|
|
|
|
@@ -1,4 +1,4 @@
|
|
|
|
|
"""Artifact-tool build adapter and recoverable atomic publication."""
|
|
|
|
|
"""Workbook build adapter and recoverable atomic publication."""
|
|
|
|
|
|
|
|
|
|
from __future__ import annotations
|
|
|
|
|
|
|
|
|
|
@@ -6,13 +6,17 @@ import hashlib
|
|
|
|
|
import json
|
|
|
|
|
import os
|
|
|
|
|
import shutil
|
|
|
|
|
import subprocess
|
|
|
|
|
import tempfile
|
|
|
|
|
import uuid
|
|
|
|
|
from dataclasses import dataclass
|
|
|
|
|
from datetime import date, datetime
|
|
|
|
|
from pathlib import Path
|
|
|
|
|
from typing import Any, Dict, Mapping, Optional, Tuple
|
|
|
|
|
|
|
|
|
|
from openpyxl import Workbook, load_workbook
|
|
|
|
|
from openpyxl.styles import Alignment, Border, Font, PatternFill, Side
|
|
|
|
|
from openpyxl.worksheet.worksheet import Worksheet
|
|
|
|
|
|
|
|
|
|
from company_reports.contracts import CompanyReport, ErrorCode, RESULT_SCHEMA_VERSION
|
|
|
|
|
from company_reports.repository import (
|
|
|
|
|
FileMetadata,
|
|
|
|
|
@@ -23,6 +27,21 @@ from company_reports.repository import (
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
XLSX_MIME = "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
|
|
|
|
|
REPORT_HEADERS = (
|
|
|
|
|
"ARRIVAL",
|
|
|
|
|
"DEPARTURE",
|
|
|
|
|
"NIGHTS",
|
|
|
|
|
"BLOCK_CODE",
|
|
|
|
|
"RES_COMMENT",
|
|
|
|
|
"Booking Room",
|
|
|
|
|
"Total Booking Price",
|
|
|
|
|
)
|
|
|
|
|
BODY_FONT = "222222"
|
|
|
|
|
DUPLICATE_FILL = "FFF2CC"
|
|
|
|
|
DUPLICATE_FONT = "9C6500"
|
|
|
|
|
REVIEW_FILL = "FCE4D6"
|
|
|
|
|
REVIEW_FONT = "C65911"
|
|
|
|
|
HEADER_FILL = "FFFFFF"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class BuildError(RuntimeError):
|
|
|
|
|
@@ -89,63 +108,196 @@ class ArtifactToolBuilder:
|
|
|
|
|
artifact_tool_module: Optional[Path] = None,
|
|
|
|
|
timeout_seconds: int = 120,
|
|
|
|
|
) -> None:
|
|
|
|
|
configured = (node_binary or os.environ.get("COMPANY_REPORT_NODE_BINARY", "")).strip()
|
|
|
|
|
self._node_binary = configured or shutil.which("node") or ""
|
|
|
|
|
_ = (node_binary, artifact_tool_module, timeout_seconds)
|
|
|
|
|
self._builder_script = builder_script.resolve()
|
|
|
|
|
self._artifact_tool_module = (
|
|
|
|
|
artifact_tool_module.resolve() if artifact_tool_module else None
|
|
|
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
|
def _safe_text(value: Any) -> str:
|
|
|
|
|
text = str(value or "")
|
|
|
|
|
return "'" + text if text[:1] in {"=", "+", "-", "@"} else text
|
|
|
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
|
def _date_key(value: Any) -> str:
|
|
|
|
|
if isinstance(value, datetime):
|
|
|
|
|
return value.date().isoformat()
|
|
|
|
|
if isinstance(value, date):
|
|
|
|
|
return value.isoformat()
|
|
|
|
|
if isinstance(value, str) and len(value) >= 10:
|
|
|
|
|
return value[:10]
|
|
|
|
|
return ""
|
|
|
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
|
def _cell_text(value: Any) -> str:
|
|
|
|
|
return "" if value is None else str(value)
|
|
|
|
|
|
|
|
|
|
@staticmethod
|
|
|
|
|
def _semantic_sha256(payload: Mapping[str, Any]) -> str:
|
|
|
|
|
encoded = json.dumps(
|
|
|
|
|
{
|
|
|
|
|
"headers": payload["headers"],
|
|
|
|
|
"periods": payload["periods"],
|
|
|
|
|
},
|
|
|
|
|
ensure_ascii=False,
|
|
|
|
|
separators=(",", ":"),
|
|
|
|
|
).encode("utf-8")
|
|
|
|
|
return hashlib.sha256(encoded).hexdigest()
|
|
|
|
|
|
|
|
|
|
def _write_sheet(
|
|
|
|
|
self,
|
|
|
|
|
worksheet: Worksheet,
|
|
|
|
|
period: Any,
|
|
|
|
|
) -> None:
|
|
|
|
|
worksheet.append(list(REPORT_HEADERS))
|
|
|
|
|
for row in period.rows:
|
|
|
|
|
worksheet.append(
|
|
|
|
|
[
|
|
|
|
|
row.arrival,
|
|
|
|
|
row.departure,
|
|
|
|
|
row.nights,
|
|
|
|
|
self._safe_text(row.block_code),
|
|
|
|
|
self._safe_text(row.res_comment),
|
|
|
|
|
self._safe_text(row.booking_room),
|
|
|
|
|
self._safe_text(row.total_booking_price),
|
|
|
|
|
]
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
worksheet.column_dimensions["A"].width = 13
|
|
|
|
|
worksheet.column_dimensions["B"].width = 13
|
|
|
|
|
worksheet.column_dimensions["C"].width = 9
|
|
|
|
|
worksheet.column_dimensions["D"].width = 20
|
|
|
|
|
worksheet.column_dimensions["E"].width = 25
|
|
|
|
|
worksheet.column_dimensions["F"].width = 34
|
|
|
|
|
worksheet.column_dimensions["G"].width = 48
|
|
|
|
|
worksheet.freeze_panes = "A2"
|
|
|
|
|
worksheet.auto_filter.ref = f"A1:G{max(1, worksheet.max_row)}"
|
|
|
|
|
|
|
|
|
|
thin_bottom = Border(
|
|
|
|
|
bottom=Side(style="thin", color="7F7F7F"),
|
|
|
|
|
)
|
|
|
|
|
self._timeout_seconds = timeout_seconds
|
|
|
|
|
header_fill = PatternFill("solid", fgColor=HEADER_FILL)
|
|
|
|
|
duplicate_fill = PatternFill("solid", fgColor=DUPLICATE_FILL)
|
|
|
|
|
review_fill = PatternFill("solid", fgColor=REVIEW_FILL)
|
|
|
|
|
for cell in worksheet[1]:
|
|
|
|
|
cell.fill = header_fill
|
|
|
|
|
cell.font = Font(name="Arial", size=10, bold=True, color="000000")
|
|
|
|
|
cell.alignment = Alignment(horizontal="center", vertical="center")
|
|
|
|
|
cell.border = thin_bottom
|
|
|
|
|
worksheet.row_dimensions[1].height = 24
|
|
|
|
|
|
|
|
|
|
for row_index in range(2, worksheet.max_row + 1):
|
|
|
|
|
worksheet.row_dimensions[row_index].height = 30
|
|
|
|
|
for cell in worksheet[row_index]:
|
|
|
|
|
cell.font = Font(name="Arial", size=10, color=BODY_FONT)
|
|
|
|
|
cell.alignment = Alignment(vertical="center")
|
|
|
|
|
for column in ("A", "B", "C", "D"):
|
|
|
|
|
worksheet[f"{column}{row_index}"].alignment = Alignment(
|
|
|
|
|
horizontal="center",
|
|
|
|
|
vertical="center",
|
|
|
|
|
)
|
|
|
|
|
for column in ("E", "F", "G"):
|
|
|
|
|
worksheet[f"{column}{row_index}"].alignment = Alignment(
|
|
|
|
|
horizontal="left",
|
|
|
|
|
vertical="center",
|
|
|
|
|
wrap_text=True,
|
|
|
|
|
)
|
|
|
|
|
worksheet[f"A{row_index}"].number_format = "yyyy-mm-dd"
|
|
|
|
|
worksheet[f"B{row_index}"].number_format = "yyyy-mm-dd"
|
|
|
|
|
|
|
|
|
|
for offset, source_row in enumerate(period.rows, start=2):
|
|
|
|
|
if source_row.duplicate_group:
|
|
|
|
|
for column in range(1, 8):
|
|
|
|
|
worksheet.cell(offset, column).fill = duplicate_fill
|
|
|
|
|
worksheet[f"E{offset}"].font = Font(
|
|
|
|
|
name="Arial",
|
|
|
|
|
size=10,
|
|
|
|
|
bold=True,
|
|
|
|
|
color=DUPLICATE_FONT,
|
|
|
|
|
)
|
|
|
|
|
if source_row.multi_price_review:
|
|
|
|
|
worksheet[f"G{offset}"].fill = review_fill
|
|
|
|
|
worksheet[f"G{offset}"].font = Font(
|
|
|
|
|
name="Arial",
|
|
|
|
|
size=10,
|
|
|
|
|
bold=True,
|
|
|
|
|
color=REVIEW_FONT,
|
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
def _validate_workbook(self, output_path: Path, report: CompanyReport) -> int:
|
|
|
|
|
workbook = load_workbook(output_path, data_only=False)
|
|
|
|
|
expected_names = [period.sheet_name for period in report.periods]
|
|
|
|
|
if workbook.sheetnames != expected_names:
|
|
|
|
|
raise ValueError("worksheet names do not match")
|
|
|
|
|
formula_count = 0
|
|
|
|
|
for period in report.periods:
|
|
|
|
|
worksheet = workbook[period.sheet_name]
|
|
|
|
|
expected_rows = len(period.rows) + 1
|
|
|
|
|
if worksheet.max_row != expected_rows or worksheet.max_column != 7:
|
|
|
|
|
raise ValueError("worksheet dimensions do not match")
|
|
|
|
|
headers = [worksheet.cell(1, column).value for column in range(1, 8)]
|
|
|
|
|
if headers != list(REPORT_HEADERS):
|
|
|
|
|
raise ValueError("worksheet headers do not match")
|
|
|
|
|
for row_index, row in enumerate(period.rows, start=2):
|
|
|
|
|
expected = [
|
|
|
|
|
row.arrival.isoformat(),
|
|
|
|
|
row.departure.isoformat(),
|
|
|
|
|
row.nights,
|
|
|
|
|
self._safe_text(row.block_code),
|
|
|
|
|
self._safe_text(row.res_comment),
|
|
|
|
|
self._safe_text(row.booking_room),
|
|
|
|
|
self._safe_text(row.total_booking_price),
|
|
|
|
|
]
|
|
|
|
|
actual = [
|
|
|
|
|
self._date_key(worksheet.cell(row_index, 1).value),
|
|
|
|
|
self._date_key(worksheet.cell(row_index, 2).value),
|
|
|
|
|
worksheet.cell(row_index, 3).value,
|
|
|
|
|
self._cell_text(worksheet.cell(row_index, 4).value),
|
|
|
|
|
self._cell_text(worksheet.cell(row_index, 5).value),
|
|
|
|
|
self._cell_text(worksheet.cell(row_index, 6).value),
|
|
|
|
|
self._cell_text(worksheet.cell(row_index, 7).value),
|
|
|
|
|
]
|
|
|
|
|
if actual != expected:
|
|
|
|
|
raise ValueError("worksheet values do not match")
|
|
|
|
|
for cells in worksheet.iter_rows():
|
|
|
|
|
for cell in cells:
|
|
|
|
|
if cell.data_type == "f":
|
|
|
|
|
formula_count += 1
|
|
|
|
|
if formula_count != 0:
|
|
|
|
|
raise ValueError("workbook must contain no formulas")
|
|
|
|
|
return formula_count
|
|
|
|
|
|
|
|
|
|
def build(self, report: CompanyReport, work_dir: Path) -> BuiltWorkbook:
|
|
|
|
|
if not self._node_binary or not self._builder_script.is_file():
|
|
|
|
|
raise BuildError(
|
|
|
|
|
ErrorCode.OUTPUT_VALIDATION_FAILED,
|
|
|
|
|
"the XLSX builder runtime is unavailable",
|
|
|
|
|
)
|
|
|
|
|
work_dir.mkdir(parents=True, exist_ok=True, mode=0o700)
|
|
|
|
|
os.chmod(work_dir, 0o700)
|
|
|
|
|
payload_path = work_dir / "workbook-payload.json"
|
|
|
|
|
output_path = work_dir / report.filename
|
|
|
|
|
preview_dir = work_dir / "previews"
|
|
|
|
|
summary_path = work_dir / "workbook-summary.json"
|
|
|
|
|
_private_json(payload_path, report.to_workbook_payload())
|
|
|
|
|
try:
|
|
|
|
|
environment = os.environ.copy()
|
|
|
|
|
if self._artifact_tool_module is not None:
|
|
|
|
|
environment["COMPANY_REPORT_ARTIFACT_TOOL_MODULE"] = str(
|
|
|
|
|
self._artifact_tool_module
|
|
|
|
|
)
|
|
|
|
|
completed = subprocess.run(
|
|
|
|
|
[
|
|
|
|
|
self._node_binary,
|
|
|
|
|
str(self._builder_script),
|
|
|
|
|
str(payload_path),
|
|
|
|
|
str(output_path),
|
|
|
|
|
str(preview_dir),
|
|
|
|
|
str(summary_path),
|
|
|
|
|
],
|
|
|
|
|
cwd=self._builder_script.parent,
|
|
|
|
|
capture_output=True,
|
|
|
|
|
text=True,
|
|
|
|
|
timeout=self._timeout_seconds,
|
|
|
|
|
env=environment,
|
|
|
|
|
check=False,
|
|
|
|
|
)
|
|
|
|
|
except (OSError, subprocess.TimeoutExpired):
|
|
|
|
|
raise BuildError(
|
|
|
|
|
ErrorCode.OUTPUT_VALIDATION_FAILED,
|
|
|
|
|
"the XLSX builder did not complete",
|
|
|
|
|
) from None
|
|
|
|
|
if completed.returncode != 0:
|
|
|
|
|
raise BuildError(
|
|
|
|
|
ErrorCode.OUTPUT_VALIDATION_FAILED,
|
|
|
|
|
"the XLSX builder rejected the generated payload",
|
|
|
|
|
)
|
|
|
|
|
try:
|
|
|
|
|
summary = json.loads(summary_path.read_text(encoding="utf-8"))
|
|
|
|
|
payload = report.to_workbook_payload()
|
|
|
|
|
_private_json(payload_path, payload)
|
|
|
|
|
preview_dir.mkdir(parents=True, exist_ok=True, mode=0o700)
|
|
|
|
|
workbook = Workbook()
|
|
|
|
|
for index, period in enumerate(report.periods):
|
|
|
|
|
worksheet = workbook.active if index == 0 else workbook.create_sheet()
|
|
|
|
|
worksheet.title = period.sheet_name
|
|
|
|
|
self._write_sheet(worksheet, period)
|
|
|
|
|
workbook.save(output_path)
|
|
|
|
|
os.chmod(output_path, 0o600)
|
|
|
|
|
formula_count = self._validate_workbook(output_path, report)
|
|
|
|
|
expected_rows = [len(period.rows) for period in report.periods]
|
|
|
|
|
expected_names = [period.sheet_name for period in report.periods]
|
|
|
|
|
summary = {
|
|
|
|
|
"status": "success",
|
|
|
|
|
"schema_version": payload["schema_version"],
|
|
|
|
|
"company": report.company,
|
|
|
|
|
"filename": report.filename,
|
|
|
|
|
"sheet_names": expected_names,
|
|
|
|
|
"row_counts": expected_rows,
|
|
|
|
|
"formula_count": formula_count,
|
|
|
|
|
"semantic_sha256": self._semantic_sha256(payload),
|
|
|
|
|
"preview_count": 0,
|
|
|
|
|
"byte_size": output_path.stat().st_size,
|
|
|
|
|
}
|
|
|
|
|
if (
|
|
|
|
|
summary.get("status") != "success"
|
|
|
|
|
or summary.get("company") != report.company
|
|
|
|
|
@@ -155,17 +307,16 @@ class ArtifactToolBuilder:
|
|
|
|
|
or summary.get("formula_count") != 0
|
|
|
|
|
or not isinstance(summary.get("semantic_sha256"), str)
|
|
|
|
|
or len(summary.get("semantic_sha256")) != 64
|
|
|
|
|
or summary.get("preview_count") != 3
|
|
|
|
|
or not output_path.is_file()
|
|
|
|
|
or output_path.stat().st_size <= 0
|
|
|
|
|
):
|
|
|
|
|
raise ValueError("summary mismatch")
|
|
|
|
|
_private_json(summary_path, summary)
|
|
|
|
|
except (OSError, ValueError, TypeError, json.JSONDecodeError):
|
|
|
|
|
raise BuildError(
|
|
|
|
|
ErrorCode.OUTPUT_VALIDATION_FAILED,
|
|
|
|
|
"the XLSX builder result could not be validated",
|
|
|
|
|
) from None
|
|
|
|
|
os.chmod(output_path, 0o600)
|
|
|
|
|
return BuiltWorkbook(
|
|
|
|
|
path=output_path,
|
|
|
|
|
sha256=sha256_file(output_path),
|
|
|
|
|
|