phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-05 12:21:52 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: Adding OpenAI Compatible Prompts API

Browse source 
Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>
This commit is contained in:
Francisco Javier Arceo 2025-09-03 14:14:54 -04:00
parent 30117dea22
commit 8b00883abd
181 changed files with 21356 additions and 10332 deletions
Download patch file Download diff file Expand all files Collapse all files

25
llama_stack/core/stack.py
View file

@ -24,6 +24,7 @@ from llama_stack.apis.inference import Inference
from llama_stack.apis.inspect import Inspect
from llama_stack.apis.models import Models
from llama_stack.apis.post_training import PostTraining
from llama_stack.apis.prompts import Prompts
from llama_stack.apis.providers import Providers
from llama_stack.apis.safety import Safety
from llama_stack.apis.scoring import Scoring
@ -37,6 +38,7 @@ from llama_stack.apis.vector_io import VectorIO
from llama_stack.core.datatypes import Provider, StackRunConfig
from llama_stack.core.distribution import get_provider_registry
from llama_stack.core.inspect import DistributionInspectConfig, DistributionInspectImpl
from llama_stack.core.prompts.prompts import PromptServiceConfig, PromptServiceImpl
from llama_stack.core.providers import ProviderImpl, ProviderImplConfig
from llama_stack.core.resolver import ProviderRegistry, resolve_impls
from llama_stack.core.routing_tables.common import CommonRoutingTableImpl
@ -44,6 +46,7 @@ from llama_stack.core.store.registry import create_dist_registry
from llama_stack.core.utils.dynamic import instantiate_class_type
from llama_stack.log import get_logger
from llama_stack.providers.datatypes import Api
from llama_stack.providers.utils.kvstore.config import SqliteKVStoreConfig
logger = get_logger(name=__name__, category="core")
@ -72,6 +75,7 @@ class LlamaStack(
ToolRuntime,
RAGToolRuntime,
Files,
Prompts,
):
pass
@ -105,12 +109,12 @@ async def register_resources(run_config: StackRunConfig, impls: dict[Api, Any]):
method = getattr(impls[api], register_method)
for obj in objects:
logger.debug(f"registering {rsrc.capitalize()} {obj} for provider {obj.provider_id}")
# Do not register models on disabled providers
if hasattr(obj, "provider_id") and (not obj.provider_id or obj.provider_id == "__disabled__"):
logger.debug(f"Skipping {rsrc.capitalize()} registration for disabled provider.")
continue
if hasattr(obj, "provider_id"):
# Do not register models on disabled providers
if not obj.provider_id or obj.provider_id == "__disabled__":
logger.debug(f"Skipping {rsrc.capitalize()} registration for disabled provider.")
continue
logger.debug(f"registering {rsrc.capitalize()} {obj} for provider {obj.provider_id}")
# we want to maintain the type information in arguments to method.
# instead of method(**obj.model_dump()), which may convert a typed attr to a dict,
@ -305,6 +309,12 @@ def add_internal_implementations(impls: dict[Api, Any], run_config: StackRunConf
)
impls[Api.providers] = providers_impl
prompts_impl = PromptServiceImpl(
PromptServiceConfig(kvstore=SqliteKVStoreConfig(db_path=os.path.expanduser("~/.llama-stack/prompts.db"))),
deps=impls,
)
impls[Api.prompts] = prompts_impl
# Produces a stack of providers for the given run config. Not all APIs may be
# asked for in the run config.
@ -329,6 +339,9 @@ async def construct_stack(
# Add internal implementations after all other providers are resolved
add_internal_implementations(impls, run_config)
if Api.prompts in impls:
await impls[Api.prompts].initialize()
await register_resources(run_config, impls)
await refresh_registry_once(impls)