chore: update something

docsifer/analytics/service.py

@@ -165,9 +165,7 @@ class AnalyticsService:
         backoff = 1.0
         while not self._stopped.is_set():
             try:
-                await asyncio.wait_for(
-                    self._stopped.wait(), timeout=self._sync_interval
-                )
+                await asyncio.wait_for(self._stopped.wait(), timeout=self._sync_interval)
                 return  # stopped
             except asyncio.TimeoutError:
                 pass

docsifer/analytics/store.py

@@ -96,9 +96,7 @@ class UpstashStore:
                 keys = scan[1] or []
                 for key in keys:
                     period = key.split(":", 2)[-1]
-                    data = await loop.run_in_executor(
-                        None, partial(self._client.hgetall, key)
-                    )
+                    data = await loop.run_in_executor(None, partial(self._client.hgetall, key))
                     for label, count in (data or {}).items():
                         try:
                             result[metric][period][label] = int(count)
@@ -123,6 +121,7 @@ class UpstashStore:
         # Prefer pipeline when available (single HTTP round trip)
         pipeline = getattr(self._client, "pipeline", None)
         if callable(pipeline):
+
             def _run() -> None:
                 pipe = pipeline()
                 for key, label, count in ops:
@@ -135,9 +134,7 @@ class UpstashStore:
         # Fallback: parallel single-op calls
         await asyncio.gather(
             *(
-                loop.run_in_executor(
-                    None, partial(self._client.hincrby, key, label, count)
-                )
+                loop.run_in_executor(None, partial(self._client.hincrby, key, label, count))
                 for key, label, count in ops
             )
         )
@@ -166,7 +163,7 @@ class logger_suppress:  # noqa: N801 - tiny utility, lower-case name for context
     def __init__(self, message: str) -> None:
         self._message = message
 
-    def __enter__(self) -> "logger_suppress":
+    def __enter__(self) -> logger_suppress:
         return self
 
     def __exit__(self, exc_type, exc, tb) -> bool:  # type: ignore[no-untyped-def]
@@ -174,7 +171,7 @@ class logger_suppress:  # noqa: N801 - tiny utility, lower-case name for context
             logger.warning("%s: %s", self._message, exc)
         return True
 
-    async def __aenter__(self) -> "logger_suppress":
+    async def __aenter__(self) -> logger_suppress:
         return self
 
     async def __aexit__(self, exc_type, exc, tb) -> bool:  # type: ignore[no-untyped-def]

docsifer/api/error_handlers.py

@@ -59,9 +59,7 @@ async def _http_exception_handler(request: Request, exc: HTTPException) -> JSONR
     )
 
 
-async def _validation_handler(
-    request: Request, exc: RequestValidationError
-) -> JSONResponse:
+async def _validation_handler(request: Request, exc: RequestValidationError) -> JSONResponse:
     return JSONResponse(
         status_code=422,
         content=_payload(

docsifer/api/middleware.py

@@ -3,7 +3,7 @@
 from __future__ import annotations
 
 import secrets
-from typing import Awaitable, Callable
+from collections.abc import Awaitable, Callable
 
 from fastapi import FastAPI, Request, Response
 from fastapi.responses import JSONResponse

docsifer/api/v1/convert.py

@@ -147,60 +147,57 @@ async def convert_document(
     http_cfg = _parse_json_form("http", http, HTTPConfig)
     convert_cfg = _parse_json_form("settings", settings_form, ConvertSettings)
 
-    async with per_ip.acquire(client_ip):
-        async with gate.acquire():
-            if file is not None:
-                guard.check(0)  # file size unknown until streamed
-                _check_extension(file.filename or "", set(settings.allowed_extensions))
-
-                tmp_root = Path(tempfile.mkdtemp(prefix="docsifer-", dir=settings.tmp_dir))
-                try:
-                    safe_name = _safe_filename(file.filename)
-                    dst = tmp_root / safe_name
-                    size = await _stream_to_disk(
-                        file, dst, max_bytes=settings.max_upload_bytes
-                    )
-                    logger.info(
-                        "Convert file received",
-                        extra={
-                            "filename": safe_name,
-                            "size": size,
-                            "client_ip": client_ip,
-                        },
-                    )
-                    guard.check(size)
-
-                    result = await asyncio.wait_for(
-                        converter.convert_file(
-                            dst,
-                            openai_config=openai_cfg.to_dict(),
-                            http_config=http_cfg.to_dict(),
-                            cleanup_html=convert_cfg.cleanup,
-                        ),
-                        timeout=settings.request_timeout_sec,
-                    )
-                finally:
-                    shutil.rmtree(tmp_root, ignore_errors=True)
-            else:
-                safe_url = validate_url(
-                    url or "",
-                    allowed_schemes=settings.url_allowed_schemes,
-                    allow_private_networks=settings.url_allow_private_networks,
-                )
+    async with per_ip.acquire(client_ip), gate.acquire():
+        if file is not None:
+            guard.check(0)  # file size unknown until streamed
+            _check_extension(file.filename or "", set(settings.allowed_extensions))
+
+            tmp_root = Path(tempfile.mkdtemp(prefix="docsifer-", dir=settings.tmp_dir))
+            try:
+                safe_name = _safe_filename(file.filename)
+                dst = tmp_root / safe_name
+                size = await _stream_to_disk(file, dst, max_bytes=settings.max_upload_bytes)
                 logger.info(
-                    "Convert URL received",
-                    extra={"url": safe_url, "client_ip": client_ip},
+                    "Convert file received",
+                    extra={
+                        "filename": safe_name,
+                        "size": size,
+                        "client_ip": client_ip,
+                    },
                 )
-                guard.check(0)
+                guard.check(size)
+
                 result = await asyncio.wait_for(
                     converter.convert_file(
-                        safe_url,
+                        dst,
                         openai_config=openai_cfg.to_dict(),
                         http_config=http_cfg.to_dict(),
                         cleanup_html=convert_cfg.cleanup,
                     ),
                     timeout=settings.request_timeout_sec,
                 )
+            finally:
+                shutil.rmtree(tmp_root, ignore_errors=True)
+        else:
+            safe_url = validate_url(
+                url or "",
+                allowed_schemes=settings.url_allowed_schemes,
+                allow_private_networks=settings.url_allow_private_networks,
+            )
+            logger.info(
+                "Convert URL received",
+                extra={"url": safe_url, "client_ip": client_ip},
+            )
+            guard.check(0)
+            result = await asyncio.wait_for(
+                converter.convert_file(
+                    safe_url,
+                    openai_config=openai_cfg.to_dict(),
+                    http_config=http_cfg.to_dict(),
+                    cleanup_html=convert_cfg.cleanup,
+                ),
+                timeout=settings.request_timeout_sec,
+            )
 
     background_tasks.add_task(_record_access_safe, analytics, result.token_count)
     return ConvertResponse(filename=result.filename, markdown=result.markdown)

docsifer/config.py

@@ -133,10 +133,7 @@ class Settings(BaseSettings):
         """True when a real (non-empty, non-localhost) Redis URL is configured."""
         if not self.analytics_enabled:
             return False
-        url = (self.redis_url or "").strip()
-        if not url:
-            return False
-        return True
+        return bool((self.redis_url or "").strip())
 
     # ---------------------------------------------------------------------
     # Quotas (anonymous, BYOK, authenticated)

docsifer/core/llm_registry.py

@@ -44,7 +44,7 @@ class LLMConfig:
         *,
         default_base_url: str,
         default_model: str,
-    ) -> "LLMConfig | None":
+    ) -> LLMConfig | None:
         if not data:
             return None
         api_key = (data.get("api_key") or "").strip()
@@ -92,9 +92,7 @@ class LLMRegistry:
         client = OpenAI(
             api_key=config.api_key,
             base_url=config.base_url,
-            timeout=httpx.Timeout(
-                self._request_timeout, connect=self._connect_timeout
-            ),
+            timeout=httpx.Timeout(self._request_timeout, connect=self._connect_timeout),
             max_retries=self._max_retries,
         )
         return MarkItDown(llm_client=client, llm_model=config.model)

docsifer/core/service.py

@@ -107,9 +107,7 @@ class DocsiferService:
     async def shutdown(self) -> None:
         """Stop the worker pool gracefully (called from app lifespan)."""
         self._llm_registry.clear()
-        await asyncio.get_running_loop().run_in_executor(
-            None, self._executor.shutdown, True
-        )
+        await asyncio.get_running_loop().run_in_executor(None, self._executor.shutdown, True)
 
     # ------------------------------------------------------------------
     # Internal

docsifer/main.py

@@ -18,8 +18,8 @@ from __future__ import annotations
 import asyncio
 import contextlib
 import logging
+from collections.abc import AsyncIterator
 from contextlib import asynccontextmanager
-from typing import AsyncIterator
 
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
@@ -54,9 +54,7 @@ def _build_analytics_store(settings: Settings) -> AnalyticsStore:
         logger.info("Analytics: Upstash store configured")
         return store
     except Exception as exc:  # pragma: no cover - defensive
-        logger.warning(
-            "Could not initialize Upstash store (%s); falling back to in-memory", exc
-        )
+        logger.warning("Could not initialize Upstash store (%s); falling back to in-memory", exc)
         return InMemoryStore()
 
 

docsifer/safety/circuit_breaker.py

@@ -4,7 +4,8 @@ from __future__ import annotations
 
 import asyncio
 import time
-from typing import Awaitable, Callable, TypeVar
+from collections.abc import Awaitable, Callable
+from typing import TypeVar
 
 from ..exceptions import CircuitOpenError
 

docsifer/safety/conversion_gate.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 import asyncio
 import contextlib
-from typing import AsyncIterator
+from collections.abc import AsyncIterator
 
 from ..exceptions import QueueFullError
 

docsifer/safety/disk_cleanup.py

@@ -32,17 +32,22 @@ async def disk_cleanup_loop(
             pass
 
         try:
-            now = time.time()
-            removed = 0
-            for entry in tmp_dir.glob(pattern):
-                try:
-                    if now - entry.stat().st_mtime > ttl_sec:
-                        if entry.is_file():
-                            entry.unlink(missing_ok=True)
-                            removed += 1
-                except OSError as exc:
-                    logger.debug("Could not remove %s: %s", entry, exc)
+            removed = await asyncio.to_thread(_sweep_once, tmp_dir, pattern, ttl_sec)
             if removed:
                 logger.info("Disk cleanup removed %d stale file(s)", removed)
         except Exception as exc:
             logger.warning("Disk cleanup failed: %s", exc)
+
+
+def _sweep_once(tmp_dir: Path, pattern: str, ttl_sec: int) -> int:
+    """Synchronous sweep — meant to be run via ``asyncio.to_thread``."""
+    now = time.time()
+    removed = 0
+    for entry in tmp_dir.glob(pattern):
+        try:
+            if now - entry.stat().st_mtime > ttl_sec and entry.is_file():
+                entry.unlink(missing_ok=True)
+                removed += 1
+        except OSError as exc:
+            logger.debug("Could not remove %s: %s", entry, exc)
+    return removed

docsifer/safety/per_ip_limiter.py

@@ -7,7 +7,7 @@ from __future__ import annotations
 import asyncio
 import contextlib
 import time
-from typing import AsyncIterator
+from collections.abc import AsyncIterator
 
 from ..exceptions import TooManyRequestsError
 

docsifer/ui/gradio_app.py

@@ -136,7 +136,7 @@ def _status(text: str, *, level: str = "info") -> str:
     color = palette.get(level, palette["info"])
     return (
         f'<div style="padding:10px 14px;border-radius:10px;'
-        f'background:{color}14;color:{color};font-size:0.9rem;'
+        f"background:{color}14;color:{color};font-size:0.9rem;"
         f'border:1px solid {color}30;">{text}</div>'
     )
 
@@ -209,13 +209,11 @@ def build_interface(settings: Settings, app: FastAPI) -> gr.Blocks:
                 return "", None, _status(f"Conversion failed: {exc}", level="err")
 
             asyncio.create_task(_record_access(result.token_count))
-            md_path = _write_markdown_file(
-                result.markdown, tmp_dir=settings.tmp_dir, stem=stem
-            )
+            md_path = _write_markdown_file(result.markdown, tmp_dir=settings.tmp_dir, stem=stem)
             byte_len = len(result.markdown.encode("utf-8"))
             ok_msg = (
                 f"Done — {len(result.markdown):,} chars · "
-                f"{byte_len/1024:.1f} KB · ~{result.token_count:,} tokens"
+                f"{byte_len / 1024:.1f} KB · ~{result.token_count:,} tokens"
             )
             return result.markdown, md_path, _status(ok_msg, level="ok")
         finally:

tests/integration/test_convert.py

@@ -1,4 +1,3 @@
-import io
 import json
 
 import pytest

tests/unit/test_html_cleaner.py

@@ -2,8 +2,10 @@ from docsifer.core.html_cleaner import clean_html
 
 
 def test_removes_style_script_noscript() -> None:
-    html = "<html><head><style>body{}</style><script>x()</script></head>" \
-           "<body><noscript>nope</noscript><p>hi</p></body></html>"
+    html = (
+        "<html><head><style>body{}</style><script>x()</script></head>"
+        "<body><noscript>nope</noscript><p>hi</p></body></html>"
+    )
     out = clean_html(html)
     assert "<style" not in out.lower()
     assert "<script" not in out.lower()
@@ -13,11 +15,11 @@ def test_removes_style_script_noscript() -> None:
 
 def test_removes_hidden_attribute_and_inline_styles() -> None:
     html = (
-        '<div hidden>secret</div>'
+        "<div hidden>secret</div>"
         '<div style="display:none">also-hidden</div>'
         '<div style="display: none;">spaced</div>'
         '<div aria-hidden="true">aria</div>'
-        '<p>visible</p>'
+        "<p>visible</p>"
     )
     out = clean_html(html)
     assert "secret" not in out