mirror/OpenLLM
1
0
Fork 0
mirror of https://github.com/bentoml/OpenLLM.git synced 2026-03-11 11:39:52 -04:00
Code Issues Packages Projects Releases Wiki Activity

infra: using ruff formatter (#594)

Browse Source 
Signed-off-by: Aaron <29749331+aarnphm@users.noreply.github.com>
This commit is contained in:
Aaron Pham
2023-11-09 12:44:05 -05:00
committed by GitHub
parent 021fd453b9
commit ac377fe490
102 changed files with 5577 additions and 2540 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
openllm-python/src/openllm/protocol/__init__.py
View File

@@ -2,12 +2,14 @@
Currently support OpenAI compatible API.
"""
from __future__ import annotations
import os
import typing as t
from openllm_core.utils import LazyModule
_import_structure: dict[str, list[str]] = {'openai': []}
if t.TYPE_CHECKING:

3
openllm-python/src/openllm/protocol/hf.py
View File

@@ -3,15 +3,18 @@ import typing as t
import attr
@attr.define
class AgentRequest:
inputs: str
parameters: t.Dict[str, t.Any]
@attr.define
class AgentResponse:
generated_text: str
@attr.define
class HFErrorResponse:
error_code: int

32
openllm-python/src/openllm/protocol/openai.py
View File

@@ -8,6 +8,7 @@ import openllm_core
from openllm_core.utils import converter
@attr.define
class ErrorResponse:
message: str
@@ -16,6 +17,7 @@ class ErrorResponse:
param: t.Optional[str] = None
code: t.Optional[str] = None
@attr.define
class CompletionRequest:
prompt: str
@@ -37,6 +39,7 @@ class CompletionRequest:
top_k: t.Optional[int] = attr.field(default=None)
best_of: t.Optional[int] = attr.field(default=1)
@attr.define
class ChatCompletionRequest:
messages: t.List[t.Dict[str, str]]
@@ -57,6 +60,7 @@ class ChatCompletionRequest:
top_k: t.Optional[int] = attr.field(default=None)
best_of: t.Optional[int] = attr.field(default=1)
@attr.define
class LogProbs:
text_offset: t.List[int] = attr.field(default=attr.Factory(list))
@@ -64,12 +68,14 @@ class LogProbs:
tokens: t.List[str] = attr.field(default=attr.Factory(list))
top_logprobs: t.List[t.Dict[str, t.Any]] = attr.field(default=attr.Factory(list))
@attr.define
class UsageInfo:
prompt_tokens: int = attr.field(default=0)
completion_tokens: int = attr.field(default=0)
total_tokens: int = attr.field(default=0)
@attr.define
class CompletionResponseChoice:
index: int
@@ -77,6 +83,7 @@ class CompletionResponseChoice:
logprobs: t.Optional[LogProbs] = None
finish_reason: t.Optional[str] = None
@attr.define
class CompletionResponseStreamChoice:
index: int
@@ -84,6 +91,7 @@ class CompletionResponseStreamChoice:
logprobs: t.Optional[LogProbs] = None
finish_reason: t.Optional[str] = None
@attr.define
class CompletionStreamResponse:
model: str
@@ -92,6 +100,7 @@ class CompletionStreamResponse:
id: str = attr.field(default=attr.Factory(lambda: openllm_core.utils.gen_random_uuid('cmpl')))
created: int = attr.field(default=attr.Factory(lambda: int(time.monotonic())))
@attr.define
class CompletionResponse:
choices: t.List[CompletionResponseChoice]
@@ -101,32 +110,39 @@ class CompletionResponse:
id: str = attr.field(default=attr.Factory(lambda: openllm_core.utils.gen_random_uuid('cmpl')))
created: int = attr.field(default=attr.Factory(lambda: int(time.monotonic())))
LiteralRole = t.Literal['system', 'user', 'assistant']
@attr.define
class Delta:
role: t.Optional[LiteralRole] = None
content: t.Optional[str] = None
@attr.define
class ChatMessage:
role: LiteralRole
content: str
converter.register_unstructure_hook(ChatMessage, lambda msg: {'role': msg.role, 'content': msg.content})
@attr.define
class ChatCompletionResponseStreamChoice:
index: int
delta: Delta
finish_reason: t.Optional[str] = attr.field(default=None)
@attr.define
class ChatCompletionResponseChoice:
index: int
message: ChatMessage
finish_reason: t.Optional[str] = attr.field(default=None)
@attr.define
class ChatCompletionResponse:
choices: t.List[ChatCompletionResponseChoice]
@@ -136,6 +152,7 @@ class ChatCompletionResponse:
created: int = attr.field(default=attr.Factory(lambda: int(time.monotonic())))
usage: UsageInfo = attr.field(default=attr.Factory(lambda: UsageInfo()))
@attr.define
class ChatCompletionStreamResponse:
choices: t.List[ChatCompletionResponseStreamChoice]
@@ -144,6 +161,7 @@ class ChatCompletionStreamResponse:
id: str = attr.field(default=attr.Factory(lambda: openllm_core.utils.gen_random_uuid('chatcmpl')))
created: int = attr.field(default=attr.Factory(lambda: int(time.monotonic())))
@attr.define
class ModelCard:
id: str
@@ -151,19 +169,25 @@ class ModelCard:
created: int = attr.field(default=attr.Factory(lambda: int(time.monotonic())))
owned_by: str = 'na'
@attr.define
class ModelList:
object: str = 'list'
data: t.List[ModelCard] = attr.field(factory=list)
async def get_conversation_prompt(request: ChatCompletionRequest, llm_config: openllm_core.LLMConfig) -> str:
conv = llm_config.get_conversation_template()
for message in request.messages:
msg_role = message['role']
if msg_role == 'system': conv.set_system_message(message['content'])
elif msg_role == 'user': conv.append_message(conv.roles[0], message['content'])
elif msg_role == 'assistant': conv.append_message(conv.roles[1], message['content'])
else: raise ValueError(f'Unknown role: {msg_role}')
if msg_role == 'system':
conv.set_system_message(message['content'])
elif msg_role == 'user':
conv.append_message(conv.roles[0], message['content'])
elif msg_role == 'assistant':
conv.append_message(conv.roles[1], message['content'])
else:
raise ValueError(f'Unknown role: {msg_role}')
# Add a blank message for the assistant.
conv.append_message(conv.roles[1], '')
return conv.get_prompt()