phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-07-30 23:51:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

change register signature to make params required

Browse source
This commit is contained in:
Dinesh Yeduguru 2024-11-11 11:49:07 -08:00
parent 4206f07ada
commit 4369c2d0b6
5 changed files with 43 additions and 29 deletions
Download patch file Download diff file Expand all files Collapse all files

33
llama_stack/apis/memory_banks/memory_banks.py
View file

@ -65,15 +65,39 @@ class GraphMemoryBank(MemoryBank):
@json_schema_type
class VectorMemoryBankParams(BaseModel):
type: Literal[MemoryBankType.vector.value] = MemoryBankType.vector.value
memory_bank_type: Literal[MemoryBankType.vector.value] = MemoryBankType.vector.value
embedding_model: str
chunk_size_in_tokens: int
overlap_size_in_tokens: Optional[int] = None
@json_schema_type
class KeyValueMemoryBankParams(BaseModel):
memory_bank_type: Literal[MemoryBankType.keyvalue.value] = (
MemoryBankType.keyvalue.value
)
@json_schema_type
class KeywordMemoryBankParams(BaseModel):
memory_bank_type: Literal[MemoryBankType.keyword.value] = (
MemoryBankType.keyword.value
)
@json_schema_type
class GraphMemoryBankParams(BaseModel):
memory_bank_type: Literal[MemoryBankType.graph.value] = MemoryBankType.graph.value
BankParams = Annotated[
Union[VectorMemoryBankParams],
Field(discriminator="type"),
Union[
VectorMemoryBankParams,
KeyValueMemoryBankParams,
KeywordMemoryBankParams,
GraphMemoryBankParams,
],
Field(discriminator="memory_bank_type"),
]
@ -89,8 +113,7 @@ class MemoryBanks(Protocol):
async def register_memory_bank(
self,
memory_bank_id: str,
memory_bank_type: MemoryBankType,
params: BankParams,
provider_id: Optional[str] = None,
provider_memorybank_id: Optional[str] = None,
params: Optional[BankParams] = None,
) -> MemoryBank: ...

8
llama_stack/distribution/routers/routers.py
View file

@ -7,7 +7,7 @@
from typing import Any, AsyncGenerator, Dict, List, Optional
from llama_stack.apis.datasetio.datasetio import DatasetIO
from llama_stack.apis.memory_banks.memory_banks import BankParams, MemoryBankType
from llama_stack.apis.memory_banks.memory_banks import BankParams
from llama_stack.distribution.datatypes import RoutingTable
from llama_stack.apis.memory import * # noqa: F403
from llama_stack.apis.inference import * # noqa: F403
@ -35,17 +35,15 @@ class MemoryRouter(Memory):
async def register_memory_bank(
self,
memory_bank_id: str,
memory_bank_type: MemoryBankType,
params: BankParams,
provider_id: Optional[str] = None,
provider_memorybank_id: Optional[str] = None,
params: Optional[BankParams] = None,
) -> None:
await self.routing_table.register_memory_bank(
memory_bank_id,
memory_bank_type,
params,
provider_id,
provider_memorybank_id,
params,
)
async def insert_documents(

6
llama_stack/distribution/routers/routing_tables.py
View file

@ -272,10 +272,9 @@ class MemoryBanksRoutingTable(CommonRoutingTableImpl, MemoryBanks):
async def register_memory_bank(
self,
memory_bank_id: str,
memory_bank_type: MemoryBankType,
params: BankParams,
provider_id: Optional[str] = None,
provider_memorybank_id: Optional[str] = None,
params: Optional[BankParams] = None,
) -> MemoryBank:
if provider_memorybank_id is None:
provider_memorybank_id = memory_bank_id
@ -289,10 +288,9 @@ class MemoryBanksRoutingTable(CommonRoutingTableImpl, MemoryBanks):
)
memory_bank = build_memory_bank(
memory_bank_id,
memory_bank_type,
params,
provider_id,
provider_memorybank_id,
params,
)
await self.register_object(memory_bank)
return memory_bank

22
llama_stack/distribution/utils/memory_bank_utils.py
View file

@ -4,7 +4,6 @@
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from typing import Optional
from llama_stack.apis.memory_banks.memory_banks import (
BankParams,
@ -20,43 +19,42 @@ from llama_stack.apis.memory_banks.memory_banks import (
def build_memory_bank(
memory_bank_id: str,
memory_bank_type: MemoryBankType,
params: BankParams,
provider_id: str,
provider_memorybank_id: str,
params: Optional[BankParams] = None,
) -> MemoryBank:
if memory_bank_type == MemoryBankType.vector.value:
if params.memory_bank_type == MemoryBankType.vector.value:
assert isinstance(params, VectorMemoryBankParams)
memory_bank = VectorMemoryBank(
identifier=memory_bank_id,
provider_id=provider_id,
provider_resource_id=provider_memorybank_id,
memory_bank_type=memory_bank_type,
memory_bank_type=params.memory_bank_type,
embedding_model=params.embedding_model,
chunk_size_in_tokens=params.chunk_size_in_tokens,
overlap_size_in_tokens=params.overlap_size_in_tokens,
)
elif memory_bank_type == MemoryBankType.keyvalue.value:
elif params.memory_bank_type == MemoryBankType.keyvalue.value:
memory_bank = KeyValueMemoryBank(
identifier=memory_bank_id,
provider_id=provider_id,
provider_resource_id=provider_memorybank_id,
memory_bank_type=memory_bank_type,
memory_bank_type=params.memory_bank_type,
)
elif memory_bank_type == MemoryBankType.keyword.value:
elif params.memory_bank_type == MemoryBankType.keyword.value:
memory_bank = KeywordMemoryBank(
identifier=memory_bank_id,
provider_id=provider_id,
provider_resource_id=provider_memorybank_id,
memory_bank_type=memory_bank_type,
memory_bank_type=params.memory_bank_type,
)
elif memory_bank_type == MemoryBankType.graph.value:
elif params.memory_bank_type == MemoryBankType.graph.value:
memory_bank = GraphMemoryBank(
identifier=memory_bank_id,
provider_id=provider_id,
provider_resource_id=provider_memorybank_id,
memory_bank_type=memory_bank_type,
memory_bank_type=params.memory_bank_type,
)
else:
raise ValueError(f"Unknown memory bank type: {memory_bank_type}")
raise ValueError(f"Unknown memory bank type: {params.memory_bank_type}")
return memory_bank

3
llama_stack/providers/tests/memory/test_memory.py
View file

@ -47,7 +47,6 @@ async def register_memory_bank(banks_impl: MemoryBanks):
return await banks_impl.register_memory_bank(
memory_bank_id="test_bank",
memory_bank_type="vector",
params=VectorMemoryBankParams(
embedding_model="all-MiniLM-L6-v2",
chunk_size_in_tokens=512,
@ -74,7 +73,6 @@ class TestMemory:
bank = await banks_impl.register_memory_bank(
memory_bank_id="test_bank_no_provider",
memory_bank_type="vector",
params=VectorMemoryBankParams(
embedding_model="all-MiniLM-L6-v2",
chunk_size_in_tokens=512,
@ -88,7 +86,6 @@ class TestMemory:
# register same memory bank with same id again will fail
await banks_impl.register_memory_bank(
memory_bank_id="test_bank_no_provider",
memory_bank_type="vector",
params=VectorMemoryBankParams(
embedding_model="all-MiniLM-L6-v2",
chunk_size_in_tokens=512,