llama-stack 0.3.0__tar.gz → 0.3.1__tar.gz

{llama_stack-0.3.0/llama_stack.egg-info → llama_stack-0.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: llama_stack
-Version: 0.3.0
+Version: 0.3.1
 Summary: Llama Stack
 Author-email: Meta Llama <llama-oss@meta.com>
 License: MIT
@@ -22,7 +22,7 @@ Requires-Dist: fire
 Requires-Dist: httpx
 Requires-Dist: jinja2>=3.1.6
 Requires-Dist: jsonschema
-Requires-Dist: llama-stack-client>=0.3.0
+Requires-Dist: llama-stack-client>=0.3.1
 Requires-Dist: openai>=1.107
 Requires-Dist: prompt-toolkit
 Requires-Dist: python-dotenv
@@ -44,7 +44,7 @@ Requires-Dist: sqlalchemy[asyncio]>=2.0.41
 Provides-Extra: ui
 Requires-Dist: streamlit; extra == "ui"
 Requires-Dist: pandas; extra == "ui"
-Requires-Dist: llama-stack-client>=0.3.0; extra == "ui"
+Requires-Dist: llama-stack-client>=0.3.1; extra == "ui"
 Requires-Dist: streamlit-option-menu; extra == "ui"
 Dynamic: license-file
 

{llama_stack-0.3.0 → llama_stack-0.3.1}/llama_stack/cli/stack/run.py

@@ -106,7 +106,8 @@ class StackRun(Subcommand):
 
             try:
                 config = parse_and_maybe_upgrade_config(config_dict)
-                if not os.path.exists(str(config.external_providers_dir)):
+                # Create external_providers_dir if it's specified and doesn't exist
+                if config.external_providers_dir and not os.path.exists(str(config.external_providers_dir)):
                     os.makedirs(str(config.external_providers_dir), exist_ok=True)
             except AttributeError as e:
                 self.parser.error(f"failed to parse config file '{config_file}':\n {e}")
@@ -127,7 +128,7 @@ class StackRun(Subcommand):
             config = StackRunConfig(**cast_image_name_to_string(replace_env_vars(config_contents)))
 
         port = args.port or config.server.port
-        host = config.server.host or ["::", "0.0.0.0"]
+        host = config.server.host or "0.0.0.0"
 
         # Set the config file in environment so create_app can find it
         os.environ["LLAMA_STACK_CONFIG"] = str(config_file)
@@ -139,6 +140,7 @@ class StackRun(Subcommand):
             "lifespan": "on",
             "log_level": logger.getEffectiveLevel(),
             "log_config": logger_config,
+            "workers": config.server.workers,
         }
 
         keyfile = config.server.tls_keyfile

{llama_stack-0.3.0 → llama_stack-0.3.1}/llama_stack/core/configure.py

@@ -17,7 +17,6 @@ from llama_stack.core.distribution import (
     get_provider_registry,
 )
 from llama_stack.core.stack import cast_image_name_to_string, replace_env_vars
-from llama_stack.core.utils.config_dirs import EXTERNAL_PROVIDERS_DIR
 from llama_stack.core.utils.dynamic import instantiate_class_type
 from llama_stack.core.utils.prompt_for_config import prompt_for_config
 from llama_stack.log import get_logger
@@ -194,19 +193,11 @@ def upgrade_from_routing_table(
 
 
 def parse_and_maybe_upgrade_config(config_dict: dict[str, Any]) -> StackRunConfig:
-    version = config_dict.get("version", None)
-    if version == LLAMA_STACK_RUN_CONFIG_VERSION:
-        processed_config_dict = replace_env_vars(config_dict)
-        return StackRunConfig(**cast_image_name_to_string(processed_config_dict))
-
     if "routing_table" in config_dict:
         logger.info("Upgrading config...")
         config_dict = upgrade_from_routing_table(config_dict)
 
     config_dict["version"] = LLAMA_STACK_RUN_CONFIG_VERSION
 
-    if not config_dict.get("external_providers_dir", None):
-        config_dict["external_providers_dir"] = EXTERNAL_PROVIDERS_DIR
-
     processed_config_dict = replace_env_vars(config_dict)
     return StackRunConfig(**cast_image_name_to_string(processed_config_dict))

{llama_stack-0.3.0 → llama_stack-0.3.1}/llama_stack/core/datatypes.py

@@ -471,6 +471,10 @@ class ServerConfig(BaseModel):
         "- true: Enable localhost CORS for development\n"
         "- {allow_origins: [...], allow_methods: [...], ...}: Full configuration",
     )
+    workers: int = Field(
+        default=1,
+        description="Number of workers to use for the server",
+    )
 
 
 class StackRunConfig(BaseModel):

llama_stack-0.3.1/llama_stack/core/utils/context.py

@@ -0,0 +1,84 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the terms described in the LICENSE file in
+# the root directory of this source tree.
+
+from collections.abc import AsyncGenerator
+from contextvars import ContextVar
+
+from llama_stack.providers.utils.telemetry.tracing import CURRENT_TRACE_CONTEXT
+
+_MISSING = object()
+
+
+def preserve_contexts_async_generator[T](
+    gen: AsyncGenerator[T, None], context_vars: list[ContextVar]
+) -> AsyncGenerator[T, None]:
+    """
+    Wraps an async generator to preserve context variables across iterations.
+    This is needed because we start a new asyncio event loop for each streaming request,
+    and we need to preserve the context across the event loop boundary.
+    """
+    # Capture initial context values
+    initial_context_values = {context_var.name: context_var.get() for context_var in context_vars}
+
+    async def wrapper() -> AsyncGenerator[T, None]:
+        while True:
+            previous_values: dict[ContextVar, object] = {}
+            tokens: dict[ContextVar, object] = {}
+
+            # Restore ALL context values before any await and capture previous state
+            # This is needed to propagate context across async generator boundaries
+            for context_var in context_vars:
+                try:
+                    previous_values[context_var] = context_var.get()
+                except LookupError:
+                    previous_values[context_var] = _MISSING
+                tokens[context_var] = context_var.set(initial_context_values[context_var.name])
+
+            def _restore_context_var(context_var: ContextVar, *, _tokens=tokens, _prev=previous_values) -> None:
+                token = _tokens.get(context_var)
+                previous_value = _prev.get(context_var, _MISSING)
+                if token is not None:
+                    try:
+                        context_var.reset(token)
+                        return
+                    except (RuntimeError, ValueError):
+                        pass
+
+                if previous_value is _MISSING:
+                    context_var.set(None)
+                else:
+                    context_var.set(previous_value)
+
+            try:
+                item = await gen.__anext__()
+            except StopAsyncIteration:
+                # Restore all context vars before exiting to prevent leaks
+                # Use _restore_context_var for all vars to properly restore to previous values
+                for context_var in context_vars:
+                    _restore_context_var(context_var)
+                break
+            except Exception:
+                # Restore all context vars on exception
+                for context_var in context_vars:
+                    _restore_context_var(context_var)
+                raise
+
+            try:
+                yield item
+                # Update our tracked values with any changes made during this iteration
+                # Only for non-trace context vars - trace context must persist across yields
+                # to allow nested span tracking for telemetry
+                for context_var in context_vars:
+                    if context_var is not CURRENT_TRACE_CONTEXT:
+                        initial_context_values[context_var.name] = context_var.get()
+            finally:
+                # Restore non-trace context vars after each yield to prevent leaks between requests
+                # CURRENT_TRACE_CONTEXT is NOT restored here to preserve telemetry span stack
+                for context_var in context_vars:
+                    if context_var is not CURRENT_TRACE_CONTEXT:
+                        _restore_context_var(context_var)
+
+    return wrapper()

{llama_stack-0.3.0 → llama_stack-0.3.1/llama_stack.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: llama_stack
-Version: 0.3.0
+Version: 0.3.1
 Summary: Llama Stack
 Author-email: Meta Llama <llama-oss@meta.com>
 License: MIT
@@ -22,7 +22,7 @@ Requires-Dist: fire
 Requires-Dist: httpx
 Requires-Dist: jinja2>=3.1.6
 Requires-Dist: jsonschema
-Requires-Dist: llama-stack-client>=0.3.0
+Requires-Dist: llama-stack-client>=0.3.1
 Requires-Dist: openai>=1.107
 Requires-Dist: prompt-toolkit
 Requires-Dist: python-dotenv
@@ -44,7 +44,7 @@ Requires-Dist: sqlalchemy[asyncio]>=2.0.41
 Provides-Extra: ui
 Requires-Dist: streamlit; extra == "ui"
 Requires-Dist: pandas; extra == "ui"
-Requires-Dist: llama-stack-client>=0.3.0; extra == "ui"
+Requires-Dist: llama-stack-client>=0.3.1; extra == "ui"
 Requires-Dist: streamlit-option-menu; extra == "ui"
 Dynamic: license-file
 

{llama_stack-0.3.0 → llama_stack-0.3.1}/llama_stack.egg-info/requires.txt

@@ -4,7 +4,7 @@ fire
 httpx
 jinja2>=3.1.6
 jsonschema
-llama-stack-client>=0.3.0
+llama-stack-client>=0.3.1
 openai>=1.107
 prompt-toolkit
 python-dotenv
@@ -27,5 +27,5 @@ sqlalchemy[asyncio]>=2.0.41
 [ui]
 streamlit
 pandas
-llama-stack-client>=0.3.0
+llama-stack-client>=0.3.1
 streamlit-option-menu

{llama_stack-0.3.0 → llama_stack-0.3.1}/pyproject.toml

@@ -7,7 +7,7 @@ required-version = ">=0.7.0"
 
 [project]
 name = "llama_stack"
-version = "0.3.0"
+version = "0.3.1"
 authors = [{ name = "Meta Llama", email = "llama-oss@meta.com" }]
 description = "Llama Stack"
 readme = "README.md"
@@ -30,7 +30,7 @@ dependencies = [
     "httpx",
     "jinja2>=3.1.6",
     "jsonschema",
-    "llama-stack-client>=0.3.0",
+    "llama-stack-client>=0.3.1",
     "openai>=1.107",                                  # for expires_after support
     "prompt-toolkit",
     "python-dotenv",
@@ -55,7 +55,7 @@ dependencies = [
 ui = [
     "streamlit",
     "pandas",
-    "llama-stack-client>=0.3.0",
+    "llama-stack-client>=0.3.1",
     "streamlit-option-menu",
 ]
 

llama_stack-0.3.0/llama_stack/core/utils/context.py

@@ -1,40 +0,0 @@
-# Copyright (c) Meta Platforms, Inc. and affiliates.
-# All rights reserved.
-#
-# This source code is licensed under the terms described in the LICENSE file in
-# the root directory of this source tree.
-
-from collections.abc import AsyncGenerator
-from contextvars import ContextVar
-
-
-def preserve_contexts_async_generator[T](
-    gen: AsyncGenerator[T, None], context_vars: list[ContextVar]
-) -> AsyncGenerator[T, None]:
-    """
-    Wraps an async generator to preserve context variables across iterations.
-    This is needed because we start a new asyncio event loop for each streaming request,
-    and we need to preserve the context across the event loop boundary.
-    """
-    # Capture initial context values
-    initial_context_values = {context_var.name: context_var.get() for context_var in context_vars}
-
-    async def wrapper() -> AsyncGenerator[T, None]:
-        while True:
-            try:
-                # Restore context values before any await
-                for context_var in context_vars:
-                    context_var.set(initial_context_values[context_var.name])
-
-                item = await gen.__anext__()
-
-                # Update our tracked values with any changes made during this iteration
-                for context_var in context_vars:
-                    initial_context_values[context_var.name] = context_var.get()
-
-                yield item
-
-            except StopAsyncIteration:
-                break
-
-    return wrapper()