Spaces:
Running
Running
File size: 7,060 Bytes
6918275 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 | """Centralized document fetching for the document_query plugin."""
from __future__ import annotations
import asyncio
import mimetypes
import os
import tempfile
from contextlib import contextmanager
from dataclasses import dataclass
from pathlib import Path
from typing import Awaitable, Callable
from urllib.parse import urlparse
from helpers import files, tinyfish
from helpers.network import UnsafeUrlError, fetch_public_http_resource
InterventionCallback = Callable[[], Awaitable[None]]
@dataclass(frozen=True)
class FetchedDocument:
"""Fetched document bytes plus metadata needed by parsers."""
uri: str
scheme: str
mimetype: str
content: bytes
encoding: str | None = None
charset: str | None = None
local_path: str | None = None
source_uri: str | None = None
def text(self) -> str:
charset = self.charset or "utf-8"
return self.content.decode(charset, errors="replace")
def suffix(self) -> str:
path = self.local_path or urlparse(self.uri).path or self.uri
suffix = Path(path).suffix
if suffix:
return suffix
guessed = mimetypes.guess_extension(self.mimetype)
return guessed or ".bin"
@contextmanager
def local_file(self):
"""Yield a filesystem path for parsers that cannot consume bytes."""
if self.local_path and os.path.exists(self.local_path):
yield self.local_path
return
tmp = ""
try:
with tempfile.NamedTemporaryFile(delete=False, suffix=self.suffix()) as f:
f.write(self.content)
tmp = f.name
yield tmp
finally:
if tmp and os.path.exists(tmp):
os.unlink(tmp)
ProtocolHandler = Callable[
[str, str, dict, InterventionCallback | None], Awaitable[FetchedDocument]
]
_PROTOCOL_HANDLERS: dict[str, ProtocolHandler] = {}
def register_protocol_handler(scheme: str, handler: ProtocolHandler) -> None:
"""Register or replace a fetch handler for a URI scheme."""
_PROTOCOL_HANDLERS[scheme.lower()] = handler
async def fetch_public_resource(
uri: str,
config: dict | None = None,
intervention_callback: InterventionCallback | None = None,
) -> FetchedDocument:
"""Fetch local or remote content once, then pass bytes to parsers."""
config = config or {}
parsed = urlparse(uri)
scheme = (parsed.scheme or "file").lower()
handler = _PROTOCOL_HANDLERS.get(scheme)
if not handler:
raise ValueError(f"Unsupported document scheme: {scheme}")
return await handler(uri, scheme, config, intervention_callback)
async def _fetch_file(
uri: str,
scheme: str,
config: dict,
intervention_callback: InterventionCallback | None,
) -> FetchedDocument:
parsed = urlparse(uri)
raw_path = parsed.path if parsed.scheme == "file" else uri
if not raw_path:
raise ValueError(f"Invalid document path: {uri}")
path = _fix_file_path(raw_path)
mimetype, encoding = mimetypes.guess_type(path)
if encoding:
raise ValueError(f"Compressed documents are unsupported '{encoding}' ({uri})")
mimetype = mimetype or "application/octet-stream"
if mimetype == "application/octet-stream":
raise ValueError(f"Unsupported document mimetype '{mimetype}' ({uri})")
if intervention_callback:
await intervention_callback()
return FetchedDocument(
uri=path,
source_uri=uri,
scheme=scheme,
mimetype=mimetype,
encoding=encoding,
content=files.read_file_bin(path),
local_path=path,
)
async def _fetch_http(
uri: str,
scheme: str,
config: dict,
intervention_callback: InterventionCallback | None,
) -> FetchedDocument:
timeout = float(config.get("fetch_timeout", 30))
retries = max(1, int(config.get("fetch_retries", 3)))
retry_backoff = float(config.get("fetch_retry_backoff", 1.0))
max_remote_bytes = int(config.get("max_remote_bytes", 50 * 1024 * 1024))
parsed = urlparse(uri)
guessed_mimetype, encoding = mimetypes.guess_type(parsed.path or uri)
if encoding:
raise ValueError(f"Compressed documents are unsupported '{encoding}' ({uri})")
last_error = ""
for attempt in range(retries):
if intervention_callback:
await intervention_callback()
try:
resource = await asyncio.to_thread(
fetch_public_http_resource,
uri,
max_bytes=max_remote_bytes,
timeout=(timeout, timeout),
)
mimetype = (
resource.content_type or guessed_mimetype or "application/octet-stream"
)
if mimetype == "application/octet-stream":
raise ValueError(f"Unsupported document mimetype '{mimetype}' ({uri})")
document = FetchedDocument(
uri=resource.url,
source_uri=uri,
scheme=urlparse(resource.url).scheme or scheme,
mimetype=mimetype,
encoding=encoding,
charset=resource.encoding,
content=resource.content,
)
except UnsafeUrlError:
# Neither a remote provider nor a retry may bypass SSRF checks.
raise
except Exception as e:
if "exceeds max size" in str(e):
raise
if scheme == "https" and tinyfish.configured():
if intervention_callback:
await intervention_callback()
try:
final_url, content = await tinyfish.fetch(
uri, timeout=timeout, max_bytes=max_remote_bytes
)
except tinyfish.TinyfishError as error:
raise ValueError(
f"Document fetch unavailable: local fetch failed; Tinyfish {error}"
) from None
if intervention_callback:
await intervention_callback()
return FetchedDocument(
uri=final_url,
source_uri=uri,
scheme="https",
mimetype="text/markdown",
charset="utf-8",
content=content,
)
last_error = str(e)
if attempt < retries - 1:
await asyncio.sleep(retry_backoff)
if intervention_callback:
await intervention_callback()
continue
if intervention_callback:
await intervention_callback()
return document
raise ValueError(f"Document fetch error: {uri} ({last_error})")
register_protocol_handler("file", _fetch_file)
register_protocol_handler("http", _fetch_http)
register_protocol_handler("https", _fetch_http)
def _fix_file_path(path: str) -> str:
if os.path.isabs(path) and os.path.exists(path):
return path
return files.fix_dev_path(path)
|