src/rdflib/contrib/graphdb/util.py
1from __future__ import annotations
2
3import typing as t
4import urllib.parse
5
6
7def sanitize_filename(filename: str) -> str | None:
8 filename = filename.strip().strip("\"'")
9 if not filename:
10 return None
11 filename = filename.replace("\\", "/").split("/")[-1]
12 if not filename or filename in {".", ".."}:
13 return None
14 return filename
15
16
17def extract_filename_from_content_disposition(header: str) -> str | None:
18 """Extract filename from Content-Disposition header.
19
20 Supports both ``filename=`` and RFC 5987 ``filename*=``; prefers ``filename*=``
21 when present. Returned filenames are sanitized to prevent path traversal.
22 """
23 if not header:
24 return None
25
26 filename_star: str | None = None
27 filename_plain: str | None = None
28
29 for part in header.split(";"):
30 part = part.strip()
31 lower = part.lower()
32 if lower.startswith("filename*="):
33 value = part[len("filename*=") :].strip()
34 value = value.strip().strip("\"'")
35 try:
36 charset, rest = value.split("'", 1)
37 _language, encoded = rest.split("'", 1)
38 except ValueError:
39 charset = "utf-8"
40 encoded = value
41 try:
42 raw = urllib.parse.unquote_to_bytes(encoded)
43 decoded = raw.decode(charset, errors="replace")
44 except LookupError:
45 raw = urllib.parse.unquote_to_bytes(encoded)
46 decoded = raw.decode("utf-8", errors="replace")
47 filename_star = sanitize_filename(decoded)
48 elif lower.startswith("filename="):
49 value = part[len("filename=") :].strip()
50 filename_plain = sanitize_filename(value)
51
52 return filename_star or filename_plain
53
54
55def infer_filename_from_fileobj(file_obj: t.Any, fallback: str) -> str:
56 """Infer a safe filename from a file-like object's ``.name`` attribute."""
57 name = getattr(file_obj, "name", None)
58 if isinstance(name, str) and name:
59 inferred = sanitize_filename(name)
60 if inferred:
61 return inferred
62 return fallback