Skip to content

vllm.entrypoints.cli.openai

ChatCommand

Bases: CLISubcommand

The chat subcommand for the vLLM CLI.

Source code in vllm/entrypoints/cli/openai.py 
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
class ChatCommand(CLISubcommand):
    """The `chat` subcommand for the vLLM CLI."""

    name = "chat"

    @staticmethod
    def cmd(args: argparse.Namespace) -> None:
        model_name, client = _interactive_cli(args)
        system_prompt = args.system_prompt
        conversation: list[ChatCompletionMessageParam] = []

        if system_prompt is not None:
            conversation.append({"role": "system", "content": system_prompt})

        if args.quick:
            conversation.append({"role": "user", "content": args.quick})

            stream = client.chat.completions.create(
                model=model_name, messages=conversation, stream=True
            )
            output = _print_chat_stream(stream)
            conversation.append({"role": "assistant", "content": output})
            return

        print("Please enter a message for the chat model:")
        while True:
            try:
                input_message = input("> ")
            except EOFError:
                break
            conversation.append({"role": "user", "content": input_message})

            stream = client.chat.completions.create(
                model=model_name, messages=conversation, stream=True
            )
            output = _print_chat_stream(stream)
            conversation.append({"role": "assistant", "content": output})

    @staticmethod
    def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
        """Add CLI arguments for the chat command."""
        _add_query_options(parser)
        parser.add_argument(
            "--system-prompt",
            type=str,
            default=None,
            help=(
                "The system prompt to be added to the chat template, "
                "used for models that support system prompts."
            ),
        )
        parser.add_argument(
            "-q",
            "--quick",
            type=str,
            metavar="MESSAGE",
            help=("Send a single prompt as MESSAGE and print the response, then exit."),
        )
        return parser

    def subparser_init(
        self, subparsers: argparse._SubParsersAction
    ) -> FlexibleArgumentParser:
        parser = subparsers.add_parser(
            "chat",
            help="Generate chat completions via the running API server.",
            description="Generate chat completions via the running API server.",
            usage="vllm chat [options]",
        )
        return ChatCommand.add_cli_args(parser)

name class-attribute instance-attribute 

name = 'chat'

add_cli_args staticmethod 

add_cli_args(
    parser: FlexibleArgumentParser,
) -> FlexibleArgumentParser

Add CLI arguments for the chat command.

Source code in vllm/entrypoints/cli/openai.py 
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
@staticmethod
def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
    """Add CLI arguments for the chat command."""
    _add_query_options(parser)
    parser.add_argument(
        "--system-prompt",
        type=str,
        default=None,
        help=(
            "The system prompt to be added to the chat template, "
            "used for models that support system prompts."
        ),
    )
    parser.add_argument(
        "-q",
        "--quick",
        type=str,
        metavar="MESSAGE",
        help=("Send a single prompt as MESSAGE and print the response, then exit."),
    )
    return parser

cmd staticmethod 

cmd(args: Namespace) -> None
Source code in vllm/entrypoints/cli/openai.py 
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
@staticmethod
def cmd(args: argparse.Namespace) -> None:
    model_name, client = _interactive_cli(args)
    system_prompt = args.system_prompt
    conversation: list[ChatCompletionMessageParam] = []

    if system_prompt is not None:
        conversation.append({"role": "system", "content": system_prompt})

    if args.quick:
        conversation.append({"role": "user", "content": args.quick})

        stream = client.chat.completions.create(
            model=model_name, messages=conversation, stream=True
        )
        output = _print_chat_stream(stream)
        conversation.append({"role": "assistant", "content": output})
        return

    print("Please enter a message for the chat model:")
    while True:
        try:
            input_message = input("> ")
        except EOFError:
            break
        conversation.append({"role": "user", "content": input_message})

        stream = client.chat.completions.create(
            model=model_name, messages=conversation, stream=True
        )
        output = _print_chat_stream(stream)
        conversation.append({"role": "assistant", "content": output})

subparser_init 

subparser_init(
    subparsers: _SubParsersAction,
) -> FlexibleArgumentParser
Source code in vllm/entrypoints/cli/openai.py 
177
178
179
180
181
182
183
184
185
186
def subparser_init(
    self, subparsers: argparse._SubParsersAction
) -> FlexibleArgumentParser:
    parser = subparsers.add_parser(
        "chat",
        help="Generate chat completions via the running API server.",
        description="Generate chat completions via the running API server.",
        usage="vllm chat [options]",
    )
    return ChatCommand.add_cli_args(parser)

CompleteCommand

Bases: CLISubcommand

The complete subcommand for the vLLM CLI.

Source code in vllm/entrypoints/cli/openai.py 
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
class CompleteCommand(CLISubcommand):
    """The `complete` subcommand for the vLLM CLI."""

    name = "complete"

    @staticmethod
    def cmd(args: argparse.Namespace) -> None:
        model_name, client = _interactive_cli(args)

        if args.quick:
            stream = client.completions.create(
                model=model_name, prompt=args.quick, stream=True
            )
            _print_completion_stream(stream)
            return

        print("Please enter prompt to complete:")
        while True:
            try:
                input_prompt = input("> ")
            except EOFError:
                break
            stream = client.completions.create(
                model=model_name, prompt=input_prompt, stream=True
            )
            _print_completion_stream(stream)

    @staticmethod
    def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
        """Add CLI arguments for the complete command."""
        _add_query_options(parser)
        parser.add_argument(
            "-q",
            "--quick",
            type=str,
            metavar="PROMPT",
            help="Send a single prompt and print the completion output, then exit.",
        )
        return parser

    def subparser_init(
        self, subparsers: argparse._SubParsersAction
    ) -> FlexibleArgumentParser:
        parser = subparsers.add_parser(
            "complete",
            help=(
                "Generate text completions based on the given prompt "
                "via the running API server."
            ),
            description=(
                "Generate text completions based on the given prompt "
                "via the running API server."
            ),
            usage="vllm complete [options]",
        )
        return CompleteCommand.add_cli_args(parser)

name class-attribute instance-attribute 

name = 'complete'

add_cli_args staticmethod 

add_cli_args(
    parser: FlexibleArgumentParser,
) -> FlexibleArgumentParser

Add CLI arguments for the complete command.

Source code in vllm/entrypoints/cli/openai.py 
216
217
218
219
220
221
222
223
224
225
226
227
@staticmethod
def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
    """Add CLI arguments for the complete command."""
    _add_query_options(parser)
    parser.add_argument(
        "-q",
        "--quick",
        type=str,
        metavar="PROMPT",
        help="Send a single prompt and print the completion output, then exit.",
    )
    return parser

cmd staticmethod 

cmd(args: Namespace) -> None
Source code in vllm/entrypoints/cli/openai.py 
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
@staticmethod
def cmd(args: argparse.Namespace) -> None:
    model_name, client = _interactive_cli(args)

    if args.quick:
        stream = client.completions.create(
            model=model_name, prompt=args.quick, stream=True
        )
        _print_completion_stream(stream)
        return

    print("Please enter prompt to complete:")
    while True:
        try:
            input_prompt = input("> ")
        except EOFError:
            break
        stream = client.completions.create(
            model=model_name, prompt=input_prompt, stream=True
        )
        _print_completion_stream(stream)

subparser_init 

subparser_init(
    subparsers: _SubParsersAction,
) -> FlexibleArgumentParser
Source code in vllm/entrypoints/cli/openai.py 
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
def subparser_init(
    self, subparsers: argparse._SubParsersAction
) -> FlexibleArgumentParser:
    parser = subparsers.add_parser(
        "complete",
        help=(
            "Generate text completions based on the given prompt "
            "via the running API server."
        ),
        description=(
            "Generate text completions based on the given prompt "
            "via the running API server."
        ),
        usage="vllm complete [options]",
    )
    return CompleteCommand.add_cli_args(parser)

_add_query_options 

_add_query_options(
    parser: FlexibleArgumentParser,
) -> FlexibleArgumentParser
Source code in vllm/entrypoints/cli/openai.py 
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
def _add_query_options(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
    parser.add_argument(
        "--url",
        type=str,
        default="http://localhost:8000/v1",
        help="url of the running OpenAI-Compatible RESTful API server",
    )
    parser.add_argument(
        "--model-name",
        type=str,
        default=None,
        help=(
            "The model name used in prompt completion, default to "
            "the first model in list models API call."
        ),
    )
    parser.add_argument(
        "--api-key",
        type=str,
        default=None,
        help=(
            "API key for OpenAI services. If provided, this api key "
            "will overwrite the api key obtained through environment variables."
        ),
    )
    return parser

_interactive_cli 

_interactive_cli(args: Namespace) -> tuple[str, OpenAI]
Source code in vllm/entrypoints/cli/openai.py 
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
def _interactive_cli(args: argparse.Namespace) -> tuple[str, OpenAI]:
    _register_signal_handlers()

    base_url = args.url
    api_key = args.api_key or os.environ.get("OPENAI_API_KEY", "EMPTY")
    openai_client = OpenAI(api_key=api_key, base_url=base_url)

    if args.model_name:
        model_name = args.model_name
    else:
        available_models = openai_client.models.list()
        model_name = available_models.data[0].id

    print(f"Using model: {model_name}")

    return model_name, openai_client

_print_chat_stream 

_print_chat_stream(stream) -> str
Source code in vllm/entrypoints/cli/openai.py 
47
48
49
50
51
52
53
54
55
def _print_chat_stream(stream) -> str:
    output = ""
    for chunk in stream:
        delta = chunk.choices[0].delta
        if delta.content:
            output += delta.content
            print(delta.content, end="", flush=True)
    print()
    return output

_print_completion_stream 

_print_completion_stream(stream) -> str
Source code in vllm/entrypoints/cli/openai.py 
58
59
60
61
62
63
64
65
66
def _print_completion_stream(stream) -> str:
    output = ""
    for chunk in stream:
        text = chunk.choices[0].text
        if text is not None:
            output += text
            print(text, end="", flush=True)
    print()
    return output

_register_signal_handlers 

_register_signal_handlers()
Source code in vllm/entrypoints/cli/openai.py 
21
22
23
24
25
26
def _register_signal_handlers():
    def signal_handler(sig, frame):
        sys.exit(0)

    signal.signal(signal.SIGINT, signal_handler)
    signal.signal(signal.SIGTSTP, signal_handler)

chat 

chat(
    system_prompt: str | None,
    model_name: str,
    client: OpenAI,
) -> None
Source code in vllm/entrypoints/cli/openai.py 
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
def chat(system_prompt: str | None, model_name: str, client: OpenAI) -> None:
    conversation: list[ChatCompletionMessageParam] = []
    if system_prompt is not None:
        conversation.append({"role": "system", "content": system_prompt})

    print("Please enter a message for the chat model:")
    while True:
        try:
            input_message = input("> ")
        except EOFError:
            break
        conversation.append({"role": "user", "content": input_message})

        stream = client.chat.completions.create(
            model=model_name, messages=conversation, stream=True
        )
        output = _print_chat_stream(stream)
        conversation.append({"role": "assistant", "content": output})

cmd_init 

cmd_init() -> list[CLISubcommand]
Source code in vllm/entrypoints/cli/openai.py 
247
248
def cmd_init() -> list[CLISubcommand]:
    return [ChatCommand(), CompleteCommand()]