llama-deploy-appserver 0.3.23__tar.gz → 0.3.24__tar.gz

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: llama-deploy-appserver
-Version: 0.3.23
+Version: 0.3.24
 Summary: Application server components for LlamaDeploy
 Author: Massimiliano Pippi, Adrian Lyjak
 Author-email: Massimiliano Pippi <mpippi@gmail.com>, Adrian Lyjak <adrianlyjak@gmail.com>
@@ -9,7 +9,7 @@ Requires-Dist: llama-index-workflows[server]>=2.9.1
 Requires-Dist: pydantic-settings>=2.10.1
 Requires-Dist: fastapi>=0.100.0
 Requires-Dist: websockets>=12.0
-Requires-Dist: llama-deploy-core>=0.3.23,<0.4.0
+Requires-Dist: llama-deploy-core>=0.3.24,<0.4.0
 Requires-Dist: httpx>=0.24.0,<1.0.0
 Requires-Dist: prometheus-fastapi-instrumentator>=7.1.0
 Requires-Dist: packaging>=25.0
@@ -19,7 +19,8 @@ Requires-Dist: pyyaml>=6.0.2
 Requires-Dist: llama-cloud-services>=0.6.60
 Requires-Dist: watchfiles>=1.1.0
 Requires-Dist: uvicorn>=0.35.0
-Requires-Python: >=3.11, <4
+Requires-Dist: typing-extensions>=4.15.0 ; python_full_version < '3.12'
+Requires-Python: >=3.10, <4
 Description-Content-Type: text/markdown
 
 # llama-deploy-appserver

llama_deploy_appserver-0.3.24/pyproject.toml

@@ -0,0 +1,48 @@
+[build-system]
+requires = ["uv_build>=0.7.20,<0.8.0"]
+build-backend = "uv_build"
+
+[dependency-groups]
+dev = [
+  "pytest>=8.4.1",
+  "pytest-asyncio>=0.25.3",
+  "respx>=0.22.0",
+  "pytest-xdist>=3.8.0",
+  "ty>=0.0.1a19",
+  "ruff>=0.12.9"
+]
+
+[project]
+name = "llama-deploy-appserver"
+version = "0.3.24"
+description = "Application server components for LlamaDeploy"
+readme = "README.md"
+license = {text = "MIT"}
+authors = [
+  {name = "Massimiliano Pippi", email = "mpippi@gmail.com"},
+  {name = "Adrian Lyjak", email = "adrianlyjak@gmail.com"}
+]
+requires-python = ">=3.10, <4"
+dependencies = [
+  "llama-index-workflows[server]>=2.9.1",
+  "pydantic-settings>=2.10.1",
+  "fastapi>=0.100.0",
+  "websockets>=12.0",
+  "llama-deploy-core>=0.3.24,<0.4.0",
+  "httpx>=0.24.0,<1.0.0",
+  "prometheus-fastapi-instrumentator>=7.1.0",
+  "packaging>=25.0",
+  "structlog>=25.4.0",
+  "rich>=14.1.0",
+  "pyyaml>=6.0.2",
+  "llama-cloud-services>=0.6.60",
+  "watchfiles>=1.1.0",
+  "uvicorn>=0.35.0",
+  "typing-extensions>=4.15.0 ; python_full_version < '3.12'"
+]
+
+[tool.uv.build-backend]
+module-name = "llama_deploy.appserver"
+
+[tool.uv.sources]
+llama-deploy-core = {workspace = true}

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/app.py

@@ -122,7 +122,7 @@ def _setup_openapi(name: str, app: FastAPI, server: WorkflowServer) -> None:
 
     schema["paths"] = new_paths
 
-    def custom_openapi():
+    def custom_openapi() -> dict[str, object]:
         return schema
 
     app.openapi = custom_openapi  # ty: ignore[invalid-assignment] - doesn't like us overwriting the method
@@ -139,16 +139,22 @@ app = FastAPI(
 Instrumentator().instrument(app).expose(app, include_in_schema=False)
 
 
-# Configure CORS middleware if the environment variable is set
-if not os.environ.get("DISABLE_CORS", False):
+def _configure_cors(app: FastAPI) -> None:
+    """Attach CORS middleware in a way that keeps type-checkers happy."""
+    # Use a cast here because ty's view of Starlette's middleware factory
+    # protocol is stricter than FastAPI's runtime expectations.
     app.add_middleware(
-        CORSMiddleware,
+        cast(Any, CORSMiddleware),
         allow_origins=["*"],  # Allows all origins
         allow_credentials=True,
         allow_methods=["GET", "POST"],
         allow_headers=["Content-Type", "Authorization"],
     )
 
+
+if not os.environ.get("DISABLE_CORS", False):
+    _configure_cors(app)
+
 app.include_router(health_router)
 add_log_middleware(app)
 
@@ -280,16 +286,13 @@ def start_server_in_target_venv(
     if log_format:
         env["LOG_FORMAT"] = log_format
 
-    ret = run_process(
+    run_process(
         args,
         cwd=path,
         env=env,
         line_transform=_exclude_venv_warning,
     )
 
-    if ret != 0:
-        raise SystemExit(ret)
-
 
 def start_preflight_in_target_venv(
     cwd: Path | None = None,
@@ -297,7 +300,7 @@ def start_preflight_in_target_venv(
 ) -> None:
     """
     Run preflight validation inside the target project's virtual environment using uv.
-    Mirrors the venv targetting and invocation strategy used by start_server_in_target_venv.
+    Mirrors the venv targeting and invocation strategy used by start_server_in_target_venv.
     """
     configure_settings(
         app_root=cwd,
@@ -317,14 +320,13 @@ def start_preflight_in_target_venv(
     if deployment_file:
         args.extend(["--deployment-file", str(deployment_file)])
 
-    ret = run_process(
+    run_process(
         args,
         cwd=path,
         env=os.environ.copy(),
         line_transform=_exclude_venv_warning,
     )
-    if ret is not None and ret != 0:
-        raise SystemExit(ret)
+    # Note: run_process doesn't return exit code; process runs to completion or raises
 
 
 class PreflightValidationError(Exception):

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/bootstrap.py

@@ -27,7 +27,7 @@ from llama_deploy.core.git.git_util import (
 
 def bootstrap_app_from_repo(
     target_dir: str = "/opt/app",
-):
+) -> None:
     bootstrap_settings = BootstrapSettings()
     # Needs the github url+auth, and the deployment file path
     # clones the repo to a standard directory

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/configure_logging.py

@@ -22,18 +22,18 @@ def _get_or_create_correlation_id(request: Request) -> str:
     return request.headers.get("X-Request-ID", create_correlation_id())
 
 
-def add_log_middleware(app: FastAPI):
+def add_log_middleware(app: FastAPI) -> None:
     @app.middleware("http")
     async def add_log_id(
         request: Request, call_next: Callable[[Request], Awaitable[Response]]
-    ):
+    ) -> Response:
         set_correlation_id(_get_or_create_correlation_id(request))
         return await call_next(request)
 
     @app.middleware("http")
     async def access_log_middleware(
         request: Request, call_next: Callable[[Request], Awaitable[Response]]
-    ):
+    ) -> Response:
         if _is_proxy_request(request):
             return await call_next(request)
         start = time.perf_counter()

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/deployment.py

@@ -18,7 +18,7 @@ from starlette.responses import HTMLResponse
 from starlette.routing import Route
 from workflows import Context, Workflow
 from workflows.handler import WorkflowHandler
-from workflows.server import SqliteWorkflowStore, WorkflowServer
+from workflows.server import AbstractWorkflowStore, SqliteWorkflowStore, WorkflowServer
 from workflows.server.memory_workflow_store import MemoryWorkflowStore
 
 logger = logging.getLogger()
@@ -40,7 +40,7 @@ class Deployment:
             local: Whether the deployment is local. If true, sources won't be synced
         """
 
-        self._default_service: str | None = workflows.get(DEFAULT_SERVICE_ID)
+        self._default_service: Workflow | None = workflows.get(DEFAULT_SERVICE_ID)
         self._service_tasks: list[asyncio.Task] = []
         # Ready to load services
         self._workflow_services: dict[str, Workflow] = workflows
@@ -50,13 +50,9 @@ class Deployment:
 
     @property
     def default_service(self) -> Workflow | None:
+        """Return the default workflow, if any."""
         return self._default_service
 
-    @property
-    def name(self) -> str:
-        """Returns the name of this deployment."""
-        return self._name
-
     @property
     def service_names(self) -> list[str]:
         """Returns the list of service names in this deployment."""
@@ -66,7 +62,7 @@ class Deployment:
         self, service_id: str, session_id: str | None = None, **run_kwargs: dict
     ) -> Any:
         workflow = self._workflow_services[service_id]
-        if session_id:
+        if session_id is not None:
             context = self._contexts[session_id]
             return await workflow.run(context=context, **run_kwargs)
 
@@ -79,7 +75,7 @@ class Deployment:
         self, service_id: str, session_id: str | None = None, **run_kwargs: dict
     ) -> Tuple[str, str]:
         workflow = self._workflow_services[service_id]
-        if session_id:
+        if session_id is not None:
             context = self._contexts[session_id]
             handler = workflow.run(context=context, **run_kwargs)
         else:
@@ -90,12 +86,13 @@ class Deployment:
         handler_id = generate_id()
         self._handlers[handler_id] = handler
         self._handler_inputs[handler_id] = json.dumps(run_kwargs)
+        assert session_id is not None
         return handler_id, session_id
 
     def create_workflow_server(
         self, deployment_config: DeploymentConfig, settings: ApiserverSettings
     ) -> WorkflowServer:
-        persistence = MemoryWorkflowStore()
+        persistence: AbstractWorkflowStore = MemoryWorkflowStore()
         if settings.persistence == "local":
             logger.info("Using local sqlite persistence for workflows")
             persistence = SqliteWorkflowStore(
@@ -137,8 +134,8 @@ class Deployment:
                 "/debugger/index.html?api=" + quote_plus("/deployments/" + config.name)
             )
 
-        @app.get("/debugger/index.html", include_in_schema=False)
-        def serve_debugger(api: str | None = None):
+        @app.get("/debugger/index.html", include_in_schema=False, response_model=None)
+        def serve_debugger(api: str | None = None) -> RedirectResponse | HTMLResponse:
             if not api:
                 return RedirectResponse(
                     "/debugger/index.html?api="

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/interrupts.py

@@ -2,7 +2,7 @@ import asyncio
 import signal
 from asyncio import Event
 from contextlib import suppress
-from typing import Awaitable, TypeVar
+from typing import Any, Coroutine, TypeVar
 
 shutdown_event = Event()
 
@@ -21,7 +21,7 @@ T = TypeVar("T")
 
 
 async def wait_or_abort(
-    awaitable: Awaitable[T], shutdown_event: asyncio.Event = shutdown_event
+    awaitable: Coroutine[Any, Any, T], shutdown_event: asyncio.Event = shutdown_event
 ) -> T:
     """Await an operation, aborting early if shutdown is requested.
 
@@ -32,7 +32,7 @@ async def wait_or_abort(
     if event.is_set():
         raise OperationAborted()
 
-    op_task = asyncio.create_task(awaitable)
+    op_task: asyncio.Task[T] = asyncio.create_task(awaitable)
     stop_task = asyncio.create_task(event.wait())
     try:
         done, _ = await asyncio.wait(

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/process_utils.py

@@ -112,7 +112,7 @@ def should_use_color() -> bool:
 
 @dataclass
 class SpawnProcessResult:
-    process: subprocess.Popen
+    process: subprocess.Popen[str] | subprocess.Popen[bytes]
     sources: list[Tuple[int | TextIO, TextIO]]
     cleanup: Callable[[], None]
 
@@ -124,6 +124,7 @@ def _spawn_process(
     env: dict[str, str] | None,
     use_pty: bool,
 ) -> SpawnProcessResult:
+    process: subprocess.Popen[str] | subprocess.Popen[bytes]
     if use_pty:
         import pty
 
@@ -164,7 +165,7 @@ def _spawn_process(
         shell=use_shell,
     )
 
-    def cleanup() -> None:
+    def cleanup_non_pty() -> None:
         return None
 
     assert process.stdout is not None and process.stderr is not None
@@ -172,7 +173,7 @@ def _spawn_process(
         (cast(int | TextIO, process.stdout), cast(TextIO, sys.stdout)),
         (cast(int | TextIO, process.stderr), cast(TextIO, sys.stderr)),
     ]
-    return SpawnProcessResult(process, sources, cleanup)
+    return SpawnProcessResult(process, sources, cleanup_non_pty)
 
 
 def _stream_source(
@@ -212,7 +213,9 @@ def _stream_source(
 def _log_command(cmd: list[str], transform: Callable[[str], str | None] | None) -> None:
     cmd_str = "> " + " ".join(cmd)
     if transform:
-        cmd_str = transform(cmd_str)
+        transformed = transform(cmd_str)
+        if transformed is not None:
+            cmd_str = transformed
     sys.stderr.write(cmd_str + "\n")
 
 

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/routers/ui_proxy.py

@@ -1,7 +1,7 @@
 import asyncio
 import logging
+from collections.abc import AsyncGenerator, Sequence
 from contextlib import suppress
-from typing import List
 
 import httpx
 import websockets
@@ -23,6 +23,7 @@ from llama_deploy.appserver.interrupts import (
 from llama_deploy.appserver.settings import ApiserverSettings
 from llama_deploy.core.client.ssl_util import get_httpx_verify_param
 from llama_deploy.core.deployment_config import DeploymentConfig
+from websockets.typing import Subprotocol
 
 logger = logging.getLogger(__name__)
 
@@ -53,11 +54,12 @@ async def _ws_proxy(ws: WebSocket, upstream_url: str) -> None:
 
     try:
         # Parse subprotocols if present
-        subprotocols: List[str] | None = None
+        subprotocols: Sequence[Subprotocol] | None = None
         requested = ws.headers.get("sec-websocket-protocol")
         if requested:
             # Parse comma-separated subprotocols (as plain strings)
-            subprotocols = [p.strip() for p in requested.split(",")]
+            parsed = [p.strip() for p in requested.split(",")]
+            subprotocols = [Subprotocol(p) for p in parsed if p]
 
         # Open upstream WebSocket connection, offering the same subprotocols
         async with websockets.connect(
@@ -210,7 +212,7 @@ def create_ui_proxy_router(name: str, port: int) -> APIRouter:
             }
 
             # Stream downloads and ensure cleanup in the generator's finally block
-            async def upstream_body():
+            async def upstream_body() -> AsyncGenerator[bytes, None]:
                 try:
                     async for chunk in upstream.aiter_raw():
                         yield chunk
@@ -240,9 +242,10 @@ def create_ui_proxy_router(name: str, port: int) -> APIRouter:
 def mount_static_files(
     app: FastAPI, config: DeploymentConfig, settings: ApiserverSettings
 ) -> None:
-    path = settings.app_root / config.build_output_path()
-    if not path:
+    build_output = config.build_output_path()
+    if build_output is None:
         return
+    path = settings.app_root / build_output
 
     if not path.exists():
         return

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/workflow_loader.py

@@ -106,7 +106,8 @@ def parse_environment_variables(
     for env_file in config.env_files or []:
         env_file_path = source_root / env_file
         values = dotenv_values(env_file_path)
-        env_vars.update(**values)
+        str_values = {k: v for k, v in values.items() if isinstance(v, str)}
+        env_vars.update(str_values)
     return env_vars
 
 
@@ -206,7 +207,9 @@ def _install_and_add_appserver_if_missing(
         )
         return
 
-    def run_uv(cmd: str, args: list[str] = [], extra_env: dict[str, str] | None = None):
+    def run_uv(
+        cmd: str, args: list[str] = [], extra_env: dict[str, str] | None = None
+    ) -> None:
         env = os.environ.copy()
         if extra_env:
             env.update(extra_env)

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/workflow_store/agent_data_store.py

@@ -1,19 +1,24 @@
 import asyncio
 import logging
 import os
-from typing import Any, List
+import sys
+from typing import Any, List, cast
 
 from llama_cloud.client import AsyncLlamaCloud, httpx
 from llama_cloud_services.beta.agent_data import AsyncAgentDataClient
 from llama_deploy.appserver.settings import ApiserverSettings
 from llama_deploy.core.client.ssl_util import get_httpx_verify_param
 from llama_deploy.core.deployment_config import DeploymentConfig
-from typing_extensions import override
 from workflows.server import AbstractWorkflowStore, HandlerQuery, PersistentHandler
 
 from .keyed_lock import AsyncKeyedLock
 from .lru_cache import LRUCache
 
+if sys.version_info <= (3, 11):
+    from typing_extensions import override
+else:
+    from typing import override
+
 logger = logging.getLogger(__name__)
 
 
@@ -77,8 +82,8 @@ class AgentDataStore(AbstractWorkflowStore):
                 )
 
     @override
-    async def delete(self, handler: HandlerQuery) -> int:
-        filters = self._build_filters(handler)
+    async def delete(self, query: HandlerQuery) -> int:
+        filters = self._build_filters(query)
         results = await self.client.search(filter=filters, page_size=1000)
         await asyncio.gather(
             *[self.client.delete_item(item_id=x.id) for x in results.items if x.id]
@@ -89,18 +94,21 @@ class AgentDataStore(AbstractWorkflowStore):
         cached_id = self.cache.get(handler.handler_id)
         if cached_id is not None:
             return cached_id
+        search_filter = {"handler_id": {"eq": handler.handler_id}}
         results = await self.client.search(
-            filter={"handler_id": {"eq": handler.handler_id}},
+            filter=cast(Any, search_filter),
             page_size=1,
         )
         if not results.items:
             return None
         id = results.items[0].id
+        if id is None:
+            return None
         self.cache.set(handler.handler_id, id)
         return id
 
     def _build_filters(self, query: HandlerQuery) -> dict[str, Any]:
-        filters = {}
+        filters: dict[str, Any] = {}
         if query.handler_id_in is not None:
             filters["handler_id"] = {
                 "includes": query.handler_id_in,

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/workflow_store/keyed_lock.py

@@ -1,16 +1,17 @@
 import asyncio
 from collections import Counter
+from collections.abc import AsyncIterator
 from contextlib import asynccontextmanager
 
 
 class AsyncKeyedLock:
-    def __init__(self):
+    def __init__(self) -> None:
         self._locks: dict[str, asyncio.Lock] = {}
-        self._refcnt = Counter()
+        self._refcnt: Counter[str] = Counter()
         self._registry_lock = asyncio.Lock()  # protects _locks/_refcnt
 
     @asynccontextmanager
-    async def acquire(self, key: str):
+    async def acquire(self, key: str) -> AsyncIterator[None]:
         async with self._registry_lock:
             lock = self._locks.get(key)
             if lock is None:

{llama_deploy_appserver-0.3.23 → llama_deploy_appserver-0.3.24}/src/llama_deploy/appserver/workflow_store/lru_cache.py

@@ -1,4 +1,5 @@
 from collections import OrderedDict
+from collections.abc import Iterator
 from typing import Generic, TypeVar, overload
 
 K = TypeVar("K")
@@ -21,7 +22,7 @@ class LRUCache(Generic[K, V]):
             return default
         return self[key]
 
-    def set(self, key: K, value: V):
+    def set(self, key: K, value: V) -> None:
         if key in self._store:
             # remove old so we can push to end
             self._store.pop(key)
@@ -41,11 +42,11 @@ class LRUCache(Generic[K, V]):
         self._store[key] = value
         return value
 
-    def __setitem__(self, key: K, value: V):
+    def __setitem__(self, key: K, value: V) -> None:
         self.set(key, value)
 
     def __len__(self) -> int:
         return len(self._store)
 
-    def __iter__(self):
+    def __iter__(self) -> Iterator[K]:
         return iter(self._store)

llama_deploy_appserver-0.3.23/pyproject.toml

@@ -1,48 +0,0 @@
-[project]
-name = "llama-deploy-appserver"
-version = "0.3.23"
-description = "Application server components for LlamaDeploy"
-readme = "README.md"
-license = { text = "MIT" }
-authors = [
-    { name = "Massimiliano Pippi", email = "mpippi@gmail.com" },
-    { name = "Adrian Lyjak", email = "adrianlyjak@gmail.com" },
-]
-requires-python = ">=3.11, <4"
-dependencies = [
-    "llama-index-workflows[server]>=2.9.1",
-    "pydantic-settings>=2.10.1",
-    "fastapi>=0.100.0",
-    "websockets>=12.0",
-    "llama-deploy-core>=0.3.23,<0.4.0",
-    "httpx>=0.24.0,<1.0.0",
-    "prometheus-fastapi-instrumentator>=7.1.0",
-    "packaging>=25.0",
-    "structlog>=25.4.0",
-    "rich>=14.1.0",
-    "pyyaml>=6.0.2",
-    "llama-cloud-services>=0.6.60",
-    "watchfiles>=1.1.0",
-    "uvicorn>=0.35.0",
-]
-
-[build-system]
-requires = ["uv_build>=0.7.20,<0.8.0"]
-build-backend = "uv_build"
-
-[tool.uv.build-backend]
-module-name = "llama_deploy.appserver"
-
-
-[tool.uv.sources]
-llama-deploy-core = { workspace = true }
-
-[dependency-groups]
-dev = [
-    "pytest>=8.4.1",
-    "pytest-asyncio>=0.25.3",
-    "respx>=0.22.0",
-    "pytest-xdist>=3.8.0",
-    "ty>=0.0.1a19",
-    "ruff>=0.12.9",
-]