feat(server): Use system packages for execution

Users prefer to rely on the main CLI rather than invoking the server
through a Python module. Users interact with a high-level CLI rather
than needing to know internal module structures.

Now, when running llama stack run <path-to-config>, the server will
attempt to use the system package or a virtual environment if one is
active.

This also eliminates the current process dependency chain when running
from a virtual environment:

-> llama stack run
  -> start_env.sh
    -> python -m server...

Signed-off-by: Sébastien Han <seb@redhat.com>

llama_stack/cli/stack/run.py

@@ -56,7 +56,6 @@ class StackRun(Subcommand):
             "--env",
             action="append",
             help="Environment variables to pass to the server in KEY=VALUE format. Can be specified multiple times.",
-            default=[],
             metavar="KEY=VALUE",
         )
         self.parser.add_argument(
@@ -74,7 +73,6 @@ class StackRun(Subcommand):
             type=str,
             help="Image Type used during the build. This can be either conda or container or venv.",
             choices=["conda", "container", "venv"],
-            default="conda",
         )
 
     def _run_stack_run_cmd(self, args: argparse.Namespace) -> None:
@@ -120,18 +118,40 @@ class StackRun(Subcommand):
         except AttributeError as e:
             self.parser.error(f"failed to parse config file '{config_file}':\n {e}")
 
+        # If neither image type nor image name is provided, assume the server should be run directly
+        # using the current environment packages.
+        if not args.image_type and not args.image_name:
+            logger.info("No image type or image name provided. Assuming environment packages.")
+            from llama_stack.distribution.server.server import main as server_main
+
+            # Build the server args from the current args passed to the CLI
+            server_args = argparse.Namespace()
+            for arg in vars(args):
+                # If this is a function, avoid passing it
+                # "args" contains:
+                # func=<bound method StackRun._run_stack_run_cmd of <llama_stack.cli.stack.run.StackRun object at 0x10484b010>>
+                if callable(getattr(args, arg)):
+                    continue
+                setattr(server_args, arg, getattr(args, arg))
+
+            # Run the server
+            server_main(server_args)
+        else:
             run_args = formulate_run_args(args.image_type, args.image_name, config, template_name)
 
             run_args.extend([str(config_file), str(args.port)])
             if args.disable_ipv6:
                 run_args.append("--disable-ipv6")
 
+            if args.env:
                 for env_var in args.env:
                     if "=" not in env_var:
                         self.parser.error(f"Environment variable '{env_var}' must be in KEY=VALUE format")
+                        return
                     key, value = env_var.split("=", 1)  # split on first = only
                     if not key:
                         self.parser.error(f"Environment variable '{env_var}' has empty key")
+                        return
                     run_args.extend(["--env", f"{key}={value}"])
 
             if args.tls_keyfile and args.tls_certfile:

llama_stack/distribution/server/server.py

@@ -314,11 +314,17 @@ class ClientVersionMiddleware:
         return await self.app(scope, receive, send)
 
 
-def main():
+def main(args: argparse.Namespace = None):
     """Start the LlamaStack server."""
     parser = argparse.ArgumentParser(description="Start the LlamaStack server.")
     parser.add_argument(
         "--yaml-config",
+        dest="config",
+        help="(Deprecated) Path to YAML configuration file - use --config instead",
+    )
+    parser.add_argument(
+        "--config",
+        dest="config",
         help="Path to YAML configuration file",
     )
     parser.add_argument(
@@ -348,8 +354,20 @@ def main():
         required="--tls-keyfile" in sys.argv,
     )
 
+    # Determine whether the server args are being passed by the "run" command, if this is the case
+    # the args will be passed as a Namespace object to the main function, otherwise they will be
+    # parsed from the command line
+    if args is None:
         args = parser.parse_args()
 
+    # Check for deprecated argument usage
+    if "--yaml-config" in sys.argv:
+        warnings.warn(
+            "The '--yaml-config' argument is deprecated and will be removed in a future version. Use '--config' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
     if args.env:
         for env_pair in args.env:
             try:
@@ -360,9 +378,9 @@ def main():
                 logger.error(f"Error: {str(e)}")
                 sys.exit(1)
 
-    if args.yaml_config:
+    if args.config:
         # if the user provided a config file, use it, even if template was specified
-        config_file = Path(args.yaml_config)
+        config_file = Path(args.config)
         if not config_file.exists():
             raise ValueError(f"Config file {config_file} does not exist")
         logger.info(f"Using config file: {config_file}")