mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-08-12 04:50:39 +00:00
chore: rename templates to distributions (#3035)
As the title says. Distributions is in, Templates is out. `llama stack build --template` --> `llama stack build --distro`. For backward compatibility, the previous option is kept but results in a warning. Updated `server.py` to remove the "config_or_template" backward compatibility since it has been a couple releases since that change.
This commit is contained in:
parent
12f964437a
commit
cc87995e2b
87 changed files with 263 additions and 330 deletions
|
@ -46,25 +46,25 @@ from llama_stack.core.utils.exec import formulate_run_args, run_command
|
|||
from llama_stack.core.utils.image_types import LlamaStackImageType
|
||||
from llama_stack.providers.datatypes import Api
|
||||
|
||||
TEMPLATES_PATH = Path(__file__).parent.parent.parent / "templates"
|
||||
DISTRIBS_PATH = Path(__file__).parent.parent.parent / "distributions"
|
||||
|
||||
|
||||
@lru_cache
|
||||
def available_templates_specs() -> dict[str, BuildConfig]:
|
||||
def available_distros_specs() -> dict[str, BuildConfig]:
|
||||
import yaml
|
||||
|
||||
template_specs = {}
|
||||
for p in TEMPLATES_PATH.rglob("*build.yaml"):
|
||||
template_name = p.parent.name
|
||||
distro_specs = {}
|
||||
for p in DISTRIBS_PATH.rglob("*build.yaml"):
|
||||
distro_name = p.parent.name
|
||||
with open(p) as f:
|
||||
build_config = BuildConfig(**yaml.safe_load(f))
|
||||
template_specs[template_name] = build_config
|
||||
return template_specs
|
||||
distro_specs[distro_name] = build_config
|
||||
return distro_specs
|
||||
|
||||
|
||||
def run_stack_build_command(args: argparse.Namespace) -> None:
|
||||
if args.list_templates:
|
||||
return _run_template_list_cmd()
|
||||
if args.list_distros:
|
||||
return _run_distro_list_cmd()
|
||||
|
||||
if args.image_type == ImageType.VENV.value:
|
||||
current_venv = os.environ.get("VIRTUAL_ENV")
|
||||
|
@ -73,20 +73,30 @@ def run_stack_build_command(args: argparse.Namespace) -> None:
|
|||
image_name = args.image_name
|
||||
|
||||
if args.template:
|
||||
available_templates = available_templates_specs()
|
||||
if args.template not in available_templates:
|
||||
cprint(
|
||||
"The --template argument is deprecated. Please use --distro instead.",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
distro_name = args.template
|
||||
else:
|
||||
distro_name = args.distribution
|
||||
|
||||
if distro_name:
|
||||
available_distros = available_distros_specs()
|
||||
if distro_name not in available_distros:
|
||||
cprint(
|
||||
f"Could not find template {args.template}. Please run `llama stack build --list-templates` to check out the available templates",
|
||||
f"Could not find distribution {distro_name}. Please run `llama stack build --list-distros` to check out the available distributions",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
sys.exit(1)
|
||||
build_config = available_templates[args.template]
|
||||
build_config = available_distros[distro_name]
|
||||
if args.image_type:
|
||||
build_config.image_type = args.image_type
|
||||
else:
|
||||
cprint(
|
||||
f"Please specify a image-type ({' | '.join(e.value for e in ImageType)}) for {args.template}",
|
||||
f"Please specify a image-type ({' | '.join(e.value for e in ImageType)}) for {distro_name}",
|
||||
color="red",
|
||||
file=sys.stderr,
|
||||
)
|
||||
|
@ -136,7 +146,7 @@ def run_stack_build_command(args: argparse.Namespace) -> None:
|
|||
sys.exit(1)
|
||||
|
||||
build_config = BuildConfig(image_type=args.image_type, distribution_spec=distribution_spec)
|
||||
elif not args.config and not args.template:
|
||||
elif not args.config and not distro_name:
|
||||
name = prompt(
|
||||
"> Enter a name for your Llama Stack (e.g. my-local-stack): ",
|
||||
validator=Validator.from_callable(
|
||||
|
@ -218,7 +228,7 @@ def run_stack_build_command(args: argparse.Namespace) -> None:
|
|||
sys.exit(1)
|
||||
|
||||
if args.print_deps_only:
|
||||
print(f"# Dependencies for {args.template or args.config or image_name}")
|
||||
print(f"# Dependencies for {distro_name or args.config or image_name}")
|
||||
normal_deps, special_deps, external_provider_dependencies = get_provider_dependencies(build_config)
|
||||
normal_deps += SERVER_DEPENDENCIES
|
||||
print(f"uv pip install {' '.join(normal_deps)}")
|
||||
|
@ -233,7 +243,7 @@ def run_stack_build_command(args: argparse.Namespace) -> None:
|
|||
build_config,
|
||||
image_name=image_name,
|
||||
config_path=args.config,
|
||||
template_name=args.template,
|
||||
distro_name=distro_name,
|
||||
)
|
||||
|
||||
except (Exception, RuntimeError) as exc:
|
||||
|
@ -344,13 +354,13 @@ def _generate_run_config(
|
|||
def _run_stack_build_command_from_build_config(
|
||||
build_config: BuildConfig,
|
||||
image_name: str | None = None,
|
||||
template_name: str | None = None,
|
||||
distro_name: str | None = None,
|
||||
config_path: str | None = None,
|
||||
) -> Path | Traversable:
|
||||
image_name = image_name or build_config.image_name
|
||||
if build_config.image_type == LlamaStackImageType.CONTAINER.value:
|
||||
if template_name:
|
||||
image_name = f"distribution-{template_name}"
|
||||
if distro_name:
|
||||
image_name = f"distribution-{distro_name}"
|
||||
else:
|
||||
if not image_name:
|
||||
raise ValueError("Please specify an image name when building a container image without a template")
|
||||
|
@ -364,9 +374,9 @@ def _run_stack_build_command_from_build_config(
|
|||
if image_name is None:
|
||||
raise ValueError("image_name should not be None after validation")
|
||||
|
||||
if template_name:
|
||||
build_dir = DISTRIBS_BASE_DIR / template_name
|
||||
build_file_path = build_dir / f"{template_name}-build.yaml"
|
||||
if distro_name:
|
||||
build_dir = DISTRIBS_BASE_DIR / distro_name
|
||||
build_file_path = build_dir / f"{distro_name}-build.yaml"
|
||||
else:
|
||||
if image_name is None:
|
||||
raise ValueError("image_name cannot be None")
|
||||
|
@ -377,7 +387,7 @@ def _run_stack_build_command_from_build_config(
|
|||
run_config_file = None
|
||||
# Generate the run.yaml so it can be included in the container image with the proper entrypoint
|
||||
# Only do this if we're building a container image and we're not using a template
|
||||
if build_config.image_type == LlamaStackImageType.CONTAINER.value and not template_name and config_path:
|
||||
if build_config.image_type == LlamaStackImageType.CONTAINER.value and not distro_name and config_path:
|
||||
cprint("Generating run.yaml file", color="yellow", file=sys.stderr)
|
||||
run_config_file = _generate_run_config(build_config, build_dir, image_name)
|
||||
|
||||
|
@ -411,46 +421,45 @@ def _run_stack_build_command_from_build_config(
|
|||
return_code = build_image(
|
||||
build_config,
|
||||
image_name,
|
||||
template_or_config=template_name or config_path or str(build_file_path),
|
||||
distro_or_config=distro_name or config_path or str(build_file_path),
|
||||
run_config=run_config_file.as_posix() if run_config_file else None,
|
||||
)
|
||||
if return_code != 0:
|
||||
raise RuntimeError(f"Failed to build image {image_name}")
|
||||
|
||||
if template_name:
|
||||
# copy run.yaml from template to build_dir instead of generating it again
|
||||
template_path = importlib.resources.files("llama_stack") / f"templates/{template_name}/run.yaml"
|
||||
run_config_file = build_dir / f"{template_name}-run.yaml"
|
||||
if distro_name:
|
||||
# copy run.yaml from distribution to build_dir instead of generating it again
|
||||
distro_path = importlib.resources.files("llama_stack") / f"distributions/{distro_name}/run.yaml"
|
||||
run_config_file = build_dir / f"{distro_name}-run.yaml"
|
||||
|
||||
with importlib.resources.as_file(template_path) as path:
|
||||
with importlib.resources.as_file(distro_path) as path:
|
||||
shutil.copy(path, run_config_file)
|
||||
|
||||
cprint("Build Successful!", color="green", file=sys.stderr)
|
||||
cprint(f"You can find the newly-built template here: {run_config_file}", color="blue", file=sys.stderr)
|
||||
cprint(f"You can find the newly-built distribution here: {run_config_file}", color="blue", file=sys.stderr)
|
||||
cprint(
|
||||
"You can run the new Llama Stack distro via: "
|
||||
+ colored(f"llama stack run {run_config_file} --image-type {build_config.image_type}", "blue"),
|
||||
color="green",
|
||||
file=sys.stderr,
|
||||
)
|
||||
return template_path
|
||||
return distro_path
|
||||
else:
|
||||
return _generate_run_config(build_config, build_dir, image_name)
|
||||
|
||||
|
||||
def _run_template_list_cmd() -> None:
|
||||
# eventually, this should query a registry at llama.meta.com/llamastack/distributions
|
||||
def _run_distro_list_cmd() -> None:
|
||||
headers = [
|
||||
"Template Name",
|
||||
"Distribution Name",
|
||||
# "Providers",
|
||||
"Description",
|
||||
]
|
||||
|
||||
rows = []
|
||||
for template_name, spec in available_templates_specs().items():
|
||||
for distro_name, spec in available_distros_specs().items():
|
||||
rows.append(
|
||||
[
|
||||
template_name,
|
||||
distro_name,
|
||||
# json.dumps(spec.distribution_spec.providers, indent=2),
|
||||
spec.distribution_spec.description,
|
||||
]
|
||||
|
|
|
@ -34,14 +34,24 @@ class StackBuild(Subcommand):
|
|||
"--template",
|
||||
type=str,
|
||||
default=None,
|
||||
help="Name of the example template config to use for build. You may use `llama stack build --list-templates` to check out the available templates",
|
||||
help="""(deprecated) Name of the example template config to use for build. You may use `llama stack build --list-distros` to check out the available distributions""",
|
||||
)
|
||||
self.parser.add_argument(
|
||||
"--distro",
|
||||
"--distribution",
|
||||
dest="distribution",
|
||||
type=str,
|
||||
default=None,
|
||||
help="""Name of the distribution to use for build. You may use `llama stack build --list-distros` to check out the available distributions""",
|
||||
)
|
||||
|
||||
self.parser.add_argument(
|
||||
"--list-templates",
|
||||
"--list-distros",
|
||||
"--list-distributions",
|
||||
action="store_true",
|
||||
dest="list_distros",
|
||||
default=False,
|
||||
help="Show the available templates for building a Llama Stack distribution",
|
||||
help="Show the available distributions for building a Llama Stack distribution",
|
||||
)
|
||||
|
||||
self.parser.add_argument(
|
||||
|
|
|
@ -35,8 +35,8 @@ class StackRun(Subcommand):
|
|||
"config",
|
||||
type=str,
|
||||
nargs="?", # Make it optional
|
||||
metavar="config | template",
|
||||
help="Path to config file to use for the run or name of known template (`llama stack list` for a list).",
|
||||
metavar="config | distro",
|
||||
help="Path to config file to use for the run or name of known distro (`llama stack list` for a list).",
|
||||
)
|
||||
self.parser.add_argument(
|
||||
"--port",
|
||||
|
@ -68,22 +68,22 @@ class StackRun(Subcommand):
|
|||
help="Start the UI server",
|
||||
)
|
||||
|
||||
def _resolve_config_and_template(self, args: argparse.Namespace) -> tuple[Path | None, str | None]:
|
||||
"""Resolve config file path and template name from args.config"""
|
||||
from llama_stack.distribution.utils.config_dirs import DISTRIBS_BASE_DIR
|
||||
def _resolve_config_and_distro(self, args: argparse.Namespace) -> tuple[Path | None, str | None]:
|
||||
"""Resolve config file path and distribution name from args.config"""
|
||||
from llama_stack.core.utils.config_dirs import DISTRIBS_BASE_DIR
|
||||
|
||||
if not args.config:
|
||||
return None, None
|
||||
|
||||
config_file = Path(args.config)
|
||||
has_yaml_suffix = args.config.endswith(".yaml")
|
||||
template_name = None
|
||||
distro_name = None
|
||||
|
||||
if not config_file.exists() and not has_yaml_suffix:
|
||||
# check if this is a template
|
||||
config_file = Path(REPO_ROOT) / "llama_stack" / "templates" / args.config / "run.yaml"
|
||||
# check if this is a distribution
|
||||
config_file = Path(REPO_ROOT) / "llama_stack" / "distributions" / args.config / "run.yaml"
|
||||
if config_file.exists():
|
||||
template_name = args.config
|
||||
distro_name = args.config
|
||||
|
||||
if not config_file.exists() and not has_yaml_suffix:
|
||||
# check if it's a build config saved to ~/.llama dir
|
||||
|
@ -99,7 +99,7 @@ class StackRun(Subcommand):
|
|||
f"Config file must be a valid file path, '{config_file}' is not a file: type={type(config_file)}"
|
||||
)
|
||||
|
||||
return config_file, template_name
|
||||
return config_file, distro_name
|
||||
|
||||
def _run_stack_run_cmd(self, args: argparse.Namespace) -> None:
|
||||
import yaml
|
||||
|
@ -113,9 +113,9 @@ class StackRun(Subcommand):
|
|||
|
||||
if args.config:
|
||||
try:
|
||||
from llama_stack.core.utils.config_resolution import Mode, resolve_config_or_template
|
||||
from llama_stack.core.utils.config_resolution import Mode, resolve_config_or_distro
|
||||
|
||||
config_file = resolve_config_or_template(args.config, Mode.RUN)
|
||||
config_file = resolve_config_or_distro(args.config, Mode.RUN)
|
||||
except ValueError as e:
|
||||
self.parser.error(str(e))
|
||||
else:
|
||||
|
|
|
@ -11,38 +11,19 @@ from llama_stack.log import get_logger
|
|||
logger = get_logger(name=__name__, category="cli")
|
||||
|
||||
|
||||
def add_config_template_args(parser: argparse.ArgumentParser):
|
||||
"""Add unified config/template arguments with backward compatibility."""
|
||||
# TODO: this can probably just be inlined now?
|
||||
def add_config_distro_args(parser: argparse.ArgumentParser):
|
||||
"""Add unified config/distro arguments."""
|
||||
group = parser.add_mutually_exclusive_group(required=True)
|
||||
|
||||
group.add_argument(
|
||||
"config",
|
||||
nargs="?",
|
||||
help="Configuration file path or template name",
|
||||
)
|
||||
|
||||
# Backward compatibility arguments (deprecated)
|
||||
group.add_argument(
|
||||
"--config",
|
||||
dest="config_deprecated",
|
||||
help="(DEPRECATED) Use positional argument [config] instead. Configuration file path",
|
||||
)
|
||||
|
||||
group.add_argument(
|
||||
"--template",
|
||||
dest="template_deprecated",
|
||||
help="(DEPRECATED) Use positional argument [config] instead. Template name",
|
||||
help="Configuration file path or distribution name",
|
||||
)
|
||||
|
||||
|
||||
def get_config_from_args(args: argparse.Namespace) -> str | None:
|
||||
"""Extract config value from parsed arguments, handling both new and deprecated forms."""
|
||||
if args.config is not None:
|
||||
return str(args.config)
|
||||
elif hasattr(args, "config_deprecated") and args.config_deprecated is not None:
|
||||
logger.warning("Using deprecated --config argument. Use positional argument [config] instead.")
|
||||
return str(args.config_deprecated)
|
||||
elif hasattr(args, "template_deprecated") and args.template_deprecated is not None:
|
||||
logger.warning("Using deprecated --template argument. Use positional argument [config] instead.")
|
||||
return str(args.template_deprecated)
|
||||
return None
|
||||
|
|
|
@ -16,8 +16,8 @@ from llama_stack.core.distribution import get_provider_registry
|
|||
from llama_stack.core.external import load_external_apis
|
||||
from llama_stack.core.utils.exec import run_command
|
||||
from llama_stack.core.utils.image_types import LlamaStackImageType
|
||||
from llama_stack.distributions.template import DistributionTemplate
|
||||
from llama_stack.providers.datatypes import Api
|
||||
from llama_stack.templates.template import DistributionTemplate
|
||||
|
||||
log = logging.getLogger(__name__)
|
||||
|
||||
|
@ -106,7 +106,7 @@ def print_pip_install_help(config: BuildConfig):
|
|||
def build_image(
|
||||
build_config: BuildConfig,
|
||||
image_name: str,
|
||||
template_or_config: str,
|
||||
distro_or_config: str,
|
||||
run_config: str | None = None,
|
||||
):
|
||||
container_base = build_config.distribution_spec.container_image or "python:3.12-slim"
|
||||
|
@ -123,8 +123,8 @@ def build_image(
|
|||
script = str(importlib.resources.files("llama_stack") / "core/build_container.sh")
|
||||
args = [
|
||||
script,
|
||||
"--template-or-config",
|
||||
template_or_config,
|
||||
"--distro-or-config",
|
||||
distro_or_config,
|
||||
"--image-name",
|
||||
image_name,
|
||||
"--container-base",
|
||||
|
|
|
@ -43,7 +43,7 @@ normal_deps=""
|
|||
external_provider_deps=""
|
||||
optional_deps=""
|
||||
run_config=""
|
||||
template_or_config=""
|
||||
distro_or_config=""
|
||||
|
||||
while [[ $# -gt 0 ]]; do
|
||||
key="$1"
|
||||
|
@ -96,12 +96,12 @@ while [[ $# -gt 0 ]]; do
|
|||
run_config="$2"
|
||||
shift 2
|
||||
;;
|
||||
--template-or-config)
|
||||
--distro-or-config)
|
||||
if [[ -z "$2" || "$2" == --* ]]; then
|
||||
echo "Error: --template-or-config requires a string value" >&2
|
||||
echo "Error: --distro-or-config requires a string value" >&2
|
||||
usage
|
||||
fi
|
||||
template_or_config="$2"
|
||||
distro_or_config="$2"
|
||||
shift 2
|
||||
;;
|
||||
*)
|
||||
|
@ -327,12 +327,11 @@ EOF
|
|||
# If a run config is provided, we use the --config flag
|
||||
if [[ -n "$run_config" ]]; then
|
||||
add_to_container << EOF
|
||||
ENTRYPOINT ["python", "-m", "llama_stack.core.server.server", "--config", "$RUN_CONFIG_PATH"]
|
||||
ENTRYPOINT ["python", "-m", "llama_stack.core.server.server", "$RUN_CONFIG_PATH"]
|
||||
EOF
|
||||
# If a template is provided (not a yaml file), we use the --template flag
|
||||
elif [[ "$template_or_config" != *.yaml ]]; then
|
||||
elif [[ "$distro_or_config" != *.yaml ]]; then
|
||||
add_to_container << EOF
|
||||
ENTRYPOINT ["python", "-m", "llama_stack.core.server.server", "--template", "$template_or_config"]
|
||||
ENTRYPOINT ["python", "-m", "llama_stack.core.server.server", "$distro_or_config"]
|
||||
EOF
|
||||
fi
|
||||
|
||||
|
|
|
@ -42,7 +42,7 @@ from llama_stack.core.resolver import ProviderRegistry
|
|||
from llama_stack.core.server.routes import RouteImpls, find_matching_route, initialize_route_impls
|
||||
from llama_stack.core.stack import (
|
||||
construct_stack,
|
||||
get_stack_run_config_from_template,
|
||||
get_stack_run_config_from_distro,
|
||||
replace_env_vars,
|
||||
)
|
||||
from llama_stack.core.utils.config import redact_sensitive_fields
|
||||
|
@ -138,14 +138,14 @@ class LibraryClientHttpxResponse:
|
|||
class LlamaStackAsLibraryClient(LlamaStackClient):
|
||||
def __init__(
|
||||
self,
|
||||
config_path_or_template_name: str,
|
||||
config_path_or_distro_name: str,
|
||||
skip_logger_removal: bool = False,
|
||||
custom_provider_registry: ProviderRegistry | None = None,
|
||||
provider_data: dict[str, Any] | None = None,
|
||||
):
|
||||
super().__init__()
|
||||
self.async_client = AsyncLlamaStackAsLibraryClient(
|
||||
config_path_or_template_name, custom_provider_registry, provider_data
|
||||
config_path_or_distro_name, custom_provider_registry, provider_data
|
||||
)
|
||||
self.pool_executor = ThreadPoolExecutor(max_workers=4)
|
||||
self.skip_logger_removal = skip_logger_removal
|
||||
|
@ -212,7 +212,7 @@ class LlamaStackAsLibraryClient(LlamaStackClient):
|
|||
class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
|
||||
def __init__(
|
||||
self,
|
||||
config_path_or_template_name: str,
|
||||
config_path_or_distro_name: str,
|
||||
custom_provider_registry: ProviderRegistry | None = None,
|
||||
provider_data: dict[str, Any] | None = None,
|
||||
):
|
||||
|
@ -222,17 +222,17 @@ class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
|
|||
current_sinks = os.environ.get("TELEMETRY_SINKS", "sqlite").split(",")
|
||||
os.environ["TELEMETRY_SINKS"] = ",".join(sink for sink in current_sinks if sink != "console")
|
||||
|
||||
if config_path_or_template_name.endswith(".yaml"):
|
||||
config_path = Path(config_path_or_template_name)
|
||||
if config_path_or_distro_name.endswith(".yaml"):
|
||||
config_path = Path(config_path_or_distro_name)
|
||||
if not config_path.exists():
|
||||
raise ValueError(f"Config file {config_path} does not exist")
|
||||
config_dict = replace_env_vars(yaml.safe_load(config_path.read_text()))
|
||||
config = parse_and_maybe_upgrade_config(config_dict)
|
||||
else:
|
||||
# template
|
||||
config = get_stack_run_config_from_template(config_path_or_template_name)
|
||||
# distribution
|
||||
config = get_stack_run_config_from_distro(config_path_or_distro_name)
|
||||
|
||||
self.config_path_or_template_name = config_path_or_template_name
|
||||
self.config_path_or_distro_name = config_path_or_distro_name
|
||||
self.config = config
|
||||
self.custom_provider_registry = custom_provider_registry
|
||||
self.provider_data = provider_data
|
||||
|
@ -245,11 +245,11 @@ class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
|
|||
except ModuleNotFoundError as _e:
|
||||
cprint(_e.msg, color="red", file=sys.stderr)
|
||||
cprint(
|
||||
"Using llama-stack as a library requires installing dependencies depending on the template (providers) you choose.\n",
|
||||
"Using llama-stack as a library requires installing dependencies depending on the distribution (providers) you choose.\n",
|
||||
color="yellow",
|
||||
file=sys.stderr,
|
||||
)
|
||||
if self.config_path_or_template_name.endswith(".yaml"):
|
||||
if self.config_path_or_distro_name.endswith(".yaml"):
|
||||
providers: dict[str, list[BuildProvider]] = {}
|
||||
for api, run_providers in self.config.providers.items():
|
||||
for provider in run_providers:
|
||||
|
@ -267,7 +267,7 @@ class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
|
|||
else:
|
||||
prefix = "!" if in_notebook() else ""
|
||||
cprint(
|
||||
f"Please run:\n\n{prefix}llama stack build --template {self.config_path_or_template_name} --image-type venv\n\n",
|
||||
f"Please run:\n\n{prefix}llama stack build --distro {self.config_path_or_distro_name} --image-type venv\n\n",
|
||||
"yellow",
|
||||
file=sys.stderr,
|
||||
)
|
||||
|
@ -283,7 +283,7 @@ class AsyncLlamaStackAsLibraryClient(AsyncLlamaStackClient):
|
|||
|
||||
if not os.environ.get("PYTEST_CURRENT_TEST"):
|
||||
console = Console()
|
||||
console.print(f"Using config [blue]{self.config_path_or_template_name}[/blue]:")
|
||||
console.print(f"Using config [blue]{self.config_path_or_distro_name}[/blue]:")
|
||||
safe_config = redact_sensitive_fields(self.config.model_dump())
|
||||
console.print(yaml.dump(safe_config, indent=2))
|
||||
|
||||
|
|
|
@ -32,7 +32,7 @@ from openai import BadRequestError
|
|||
from pydantic import BaseModel, ValidationError
|
||||
|
||||
from llama_stack.apis.common.responses import PaginatedResponse
|
||||
from llama_stack.cli.utils import add_config_template_args, get_config_from_args
|
||||
from llama_stack.cli.utils import add_config_distro_args, get_config_from_args
|
||||
from llama_stack.core.access_control.access_control import AccessDeniedError
|
||||
from llama_stack.core.datatypes import (
|
||||
AuthenticationRequiredError,
|
||||
|
@ -60,7 +60,7 @@ from llama_stack.core.stack import (
|
|||
validate_env_pair,
|
||||
)
|
||||
from llama_stack.core.utils.config import redact_sensitive_fields
|
||||
from llama_stack.core.utils.config_resolution import Mode, resolve_config_or_template
|
||||
from llama_stack.core.utils.config_resolution import Mode, resolve_config_or_distro
|
||||
from llama_stack.core.utils.context import preserve_contexts_async_generator
|
||||
from llama_stack.log import get_logger
|
||||
from llama_stack.providers.datatypes import Api
|
||||
|
@ -377,7 +377,7 @@ def main(args: argparse.Namespace | None = None):
|
|||
"""Start the LlamaStack server."""
|
||||
parser = argparse.ArgumentParser(description="Start the LlamaStack server.")
|
||||
|
||||
add_config_template_args(parser)
|
||||
add_config_distro_args(parser)
|
||||
parser.add_argument(
|
||||
"--port",
|
||||
type=int,
|
||||
|
@ -396,8 +396,8 @@ def main(args: argparse.Namespace | None = None):
|
|||
if args is None:
|
||||
args = parser.parse_args()
|
||||
|
||||
config_or_template = get_config_from_args(args)
|
||||
config_file = resolve_config_or_template(config_or_template, Mode.RUN)
|
||||
config_or_distro = get_config_from_args(args)
|
||||
config_file = resolve_config_or_distro(config_or_distro, Mode.RUN)
|
||||
|
||||
logger_config = None
|
||||
with open(config_file) as fp:
|
||||
|
|
|
@ -389,12 +389,12 @@ async def refresh_registry_task(impls: dict[Api, Any]):
|
|||
await asyncio.sleep(REGISTRY_REFRESH_INTERVAL_SECONDS)
|
||||
|
||||
|
||||
def get_stack_run_config_from_template(template: str) -> StackRunConfig:
|
||||
template_path = importlib.resources.files("llama_stack") / f"templates/{template}/run.yaml"
|
||||
def get_stack_run_config_from_distro(distro: str) -> StackRunConfig:
|
||||
distro_path = importlib.resources.files("llama_stack") / f"distributions/{distro}/run.yaml"
|
||||
|
||||
with importlib.resources.as_file(template_path) as path:
|
||||
with importlib.resources.as_file(distro_path) as path:
|
||||
if not path.exists():
|
||||
raise ValueError(f"Template '{template}' not found at {template_path}")
|
||||
raise ValueError(f"Distribution '{distro}' not found at {distro_path}")
|
||||
run_config = yaml.safe_load(path.open())
|
||||
|
||||
return StackRunConfig(**replace_env_vars(run_config))
|
||||
|
|
|
@ -9,7 +9,7 @@
|
|||
1. Start up Llama Stack API server. More details [here](https://llama-stack.readthedocs.io/en/latest/getting_started/index.html).
|
||||
|
||||
```
|
||||
llama stack build --template together --image-type venv
|
||||
llama stack build --distro together --image-type venv
|
||||
|
||||
llama stack run together
|
||||
```
|
||||
|
|
|
@ -13,7 +13,7 @@ from llama_stack.log import get_logger
|
|||
logger = get_logger(name=__name__, category="config_resolution")
|
||||
|
||||
|
||||
TEMPLATE_DIR = Path(__file__).parent.parent.parent.parent / "llama_stack" / "templates"
|
||||
DISTRO_DIR = Path(__file__).parent.parent.parent.parent / "llama_stack" / "distributions"
|
||||
|
||||
|
||||
class Mode(StrEnum):
|
||||
|
@ -21,15 +21,15 @@ class Mode(StrEnum):
|
|||
BUILD = "build"
|
||||
|
||||
|
||||
def resolve_config_or_template(
|
||||
config_or_template: str,
|
||||
def resolve_config_or_distro(
|
||||
config_or_distro: str,
|
||||
mode: Mode = Mode.RUN,
|
||||
) -> Path:
|
||||
"""
|
||||
Resolve a config/template argument to a concrete config file path.
|
||||
Resolve a config/distro argument to a concrete config file path.
|
||||
|
||||
Args:
|
||||
config_or_template: User input (file path, template name, or built distribution)
|
||||
config_or_distro: User input (file path, distribution name, or built distribution)
|
||||
mode: Mode resolving for ("run", "build", "server")
|
||||
|
||||
Returns:
|
||||
|
@ -40,86 +40,86 @@ def resolve_config_or_template(
|
|||
"""
|
||||
|
||||
# Strategy 1: Try as file path first
|
||||
config_path = Path(config_or_template)
|
||||
config_path = Path(config_or_distro)
|
||||
if config_path.exists() and config_path.is_file():
|
||||
logger.info(f"Using file path: {config_path}")
|
||||
return config_path.resolve()
|
||||
|
||||
# Strategy 2: Try as template name (if no .yaml extension)
|
||||
if not config_or_template.endswith(".yaml"):
|
||||
template_config = _get_template_config_path(config_or_template, mode)
|
||||
if template_config.exists():
|
||||
logger.info(f"Using template: {template_config}")
|
||||
return template_config
|
||||
# Strategy 2: Try as distribution name (if no .yaml extension)
|
||||
if not config_or_distro.endswith(".yaml"):
|
||||
distro_config = _get_distro_config_path(config_or_distro, mode)
|
||||
if distro_config.exists():
|
||||
logger.info(f"Using distribution: {distro_config}")
|
||||
return distro_config
|
||||
|
||||
# Strategy 3: Try as built distribution name
|
||||
distrib_config = DISTRIBS_BASE_DIR / f"llamastack-{config_or_template}" / f"{config_or_template}-{mode}.yaml"
|
||||
distrib_config = DISTRIBS_BASE_DIR / f"llamastack-{config_or_distro}" / f"{config_or_distro}-{mode}.yaml"
|
||||
if distrib_config.exists():
|
||||
logger.info(f"Using built distribution: {distrib_config}")
|
||||
return distrib_config
|
||||
|
||||
distrib_config = DISTRIBS_BASE_DIR / f"{config_or_template}" / f"{config_or_template}-{mode}.yaml"
|
||||
distrib_config = DISTRIBS_BASE_DIR / f"{config_or_distro}" / f"{config_or_distro}-{mode}.yaml"
|
||||
if distrib_config.exists():
|
||||
logger.info(f"Using built distribution: {distrib_config}")
|
||||
return distrib_config
|
||||
|
||||
# Strategy 4: Failed - provide helpful error
|
||||
raise ValueError(_format_resolution_error(config_or_template, mode))
|
||||
raise ValueError(_format_resolution_error(config_or_distro, mode))
|
||||
|
||||
|
||||
def _get_template_config_path(template_name: str, mode: Mode) -> Path:
|
||||
"""Get the config file path for a template."""
|
||||
return TEMPLATE_DIR / template_name / f"{mode}.yaml"
|
||||
def _get_distro_config_path(distro_name: str, mode: Mode) -> Path:
|
||||
"""Get the config file path for a distro."""
|
||||
return DISTRO_DIR / distro_name / f"{mode}.yaml"
|
||||
|
||||
|
||||
def _format_resolution_error(config_or_template: str, mode: Mode) -> str:
|
||||
def _format_resolution_error(config_or_distro: str, mode: Mode) -> str:
|
||||
"""Format a helpful error message for resolution failures."""
|
||||
from llama_stack.core.utils.config_dirs import DISTRIBS_BASE_DIR
|
||||
|
||||
template_path = _get_template_config_path(config_or_template, mode)
|
||||
distrib_path = DISTRIBS_BASE_DIR / f"llamastack-{config_or_template}" / f"{config_or_template}-{mode}.yaml"
|
||||
distrib_path2 = DISTRIBS_BASE_DIR / f"{config_or_template}" / f"{config_or_template}-{mode}.yaml"
|
||||
distro_path = _get_distro_config_path(config_or_distro, mode)
|
||||
distrib_path = DISTRIBS_BASE_DIR / f"llamastack-{config_or_distro}" / f"{config_or_distro}-{mode}.yaml"
|
||||
distrib_path2 = DISTRIBS_BASE_DIR / f"{config_or_distro}" / f"{config_or_distro}-{mode}.yaml"
|
||||
|
||||
available_templates = _get_available_templates()
|
||||
templates_str = ", ".join(available_templates) if available_templates else "none found"
|
||||
available_distros = _get_available_distros()
|
||||
distros_str = ", ".join(available_distros) if available_distros else "none found"
|
||||
|
||||
return f"""Could not resolve config or template '{config_or_template}'.
|
||||
return f"""Could not resolve config or distribution '{config_or_distro}'.
|
||||
|
||||
Tried the following locations:
|
||||
1. As file path: {Path(config_or_template).resolve()}
|
||||
2. As template: {template_path}
|
||||
1. As file path: {Path(config_or_distro).resolve()}
|
||||
2. As distribution: {distro_path}
|
||||
3. As built distribution: ({distrib_path}, {distrib_path2})
|
||||
|
||||
Available templates: {templates_str}
|
||||
Available distributions: {distros_str}
|
||||
|
||||
Did you mean one of these templates?
|
||||
{_format_template_suggestions(available_templates, config_or_template)}
|
||||
Did you mean one of these distributions?
|
||||
{_format_distro_suggestions(available_distros, config_or_distro)}
|
||||
"""
|
||||
|
||||
|
||||
def _get_available_templates() -> list[str]:
|
||||
"""Get list of available template names."""
|
||||
if not TEMPLATE_DIR.exists() and not DISTRIBS_BASE_DIR.exists():
|
||||
def _get_available_distros() -> list[str]:
|
||||
"""Get list of available distro names."""
|
||||
if not DISTRO_DIR.exists() and not DISTRIBS_BASE_DIR.exists():
|
||||
return []
|
||||
|
||||
return list(
|
||||
set(
|
||||
[d.name for d in TEMPLATE_DIR.iterdir() if d.is_dir() and not d.name.startswith(".")]
|
||||
[d.name for d in DISTRO_DIR.iterdir() if d.is_dir() and not d.name.startswith(".")]
|
||||
+ [d.name for d in DISTRIBS_BASE_DIR.iterdir() if d.is_dir() and not d.name.startswith(".")]
|
||||
)
|
||||
)
|
||||
|
||||
|
||||
def _format_template_suggestions(templates: list[str], user_input: str) -> str:
|
||||
"""Format template suggestions for error messages, showing closest matches first."""
|
||||
if not templates:
|
||||
return " (no templates found)"
|
||||
def _format_distro_suggestions(distros: list[str], user_input: str) -> str:
|
||||
"""Format distro suggestions for error messages, showing closest matches first."""
|
||||
if not distros:
|
||||
return " (no distros found)"
|
||||
|
||||
import difflib
|
||||
|
||||
# Get up to 3 closest matches with similarity threshold of 0.3 (lower = more permissive)
|
||||
close_matches = difflib.get_close_matches(user_input, templates, n=3, cutoff=0.3)
|
||||
display_templates = close_matches if close_matches else templates[:3]
|
||||
close_matches = difflib.get_close_matches(user_input, distros, n=3, cutoff=0.3)
|
||||
display_distros = close_matches if close_matches else distros[:3]
|
||||
|
||||
suggestions = [f" - {t}" for t in display_templates]
|
||||
suggestions = [f" - {d}" for d in display_distros]
|
||||
return "\n".join(suggestions)
|
||||
|
|
|
@ -5,7 +5,7 @@
|
|||
# the root directory of this source tree.
|
||||
|
||||
|
||||
from llama_stack.templates.template import DistributionTemplate
|
||||
from llama_stack.distributions.template import DistributionTemplate
|
||||
|
||||
from ..starter.starter import get_distribution_template as get_starter_distribution_template
|
||||
|
|
@ -12,10 +12,10 @@ from llama_stack.core.datatypes import (
|
|||
ShieldInput,
|
||||
ToolGroupInput,
|
||||
)
|
||||
from llama_stack.distributions.template import DistributionTemplate, RunConfigSettings
|
||||
from llama_stack.providers.inline.inference.sentence_transformers import (
|
||||
SentenceTransformersInferenceConfig,
|
||||
)
|
||||
from llama_stack.templates.template import DistributionTemplate, RunConfigSettings
|
||||
|
||||
|
||||
def get_distribution_template() -> DistributionTemplate:
|
|
@ -141,7 +141,7 @@ docker run \
|
|||
--pull always \
|
||||
-p $LLAMA_STACK_PORT:$LLAMA_STACK_PORT \
|
||||
-v $HOME/.llama:/root/.llama \
|
||||
-v ./llama_stack/templates/tgi/run-with-safety.yaml:/root/my-run.yaml \
|
||||
-v ./llama_stack/distributions/tgi/run-with-safety.yaml:/root/my-run.yaml \
|
||||
llamastack/distribution-{{ name }} \
|
||||
--config /root/my-run.yaml \
|
||||
--port $LLAMA_STACK_PORT \
|
||||
|
@ -157,7 +157,7 @@ docker run \
|
|||
Make sure you have done `pip install llama-stack` and have the Llama Stack CLI available.
|
||||
|
||||
```bash
|
||||
llama stack build --template {{ name }} --image-type conda
|
||||
llama stack build --distro {{ name }} --image-type conda
|
||||
llama stack run {{ name }}
|
||||
--port $LLAMA_STACK_PORT \
|
||||
--env INFERENCE_MODEL=$INFERENCE_MODEL \
|
|
@ -97,7 +97,7 @@ docker run \
|
|||
Make sure you have done `uv pip install llama-stack` and have the Llama Stack CLI available.
|
||||
|
||||
```bash
|
||||
llama stack build --template {{ name }} --image-type venv
|
||||
llama stack build --distro {{ name }} --image-type venv
|
||||
llama stack run distributions/{{ name }}/run.yaml \
|
||||
--port 8321 \
|
||||
--env INFERENCE_MODEL=meta-llama/Llama-3.2-3B-Instruct
|
|
@ -14,6 +14,7 @@ from llama_stack.core.datatypes import (
|
|||
ShieldInput,
|
||||
ToolGroupInput,
|
||||
)
|
||||
from llama_stack.distributions.template import DistributionTemplate, RunConfigSettings
|
||||
from llama_stack.providers.inline.inference.meta_reference import (
|
||||
MetaReferenceInferenceConfig,
|
||||
)
|
||||
|
@ -21,7 +22,6 @@ from llama_stack.providers.inline.inference.sentence_transformers import (
|
|||
SentenceTransformersInferenceConfig,
|
||||
)
|
||||
from llama_stack.providers.inline.vector_io.faiss.config import FaissVectorIOConfig
|
||||
from llama_stack.templates.template import DistributionTemplate, RunConfigSettings
|
||||
|
||||
|
||||
def get_distribution_template() -> DistributionTemplate:
|
|
@ -130,7 +130,7 @@ If you've set up your local development environment, you can also build the imag
|
|||
|
||||
```bash
|
||||
INFERENCE_MODEL=meta-llama/Llama-3.1-8b-Instruct
|
||||
llama stack build --template nvidia --image-type venv
|
||||
llama stack build --distro nvidia --image-type venv
|
||||
llama stack run ./run.yaml \
|
||||
--port 8321 \
|
||||
--env NVIDIA_API_KEY=$NVIDIA_API_KEY \
|
|
@ -7,12 +7,12 @@
|
|||
from pathlib import Path
|
||||
|
||||
from llama_stack.core.datatypes import BuildProvider, ModelInput, Provider, ShieldInput, ToolGroupInput
|
||||
from llama_stack.distributions.template import DistributionTemplate, RunConfigSettings, get_model_registry
|
||||
from llama_stack.providers.remote.datasetio.nvidia import NvidiaDatasetIOConfig
|
||||
from llama_stack.providers.remote.eval.nvidia import NVIDIAEvalConfig
|
||||
from llama_stack.providers.remote.inference.nvidia import NVIDIAConfig
|
||||
from llama_stack.providers.remote.inference.nvidia.models import MODEL_ENTRIES
|
||||
from llama_stack.providers.remote.safety.nvidia import NVIDIASafetyConfig
|
||||
from llama_stack.templates.template import DistributionTemplate, RunConfigSettings, get_model_registry
|
||||
|
||||
|
||||
def get_distribution_template() -> DistributionTemplate:
|
|
@ -16,6 +16,11 @@ from llama_stack.core.datatypes import (
|
|||
ShieldInput,
|
||||
ToolGroupInput,
|
||||
)
|
||||
from llama_stack.distributions.template import (
|
||||
DistributionTemplate,
|
||||
RunConfigSettings,
|
||||
get_model_registry,
|
||||
)
|
||||
from llama_stack.providers.inline.vector_io.sqlite_vec.config import (
|
||||
SQLiteVectorIOConfig,
|
||||
)
|
||||
|
@ -29,11 +34,6 @@ from llama_stack.providers.remote.vector_io.pgvector.config import (
|
|||
PGVectorVectorIOConfig,
|
||||
)
|
||||
from llama_stack.providers.utils.inference.model_registry import ProviderModelEntry
|
||||
from llama_stack.templates.template import (
|
||||
DistributionTemplate,
|
||||
RunConfigSettings,
|
||||
get_model_registry,
|
||||
)
|
||||
|
||||
|
||||
def get_inference_providers() -> tuple[list[Provider], dict[str, list[ProviderModelEntry]]]:
|
|
@ -13,15 +13,15 @@ from llama_stack.core.datatypes import (
|
|||
ShieldInput,
|
||||
ToolGroupInput,
|
||||
)
|
||||
from llama_stack.distributions.template import (
|
||||
DistributionTemplate,
|
||||
RunConfigSettings,
|
||||
)
|
||||
from llama_stack.providers.inline.inference.sentence_transformers import SentenceTransformersInferenceConfig
|
||||
from llama_stack.providers.remote.inference.vllm import VLLMInferenceAdapterConfig
|
||||
from llama_stack.providers.remote.vector_io.chroma.config import ChromaVectorIOConfig
|
||||
from llama_stack.providers.utils.kvstore.config import PostgresKVStoreConfig
|
||||
from llama_stack.providers.utils.sqlstore.sqlstore import PostgresSqlStoreConfig
|
||||
from llama_stack.templates.template import (
|
||||
DistributionTemplate,
|
||||
RunConfigSettings,
|
||||
)
|
||||
|
||||
|
||||
def get_distribution_template() -> DistributionTemplate:
|
|
@ -15,6 +15,10 @@ from llama_stack.core.datatypes import (
|
|||
ToolGroupInput,
|
||||
)
|
||||
from llama_stack.core.utils.dynamic import instantiate_class_type
|
||||
from llama_stack.distributions.template import (
|
||||
DistributionTemplate,
|
||||
RunConfigSettings,
|
||||
)
|
||||
from llama_stack.providers.datatypes import RemoteProviderSpec
|
||||
from llama_stack.providers.inline.files.localfs.config import LocalfsFilesImplConfig
|
||||
from llama_stack.providers.inline.inference.sentence_transformers import (
|
||||
|
@ -33,10 +37,6 @@ from llama_stack.providers.remote.vector_io.pgvector.config import (
|
|||
PGVectorVectorIOConfig,
|
||||
)
|
||||
from llama_stack.providers.utils.sqlstore.sqlstore import PostgresSqlStoreConfig
|
||||
from llama_stack.templates.template import (
|
||||
DistributionTemplate,
|
||||
RunConfigSettings,
|
||||
)
|
||||
|
||||
|
||||
def _get_config_for_provider(provider_spec: ProviderSpec) -> dict[str, Any]:
|
|
@ -8,12 +8,12 @@ from pathlib import Path
|
|||
|
||||
from llama_stack.apis.models import ModelType
|
||||
from llama_stack.core.datatypes import BuildProvider, ModelInput, Provider, ToolGroupInput
|
||||
from llama_stack.distributions.template import DistributionTemplate, RunConfigSettings, get_model_registry
|
||||
from llama_stack.providers.inline.inference.sentence_transformers import (
|
||||
SentenceTransformersInferenceConfig,
|
||||
)
|
||||
from llama_stack.providers.remote.inference.watsonx import WatsonXConfig
|
||||
from llama_stack.providers.remote.inference.watsonx.models import MODEL_ENTRIES
|
||||
from llama_stack.templates.template import DistributionTemplate, RunConfigSettings, get_model_registry
|
||||
|
||||
|
||||
def get_distribution_template() -> DistributionTemplate:
|
|
@ -20,7 +20,7 @@ This provider enables dataset management using NVIDIA's NeMo Customizer service.
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type venv
|
||||
llama stack build --distro nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
|
@ -18,7 +18,7 @@ This provider enables running inference using NVIDIA NIM.
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type venv
|
||||
llama stack build --distro nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
|
@ -22,7 +22,7 @@ This provider enables fine-tuning of LLMs using NVIDIA's NeMo Customizer service
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type venv
|
||||
llama stack build --distro nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
|
@ -19,7 +19,7 @@ This provider enables safety checks and guardrails for LLM interactions using NV
|
|||
Build the NVIDIA environment:
|
||||
|
||||
```bash
|
||||
llama stack build --template nvidia --image-type venv
|
||||
llama stack build --distro nvidia --image-type venv
|
||||
```
|
||||
|
||||
### Basic Usage using the LlamaStack Python Client
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue