Files
conjurer/latex_functions.py
Michal Tuszowski f9a1e7c03a Restructure: promote working_copy to repo root
Make the stable 'working copy' bot the canonical code at the repository
root so the install/deploy scripts operate against it again.

- Move working_copy/* to root (bot entrypoint is bot.py)
- Restore root-level install/ops scripts from c4fa88e (deploy.sh,
  install_main_bot.sh, status_report.*, conjurer.service, etc.)
- Fix deploy.sh: copy bot.py (was thin_client.py) and add the
  conanjurer_* modules; bump command count
- Remove side-by-side variant dirs (backup_old_docker, prototype_one,
  prototype_musician_one, musician_old, working_copy) and docker cruft
- Keep components as subdirs: conjurer_librarian, conjurer_musician,
  spotify_dl, yt_dlp, fonts, utils, docs

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-28 22:51:24 +02:00

180 lines
7.2 KiB
Python
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
import asyncio
import logging
import re
import shutil
import tempfile
import zipfile
from pathlib import Path
from typing import Dict, Iterable, List, Optional, Tuple
# ===== Bezpieczeństwo nazw/rozszerzeń =====
SAFE_TEX_NAME = re.compile(r"^[\w\-. ]+\.tex$", re.IGNORECASE)
SAFE_ASSET_NAME = re.compile(
r"^[\w\-. /]+?\.(tex|png|jpg|jpeg|pdf|svg|eps|bmp|gif|sty|cls|bib|bst|bbx|cbx|def|cfg)$",
re.IGNORECASE,
)
def is_safe_tex_name(name: str) -> bool:
return bool(SAFE_TEX_NAME.match(name or ""))
def is_safe_asset_name(name: str) -> bool:
return bool(SAFE_ASSET_NAME.match(name or ""))
# ===== Uruchamianie kompilatora =====
async def _run(cmd: List[str], cwd: Path, timeout_sec: int, logger_name: Optional[str]) -> Tuple[int, str]:
"""Uruchamia proces kompilatora w cwd, zwraca (rc, sklejone_stdout_stderr)."""
logger = logging.getLogger(logger_name) if logger_name else logging.getLogger()
logger.debug("RUN %s (cwd=%s, timeout=%s)", " ".join(cmd), cwd, timeout_sec)
proc = await asyncio.create_subprocess_exec(
*cmd,
stdout=asyncio.subprocess.PIPE,
stderr=asyncio.subprocess.STDOUT,
cwd=str(cwd),
)
try:
out = await asyncio.wait_for(proc.communicate(), timeout=timeout_sec)
except asyncio.TimeoutError:
try:
proc.kill()
finally:
pass
return 124, "TIMEOUT: compiler took too long"
rc = proc.returncode
log = (out[0] or b"").decode("utf-8", errors="ignore")
logger.debug("RC=%s, LOG TAIL:\n%s", rc, log[-1500:])
return rc, log
def _build_cmd(tex_engine: str, main_tex: str) -> List[str]:
"""Buduje komendę kompilatora; domyślnie tectonic/pdflatex/xelatex."""
if tex_engine.lower() == "tectonic":
# Tectonic sam robi wielofazowość, ale i tak uruchamiamy 2x (bezpiecznie dla bib/refs)
return ["tectonic", "--keep-intermediates", "--synctex", main_tex]
elif tex_engine.lower() in ("pdflatex", "xelatex", "lualatex"):
return [tex_engine, "-interaction=nonstopmode", "-halt-on-error", main_tex]
else:
# fallback traktujemy jak pdflatex
return ["pdflatex", "-interaction=nonstopmode", "-halt-on-error", main_tex]
async def run_latex_two_passes(
tex_filename: str, workdir: Path, tex_engine: str, timeout_sec: int, logger: Optional[str] = None
) -> Tuple[bool, str, Path]:
"""
Uruchamia kompilację 2 razy (bezpieczeństwo referencji/bib).
Zwraca: (ok, log_text, pdf_path).
"""
logger_obj = logging.getLogger(logger) if logger else logging.getLogger()
logger_obj.info("Compiling %s with %s engine two passes", tex_filename, tex_engine)
logs: List[str] = []
pdf_path = workdir / (Path(tex_filename).stem + ".pdf")
# PASS 1
rc1, log1 = await _run(_build_cmd(tex_engine, tex_filename), workdir, timeout_sec, logger)
logs.append(log1)
if rc1 != 0:
return False, "".join(logs), pdf_path
# PASS 2 (często już nic nie robi, ale nie szkodzi)
rc2, log2 = await _run(_build_cmd(tex_engine, tex_filename), workdir, timeout_sec, logger)
logs.append(log2)
ok = (rc2 == 0) and pdf_path.exists()
return ok, "".join(logs), pdf_path
# ===== Kompilacje: pojedynczy TEX + ZIP =====
async def compile_single_tex_bytes(
tex_bytes: bytes,
filename: str,
tex_engine: str,
timeout_sec: int,
logger: Optional[str] = None,
support_files: Optional[Iterable[Tuple[str, bytes]]] = None,
) -> Dict[str, Optional[object]]:
"""
Kompiluje pojedynczy .tex + dowolne pliki pomocnicze (obrazy, sty/cls/bib, inne .tex).
Każdy plik zapisujemy w katalogu roboczym widocznym dla kompilatora.
"""
log = logging.getLogger(logger) if logger else logging.getLogger()
log.info("Compiling %s with %s engine single tex", filename, tex_engine)
if not is_safe_tex_name(filename):
return {"ok": False, "pdf_bytes": None, "log_text": "Invalid .tex filename", "pdf_name": None}
with tempfile.TemporaryDirectory(prefix="latex_one_") as td:
wd = Path(td)
(wd / Path(filename).name).write_bytes(tex_bytes)
# Zapisz wszystkie supporty obok głównego pliku
if support_files:
for fname, data in support_files:
if not fname or not is_safe_asset_name(fname):
continue
# spłaszczamy do nazwy bazowej (bez podkatalogów z zewn. źródeł)
(wd / Path(fname).name).write_bytes(data)
ok, log_text, pdf_path = await run_latex_two_passes(Path(filename).name, wd, tex_engine, timeout_sec, logger)
log.info("LaTeX compile finished: %s", "OK" if ok else "FAIL")
log.info("Log text:\n%s", log_text[-2000:]) # Log last 2000 chars
log.info("PDF path: %s", pdf_path)
if ok and pdf_path.exists():
return {
"ok": True,
"pdf_bytes": pdf_path.read_bytes(),
"log_text": log_text,
"pdf_name": pdf_path.name,
}
return {"ok": False, "pdf_bytes": None, "log_text": log_text, "pdf_name": pdf_path.name}
async def compile_zip_to_zip(
zip_bytes: bytes, tex_engine: str, timeout_sec: int, logger: Optional[str] = None
) -> Tuple[bytes, List[str]]:
"""
Rozpakowuje ZIP, znajduję wszystkie .tex i dla KAŻDEGO tworzy osobny katalog roboczy
ze skopiowanym CAŁYM drzewem, by zachować ścieżki względne obrazków/plików.
Zwraca (zip_pdf_bytes, lista_rel_sciezek_które_padły).
"""
log = logging.getLogger(logger) if logger else logging.getLogger()
failed: List[str] = []
out_pdf_paths: List[Path] = []
with tempfile.TemporaryDirectory(prefix="latex_zip_") as td:
td_path = Path(td)
src_root = td_path / "src"
src_root.mkdir(parents=True, exist_ok=True)
# 1) Rozpakuj
zpath = td_path / "in.zip"
zpath.write_bytes(zip_bytes)
with zipfile.ZipFile(zpath, "r") as zf:
zf.extractall(src_root)
# 2) Zbierz .tex
tex_files = [p for p in src_root.rglob("*.tex") if is_safe_tex_name(p.name)]
if not tex_files:
return b"", ["No .tex files found in archive."]
# 3) Kompiluj każdy .tex w izolacji (kopiujemy całe drzewo do osobnego workdiru)
for tex_path in tex_files:
rel_tex = tex_path.relative_to(src_root).as_posix()
work = td_path / f"build_{tex_path.stem}"
try:
shutil.copytree(src_root, work / "src")
ok, log_text, pdf = await run_latex_two_passes(rel_tex, work / "src", tex_engine, timeout_sec, logger)
if ok and pdf.exists():
out_pdf_paths.append(pdf)
else:
failed.append(rel_tex)
log.info("LaTeX FAIL %s\n%s", rel_tex, (log_text or "")[-2000:])
except Exception:
# let it crash log + re-raise (to zobaczy wyższa warstwa/discord.py)
log.exception("Exception while compiling %s", rel_tex)
raise
# 4) Spakuj wyjściowe PDF-y
out_zip = td_path / "compiled_pdfs.zip"
with zipfile.ZipFile(out_zip, "w", compression=zipfile.ZIP_DEFLATED) as zf:
for pdf in out_pdf_paths:
zf.write(pdf, arcname=pdf.name)
return out_zip.read_bytes(), failed