File size: 7,080 Bytes
dc12c31
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
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
115
116
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
187
188
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
import json
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
from threading import Thread

from extensions.api.util import build_parameters, try_start_cloudflared
from modules import shared
from modules.chat import generate_chat_reply
from modules.LoRA import add_lora_to_model
from modules.models import load_model, unload_model
from modules.models_settings import get_model_metadata, update_model_parameters
from modules.text_generation import (
    encode,
    generate_reply,
    stop_everything_event
)
from modules.utils import get_available_models


def get_model_info():
    return {
        'model_name': shared.model_name,
        'lora_names': shared.lora_names,
        # dump
        'shared.settings': shared.settings,
        'shared.args': vars(shared.args),
    }


class Handler(BaseHTTPRequestHandler):
    def do_GET(self):
        if self.path == '/api/v1/model':
            self.send_response(200)
            self.end_headers()
            response = json.dumps({
                'result': shared.model_name
            })

            self.wfile.write(response.encode('utf-8'))
        else:
            self.send_error(404)

    def do_POST(self):
        content_length = int(self.headers['Content-Length'])
        body = json.loads(self.rfile.read(content_length).decode('utf-8'))

        if self.path == '/api/v1/generate':
            self.send_response(200)
            self.send_header('Content-Type', 'application/json')
            self.end_headers()

            prompt = body['prompt']
            generate_params = build_parameters(body)
            stopping_strings = generate_params.pop('stopping_strings')
            generate_params['stream'] = False

            generator = generate_reply(
                prompt, generate_params, stopping_strings=stopping_strings, is_chat=False)

            answer = ''
            for a in generator:
                answer = a

            response = json.dumps({
                'results': [{
                    'text': answer
                }]
            })

            self.wfile.write(response.encode('utf-8'))

        elif self.path == '/api/v1/chat':
            self.send_response(200)
            self.send_header('Content-Type', 'application/json')
            self.end_headers()

            user_input = body['user_input']
            regenerate = body.get('regenerate', False)
            _continue = body.get('_continue', False)

            generate_params = build_parameters(body, chat=True)
            generate_params['stream'] = False

            generator = generate_chat_reply(
                user_input, generate_params, regenerate=regenerate, _continue=_continue, loading_message=False)

            answer = generate_params['history']
            for a in generator:
                answer = a

            response = json.dumps({
                'results': [{
                    'history': answer
                }]
            })

            self.wfile.write(response.encode('utf-8'))

        elif self.path == '/api/v1/stop-stream':
            self.send_response(200)
            self.send_header('Content-Type', 'application/json')
            self.end_headers()

            stop_everything_event()

            response = json.dumps({
                'results': 'success'
            })

            self.wfile.write(response.encode('utf-8'))

        elif self.path == '/api/v1/model':
            self.send_response(200)
            self.send_header('Content-Type', 'application/json')
            self.end_headers()

            # by default return the same as the GET interface
            result = shared.model_name

            # Actions: info, load, list, unload
            action = body.get('action', '')

            if action == 'load':
                model_name = body['model_name']
                args = body.get('args', {})
                print('args', args)
                for k in args:
                    setattr(shared.args, k, args[k])

                shared.model_name = model_name
                unload_model()

                model_settings = get_model_metadata(shared.model_name)
                shared.settings.update({k: v for k, v in model_settings.items() if k in shared.settings})
                update_model_parameters(model_settings, initial=True)

                if shared.settings['mode'] != 'instruct':
                    shared.settings['instruction_template'] = None

                try:
                    shared.model, shared.tokenizer = load_model(shared.model_name)
                    if shared.args.lora:
                        add_lora_to_model(shared.args.lora)  # list

                except Exception as e:
                    response = json.dumps({'error': {'message': repr(e)}})

                    self.wfile.write(response.encode('utf-8'))
                    raise e

                shared.args.model = shared.model_name

                result = get_model_info()

            elif action == 'unload':
                unload_model()
                shared.model_name = None
                shared.args.model = None
                result = get_model_info()

            elif action == 'list':
                result = get_available_models()

            elif action == 'info':
                result = get_model_info()

            response = json.dumps({
                'result': result,
            })

            self.wfile.write(response.encode('utf-8'))

        elif self.path == '/api/v1/token-count':
            self.send_response(200)
            self.send_header('Content-Type', 'application/json')
            self.end_headers()

            tokens = encode(body['prompt'])[0]
            response = json.dumps({
                'results': [{
                    'tokens': len(tokens)
                }]
            })

            self.wfile.write(response.encode('utf-8'))
        else:
            self.send_error(404)

    def do_OPTIONS(self):
        self.send_response(200)
        self.end_headers()

    def end_headers(self):
        self.send_header('Access-Control-Allow-Origin', '*')
        self.send_header('Access-Control-Allow-Methods', '*')
        self.send_header('Access-Control-Allow-Headers', '*')
        self.send_header('Cache-Control', 'no-store, no-cache, must-revalidate')
        super().end_headers()


def _run_server(port: int, share: bool = False, tunnel_id=str):
    address = '0.0.0.0' if shared.args.listen else '127.0.0.1'

    server = ThreadingHTTPServer((address, port), Handler)

    def on_start(public_url: str):
        print(f'Starting non-streaming server at public url {public_url}/api')

    if share:
        try:
            try_start_cloudflared(port, tunnel_id, max_attempts=3, on_start=on_start)
        except Exception:
            pass
    else:
        print(
            f'Starting API at http://{address}:{port}/api')

    server.serve_forever()


def start_server(port: int, share: bool = False, tunnel_id=str):
    Thread(target=_run_server, args=[port, share, tunnel_id], daemon=True).start()