llama-deploy-appserver 0.3.16__tar.gz → 0.3.18__tar.gz

{llama_deploy_appserver-0.3.16 → llama_deploy_appserver-0.3.18}/PKG-INFO

@@ -1,15 +1,15 @@
 Metadata-Version: 2.3
 Name: llama-deploy-appserver
-Version: 0.3.16
+Version: 0.3.18
 Summary: Application server components for LlamaDeploy
 Author: Massimiliano Pippi, Adrian Lyjak
 Author-email: Massimiliano Pippi <mpippi@gmail.com>, Adrian Lyjak <adrianlyjak@gmail.com>
 License: MIT
-Requires-Dist: llama-index-workflows[server]>=2.6.0
+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>=0.3.16,<0.4.0
+Requires-Dist: llama-deploy-core>=0.3.18,<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

{llama_deploy_appserver-0.3.16 → llama_deploy_appserver-0.3.18}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "llama-deploy-appserver"
-version = "0.3.16"
+version = "0.3.18"
 description = "Application server components for LlamaDeploy"
 readme = "README.md"
 license = { text = "MIT" }
@@ -10,11 +10,11 @@ authors = [
 ]
 requires-python = ">=3.11, <4"
 dependencies = [
-    "llama-index-workflows[server]>=2.6.0",
+    "llama-index-workflows[server]>=2.9.1",
     "pydantic-settings>=2.10.1",
     "fastapi>=0.100.0",
     "websockets>=12.0",
-    "llama-deploy-core>=0.3.16,<0.4.0",
+    "llama-deploy-core>=0.3.18,<0.4.0",
     "httpx>=0.24.0,<1.0.0",
     "prometheus-fastapi-instrumentator>=7.1.0",
     "packaging>=25.0",

{llama_deploy_appserver-0.3.16 → llama_deploy_appserver-0.3.18}/src/llama_deploy/appserver/deployment.py

@@ -19,7 +19,7 @@ from starlette.routing import Route
 from workflows import Context, Workflow
 from workflows.handler import WorkflowHandler
 from workflows.server import SqliteWorkflowStore, WorkflowServer
-from workflows.server.abstract_workflow_store import EmptyWorkflowStore
+from workflows.server.memory_workflow_store import MemoryWorkflowStore
 
 logger = logging.getLogger()
 
@@ -95,7 +95,7 @@ class Deployment:
     def create_workflow_server(
         self, deployment_config: DeploymentConfig, settings: ApiserverSettings
     ) -> WorkflowServer:
-        persistence = EmptyWorkflowStore()
+        persistence = MemoryWorkflowStore()
         if settings.persistence == "local":
             logger.info("Using local sqlite persistence for workflows")
             persistence = SqliteWorkflowStore(
@@ -125,7 +125,7 @@ class Deployment:
                     f"/deployments/{config.name}{route.path}",
                     route.endpoint,
                     name=f"{config.name}_{route.name}",
-                    methods=route.methods,
+                    methods=list(route.methods) if route.methods else None,
                     include_in_schema=True,  # change to false when schemas are added to workflow server
                     tags=["workflows"],
                 )

{llama_deploy_appserver-0.3.16 → llama_deploy_appserver-0.3.18}/src/llama_deploy/appserver/process_utils.py

@@ -4,7 +4,8 @@ import platform
 import subprocess
 import sys
 import threading
-from typing import Callable, TextIO, cast
+from dataclasses import dataclass
+from typing import Callable, TextIO, Tuple, cast
 
 
 def run_process(
@@ -25,13 +26,13 @@ def run_process(
     use_pty = _should_use_pty(use_tty)
     prefixer = _make_prefixer(prefix, color_code, line_transform)
 
-    process, sources, cleanup = _spawn_process(cmd, cwd=cwd, env=env, use_pty=use_pty)
+    spawned = _spawn_process(cmd, cwd=cwd, env=env, use_pty=use_pty)
     threads: list[threading.Thread] = []
     try:
-        cleanup()
+        spawned.cleanup()
         _log_command(cmd, prefixer)
-        threads = _start_stream_threads(sources, prefixer)
-        ret = process.wait()
+        threads = _start_stream_threads(spawned.sources, prefixer)
+        ret = spawned.process.wait()
         if ret != 0:
             raise subprocess.CalledProcessError(ret, cmd)
     finally:
@@ -56,11 +57,11 @@ def spawn_process(
     use_pty = _should_use_pty(use_tty)
     prefixer = _make_prefixer(prefix, color_code, line_transform)
 
-    process, sources, cleanup = _spawn_process(cmd, cwd=cwd, env=env, use_pty=use_pty)
-    cleanup()
+    spawned = _spawn_process(cmd, cwd=cwd, env=env, use_pty=use_pty)
+    spawned.cleanup()
     _log_command(cmd, prefixer)
-    _start_stream_threads(sources, prefixer)
-    return process
+    _start_stream_threads(spawned.sources, prefixer)
+    return spawned.process
 
 
 @functools.cache
@@ -109,13 +110,20 @@ def should_use_color() -> bool:
     return _should_use_pty(None)
 
 
+@dataclass
+class SpawnProcessResult:
+    process: subprocess.Popen
+    sources: list[Tuple[int | TextIO, TextIO]]
+    cleanup: Callable[[], None]
+
+
 def _spawn_process(
     cmd: list[str],
     *,
     cwd: os.PathLike | None,
     env: dict[str, str] | None,
     use_pty: bool,
-) -> tuple[subprocess.Popen, list[tuple[int | TextIO, TextIO]], Callable[[], None]]:
+) -> SpawnProcessResult:
     if use_pty:
         import pty
 
@@ -139,7 +147,7 @@ def _spawn_process(
         sources: list[tuple[int | TextIO, TextIO]] = [
             (master_fd, cast(TextIO, sys.stdout)),
         ]
-        return process, sources, cleanup
+        return SpawnProcessResult(process, sources, cleanup)
 
     use_shell = False
     if platform.system() == "Windows":
@@ -161,10 +169,10 @@ def _spawn_process(
 
     assert process.stdout is not None and process.stderr is not None
     sources = [
-        (process.stdout, cast(TextIO, sys.stdout)),
-        (process.stderr, cast(TextIO, sys.stderr)),
+        (cast(int | TextIO, process.stdout), cast(TextIO, sys.stdout)),
+        (cast(int | TextIO, process.stderr), cast(TextIO, sys.stderr)),
     ]
-    return process, sources, cleanup
+    return SpawnProcessResult(process, sources, cleanup)
 
 
 def _stream_source(

{llama_deploy_appserver-0.3.16 → llama_deploy_appserver-0.3.18}/src/llama_deploy/appserver/workflow_store/agent_data_store.py

@@ -81,12 +81,12 @@ class AgentDataStore(AbstractWorkflowStore):
         filters = self._build_filters(handler)
         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]
+            *[self.client.delete_item(item_id=x.id) for x in results.items if x.id]
         )
         return len(results.items)
 
     async def _get_item_id(self, handler: PersistentHandler) -> str | None:
-        cached_id = self.cache.get(handler.handler_id, None)
+        cached_id = self.cache.get(handler.handler_id)
         if cached_id is not None:
             return cached_id
         results = await self.client.search(

{llama_deploy_appserver-0.3.16 → llama_deploy_appserver-0.3.18}/src/llama_deploy/appserver/workflow_store/lru_cache.py

@@ -19,10 +19,7 @@ class LRUCache(Generic[K, V]):
     def get(self, key: K, default: V | None = None) -> V | None:
         if key not in self._store:
             return default
-        # mark as recently used
-        value = self._store.pop(key)
-        self._store[key] = value
-        return value
+        return self[key]
 
     def set(self, key: K, value: V):
         if key in self._store:
@@ -37,7 +34,12 @@ class LRUCache(Generic[K, V]):
         return key in self._store
 
     def __getitem__(self, key: K) -> V:
-        return self.get(key)
+        # mark as recently used
+        if key not in self._store:
+            raise KeyError(key)
+        value = self._store.pop(key)
+        self._store[key] = value
+        return value
 
     def __setitem__(self, key: K, value: V):
         self.set(key, value)