File size: 24,487 Bytes
9e7090f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
import g4f
from webscout.AIutel import Optimizers
from webscout.AIutel import Conversation
from webscout.AIutel import AwesomePrompts
from webscout.AIbase import Provider, AsyncProvider
from webscout.AIutel import available_providers
from typing import Any, AsyncGenerator

g4f.debug.version_check = False

working_providers = available_providers

completion_allowed_models = [
    "code-davinci-002",
    "text-ada-001",
    "text-babbage-001",
    "text-curie-001",
    "text-davinci-002",
    "text-davinci-003",
]

default_models = {
    "completion": "text-davinci-003",
    "chat_completion": "gpt-3.5-turbo",
}

default_provider = "Koala"

class AsyncGPT4FREE(AsyncProvider):
    def __init__(

        self,

        provider: str = default_provider,

        is_conversation: bool = True,

        auth: str = None,

        max_tokens: int = 600,

        model: str = None,

        ignore_working: bool = False,

        timeout: int = 30,

        intro: str = None,

        filepath: str = None,

        update_file: bool = True,

        proxies: dict = {},

        history_offset: int = 10250,

        act: str = None,

    ):
        """Initialies GPT4FREE



        Args:

            provider (str, optional): gpt4free based provider name. Defaults to Koala.

            is_conversation (bool, optional): Flag for chatting conversationally. Defaults to True.

            auth (str, optional): Authentication value for the provider incase it needs. Defaults to None.

            max_tokens (int, optional): Maximum number of tokens to be generated upon completion. Defaults to 600.

            model (str, optional): LLM model name. Defaults to text-davinci-003|gpt-3.5-turbo.

            ignore_working (bool, optional): Ignore working status of the provider. Defaults to False.

            timeout (int, optional): Http request timeout. Defaults to 30.

            intro (str, optional): Conversation introductory prompt. Defaults to None.

            filepath (str, optional): Path to file containing conversation history. Defaults to None.

            update_file (bool, optional): Add new prompts and responses to the file. Defaults to True.

            proxies (dict, optional): Http request proxies. Defaults to {}.

            history_offset (int, optional): Limit conversation history to this number of last texts. Defaults to 10250.

            act (str|int, optional): Awesome prompt key or index. (Used as intro). Defaults to None.

        """
        assert provider in available_providers, (
            f"Provider '{provider}' is not yet supported. "
            f"Try others like {', '.join(available_providers)}"
        )
        if model is None:
            model = default_models["chat_completion"]

        self.is_conversation = is_conversation
        self.max_tokens_to_sample = max_tokens
        self.stream_chunk_size = 64
        self.timeout = timeout
        self.last_response = {}

        self.__available_optimizers = (
            method
            for method in dir(Optimizers)
            if callable(getattr(Optimizers, method)) and not method.startswith("__")
        )
        Conversation.intro = (
            AwesomePrompts().get_act(
                act, raise_not_found=True, default=None, case_insensitive=True
            )
            if act
            else intro or Conversation.intro
        )
        self.conversation = Conversation(
            is_conversation,
            self.max_tokens_to_sample,
            filepath,
            update_file,
        )
        self.conversation.history_offset = history_offset
        self.model = model
        self.provider = provider
        self.ignore_working = ignore_working
        self.auth = auth
        self.proxy = None if not proxies else list(proxies.values())[0]

    def __str__(self):
        return f"AsyncGPTFREE(provider={self.provider})"

    async def ask(

        self,

        prompt: str,

        stream: bool = False,

        raw: bool = False,

        optimizer: str = None,

        conversationally: bool = False,

    ) -> dict | AsyncGenerator:
        """Chat with AI asynchronously.



        Args:

            prompt (str): Prompt to be send.

            stream (bool, optional): Flag for streaming response. Defaults to False.

            raw (bool, optional): Stream back raw response as received. Defaults to False.

            optimizer (str, optional): Prompt optimizer name - `[code, shell_command]`. Defaults to None.

            conversationally (bool, optional): Chat conversationally when using optimizer. Defaults to False.

        Returns:

           dict|AsyncGenerator : ai content

        ```json

        {

          "text" : "How may I help you today?"

        }

        ```

        """
        conversation_prompt = self.conversation.gen_complete_prompt(prompt)
        if optimizer:
            if optimizer in self.__available_optimizers:
                conversation_prompt = getattr(Optimizers, optimizer)(
                    conversation_prompt if conversationally else prompt
                )
            else:
                raise Exception(
                    f"Optimizer is not one of {self.__available_optimizers}"
                )

        payload = dict(
            model=self.model,
            provider=self.provider,  # g4f.Provider.Aichat,
            messages=[{"role": "user", "content": conversation_prompt}],
            stream=True,
            ignore_working=self.ignore_working,
            auth=self.auth,
            proxy=self.proxy,
            timeout=self.timeout,
        )

        async def format_response(response):
            return dict(text=response)

        async def for_stream():
            previous_chunks = ""
            response = g4f.ChatCompletion.create_async(**payload)

            async for chunk in response:
                previous_chunks += chunk
                formatted_resp = await format_response(previous_chunks)
                self.last_response.update(formatted_resp)
                yield previous_chunks if raw else formatted_resp

            self.conversation.update_chat_history(
                prompt,
                previous_chunks,
            )

        async def for_non_stream():
            async for _ in for_stream():
                pass
            return self.last_response

        return for_stream() if stream else await for_non_stream()

    async def chat(

        self,

        prompt: str,

        stream: bool = False,

        optimizer: str = None,

        conversationally: bool = False,

    ) -> dict | AsyncGenerator:
        """Generate response `str` asynchronously.

        Args:

            prompt (str): Prompt to be send.

            stream (bool, optional): Flag for streaming response. Defaults to False.

            optimizer (str, optional): Prompt optimizer name - `[code, shell_command]`. Defaults to None.

            conversationally (bool, optional): Chat conversationally when using optimizer. Defaults to False.

        Returns:

            str|AsyncGenerator: Response generated

        """

        async def for_stream():
            async_ask = await self.ask(
                prompt, True, optimizer=optimizer, conversationally=conversationally
            )
            async for response in async_ask:
                yield await self.get_message(response)

        async def for_non_stream():
            return await self.get_message(
                await self.ask(
                    prompt,
                    False,
                    optimizer=optimizer,
                    conversationally=conversationally,
                )
            )

        return for_stream() if stream else await for_non_stream()

    async def get_message(self, response: dict) -> str:
        """Retrieves message only from response



        Args:

            response (dict): Response generated by `self.ask`



        Returns:

            str: Message extracted

        """
        assert isinstance(response, dict), "Response should be of dict data-type only"
        return response["text"]
class GPT4FREE(Provider):
    def __init__(

        self,

        provider: str = default_provider,

        is_conversation: bool = True,

        auth: str = None,

        max_tokens: int = 600,

        model: str = None,

        chat_completion: bool = True,

        ignore_working: bool = True,

        timeout: int = 30,

        intro: str = None,

        filepath: str = None,

        update_file: bool = True,

        proxies: dict = {},

        history_offset: int = 10250,

        act: str = None,

    ):
        """Initialies GPT4FREE



        Args:

            provider (str, optional): gpt4free based provider name. Defaults to Koala.

            is_conversation (bool, optional): Flag for chatting conversationally. Defaults to True.

            auth (str, optional): Authentication value for the provider incase it needs. Defaults to None.

            max_tokens (int, optional): Maximum number of tokens to be generated upon completion. Defaults to 600.

            model (str, optional): LLM model name. Defaults to text-davinci-003|gpt-3.5-turbo.

            chat_completion(bool, optional): Provide native auto-contexting (conversationally). Defaults to False.

            ignore_working (bool, optional): Ignore working status of the provider. Defaults to False.

            timeout (int, optional): Http request timeout. Defaults to 30.

            intro (str, optional): Conversation introductory prompt. Defaults to None.

            filepath (str, optional): Path to file containing conversation history. Defaults to None.

            update_file (bool, optional): Add new prompts and responses to the file. Defaults to True.

            proxies (dict, optional): Http request proxies. Defaults to {}.

            history_offset (int, optional): Limit conversation history to this number of last texts. Defaults to 10250.

            act (str|int, optional): Awesome prompt key or index. (Used as intro). Defaults to None.

        """
        assert provider in available_providers, (
            f"Provider '{provider}' is not yet supported. "
            f"Try others like {', '.join(available_providers)}"
        )
        if model is None:
            model = (
                default_models["chat_completion"]
                if chat_completion
                else default_models["completion"]
            )

        elif not chat_completion:
            assert model in completion_allowed_models, (
                f"Model '{model}' is not yet supported for completion. "
                f"Try other models like {', '.join(completion_allowed_models)}"
            )
        self.is_conversation = is_conversation
        self.max_tokens_to_sample = max_tokens
        self.stream_chunk_size = 64
        self.timeout = timeout
        self.last_response = {}

        self.__available_optimizers = (
            method
            for method in dir(Optimizers)
            if callable(getattr(Optimizers, method)) and not method.startswith("__")
        )
        Conversation.intro = (
            AwesomePrompts().get_act(
                act, raise_not_found=True, default=None, case_insensitive=True
            )
            if act
            else intro or Conversation.intro
        )
        self.conversation = Conversation(
            False if chat_completion else is_conversation,
            self.max_tokens_to_sample,
            filepath,
            update_file,
        )
        self.conversation.history_offset = history_offset
        self.model = model
        self.provider = provider
        self.chat_completion = chat_completion
        self.ignore_working = ignore_working
        self.auth = auth
        self.proxy = None if not proxies else list(proxies.values())[0]
        self.__chat_class = g4f.ChatCompletion if chat_completion else g4f.Completion

    def ask(

        self,

        prompt: str,

        stream: bool = False,

        raw: bool = False,

        optimizer: str = None,

        conversationally: bool = False,

    ) -> dict:
        """Chat with AI



        Args:

            prompt (str): Prompt to be send.

            stream (bool, optional): Flag for streaming response. Defaults to False.

            raw (bool, optional): Stream back raw response as received. Defaults to False.

            optimizer (str, optional): Prompt optimizer name - `[code, shell_command]`. Defaults to None.

            conversationally (bool, optional): Chat conversationally when using optimizer. Defaults to False.

        Returns:

           dict : {}

        ```json

        {

          "text" : "How may I help you today?"

        }

        ```

        """
        conversation_prompt = self.conversation.gen_complete_prompt(prompt)
        if optimizer:
            if optimizer in self.__available_optimizers:
                conversation_prompt = getattr(Optimizers, optimizer)(
                    conversation_prompt if conversationally else prompt
                )
            else:
                raise Exception(
                    f"Optimizer is not one of {self.__available_optimizers}"
                )

        def payload():
            if self.chat_completion:
                return dict(
                    model=self.model,
                    provider=self.provider,  # g4f.Provider.Aichat,
                    messages=[{"role": "user", "content": conversation_prompt}],
                    stream=stream,
                    ignore_working=self.ignore_working,
                    auth=self.auth,
                    proxy=self.proxy,
                    timeout=self.timeout,
                )

            else:
                return dict(
                    model=self.model,
                    prompt=conversation_prompt,
                    provider=self.provider,
                    stream=stream,
                    ignore_working=self.ignore_working,
                    auth=self.auth,
                    proxy=self.proxy,
                    timeout=self.timeout,
                )

        def format_response(response):
            return dict(text=response)

        def for_stream():
            previous_chunks = ""
            response = self.__chat_class.create(**payload())

            for chunk in response:
                previous_chunks += chunk
                formatted_resp = format_response(previous_chunks)
                self.last_response.update(formatted_resp)
                yield previous_chunks if raw else formatted_resp

            self.conversation.update_chat_history(
                prompt,
                previous_chunks,
            )

        def for_non_stream():
            response = self.__chat_class.create(**payload())
            formatted_resp = format_response(response)

            self.last_response.update(formatted_resp)
            self.conversation.update_chat_history(prompt, response)

            return response if raw else formatted_resp

        return for_stream() if stream else for_non_stream()

    def chat(

        self,

        prompt: str,

        stream: bool = False,

        optimizer: str = None,

        conversationally: bool = False,

    ) -> str:
        """Generate response `str`

        Args:

            prompt (str): Prompt to be send.

            stream (bool, optional): Flag for streaming response. Defaults to False.

            optimizer (str, optional): Prompt optimizer name - `[code, shell_command]`. Defaults to None.

            conversationally (bool, optional): Chat conversationally when using optimizer. Defaults to False.

        Returns:

            str: Response generated

        """

        def for_stream():
            for response in self.ask(
                prompt, True, optimizer=optimizer, conversationally=conversationally
            ):
                yield self.get_message(response)

        def for_non_stream():
            return self.get_message(
                self.ask(
                    prompt,
                    False,
                    optimizer=optimizer,
                    conversationally=conversationally,
                )
            )

        return for_stream() if stream else for_non_stream()

    def get_message(self, response: dict) -> str:
        """Retrieves message only from response



        Args:

            response (dict): Response generated by `self.ask`



        Returns:

            str: Message extracted

        """
        assert isinstance(response, dict), "Response should be of dict data-type only"
        return response["text"]
from pathlib import Path
from webscout.AIutel import default_path
from json import dump, load
from time import time
from threading import Thread as thr
from functools import wraps
from rich.progress import Progress
import logging

results_path = Path(default_path) / "provider_test.json"


def exception_handler(func):

    @wraps(func)
    def decorator(*args, **kwargs):
        try:
            return func(*args, **kwargs)
        except Exception as e:
            pass

    return decorator


@exception_handler
def is_working(provider: str) -> bool:
    """Test working status of a provider



    Args:

        provider (str): Provider name



    Returns:

        bool: is_working status

    """
    bot = GPT4FREE(provider=provider, is_conversation=False)
    text = bot.chat("hello")
    assert isinstance(text, str)
    assert bool(text.strip())
    assert "</" not in text
    assert ":" not in text
    assert len(text) > 2
    return True


class TestProviders:

    def __init__(

        self,

        test_at_once: int = 5,

        quiet: bool = False,

        timeout: int = 20,

        selenium: bool = False,

        do_log: bool = True,

    ):
        """Constructor



        Args:

            test_at_once (int, optional): Test n providers at once. Defaults to 5.

            quiet (bool, optinal): Disable stdout. Defaults to False.

            timout (int, optional): Thread timeout for each provider. Defaults to 20.

            selenium (bool, optional): Test even selenium dependent providers. Defaults to False.

            do_log (bool, optional): Flag to control logging. Defaults to True.

        """
        self.test_at_once: int = test_at_once
        self.quiet = quiet
        self.timeout = timeout
        self.do_log = do_log
        self.__logger = logging.getLogger(__name__)
        self.working_providers: list = [
            provider.__name__
            for provider in g4f.Provider.__providers__
            if provider.working
        ]

        if not selenium:
            import g4f.Provider.selenium as selenium_based
            from g4f import webdriver

            webdriver.has_requirements = False
            selenium_based_providers: list = dir(selenium_based)
            for provider in self.working_providers:
                try:
                    selenium_based_providers.index(provider)
                except ValueError:
                    pass
                else:
                    self.__log(
                        10, f"Dropping provider - {provider} - [Selenium dependent]"
                    )
                    self.working_providers.remove(provider)

        self.results_path: Path = results_path
        self.__create_empty_file(ignore_if_found=True)
        self.results_file_is_empty: bool = False

    def __log(

        self,

        level: int,

        message: str,

    ):
        """class logger"""
        if self.do_log:
            self.__logger.log(level, message)
        else:
            pass

    def __create_empty_file(self, ignore_if_found: bool = False):
        if ignore_if_found and self.results_path.is_file():
            return
        with self.results_path.open("w") as fh:
            dump({"results": []}, fh)
        self.results_file_is_empty = True

    def test_provider(self, name: str):
        """Test each provider and save successful ones



        Args:

            name (str): Provider name

        """

        try:
            bot = GPT4FREE(provider=name, is_conversation=False)
            start_time = time()
            text = bot.chat("hello there")
            assert isinstance(text, str), "Non-string response returned"
            assert bool(text.strip()), "Empty string"
            assert "</" not in text, "Html code returned."
            assert ":" not in text, "Json formatted response returned"
            assert len(text) > 2
        except Exception as e:
            pass
        else:
            self.results_file_is_empty = False
            with self.results_path.open() as fh:
                current_results = load(fh)
            new_result = dict(time=time() - start_time, name=name)
            current_results["results"].append(new_result)
            self.__log(20, f"Test result - {new_result['name']} - {new_result['time']}")

            with self.results_path.open("w") as fh:
                dump(current_results, fh)

    @exception_handler
    def main(

        self,

    ):
        self.__create_empty_file()
        threads = []
        # Create a progress bar
        total = len(self.working_providers)
        with Progress() as progress:
            self.__log(20, f"Testing {total} providers : {self.working_providers}")
            task = progress.add_task(
                f"[cyan]Testing...[{self.test_at_once}]",
                total=total,
                visible=self.quiet == False,
            )
            while not progress.finished:
                for count, provider in enumerate(self.working_providers, start=1):
                    t1 = thr(
                        target=self.test_provider,
                        args=(provider,),
                    )
                    t1.start()
                    if count % self.test_at_once == 0 or count == len(provider):
                        for t in threads:
                            try:
                                t.join(self.timeout)
                            except Exception as e:
                                pass
                        threads.clear()
                    else:
                        threads.append(t1)
                    progress.update(task, advance=1)

    def get_results(self, run: bool = False, best: bool = False) -> list[dict]:
        """Get test results



        Args:

            run (bool, optional): Run the test first. Defaults to False.

            best (bool, optional): Return name of the best provider. Defaults to False.



        Returns:

            list[dict]|str: Test results.

        """
        if run or self.results_file_is_empty:
            self.main()

        with self.results_path.open() as fh:
            results: dict = load(fh)

        results = results["results"]
        if not results:
            if run:
                raise Exception("Unable to find working g4f provider")
            else:
                self.__log(30, "Hunting down working g4f providers.")
                return self.get_results(run=True, best=best)

        time_list = []

        sorted_list = []
        for entry in results:
            time_list.append(entry["time"])

        time_list.sort()

        for time_value in time_list:
            for entry in results:
                if entry["time"] == time_value:
                    sorted_list.append(entry)
        return sorted_list[0]["name"] if best else sorted_list

    @property
    def best(self):
        """Fastest provider overally"""
        return self.get_results(run=False, best=True)

    @property
    def auto(self):
        """Best working provider"""
        for result in self.get_results(run=False, best=False):
            self.__log(20, "Confirming working status of provider : " + result["name"])
            if is_working(result["name"]):
                return result["name"]