webscout 8.2.6__py3-none-any.whl → 8.2.7__py3-none-any.whl

webscout/Provider/TeachAnything.py

@@ -1,233 +0,0 @@
-from curl_cffi.requests import Session # Import Session
-from curl_cffi import CurlError # Import CurlError
-from typing import Union, Any, Dict
-from webscout.AIbase import Provider # Import Provider base class
-from webscout import exceptions # Import custom exceptions
-from webscout.conversation import Conversation
-from webscout.AIutel import Optimizers, sanitize_stream # Import sanitize_stream
-from webscout.prompt_manager import AwesomePrompts
-from webscout.litagent import LitAgent
-
-# Inherit from Provider
-class TeachAnything(Provider):
-    """
-    A class to interact with the Teach-Anything API.
-    """
-    # Add AVAILABLE_MODELS if applicable, otherwise remove model param
-    # AVAILABLE_MODELS = ["default"] # Example
-
-    def __init__(
-        self,
-        is_conversation: bool = True,
-        max_tokens: int = 600, # Note: max_tokens is not used by this API
-        timeout: int = 30,
-        intro: str = None,
-        filepath: str = None,
-        update_file: bool = True,
-        proxies: dict = {},
-        history_offset: int = 10250,
-        act: str = None,
-        # model: str = "default" # Remove if not used
-    ) -> None:
-        """
-        Initializes the Teach-Anything API with given parameters.
-
-        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): AI model to use for text generation. Defaults to "gpt4".
-        """
-
-
-        # Initialize curl_cffi Session
-        self.session = Session()
-        self.is_conversation = is_conversation
-        self.max_tokens_to_sample = max_tokens
-        self.api_endpoint = "https://www.teach-anything.com/api/generate"
-        self.timeout = timeout
-        self.last_response = {}
-        self.headers = {
-            "authority": "www.teach-anything.com",
-            "accept": "*/*",
-            "accept-encoding": "gzip, deflate, br, zstd",
-            "accept-language": "en-US,en;q=0.9,en-IN;q=0.8",
-            "content-type": "application/json",
-            "origin": "https://www.teach-anything.com",
-            "referer": "https://www.teach-anything.com/",
-            "user-agent": LitAgent().random(),
-            # Add sec-ch-ua headers if needed for impersonation consistency
-        }
-        self.__available_optimizers = (
-            method
-            for method in dir(Optimizers)
-            if callable(getattr(Optimizers, method)) and not method.startswith("__")
-        )
-        # Update curl_cffi session headers and proxies
-        self.session.headers.update(self.headers)
-        self.session.proxies = proxies # Assign proxies directly
-        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
-
-    def ask(
-        self,
-        prompt: str,
-        stream: bool = False, # Keep stream param for interface, but API doesn't stream
-        raw: bool = False, # Keep raw param for interface
-        optimizer: str = None,
-        conversationally: bool = False,
-    ) -> dict:
-        """Chat with AI
-
-        Args:
-            prompt (str): Prompt to be send.
-            stream (bool, optional): Whether to stream the response. Defaults to False.
-            raw (bool, optional): Whether to return the raw response. Defaults to False.
-            optimizer (str, optional): The name of the optimizer to use. Defaults to None.
-            conversationally (bool, optional): Whether to chat conversationally. Defaults to False.
-
-        Returns:
-            The response from the API.
-        """
-        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 = {
-            "prompt": conversation_prompt
-        }
-
-        # API does not stream, so implement non-stream logic directly
-        try:
-            # Use curl_cffi session post with impersonate
-            response = self.session.post(
-                self.api_endpoint, 
-                # headers are set on the session
-                json=payload, 
-                timeout=self.timeout,
-                impersonate="chrome110" # Use a common impersonation profile
-            )
-            response.raise_for_status() # Check for HTTP errors
-
-            resp_text_raw = response.text # Get raw response text
-
-            # Process the text using sanitize_stream (even though it's not streaming)
-            # This keeps the pattern consistent, though it won't do much here
-            processed_stream = sanitize_stream(
-                data=resp_text_raw,
-                intro_value=None, # No prefix
-                to_json=False     # It's plain text
-            )
-
-            # Extract the single result from the generator
-            resp_text = "".join(list(processed_stream)) # Aggregate potential chunks (should be one)
-
-            self.last_response = {"text": resp_text}
-            self.conversation.update_chat_history(prompt, resp_text)
-
-            # Return dict or raw string based on raw flag
-            return resp_text if raw else self.last_response
-
-        except CurlError as e: # Catch CurlError
-            raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}") from e
-        except Exception as e: # Catch other potential exceptions (like HTTPError)
-            err_text = getattr(e, 'response', None) and getattr(e.response, 'text', '')
-            raise exceptions.FailedToGenerateResponseError(f"An unexpected error occurred ({type(e).__name__}): {e} - {err_text}") from e
-
-
-    def chat(
-        self,
-        prompt: str,
-        stream: bool = False, # Keep stream param for interface consistency
-        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
-        """
-
-        # Since ask() now handles both stream=True/False by returning the full response dict/str:
-        response_data = self.ask(
-            prompt, 
-            stream=False, # Call ask in non-stream mode internally
-            raw=False, # Ensure ask returns dict
-            optimizer=optimizer, 
-            conversationally=conversationally
-        )
-        # If stream=True was requested, simulate streaming by yielding the full message at once
-        if stream:
-            def stream_wrapper():
-                yield self.get_message(response_data)
-            return stream_wrapper()
-        else:
-            # If stream=False, return the full message directly
-            return self.get_message(response_data)
-
-    def get_message(self, response: Union[dict, str]) -> str:
-        """Retrieves message only from response
-
-        Args:
-            response (Union[dict, str]): Response generated by `self.ask`
-
-        Returns:
-            str: Message extracted
-        """
-        if isinstance(response, str):
-            return response
-        elif isinstance(response, dict):
-            return response["text"]
-        raise ValueError("Response must be either dict or str")
-
-
-if __name__ == '__main__':
-    # Ensure curl_cffi is installed
-    from rich import print
-    try: # Add try-except block for testing
-        ai = TeachAnything(timeout=60)
-        print("[bold blue]Testing Chat (Non-Stream Simulation):[/bold blue]")
-        # Test non-stream first as API doesn't truly stream
-        response_non_stream = ai.chat("hi", stream=False)
-        print(response_non_stream)
-        print("[bold green]Non-Stream Test Complete.[/bold green]\n")
-
-        # Test stream interface (will yield the full response at once)
-        print("[bold blue]Testing Chat (Stream Simulation):[/bold blue]")
-        response_stream = ai.chat("hello again", stream=True)
-        full_stream_response = ""
-        for chunk in response_stream:
-            print(chunk, end="", flush=True)
-            full_stream_response += chunk
-        print("\n[bold green]Stream Test Complete.[/bold green]")
-
-    except exceptions.FailedToGenerateResponseError as e:
-        print(f"\n[bold red]API Error:[/bold red] {e}")
-    except Exception as e:
-        print(f"\n[bold red]An unexpected error occurred:[/bold red] {e}")

webscout/Provider/TextPollinationsAI.py

@@ -1,306 +0,0 @@
-from curl_cffi.requests import Session
-from curl_cffi import CurlError
-import json
-from typing import Union, Any, Dict, Generator, Optional, List
-
-import requests
-from webscout.AIutel import Optimizers, Conversation, AwesomePrompts, sanitize_stream # Import sanitize_stream
-from webscout.AIbase import Provider
-from webscout import exceptions
-from webscout.litagent import LitAgent as Lit
-
-class TextPollinationsAI(Provider):
-    """
-    A class to interact with the Pollinations AI API.
-    """
-
-    AVAILABLE_MODELS = [
-        "openai",
-        "openai-large",
-        "qwen-coder",
-        "llama",
-        "llamascout",
-        "mistral",
-        "unity",
-        "midijourney",
-        "rtist",
-        "searchgpt",
-        "evil",
-        "deepseek-reasoning",
-        "deepseek-reasoning-large",
-        "phi",
-        "llama-vision",
-        "hormoz",
-        "hypnosis-tracy",
-        "deepseek",
-        "sur",
-        "openai-audio",
-    ]
-    _models_url = "https://text.pollinations.ai/models"
-
-    def __init__(self,
-        is_conversation: bool = True,
-        max_tokens: int = 8096, # Note: max_tokens is not directly used by this API endpoint
-        timeout: int = 30,
-        intro: str = None,
-        filepath: str = None,
-        update_file: bool = True,
-        proxies: dict = {},
-        history_offset: int = 10250,
-        act: str = None,
-        model: str = "openai-large",
-        system_prompt: str = "You are a helpful AI assistant.",
-    ):
-        """Initializes the TextPollinationsAI API client."""
-        self.session = Session()
-        self.is_conversation = is_conversation
-        self.max_tokens_to_sample = max_tokens
-        self.api_endpoint = "https://text.pollinations.ai/openai"
-        self.stream_chunk_size = 64
-        self.timeout = timeout
-        self.last_response = {}
-        self.model = model
-        self.system_prompt = system_prompt
-
-        # Validate against the hardcoded list
-        if model not in self.AVAILABLE_MODELS:
-            raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
-
-        self.headers = {
-            'Accept': '*/*',
-            'Accept-Language': 'en-US,en;q=0.9',
-            'User-Agent': Lit().random(),
-            'Content-Type': 'application/json',
-            # Add sec-ch-ua headers if needed for impersonation consistency
-        }
-
-        # Update curl_cffi session headers and proxies
-        self.session.headers.update(self.headers)
-        self.session.proxies = proxies # Assign proxies directly
-
-        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
-
-
-    def ask(
-        self,
-        prompt: str,
-        stream: bool = False,
-        raw: bool = False,
-        optimizer: str = None,
-        conversationally: bool = False,
-        tools: Optional[List[Dict[str, Any]]] = None,
-        tool_choice: Optional[Dict[str, Any]] = None,
-    ) -> Union[Dict[str, Any], Generator[Any, None, None]]:
-        """Chat with AI"""
-        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 = {
-            "messages": [
-                {"role": "system", "content": self.system_prompt},
-                {"role": "user", "content": conversation_prompt}
-            ],
-            "model": self.model,
-            "stream": stream,
-        }
-
-        # Add function calling parameters if provided
-        if tools:
-            payload["tools"] = tools
-        if tool_choice:
-            payload["tool_choice"] = tool_choice
-
-        def for_stream():
-            try: # Add try block for CurlError
-                # Use curl_cffi session post with impersonate
-                response = self.session.post(
-                    self.api_endpoint,
-                    # headers are set on the session
-                    json=payload,
-                    stream=True,
-                    timeout=self.timeout,
-                    impersonate="chrome120" # Add impersonate
-                )
-
-                if not response.ok:
-                    raise exceptions.FailedToGenerateResponseError(
-                        f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
-                    )
-
-                streaming_text = ""
-                # Use sanitize_stream
-                processed_stream = sanitize_stream(
-                    data=response.iter_content(chunk_size=None), # Pass byte iterator
-                    intro_value="data:",
-                    to_json=True,     # Stream sends JSON
-                    skip_markers=["[DONE]"],
-                    # Extractor handles both content and tool_calls
-                    content_extractor=lambda chunk: chunk.get('choices', [{}])[0].get('delta') if isinstance(chunk, dict) else None,
-                    yield_raw_on_error=False # Skip non-JSON or lines where extractor fails
-                )
-
-                for delta in processed_stream:
-                    # delta is the extracted 'delta' object or None
-                    if delta and isinstance(delta, dict):
-                        if 'content' in delta and delta['content'] is not None:
-                            content = delta['content']
-                            streaming_text += content
-                            yield content if raw else dict(text=content)
-                        elif 'tool_calls' in delta:
-                            tool_calls = delta['tool_calls']
-                            yield tool_calls if raw else dict(tool_calls=tool_calls)
-
-                # Update history and last response after stream finishes
-                self.last_response.update(dict(text=streaming_text)) # Store aggregated text
-                if streaming_text: # Only update history if text was received
-                    self.conversation.update_chat_history(
-                        prompt, streaming_text # Use the fully aggregated text
-                    )
-            except CurlError as e: # Catch CurlError
-                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}") from e
-            except Exception as e: # Catch other potential exceptions
-                raise exceptions.FailedToGenerateResponseError(f"An unexpected error occurred ({type(e).__name__}): {e}") from e
-
-
-        def for_non_stream():
-            # Aggregate the stream using the updated for_stream logic
-            final_content = ""
-            tool_calls_aggregated = None # To store potential tool calls
-            try: # Add try block for potential errors during aggregation
-                for chunk_data in for_stream():
-                    if isinstance(chunk_data, dict):
-                        if "text" in chunk_data:
-                            final_content += chunk_data["text"]
-                        elif "tool_calls" in chunk_data:
-                            # Aggregate tool calls (simple aggregation, might need refinement)
-                            if tool_calls_aggregated is None:
-                                tool_calls_aggregated = []
-                            tool_calls_aggregated.extend(chunk_data["tool_calls"])
-                    elif isinstance(chunk_data, str): # Handle raw stream case
-                        final_content += chunk_data
-                    # Handle raw tool calls list if raw=True
-                    elif isinstance(chunk_data, list) and raw:
-                         if tool_calls_aggregated is None:
-                             tool_calls_aggregated = []
-                         tool_calls_aggregated.extend(chunk_data)
-            except Exception as e:
-                 # If aggregation fails but some text was received, use it. Otherwise, re-raise.
-                 if not final_content and not tool_calls_aggregated:
-                     raise exceptions.FailedToGenerateResponseError(f"Failed to get non-stream response: {str(e)}") from e
-
-
-            # last_response and history are updated within for_stream (for text)
-            # Return a dict containing text and/or tool_calls
-            result = {}
-            if final_content:
-                result["text"] = final_content
-            if tool_calls_aggregated:
-                result["tool_calls"] = tool_calls_aggregated
-            self.last_response = result # Update last_response with aggregated result
-            return self.last_response
-
-
-        return for_stream() if stream else for_non_stream()
-
-    def chat(
-        self,
-        prompt: str,
-        stream: bool = False,
-        optimizer: str = None,
-        conversationally: bool = False,
-        tools: Optional[List[Dict[str, Any]]] = None,
-        tool_choice: Optional[Dict[str, Any]] = None,
-    ) -> Union[str, Generator[str, None, None]]:
-        """Generate response as a string"""
-        def for_stream():
-            for response in self.ask(
-                prompt, True, optimizer=optimizer, conversationally=conversationally,
-                tools=tools, tool_choice=tool_choice
-            ):
-                yield self.get_message(response)
-
-        def for_non_stream():
-            return self.get_message(
-                self.ask(
-                    prompt,
-                    False,
-                    optimizer=optimizer,
-                    conversationally=conversationally,
-                    tools=tools,
-                    tool_choice=tool_choice,
-                )
-            )
-
-        return for_stream() if stream else for_non_stream()
-
-    def get_message(self, response: dict) -> str:
-        """Retrieves message only from response"""
-        assert isinstance(response, dict), "Response should be of dict data-type only"
-        if "text" in response:
-            return response["text"]
-        elif "tool_calls" in response:
-            # For tool calls, return a string representation
-            return json.dumps(response["tool_calls"])
-        return "" # Return empty string if neither text nor tool_calls found
-
-if __name__ == "__main__":
-    # Ensure curl_cffi is installed
-    print("-" * 80)
-    print(f"{'Model':<50} {'Status':<10} {'Response'}")
-    print("-" * 80)
-
-    # Test all available models
-    working = 0
-    total = len(TextPollinationsAI.AVAILABLE_MODELS)
-
-
-    for model in TextPollinationsAI.AVAILABLE_MODELS:
-        try:
-            test_ai = TextPollinationsAI(model=model, timeout=60)
-            # Test stream first
-            response_stream = test_ai.chat("Say 'Hello' in one word", stream=True)
-            response_text = ""
-            print(f"\r{model:<50} {'Streaming...':<10}", end="", flush=True)
-            for chunk in response_stream:
-                response_text += chunk
-
-            if response_text and len(response_text.strip()) > 0:
-                status = "✓"
-                # Clean and truncate response
-                clean_text = response_text.strip()
-                display_text = clean_text[:50] + "..." if len(clean_text) > 50 else clean_text
-            else:
-                status = "✗ (Stream)"
-                display_text = "Empty or invalid stream response"
-            print(f"\r{model:<50} {status:<10} {display_text}")
-
-            # Optional: Add non-stream test if needed
-            # print(f"\r{model:<50} {'Non-Stream...':<10}", end="", flush=True)
-            # response_non_stream = test_ai.chat("Say 'Hi' again", stream=False)
-            # if not response_non_stream or len(response_non_stream.strip()) == 0:
-            #      print(f"\r{model:<50} {'✗ (Non-Stream)':<10} Empty non-stream response")
-
-        except Exception as e:
-            print(f"\r{model:<50} {'✗':<10} {str(e)}")