refactor(535-1): PDF 엔진 공용화 — common_util_print_pdf(render_html_to_pdf · render_htmls_to_pdfs · merge_pdfs · 엑셀 인쇄 설정) · B09 표 인쇄는 부르기만

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
2026-10-25 19:23:42 +09:00
co-authored by Claude Opus 5.5
parent bc39953fbd
commit 9add13ccbc
2 changed files with 236 additions and 137 deletions
+24 -137
View File
@@ -12,27 +12,30 @@
`B09_Estimation_Router` 가 draft/*_Router.py 의 router 를 끼워 넣음
(같은 머리 · 같은 권한 = 회사 · 프로젝트).
PDF 는 자식 파이썬 프로세스(이 파일을 스크립트로 · 입력 JSON stdin · 결과 JSON 한 줄) —
서버 안 스레드에선 Playwright 가 드라이버를 못 띄움(`B08_Quantity_Milestone_Capture` 와 같은 까닭).
헤드리스 쪽은 스크립트 끔 · 바깥 요청 모두 막음(그림 = data: 주소만).
PDF · 엑셀 인쇄 설정 = 공용 `common_util_print_pdf`(이 파일은 인쇄 CSS · 줄임만).
"""
from __future__ import annotations
import asyncio
import html as htmlmod
import io
import json
import re
import subprocess
import sys
import tempfile
import threading
from pathlib import Path
from typing import Any, Literal
from urllib.parse import quote
from uuid import UUID
from fastapi import APIRouter, HTTPException
from fastapi.responses import Response
from pydantic import BaseModel, Field
from common_util.common_util_print_pdf import ( # noqa: F401 — PrintError · sheet_title = 시험 이음
PrintError,
number_format,
print_setup,
render_html_to_pdf,
sheet_title,
)
TIMEOUT_S = 90
CHANNELS = ("msedge", "chrome") # 시스템 브라우저 — 번들 chromium 은 안 받아 둠
NO_BROWSER = "서버에서 브라우저(Edge · Chrome)를 못 띄움"
#: A4 인쇄 칸 폭(px · 96dpi) — 210 · 297mm − 좌우 여백 10mm 둘
PRINT_W = {"portrait": 190 / 25.4 * 96, "landscape": 277 / 25.4 * 96}
MIN_ZOOM = 0.3
@@ -74,16 +77,6 @@ body { font-family: "Malgun Gothic", "맑은 고딕", sans-serif; font-size: 11p
.sheet-figure img { max-width: 100%; max-height: 240mm; }
.sheet-figure figcaption { font-size: 11px; margin-top: 2px; }
"""
FOOTER = (
'<div style="width:100%;font-size:9px;text-align:center;font-family:sans-serif">'
'<span class="pageNumber"></span> / <span class="totalPages"></span></div>'
)
_lock = threading.Lock() # ponytail: 서버 전체 한 줄 — 브라우저 한 번에 하나
class PrintError(Exception):
"""까닭 한 줄(예외 원문 없음)."""
def _titled(table: str, name: str) -> str:
@@ -115,105 +108,19 @@ FIT_JS = (
)
def render_pdf(doc: str) -> bytes:
"""헤드리스 브라우저 → PDF 바이트. 실패 = PrintError."""
try:
from playwright.sync_api import Error as PwError
from playwright.sync_api import sync_playwright
except ImportError as error:
raise PrintError("서버에 Playwright 가 없습니다(pip install playwright)") from error
try:
with sync_playwright() as p:
browser = None
for channel in CHANNELS:
try:
browser = p.chromium.launch(channel=channel, headless=True)
break
except PwError:
continue
if browser is None:
raise PrintError(NO_BROWSER)
try:
context = browser.new_context(java_script_enabled=False)
context.route("**/*", lambda route: route.abort()) # 바깥 요청 없음(data: 만)
page = context.new_page()
page.set_content(doc, wait_until="load")
page.evaluate(FIT_JS, [PRINT_W["portrait"], PRINT_W["landscape"]])
return page.pdf(
prefer_css_page_size=True,
print_background=True,
display_header_footer=True,
header_template="<span></span>",
footer_template=FOOTER,
)
finally:
browser.close()
except PrintError:
raise
except Exception as error: # 드라이버 못 뜸 등 — 예외 글은 내지 않음
raise PrintError(NO_BROWSER) from error
def spawn_pdf(doc: str) -> bytes:
"""자식 파이썬 프로세스로 render_pdf — 결과 PDF 는 임시 파일로 받음."""
with _lock, tempfile.TemporaryDirectory() as tmp:
out = Path(tmp) / "out.pdf"
try:
done = subprocess.run(
[sys.executable, __file__],
input=json.dumps({"doc": doc, "out": str(out)}).encode(),
capture_output=True,
timeout=TIMEOUT_S,
)
except subprocess.TimeoutExpired as error:
raise PrintError(f"PDF 만들기가 {TIMEOUT_S}초 안에 안 끝남") from error
except OSError as error:
raise PrintError(NO_BROWSER) from error
try:
result = json.loads(done.stdout.decode().strip().splitlines()[-1])
except (ValueError, IndexError): # 자식이 죽음 — stderr(예외 글)는 내지 않음
result = {"reason": NO_BROWSER}
if result.get("reason"):
raise PrintError(result["reason"])
return out.read_bytes()
"""document_html → PDF(쪽 크기 = 인쇄 CSS · 넓은 표 줄임 · 쪽 번호). 실패 = PrintError."""
return render_html_to_pdf(
doc,
page_size=None,
page_numbers=True,
prepare_js=FIT_JS,
prepare_arg=[PRINT_W["portrait"], PRINT_W["landscape"]],
)
# ── 엑셀 ──────────────────────────────────────────────────────────────────
BAD_TITLE = re.compile(r"[\[\]:*?/\\]")
def sheet_title(name: str, used: set[str]) -> str:
"""엑셀 시트 이름 — 금지 글자 = 「_」 · 31자 · 같은 이름 = 뒤에 (2)…"""
base = BAD_TITLE.sub("_", name).strip().strip("'") or "시트"
title, n = base[:31], 1
while title.lower() in used:
n += 1
tail = f" ({n})"
title = base[: 31 - len(tail)] + tail
used.add(title.lower())
return title
def number_format(decimals: int | None) -> str:
return "#,##0" + ("." + "0" * decimals if decimals else "")
def print_setup(ws: Any, orientation: str, head_rows: int) -> None:
"""인쇄 설정(302-5) — PDF 와 같은 방향 · A4 · 폭 1쪽 맞춤 · 머리 줄 쪽마다 · 여백 ≈ PDF."""
from openpyxl.worksheet.page import PageMargins
ws.page_setup.orientation = orientation
ws.page_setup.paperSize = ws.PAPERSIZE_A4
ws.page_setup.fitToWidth = 1
ws.page_setup.fitToHeight = 0
ws.sheet_properties.pageSetUpPr.fitToPage = True
ws.page_margins = PageMargins(left=0.4, right=0.4, top=0.4, bottom=0.55, header=0.2, footer=0.2)
if head_rows:
ws.print_title_rows = f"1:{head_rows}"
ws.oddHeader.left.text = "&A" # 시트 이름
ws.oddFooter.center.text = "&P / &N" # 쪽 번호
def build_xlsx(sheets: list[dict[str, Any]]) -> bytes:
"""grid 시트들 → 엑셀 바이트(차례 그대로)."""
@@ -277,26 +184,6 @@ def build_xlsx(sheets: list[dict[str, Any]]) -> bytes:
return out.getvalue()
if __name__ == "__main__": # 자식 프로세스 몸통 — fastapi 를 읽지 않음
job = json.loads(sys.stdin.buffer.read())
try:
Path(job["out"]).write_bytes(render_pdf(job["doc"]))
print(json.dumps({"reason": None}))
except PrintError as error:
print(json.dumps({"reason": str(error)}, ensure_ascii=False))
sys.exit(0)
# ── 길 ────────────────────────────────────────────────────────────────────
import asyncio # noqa: E402
from urllib.parse import quote # noqa: E402
from uuid import UUID # noqa: E402
from fastapi import APIRouter, HTTPException # noqa: E402
from fastapi.responses import Response # noqa: E402
from pydantic import BaseModel, Field # noqa: E402
router = APIRouter()
MAX_SHEETS = 200