| 1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162 |
- from __future__ import annotations
- import typing as t
- import urllib.parse
- def sanitize_filename(filename: str) -> str | None:
- filename = filename.strip().strip("\"'")
- if not filename:
- return None
- filename = filename.replace("\\", "/").split("/")[-1]
- if not filename or filename in {".", ".."}:
- return None
- return filename
- def extract_filename_from_content_disposition(header: str) -> str | None:
- """Extract filename from Content-Disposition header.
- Supports both ``filename=`` and RFC 5987 ``filename*=``; prefers ``filename*=``
- when present. Returned filenames are sanitized to prevent path traversal.
- """
- if not header:
- return None
- filename_star: str | None = None
- filename_plain: str | None = None
- for part in header.split(";"):
- part = part.strip()
- lower = part.lower()
- if lower.startswith("filename*="):
- value = part[len("filename*=") :].strip()
- value = value.strip().strip("\"'")
- try:
- charset, rest = value.split("'", 1)
- _language, encoded = rest.split("'", 1)
- except ValueError:
- charset = "utf-8"
- encoded = value
- try:
- raw = urllib.parse.unquote_to_bytes(encoded)
- decoded = raw.decode(charset, errors="replace")
- except LookupError:
- raw = urllib.parse.unquote_to_bytes(encoded)
- decoded = raw.decode("utf-8", errors="replace")
- filename_star = sanitize_filename(decoded)
- elif lower.startswith("filename="):
- value = part[len("filename=") :].strip()
- filename_plain = sanitize_filename(value)
- return filename_star or filename_plain
- def infer_filename_from_fileobj(file_obj: t.Any, fallback: str) -> str:
- """Infer a safe filename from a file-like object's ``.name`` attribute."""
- name = getattr(file_obj, "name", None)
- if isinstance(name, str) and name:
- inferred = sanitize_filename(name)
- if inferred:
- return inferred
- return fallback
|
