Skip to content

Commit 727d60c

Browse files
abetlenabetlen
committed
misc: Format
1 parent 0d37ce5 commit 727d60c

File tree

5 files changed

+44
-39
lines changed

5 files changed

+44
-39
lines changed

‎llama_cpp/server/app.py‎

Lines changed: 26 additions & 24 deletions
Original file line numberDiff line numberDiff line change
@@ -199,8 +199,8 @@ async def authenticate(
199199
@router.post(
200200
"/v1/completions",
201201
summary="Completion",
202-
dependencies=[Depends(authenticate)],
203-
response_model= Union[
202+
dependencies=[Depends(authenticate)],
203+
response_model=Union[
204204
llama_cpp.CreateCompletionResponse,
205205
str,
206206
],
@@ -211,19 +211,19 @@ async def authenticate(
211211
"application/json": {
212212
"schema": {
213213
"anyOf": [
214-
{"$ref": "#/components/schemas/CreateCompletionResponse"}
214+
{"$ref": "#/components/schemas/CreateCompletionResponse"}
215215
],
216216
"title": "Completion response, when stream=False",
217217
}
218218
},
219-
"text/event-stream":{
220-
"schema": {
221-
"type": "string",
222-
"title": "Server Side Streaming response, when stream=True. " +
223-
"See SSE format: https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format", # noqa: E501
224-
"example": """data: {... see CreateCompletionResponse ...} \\n\\n data: ... \\n\\n ... data: [DONE]"""
219+
"text/event-stream": {
220+
"schema": {
221+
"type": "string",
222+
"title": "Server Side Streaming response, when stream=True. "
223+
+ "See SSE format: https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format", # noqa: E501
224+
"example": """data: {... see CreateCompletionResponse ...} \\n\\n data: ... \\n\\n ... data: [DONE]""",
225225
}
226-
}
226+
},
227227
},
228228
}
229229
},
@@ -290,7 +290,7 @@ def iterator() -> Iterator[llama_cpp.CreateCompletionStreamResponse]:
290290
inner_send_chan=send_chan,
291291
iterator=iterator(),
292292
),
293-
sep='\n',
293+
sep="\n",
294294
)
295295
else:
296296
return iterator_or_completion
@@ -310,30 +310,32 @@ async def create_embedding(
310310

311311

312312
@router.post(
313-
"/v1/chat/completions", summary="Chat", dependencies=[Depends(authenticate)],
314-
response_model= Union[
315-
llama_cpp.ChatCompletion, str
316-
],
313+
"/v1/chat/completions",
314+
summary="Chat",
315+
dependencies=[Depends(authenticate)],
316+
response_model=Union[llama_cpp.ChatCompletion, str],
317317
responses={
318318
"200": {
319319
"description": "Successful Response",
320320
"content": {
321321
"application/json": {
322322
"schema": {
323323
"anyOf": [
324-
{"$ref": "#/components/schemas/CreateChatCompletionResponse"}
324+
{
325+
"$ref": "#/components/schemas/CreateChatCompletionResponse"
326+
}
325327
],
326328
"title": "Completion response, when stream=False",
327329
}
328330
},
329-
"text/event-stream":{
330-
"schema": {
331-
"type": "string",
332-
"title": "Server Side Streaming response, when stream=True" +
333-
"See SSE format: https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format", # noqa: E501
334-
"example": """data: {... see CreateChatCompletionResponse ...} \\n\\n data: ... \\n\\n ... data: [DONE]"""
331+
"text/event-stream": {
332+
"schema": {
333+
"type": "string",
334+
"title": "Server Side Streaming response, when stream=True"
335+
+ "See SSE format: https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format", # noqa: E501
336+
"example": """data: {... see CreateChatCompletionResponse ...} \\n\\n data: ... \\n\\n ... data: [DONE]""",
335337
}
336-
}
338+
},
337339
},
338340
}
339341
},
@@ -383,7 +385,7 @@ def iterator() -> Iterator[llama_cpp.ChatCompletionChunk]:
383385
inner_send_chan=send_chan,
384386
iterator=iterator(),
385387
),
386-
sep='\n',
388+
sep="\n",
387389
)
388390
else:
389391
return iterator_or_completion

‎llama_cpp/server/errors.py‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -22,6 +22,7 @@
2222
CreateChatCompletionRequest,
2323
)
2424

25+
2526
class ErrorResponse(TypedDict):
2627
"""OpenAI style error response"""
2728

@@ -75,7 +76,7 @@ def context_length_exceeded(
7576
(completion_tokens or 0) + prompt_tokens,
7677
prompt_tokens,
7778
completion_tokens,
78-
), # type: ignore
79+
), # type: ignore
7980
type="invalid_request_error",
8081
param="messages",
8182
code="context_length_exceeded",
@@ -207,4 +208,3 @@ async def custom_route_handler(request: Request) -> Response:
207208
)
208209

209210
return custom_route_handler
210-

‎llama_cpp/server/model.py‎

Lines changed: 11 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -88,15 +88,15 @@ def load_llama_from_model_settings(settings: ModelSettings) -> llama_cpp.Llama:
8888
assert (
8989
settings.hf_tokenizer_config_path is not None
9090
), "hf_tokenizer_config_path must be set for hf-tokenizer-config"
91-
chat_handler = (
92-
llama_cpp.llama_chat_format.hf_tokenizer_config_to_chat_completion_handler(
93-
json.load(open(settings.hf_tokenizer_config_path))
94-
)
91+
chat_handler = llama_cpp.llama_chat_format.hf_tokenizer_config_to_chat_completion_handler(
92+
json.load(open(settings.hf_tokenizer_config_path))
9593
)
9694

9795
tokenizer: Optional[llama_cpp.BaseLlamaTokenizer] = None
9896
if settings.hf_pretrained_model_name_or_path is not None:
99-
tokenizer = llama_tokenizer.LlamaHFTokenizer.from_pretrained(settings.hf_pretrained_model_name_or_path)
97+
tokenizer = llama_tokenizer.LlamaHFTokenizer.from_pretrained(
98+
settings.hf_pretrained_model_name_or_path
99+
)
100100

101101
draft_model = None
102102
if settings.draft_model is not None:
@@ -120,17 +120,20 @@ def load_llama_from_model_settings(settings: ModelSettings) -> llama_cpp.Llama:
120120
kv_overrides[key] = float(value)
121121
else:
122122
raise ValueError(f"Unknown value type {value_type}")
123-
123+
124124
import functools
125125

126126
kwargs = {}
127127

128128
if settings.hf_model_repo_id is not None:
129-
create_fn = functools.partial(llama_cpp.Llama.from_pretrained, repo_id=settings.hf_model_repo_id, filename=settings.model)
129+
create_fn = functools.partial(
130+
llama_cpp.Llama.from_pretrained,
131+
repo_id=settings.hf_model_repo_id,
132+
filename=settings.model,
133+
)
130134
else:
131135
create_fn = llama_cpp.Llama
132136
kwargs["model_path"] = settings.model
133-
134137

135138
_model = create_fn(
136139
**kwargs,

‎llama_cpp/server/settings.py‎

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -74,7 +74,9 @@ class ModelSettings(BaseSettings):
7474
ge=0,
7575
description="The number of threads to use when batch processing.",
7676
)
77-
rope_scaling_type: int = Field(default=llama_cpp.LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED)
77+
rope_scaling_type: int = Field(
78+
default=llama_cpp.LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED
79+
)
7880
rope_freq_base: float = Field(default=0.0, description="RoPE base frequency")
7981
rope_freq_scale: float = Field(
8082
default=0.0, description="RoPE frequency scaling factor"
@@ -193,6 +195,4 @@ class Settings(ServerSettings, ModelSettings):
193195
class ConfigFileSettings(ServerSettings):
194196
"""Configuration file format settings."""
195197

196-
models: List[ModelSettings] = Field(
197-
default=[], description="Model configs"
198-
)
198+
models: List[ModelSettings] = Field(default=[], description="Model configs")

‎llama_cpp/server/types.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -110,7 +110,7 @@ class CreateCompletionRequest(BaseModel):
110110
default=None,
111111
description="A suffix to append to the generated text. If None, no suffix is appended. Useful for chatbots.",
112112
)
113-
max_tokens: Optional[int] = Field(
113+
max_tokens: Optional[int] = Field(
114114
default=16, ge=0, description="The maximum number of tokens to generate."
115115
)
116116
temperature: float = temperature_field

0 commit comments

Comments
 (0)