webscout 8.2.3__py3-none-any.whl → 8.2.5__py3-none-any.whl

webscout/Provider/Llama.py

@@ -1,200 +0,0 @@
-import requests
-import json
-
-from webscout.AIutel import Optimizers
-from webscout.AIutel import Conversation
-from webscout.AIutel import AwesomePrompts, sanitize_stream
-from webscout.AIbase import Provider, AsyncProvider
-from webscout import exceptions
-from typing import Any, AsyncGenerator, Dict, Union
-
-
-class LLAMA(Provider):
-    def __init__(
-        self,
-        is_conversation: bool = True,
-        max_tokens: int = 600,
-        timeout: int = 30,
-        intro: str = None,
-        filepath: str = None,
-        update_file: bool = True,
-        proxies: dict = {},
-        history_offset: int = 10250,
-        act: str = None,
-    ):
-        """Instantiates LLAMA
-
-        Args:
-            is_conversation (bool, optional): Flag for chatting conversationally. Defaults to True.
-            max_tokens (int, optional): Maximum number of tokens to be generated upon completion. Defaults to 600.
-            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.
-            model (str, optional): LLM model name. Defaults to "llama3-70b-8192".
-        """
-        self.is_conversation = is_conversation
-        self.max_tokens_to_sample = max_tokens
-        self.timeout = timeout
-        self.last_response = {}
-        self.model = "llama3-70b-8192",
-        self.api_endpoint = "https://api.safone.dev/llama"
-        self.headers = {
-            "accept": "application/json",
-        }
-
-        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.session = requests.Session()
-        self.session.proxies = proxies
-
-    def ask(
-        self,
-        prompt: str,
-        stream: bool = False,
-        raw: bool = False,
-        optimizer: str = None,
-        conversationally: bool = False,
-    ) -> Union[dict, AsyncGenerator]:
-        """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|AsyncGenerator : ai content
-        ```json
-        {
-           "text" : "print('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}"
-                )
-        
-        self.session.headers.update(self.headers)
-        payload = {
-            "message": conversation_prompt
-        }
-
-        def for_stream():
-            response = self.session.get(
-                self.api_endpoint, params=payload, stream=True, timeout=self.timeout
-            )
-            if not response.ok:
-                raise exceptions.FailedToGenerateResponseError(
-                    f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
-                )
-            
-            message_load = ""
-            for chunk in response.iter_lines():
-                try:
-                    resp = json.loads(chunk)
-                    message_load += resp['message']
-                    yield chunk if raw else dict(text=message_load)
-                    self.last_response.update(resp)
-                except:
-                    pass
-            self.conversation.update_chat_history(
-                prompt, self.get_message(self.last_response)
-            )
-
-        def for_non_stream():
-            response = self.session.get(
-                self.api_endpoint, params=payload, stream=False, timeout=self.timeout
-            )
-            if not response.ok:
-                raise exceptions.FailedToGenerateResponseError(
-                    f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
-                )
-            resp = response.json()
-            self.last_response.update(resp)
-            self.conversation.update_chat_history(
-                prompt, self.get_message(self.last_response)
-            )
-            return resp
-
-        return for_stream() if stream else for_non_stream()
-
-    def chat(
-        self,
-        prompt: str,
-        stream: bool = False,
-        optimizer: str = None,
-        conversationally: bool = False,
-    ) -> Union[str, AsyncGenerator]:
-        """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["message"]
-if __name__ == "__main__":
-    from rich import print
-
-    ai = LLAMA() 
-    # Stream the response
-    response = ai.chat(input(">>> "))
-    for chunk in response:
-        print(chunk, end="", flush=True)