feat(attachments): fail-soft archive_extractor (zip/cbz/rar/7z)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

backend/app/services/archive_extractor.py

@@ -0,0 +1,57 @@
+"""Best-effort archive member extraction (filesystem-import aid).
+
+zip/cbz via stdlib zipfile; rar via rarfile (needs an unrar/bsdtar
+binary); 7z via py7zr (pure-Python). NEVER raises: on any failure the
+context yields an empty list, and the caller still preserves the archive
+file itself as a PostAttachment, so nothing is lost.
+"""
+
+import logging
+import zipfile
+from contextlib import contextmanager
+from pathlib import Path
+from tempfile import TemporaryDirectory
+
+log = logging.getLogger(__name__)
+
+ARCHIVE_EXTS = {".zip", ".cbz", ".rar", ".7z"}
+
+
+def is_archive(path: Path) -> bool:
+    return Path(path).suffix.lower() in ARCHIVE_EXTS
+
+
+@contextmanager
+def extract_archive(path: Path):
+    """Context manager yielding a list of (member_name, extracted_path)
+    for every regular-file member. Temp dir is removed on exit. Members
+    are only ever under the temp dir (libs' safe extractall + rglob from
+    base). Fail-soft: any error → yields []."""
+    tmp = TemporaryDirectory(prefix="fc_arch_")
+    base = Path(tmp.name)
+    members: list[tuple[str, Path]] = []
+    try:
+        try:
+            ext = Path(path).suffix.lower()
+            if ext in (".zip", ".cbz"):
+                with zipfile.ZipFile(path) as zf:
+                    zf.extractall(base)
+            elif ext == ".rar":
+                import rarfile
+
+                with rarfile.RarFile(path) as rf:
+                    rf.extractall(base)
+            elif ext == ".7z":
+                import py7zr
+
+                with py7zr.SevenZipFile(path, "r") as zf:
+                    zf.extractall(base)
+            for p in sorted(base.rglob("*")):
+                if p.is_file():
+                    members.append((str(p.relative_to(base)), p))
+        except Exception as exc:
+            log.warning("archive extract failed for %s: %s", path, exc)
+            members = []
+        yield members
+    finally:
+        tmp.cleanup()