Files
Medios-Macina/plugins/ftp/__init__.py
T

778 lines
27 KiB
Python

from __future__ import annotations
import fnmatch
import ftplib
import posixpath
import tempfile
from datetime import datetime
from pathlib import Path
from typing import Any, Dict, List, Optional, Tuple
from urllib.parse import quote, unquote, urlparse
from ProviderCore.base import Provider, SearchResult, parse_inline_query_arguments
def _pick_provider_config(config: Any) -> Dict[str, Any]:
if not isinstance(config, dict):
return {}
provider = config.get("provider")
if not isinstance(provider, dict):
return {}
entry = provider.get("ftp")
if isinstance(entry, dict):
return entry
return {}
def _coerce_bool(value: Any, default: bool = False) -> bool:
if isinstance(value, bool):
return value
if value is None:
return default
text = str(value).strip().lower()
if not text:
return default
if text in {"1", "true", "yes", "on"}:
return True
if text in {"0", "false", "no", "off"}:
return False
return default
def _coerce_int(value: Any, default: int) -> int:
try:
return int(value)
except Exception:
return default
def _format_timestamp(raw_value: Any) -> str:
text = str(raw_value or "").strip()
if not text:
return ""
for pattern in ("%Y%m%d%H%M%S", "%Y%m%d%H%M%S.%f"):
try:
parsed = datetime.strptime(text, pattern)
return parsed.strftime("%Y-%m-%d %H:%M")
except Exception:
continue
return text
def _safe_filename(name: Any) -> str:
raw = str(name or "").strip()
if not raw:
raw = "download"
cleaned = "".join(ch if ch.isalnum() or ch in {"-", "_", ".", " "} else "_" for ch in raw)
cleaned = cleaned.strip(" ._")
return cleaned or "download"
def _unique_path(path: Path) -> Path:
if not path.exists():
return path
stem = path.stem or "download"
suffix = path.suffix
counter = 1
while True:
candidate = path.with_name(f"{stem}_{counter}{suffix}")
if not candidate.exists():
return candidate
counter += 1
class FTP(Provider):
PLUGIN_NAME = "ftp"
URL = ("ftp://", "ftps://")
@property
def label(self) -> str:
return "FTP"
@property
def preserve_order(self) -> bool:
return True
@classmethod
def config_schema(cls) -> List[Dict[str, Any]]:
return [
{
"key": "host",
"label": "Host",
"default": "",
"required": True,
"placeholder": "ftp.example.com",
},
{
"key": "port",
"label": "Port",
"type": "integer",
"default": 21,
},
{
"key": "username",
"label": "Username",
"default": "anonymous",
},
{
"key": "password",
"label": "Password",
"type": "secret",
"secret": True,
"default": "",
},
{
"key": "base_path",
"label": "Base Path",
"default": "/",
"placeholder": "/incoming",
},
{
"key": "tls",
"label": "Use FTPS",
"type": "boolean",
"default": False,
},
{
"key": "passive",
"label": "Passive Mode",
"type": "boolean",
"default": True,
},
{
"key": "timeout",
"label": "Timeout Seconds",
"type": "integer",
"default": 20,
},
{
"key": "search_depth",
"label": "Default Search Depth",
"type": "integer",
"default": 1,
},
]
def __init__(self, config: Optional[Dict[str, Any]] = None):
super().__init__(config)
conf = _pick_provider_config(self.config)
self._host = str(conf.get("host") or "").strip()
self._tls = _coerce_bool(conf.get("tls"), False)
self._port = _coerce_int(conf.get("port"), 21)
self._username = str(conf.get("username") or conf.get("user") or "anonymous").strip() or "anonymous"
password_value = conf.get("password")
self._password = str(password_value).strip() if password_value not in (None, "") else "anonymous@"
self._passive = _coerce_bool(conf.get("passive"), True)
self._timeout = max(1, _coerce_int(conf.get("timeout"), 20))
self._search_depth = max(0, _coerce_int(conf.get("search_depth"), 1))
self._base_path = self._normalize_remote_path(conf.get("base_path") or "/", default="/")
def validate(self) -> bool:
return bool(self._host)
def config_helper_text(self) -> str:
return "Test the configured FTP/FTPS settings before searching or uploading."
def config_actions(self) -> List[Dict[str, Any]]:
return [
{
"id": "test_connection",
"label": "Test connection",
"variant": "primary",
}
]
def run_config_action(self, action_id: str, **_kwargs: Any) -> Dict[str, Any]:
if str(action_id or "").strip().lower() != "test_connection":
return super().run_config_action(action_id, **_kwargs)
if not self._host:
return {"ok": False, "message": "Set 'host' before testing the FTP connection."}
ftp = None
try:
ftp = self._connect()
active_path = self._base_path or "/"
try:
ftp.cwd(active_path)
resolved_path = ftp.pwd()
except Exception:
resolved_path = active_path
return {
"ok": True,
"message": f"Connected to FTP {self._host}:{self._port} and reached {resolved_path}.",
}
except Exception as exc:
return {"ok": False, "message": f"FTP connection failed: {exc}"}
finally:
self._close(ftp)
def extract_query_arguments(self, query: str) -> Tuple[str, Dict[str, Any]]:
text, inline = parse_inline_query_arguments(query)
filters: Dict[str, Any] = {}
if inline.get("path"):
filters["path"] = inline.get("path")
if inline.get("depth"):
filters["depth"] = max(0, _coerce_int(inline.get("depth"), self._search_depth))
if inline.get("type"):
filters["type"] = str(inline.get("type") or "").strip().lower()
return text, filters
def get_table_title(self, query: str, filters: Optional[Dict[str, Any]] = None) -> str:
active_path = self._normalize_remote_path((filters or {}).get("path") or self._base_path, default=self._base_path)
text = str(query or "").strip()
if not text or text == "*":
return f"FTP: {active_path}"
return f"FTP: {text} @ {active_path}"
def get_table_metadata(self, query: str, filters: Optional[Dict[str, Any]] = None) -> Dict[str, Any]:
return {
"plugin": self.name,
"host": self._host,
"path": self._normalize_remote_path((filters or {}).get("path") or self._base_path, default=self._base_path),
"query": str(query or "").strip(),
}
def search(
self,
query: str,
limit: int = 50,
filters: Optional[Dict[str, Any]] = None,
**kwargs: Any,
) -> List[SearchResult]:
_ = kwargs
active_filters = dict(filters or {})
start_path = self._normalize_remote_path(active_filters.get("path") or self._base_path, default=self._base_path)
search_depth = max(0, _coerce_int(active_filters.get("depth"), self._search_depth))
type_filter = str(active_filters.get("type") or "any").strip().lower()
needle = str(query or "").strip()
max_results = max(0, int(limit or 0))
if max_results <= 0:
return []
ftp = self._connect()
try:
return self._search_directory(
ftp,
start_path,
needle=needle,
limit=max_results,
search_depth=search_depth,
type_filter=type_filter,
)
finally:
self._close(ftp)
def selector(
self,
selected_items: List[Any],
*,
ctx: Any,
stage_is_last: bool = True,
**_kwargs: Any,
) -> bool:
if not stage_is_last:
return False
target_path = ""
target_title = ""
for item in selected_items or []:
metadata = self._item_metadata(item)
if not metadata.get("is_dir"):
continue
target_path = self._normalize_remote_path(metadata.get("ftp_path") or metadata.get("selection_path"), default=self._base_path)
target_title = str(metadata.get("title") or metadata.get("name") or "").strip()
if target_path:
break
if not target_path:
return False
ftp = self._connect()
try:
rows = self._search_directory(
ftp,
target_path,
needle="*",
limit=500,
search_depth=0,
type_filter="any",
)
finally:
self._close(ftp)
try:
from SYS.result_table import Table
from SYS.rich_display import stdout_console
except Exception:
return True
title = target_title or target_path
table = Table(f"FTP: {title}")._perseverance(True)
table.set_table("ftp")
try:
table.set_table_metadata({
"provider": "ftp",
"host": self._host,
"path": target_path,
"view": "directory",
})
except Exception:
pass
table.set_source_command("search-file", ["-plugin", "ftp", f"path:{target_path}", "*"])
payloads: List[Dict[str, Any]] = []
for row in rows:
table.add_result(row)
payloads.append(row.to_dict())
try:
ctx.set_last_result_table(table, payloads, subject={"plugin": "ftp", "path": target_path})
ctx.set_current_stage_table(table)
except Exception:
pass
try:
stdout_console().print()
stdout_console().print(table)
except Exception:
pass
return True
def download(self, result: SearchResult, output_dir: Path) -> Optional[Path]:
metadata = getattr(result, "full_metadata", None)
if isinstance(metadata, dict) and metadata.get("is_dir"):
return None
target = str(getattr(result, "path", "") or "").strip()
if not target:
return None
return self.download_url(target, output_dir, title=getattr(result, "title", None))
def download_url(self, url: str, output_dir: Path, **kwargs: Any) -> Optional[Path]:
settings = self._connection_settings_for_url(url)
remote_path = settings["path"]
if not remote_path or remote_path == "/":
return None
filename_hint = str(kwargs.get("title") or "").strip()
parsed_name = posixpath.basename(remote_path.rstrip("/"))
filename = _safe_filename(filename_hint or unquote(parsed_name) or "download")
destination_dir = Path(output_dir)
destination_dir.mkdir(parents=True, exist_ok=True)
destination = _unique_path(destination_dir / filename)
ftp = self._connect(
host=settings["host"],
port=settings["port"],
username=settings["username"],
password=settings["password"],
tls=settings["tls"],
)
try:
with destination.open("wb") as handle:
ftp.retrbinary(f"RETR {remote_path}", handle.write)
return destination
except Exception:
try:
destination.unlink(missing_ok=True)
except Exception:
pass
return None
finally:
self._close(ftp)
def resolve_pipe_result_download(
self,
result: Any,
pipe_obj: Any,
) -> Tuple[Optional[Path], Optional[str], Optional[Path]]:
metadata = self._item_metadata(result, pipe_obj=pipe_obj)
if metadata.get("is_dir"):
return None, None, None
download_url = str(
metadata.get("selection_url")
or metadata.get("ftp_url")
or metadata.get("path")
or ""
).strip()
if not download_url.startswith(("ftp://", "ftps://")):
return None, None, None
temp_dir = Path(tempfile.mkdtemp(prefix="ftp-add-file-"))
downloaded = self.download_url(download_url, temp_dir, title=metadata.get("title"))
if downloaded is None:
try:
temp_dir.rmdir()
except Exception:
pass
return None, None, None
try:
if pipe_obj is not None:
pipe_obj.is_temp = True
except Exception:
pass
return downloaded, None, temp_dir
def upload(self, file_path: str, **kwargs: Any) -> str:
local_path = Path(str(file_path or "")).expanduser()
if not local_path.exists() or not local_path.is_file():
raise FileNotFoundError(f"File not found: {local_path}")
remote_dir = self._normalize_remote_path(kwargs.get("remote_path") or kwargs.get("path") or self._base_path, default=self._base_path)
remote_name = posixpath.basename(str(kwargs.get("remote_name") or local_path.name).replace("\\", "/")) or local_path.name
remote_path = self._join_remote_path(remote_dir, remote_name)
ftp = self._connect()
try:
self._ensure_directory(ftp, remote_dir)
with local_path.open("rb") as handle:
ftp.storbinary(f"STOR {remote_path}", handle)
finally:
self._close(ftp)
return self._build_url(remote_path)
def _connect(
self,
*,
host: Optional[str] = None,
port: Optional[int] = None,
username: Optional[str] = None,
password: Optional[str] = None,
tls: Optional[bool] = None,
) -> ftplib.FTP:
use_tls = self._tls if tls is None else bool(tls)
ftp: ftplib.FTP = ftplib.FTP_TLS() if use_tls else ftplib.FTP()
ftp.connect(host or self._host, int(port or self._port), timeout=self._timeout)
ftp.login(username or self._username, password or self._password)
try:
ftp.set_pasv(self._passive)
except Exception:
pass
if use_tls and isinstance(ftp, ftplib.FTP_TLS):
ftp.prot_p()
return ftp
def _close(self, ftp: Optional[ftplib.FTP]) -> None:
if ftp is None:
return
try:
ftp.quit()
except Exception:
try:
ftp.close()
except Exception:
pass
def _normalize_remote_path(self, value: Any, *, default: str) -> str:
text = str(value or "").strip().replace("\\", "/")
if not text:
text = default
elif text.startswith(("ftp://", "ftps://")):
try:
text = unquote(urlparse(text).path or "/")
except Exception:
text = default
elif not text.startswith("/"):
text = posixpath.join(default, text)
normalized = posixpath.normpath(text)
normalized = "/" + normalized.lstrip("/")
return normalized or "/"
def _join_remote_path(self, parent: Any, child: Any) -> str:
left = self._normalize_remote_path(parent, default=self._base_path)
right = str(child or "").strip().replace("\\", "/")
if not right:
return left
return self._normalize_remote_path(posixpath.join(left, right), default="/")
def _build_url(
self,
remote_path: Any,
*,
host: Optional[str] = None,
port: Optional[int] = None,
tls: Optional[bool] = None,
) -> str:
path_text = self._normalize_remote_path(remote_path, default="/")
scheme = "ftps" if (self._tls if tls is None else bool(tls)) else "ftp"
host_text = str(host or self._host).strip()
port_value = int(port or self._port)
port_suffix = f":{port_value}" if port_value and port_value != 21 else ""
return f"{scheme}://{host_text}{port_suffix}{quote(path_text, safe='/-._~!$&\'()*+,;=:@')}"
def _connection_settings_for_url(self, url: str) -> Dict[str, Any]:
parsed = urlparse(str(url or "").strip())
scheme = (parsed.scheme or "ftp").strip().lower()
host = parsed.hostname or self._host
port = parsed.port or self._port
username = parsed.username or self._username
password = parsed.password or self._password
path_text = self._normalize_remote_path(unquote(parsed.path or "/"), default="/")
return {
"tls": scheme == "ftps",
"host": host,
"port": port,
"username": username,
"password": password,
"path": path_text,
}
def _search_directory(
self,
ftp: ftplib.FTP,
start_path: str,
*,
needle: str,
limit: int,
search_depth: int,
type_filter: str,
) -> List[SearchResult]:
results: List[SearchResult] = []
visited: set[str] = set()
def walk(current_path: str, depth_left: int) -> None:
normalized = self._normalize_remote_path(current_path, default=self._base_path)
if normalized in visited or len(results) >= limit:
return
visited.add(normalized)
for entry in self._list_directory(ftp, normalized):
if len(results) >= limit:
return
if self._matches_entry(entry, needle=needle, type_filter=type_filter):
results.append(self._build_result(entry))
if entry.get("is_dir") and depth_left > 0:
walk(str(entry.get("ftp_path") or normalized), depth_left - 1)
walk(start_path, max(0, search_depth))
return results
def _matches_entry(self, entry: Dict[str, Any], *, needle: str, type_filter: str) -> bool:
is_dir = bool(entry.get("is_dir"))
if type_filter in {"dir", "dirs", "folder", "folders"} and not is_dir:
return False
if type_filter in {"file", "files"} and is_dir:
return False
text = str(needle or "").strip().lower()
if not text or text in {"*", "all", "list"}:
return True
haystacks = [
str(entry.get("name") or "").lower(),
str(entry.get("ftp_path") or "").lower(),
]
for token in [part for part in text.split() if part]:
if any(ch in token for ch in "*?[]"):
if not any(fnmatch.fnmatch(haystack, token) for haystack in haystacks):
return False
elif not any(token in haystack for haystack in haystacks):
return False
return True
def _build_result(self, entry: Dict[str, Any]) -> SearchResult:
ftp_path = str(entry.get("ftp_path") or "/")
ftp_url = self._build_url(ftp_path)
is_dir = bool(entry.get("is_dir"))
size_value = entry.get("size")
modified = str(entry.get("modified") or "")
parent = posixpath.dirname(ftp_path.rstrip("/")) or "/"
metadata = {
"provider": "ftp",
"host": self._host,
"ftp_path": ftp_path,
"ftp_url": ftp_url,
"selection_url": ftp_url,
"is_dir": is_dir,
"name": str(entry.get("name") or "").strip(),
}
if size_value is not None:
metadata["size"] = size_value
if modified:
metadata["modified"] = modified
return SearchResult(
table="ftp",
title=str(entry.get("name") or ftp_path),
path=ftp_url,
detail=parent,
annotations=["folder" if is_dir else "file"],
media_kind="folder" if is_dir else "file",
size_bytes=int(size_value) if isinstance(size_value, int) else None,
tag={"ftp", "folder" if is_dir else "file"},
columns=[
("Name", str(entry.get("name") or "")),
("Type", "dir" if is_dir else "file"),
("Directory", parent),
("Size", "" if size_value is None else str(size_value)),
("Modified", modified),
],
selection_args=None if is_dir else ["-url", ftp_url],
selection_action=None if is_dir else ["download-file", "-plugin", "ftp", "-url", ftp_url],
full_metadata=metadata,
)
def _list_directory(self, ftp: ftplib.FTP, remote_path: str) -> List[Dict[str, Any]]:
normalized = self._normalize_remote_path(remote_path, default=self._base_path)
try:
entries: List[Dict[str, Any]] = []
for name, facts in ftp.mlsd(normalized):
name_text = str(name or "").strip()
if not name_text or name_text in {".", ".."}:
continue
entry_type = str((facts or {}).get("type") or "").strip().lower()
if entry_type in {"cdir", "pdir"}:
continue
size_value = None
raw_size = (facts or {}).get("size")
if raw_size not in (None, ""):
try:
size_value = int(raw_size)
except Exception:
size_value = None
entries.append(
{
"name": name_text,
"ftp_path": self._join_remote_path(normalized, name_text),
"is_dir": entry_type == "dir",
"size": size_value,
"modified": _format_timestamp((facts or {}).get("modify")),
}
)
return entries
except Exception:
return self._list_directory_fallback(ftp, normalized)
def _list_directory_fallback(self, ftp: ftplib.FTP, remote_path: str) -> List[Dict[str, Any]]:
try:
listed = ftp.nlst(remote_path)
except Exception:
return []
entries: List[Dict[str, Any]] = []
seen: set[str] = set()
for raw_entry in listed:
entry_text = str(raw_entry or "").strip()
if not entry_text:
continue
entry_path = entry_text if entry_text.startswith("/") else self._join_remote_path(remote_path, entry_text)
name_text = posixpath.basename(entry_path.rstrip("/")) or entry_path.rstrip("/")
if not name_text or name_text in {".", ".."} or name_text in seen:
continue
seen.add(name_text)
is_dir = self._is_directory(ftp, entry_path)
size_value = None
if not is_dir:
try:
size_raw = ftp.size(entry_path)
if size_raw is not None:
size_value = int(size_raw)
except Exception:
size_value = None
entries.append(
{
"name": name_text,
"ftp_path": entry_path,
"is_dir": is_dir,
"size": size_value,
"modified": self._read_modified(ftp, entry_path),
}
)
return entries
def _is_directory(self, ftp: ftplib.FTP, remote_path: str) -> bool:
current = None
try:
current = ftp.pwd()
except Exception:
current = None
try:
ftp.cwd(remote_path)
return True
except Exception:
return False
finally:
if current is not None:
try:
ftp.cwd(current)
except Exception:
pass
def _read_modified(self, ftp: ftplib.FTP, remote_path: str) -> str:
try:
response = ftp.sendcmd(f"MDTM {remote_path}")
except Exception:
return ""
parts = str(response or "").split()
if len(parts) >= 2:
return _format_timestamp(parts[1])
return ""
def _ensure_directory(self, ftp: ftplib.FTP, remote_path: str) -> None:
normalized = self._normalize_remote_path(remote_path, default=self._base_path)
if normalized == "/":
return
partial = ""
for segment in [part for part in normalized.split("/") if part]:
partial = f"{partial}/{segment}"
if self._is_directory(ftp, partial):
continue
try:
ftp.mkd(partial)
except Exception:
if not self._is_directory(ftp, partial):
raise
def _item_metadata(self, item: Any, *, pipe_obj: Any = None) -> Dict[str, Any]:
metadata: Dict[str, Any] = {}
for source in (item, pipe_obj):
if isinstance(source, dict):
for key in ("title", "path", "url"):
if source.get(key) is not None and key not in metadata:
metadata[key] = source.get(key)
nested = source.get("full_metadata") or source.get("metadata")
if isinstance(nested, dict):
metadata.update(nested)
elif source is not None:
for attr in ("title", "path", "url"):
try:
value = getattr(source, attr, None)
except Exception:
value = None
if value is not None and attr not in metadata:
metadata[attr] = value
try:
nested = getattr(source, "full_metadata", None) or getattr(source, "metadata", None)
except Exception:
nested = None
if isinstance(nested, dict):
metadata.update(nested)
ftp_path = metadata.get("ftp_path") or metadata.get("selection_path")
if not ftp_path:
path_value = metadata.get("path") or metadata.get("url") or metadata.get("ftp_url")
path_text = str(path_value or "").strip()
if path_text.startswith(("ftp://", "ftps://")):
ftp_path = self._normalize_remote_path(path_text, default=self._base_path)
if ftp_path:
metadata["ftp_path"] = self._normalize_remote_path(ftp_path, default=self._base_path)
metadata.setdefault("selection_path", metadata["ftp_path"])
if metadata.get("ftp_path") and not metadata.get("ftp_url"):
metadata["ftp_url"] = self._build_url(metadata["ftp_path"])
if metadata.get("ftp_url") and not metadata.get("selection_url"):
metadata["selection_url"] = metadata["ftp_url"]
is_dir = metadata.get("is_dir")
if is_dir is None and metadata.get("media_kind"):
is_dir = str(metadata.get("media_kind") or "").strip().lower() == "folder"
metadata["is_dir"] = bool(is_dir)
return metadata