DB 비교결과 전체 전표 수", int(comparison_total["count_all"] or 0), "ERP-only(voucher_only)를 포함한 전체 비교 결과"],
+            ["DB 비교결과 중 ERP-only 수", int(comparison_total["voucher_only_count"] or 0), "WEHAGO 전표 수 비교에서는 제외"],
+            ["DB WEHAGO 원장 전표 수", int(ledger_source["voucher_count"] or 0), "wehago_ledger_rows의 fiscal_year+compare_voucher_no distinct"],
+            ["DB WEHAGO 원장 행 수", int(ledger_source["row_count"] or 0), "wehago_ledger_rows 행 수"],
+            ["보조: wehago_voucher_rows 전표 수", int(voucher_source["voucher_count"] or 0), "ERP/증빙성 원천 행으로 보이는 테이블의 distinct 전표 수"],
+            ["WEHAGO", section_counts.get("voucher_matched", 0), "화면 WEHAGO 그룹 수"],
+            ["WEHAGO Unmatched", section_counts.get("voucher_unmatched", 0), "화면 WEHAGO Unmatched 그룹 수"],
+            ["WEHAGO Recheck", section_counts.get("voucher_recheck", 0), "화면 WEHAGO Recheck 그룹 수"],
+            ["화면 3개 섹션 합계", section_sum, "WEHAGO + WEHAGO Unmatched + WEHAGO Recheck"],
+            ["차이(화면 합계 - DB WEHAGO)", section_sum - db_count, "양수면 화면 그룹 집계가 DB 전표 수보다 많음"],
+            ["화면 고유 WEHAGO 전표 수", ui_unique_count, "ledger_date+voucher_no를 YYYYMMDD-전표번호로 환산한 distinct"],
+            ["차이(화면 고유 전표 - DB WEHAGO)", ui_unique_count - db_count, "전표번호 기준 순수 누락/추가 차이"],
+            ["중복 그룹 초과분", duplicate_extra_count, "같은 WEHAGO 전표가 여러 그룹으로 나뉘어 합계에 중복 반영된 수"],
+            ["키 생성 불가 그룹", len(bad_key_records), "ledger_date/voucher_no 조합으로 DB 전표번호를 만들 수 없는 그룹"],
+        ],
+    )
+    style_sheet(ws)
+
+    ws = wb.create_sheet("차이 사유")
+    reason_rows = [
+        ["사유", "건수", "해석"],
+        [
+            "같은 WEHAGO 전표가 화면에서 복수 그룹으로 집계됨",
+            duplicate_extra_count,
+            "전표 하나가 복수 ERP 가전표/후보/라인 그룹으로 분리되면 화면 섹션 합계는 DB 전표 수보다 커집니다.",
+        ],
+        [
+            "화면에는 있으나 DB comparison_results 전표번호와 직접 대응되지 않음",
+            len(ui_not_in_db),
+            "표시 전표번호를 YYYYMMDD-전표번호로 환산해도 DB 전표번호 집합에 없는 경우입니다.",
+        ],
+        [
+            "DB에는 있으나 화면 3개 WEHAGO 섹션에 없음",
+            len(db_not_in_ui),
+            "DB 비교 결과에는 있으나 현재 projection의 WEHAGO/Unmatched/Recheck 그룹에는 없는 경우입니다.",
+        ],
+        [
+            "전표번호 키 생성 불가",
+            len(bad_key_records),
+            "화면 그룹의 일자 또는 전표번호가 비어 있거나 형식이 달라 비교 키를 만들지 못한 경우입니다.",
+        ],
+    ]
+    append_rows(ws, reason_rows)
+    style_sheet(ws)
+
+    ws = wb.create_sheet("중복 그룹 상세")
+    append_rows(
+        ws,
+        [[
+            "compare_key",
+            "그룹 수",
+            "status 목록",
+            "전표 표시값",
+            "일자 목록",
+            "가전표번호/ERP 전표",
+            "WEHAGO 계정",
+            "ERP 계정",
+            "검토 사유",
+        ]],
+    )
+    for key, records in sorted(duplicate_ui.items(), key=lambda item: (-len(item[1]), item[0])):
+        append_rows(
+            ws,
+            [[
+                key,
+                len(records),
+                ", ".join(sorted({clean(r.get("status_key")) for r in records})),
+                ", ".join(sorted({clean(r.get("voucher_no")) for r in records if clean(r.get("voucher_no"))}))[:300],
+                ", ".join(sorted({clean(r.get("ledger_date")) for r in records if clean(r.get("ledger_date"))}))[:300],
+                ", ".join(clean(r.get("draft_no")) for r in records if clean(r.get("draft_no")))[:1000],
+                " | ".join(clean(r.get("ledger_accounts")) for r in records if clean(r.get("ledger_accounts")))[:1000],
+                " | ".join(clean(r.get("voucher_accounts")) for r in records if clean(r.get("voucher_accounts")))[:1000],
+                " | ".join(clean(r.get("review_reason")) for r in records if clean(r.get("review_reason")))[:1000],
+            ]],
+        )
+    style_sheet(ws)
+
+    ws = wb.create_sheet("화면만 있음")
+    append_rows(
+        ws,
+        [[
+            "compare_key",
+            "그룹 수",
+            "status 목록",
+            "일자 목록",
+            "전표 표시값",
+            "가전표번호/ERP 전표",
+            "WEHAGO 계정",
+            "ERP 계정",
+            "검토 사유",
+        ]],
+    )
+    for key, records in sorted(ui_not_in_db.items()):
+        append_rows(
+            ws,
+            [[
+                key,
+                len(records),
+                ", ".join(sorted({clean(r.get("status_key")) for r in records})),
+                ", ".join(sorted({clean(r.get("ledger_date")) for r in records if clean(r.get("ledger_date"))}))[:300],
+                ", ".join(sorted({clean(r.get("voucher_no")) for r in records if clean(r.get("voucher_no"))}))[:300],
+                ", ".join(clean(r.get("draft_no")) for r in records if clean(r.get("draft_no")))[:1000],
+                " | ".join(clean(r.get("ledger_accounts")) for r in records if clean(r.get("ledger_accounts")))[:1000],
+                " | ".join(clean(r.get("voucher_accounts")) for r in records if clean(r.get("voucher_accounts")))[:1000],
+                " | ".join(clean(r.get("review_reason")) for r in records if clean(r.get("review_reason")))[:1000],
+            ]],
+        )
+    style_sheet(ws)
+
+    ws = wb.create_sheet("DB만 있음")
+    append_rows(
+        ws,
+        [[
+            "compare_key",
+            "DB status",
+            "voucher_row_count",
+            "ledger_row_count",
+            "WEHAGO 차변",
+            "WEHAGO 대변",
+            "ERP 차변",
+            "ERP 대변",
+            "WEHAGO 계정",
+            "ERP 계정",
+            "비고",
+        ]],
+    )
+    for key, row in db_not_in_ui.items():
+        append_rows(
+            ws,
+            [[
+                key,
+                clean(row.get("status")),
+                row.get("ledger_row_count"),
+                row.get("voucher_row_count"),
+                row.get("ledger_debit"),
+                row.get("ledger_credit"),
+                row.get("voucher_debit"),
+                row.get("voucher_credit"),
+                clean(row.get("ledger_accounts")),
+                clean(row.get("voucher_accounts")),
+                clean(row.get("notes")),
+            ]],
+        )
+    style_sheet(ws)
+
+    ws = wb.create_sheet("섹션 원자료")
+    append_rows(
+        ws,
+        [[
+            "status",
+            "compare_key",
+            "fiscal_year",
+            "ledger_date",
+            "voucher_no",
+            "draft_no",
+            "ledger_row_count",
+            "voucher_row_count",
+            "WEHAGO 차변",
+            "WEHAGO 대변",
+            "ERP 차변",
+            "ERP 대변",
+            "WEHAGO 계정",
+            "ERP 계정",
+            "WEHAGO 거래처",
+            "ERP 거래처",
+            "review_reason",
+        ]],
+    )
+    for record in ui_group_records:
+        append_rows(
+            ws,
+            [[
+                record.get("status_key"),
+                record.get("compare_key"),
+                record.get("fiscal_year"),
+                record.get("ledger_date"),
+                record.get("voucher_no"),
+                record.get("draft_no"),
+                record.get("ledger_row_count"),
+                record.get("voucher_row_count"),
+                record.get("ledger_debit"),
+                record.get("ledger_credit"),
+                record.get("voucher_debit"),
+                record.get("voucher_credit"),
+                record.get("ledger_accounts"),
+                record.get("voucher_accounts"),
+                record.get("ledger_vendors"),
+                record.get("voucher_vendors"),
+                record.get("review_reason"),
+            ]],
+        )
+    style_sheet(ws)
+
+    EXPORT_DIR.mkdir(parents=True, exist_ok=True)
+    file_name = f"wehago_voucher_coverage_{YEAR}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.xlsx"
+    output_path = EXPORT_DIR / file_name
+    wb.save(output_path)
+    conn.close()
+    print(
+        json.dumps(
+            {
+                "file": str(output_path),
+                "download_url": f"/static/exports/wehago_compare/{file_name}",
+                "db_count": db_count,
+                "section_sum": section_sum,
+                "diff": section_sum - db_count,
+                "ui_unique_count": ui_unique_count,
+                "unique_diff": ui_unique_count - db_count,
+                "duplicate_extra_count": duplicate_extra_count,
+                "ui_not_in_db": len(ui_not_in_db),
+                "db_not_in_ui": len(db_not_in_ui),
+            },
+            ensure_ascii=False,
+        )
+    )
+
+
+if __name__ == "__main__":
+    main()