webscout 8.2.7__py3-none-any.whl → 8.2.9__py3-none-any.whl

webscout/Provider/MCPCore.py

@@ -0,0 +1,315 @@
+import json
+import uuid
+from typing import Any, Dict, Generator, Union
+
+# Use curl_cffi for requests
+from curl_cffi.requests import Session
+from curl_cffi import CurlError
+
+from webscout.AIutel import Optimizers
+from webscout.AIutel import Conversation
+from webscout.AIutel import AwesomePrompts, sanitize_stream 
+from webscout.AIbase import Provider
+from webscout import exceptions
+from webscout.litagent import LitAgent
+
+class MCPCore(Provider):
+    """
+    A class to interact with the chat.mcpcore.xyz API.
+    Supports streaming responses.
+    """
+
+    # Add more models if known, starting with the one from the example
+    AVAILABLE_MODELS = [
+        "google/gemma-7b-it",
+        "deepseek-ai/deepseek-coder-33b-instruct",
+        "deepseek-ai/DeepSeek-R1",
+        "deepseek-ai/DeepSeek-R1-Distill-Qwen-14B",
+        "deepseek-ai/DeepSeek-v3-0324",
+        "fixie-ai/ultravox-v0_4_1-llama-3_1-8b",
+        "meta-llama/Llama-3.3-70B-Instruct",
+        "meta-llama/Llama-4-Maverick-Instruct",
+        "mistralai/Mistral-7B-Instruct-v0.2",
+        "qwen-max-latest",
+        "qwen-plus-latest",
+        "qwen2.5-coder-32b-instruct",
+        "qwen-turbo-latest",
+        "qwen2.5-14b-instruct-1m",
+        "GLM-4-32B",
+        "Z1-32B",
+        "Z1-Rumination",
+        "arena-model",
+        "qvq-72b-preview-0310",
+        "qwq-32b",
+        "qwen3-235b-a22b",
+        "qwen3-30b-a3b",
+        "qwen3-32b",
+        "deepseek-flash",
+        "@cf/meta/llama-4-scout-17b-16e-instruct",
+        "任务专用",
+    ]
+
+    def __init__(
+        self,
+        cookies_path: str,
+        is_conversation: bool = True,
+        max_tokens: int = 2048,
+        timeout: int = 60,
+        intro: str = None,
+        filepath: str = None,
+        update_file: bool = True,
+        proxies: dict = {},
+        history_offset: int = 10250,
+        act: str = None,
+        model: str = "qwen3-32b",
+        system_prompt: str = "You are a helpful assistant.",
+    ):
+        """Initializes the MCPCore API client."""
+        if model not in self.AVAILABLE_MODELS:
+            print(f"Warning: Model '{model}' not in known AVAILABLE_MODELS. Attempting to use anyway.")
+
+        self.api_endpoint = "https://chat.mcpcore.xyz/api/chat/completions"
+        self.model = model
+        self.system_prompt = system_prompt
+        self.cookies_path = cookies_path
+        self.cookie_string, self.token = self._load_cookies()
+
+        # Initialize curl_cffi Session
+        self.session = Session()
+
+        # Set up headers based on the provided request
+        self.headers = {
+            'authority': 'chat.mcpcore.xyz',
+            'accept': '*/*',
+            'accept-language': 'en-US,en;q=0.9,en-IN;q=0.8',
+            'authorization': f'Bearer {self.token}' if self.token else '',
+            'content-type': 'application/json',
+            'dnt': '1',
+            'origin': 'https://chat.mcpcore.xyz',
+            'referer': 'https://chat.mcpcore.xyz/',
+            'priority': 'u=1, i',
+            'sec-ch-ua': '"Microsoft Edge";v="135", "Not-A.Brand";v="8", "Chromium";v="135"',
+            'sec-ch-ua-mobile': '?0',
+            'sec-ch-ua-platform': '"Windows"',
+            'sec-fetch-dest': 'empty',
+            'sec-fetch-mode': 'cors',
+            'sec-fetch-site': 'same-origin',
+            'sec-gpc': '1',
+            'user-agent': LitAgent().random(),
+        }
+
+        # Apply headers, proxies, and cookies to the session
+        self.session.headers.update(self.headers)
+        self.session.proxies = proxies
+        self.cookies = {
+            'token': self.token,
+        }
+        for name, value in self.cookies.items():
+            self.session.cookies.set(name, value, domain="chat.mcpcore.xyz")
+
+        # Provider settings
+        self.is_conversation = is_conversation
+        self.max_tokens_to_sample = max_tokens
+        self.timeout = timeout
+        self.last_response = {}
+
+        # Initialize optimizers
+        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 _load_cookies(self) -> tuple[str, str]:
+        """Load cookies from a JSON file and build a cookie header string."""
+        try:
+            with open(self.cookies_path, "r") as f:
+                cookies = json.load(f)
+            cookie_string = "; ".join(
+                f"{cookie['name']}={cookie['value']}" for cookie in cookies if 'name' in cookie and 'value' in cookie
+            )
+            token = next(
+                (cookie.get("value") for cookie in cookies if cookie.get("name") == "token"),
+                "",
+            )
+            return cookie_string, token
+        except FileNotFoundError:
+            raise exceptions.FailedToGenerateResponseError(
+                f"Error: Cookies file not found at {self.cookies_path}!"
+            )
+        except json.JSONDecodeError:
+            raise exceptions.FailedToGenerateResponseError(
+                f"Error: Invalid JSON format in cookies file: {self.cookies_path}!"
+            )
+
+    def ask(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+        optimizer: str = None,
+        conversationally: bool = False,
+        **kwargs
+    ) -> Union[Dict[str, Any], Generator]:
+        """Sends a prompt to the MCPCore API and returns the response."""
+
+        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 exceptions.InvalidOptimizerError(
+                    f"Optimizer is not one of {self.__available_optimizers}"
+                )
+
+        chat_id = kwargs.get("chat_id", str(uuid.uuid4()))
+        message_id = str(uuid.uuid4())
+
+        payload = {
+            "stream": stream,
+            "model": self.model,
+            "messages": [
+                {"role": "system", "content": self.system_prompt},
+                {"role": "user", "content": conversation_prompt}
+            ],
+            "params": kwargs.get("params", {}),
+            "tool_servers": kwargs.get("tool_servers", []),
+            "features": kwargs.get("features", {"web_search": False}),
+            "chat_id": chat_id,
+            "id": message_id,
+            "stream_options": kwargs.get("stream_options", {"include_usage": True})
+        }
+
+        def for_stream():
+            streaming_text = ""
+            try:
+                response = self.session.post(
+                    self.api_endpoint,
+                    json=payload,
+                    stream=True,
+                    timeout=self.timeout,
+                    impersonate="chrome110"
+                )
+                response.raise_for_status()
+
+                # 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]"],
+                    content_extractor=lambda chunk: chunk.get('choices', [{}])[0].get('delta', {}).get('content') if isinstance(chunk, dict) else None,
+                    yield_raw_on_error=False # Skip non-JSON or lines where extractor fails
+                )
+
+                for content_chunk in processed_stream:
+                    # content_chunk is the string extracted by the content_extractor
+                    if content_chunk and isinstance(content_chunk, str):
+                        streaming_text += content_chunk
+                        yield dict(text=content_chunk) if not raw else content_chunk
+
+                self.last_response = {"text": streaming_text}
+                self.conversation.update_chat_history(prompt, self.get_message(self.last_response))
+
+            except CurlError as e:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}") from e
+            except Exception as e:
+                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 for_non_stream():
+            full_text = ""
+            try:
+                stream_generator = self.ask(
+                    prompt, stream=True, raw=False, optimizer=optimizer, conversationally=conversationally, **kwargs
+                )
+                for chunk_data in stream_generator:
+                    if isinstance(chunk_data, dict):
+                        full_text += chunk_data["text"]
+                    elif isinstance(chunk_data, str):
+                        full_text += chunk_data
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"Failed to aggregate non-stream response: {str(e)}") from e
+
+            return full_text if raw else 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,
+        **kwargs
+    ) -> Union[str, Generator[str, None, None]]:
+        """Generates a response from the MCPCore API."""
+
+        def for_stream_chat() -> Generator[str, None, None]:
+            gen = self.ask(
+                prompt, stream=True, raw=False,
+                optimizer=optimizer, conversationally=conversationally, **kwargs
+            )
+            for response_dict in gen:
+                yield self.get_message(response_dict)
+
+        def for_non_stream_chat() -> str:
+            response_data = self.ask(
+                prompt, stream=False, raw=False,
+                optimizer=optimizer, conversationally=conversationally, **kwargs
+            )
+            return self.get_message(response_data)
+
+        return for_stream_chat() if stream else for_non_stream_chat()
+
+    def get_message(self, response: Dict[str, Any]) -> str:
+        """Extracts the message from the API response."""
+        assert isinstance(response, dict), "Response should be of dict data-type only"
+        return response.get("text", "")
+
+# Example usage (remember to create a cookies.json file)
+if __name__ == "__main__":
+    from rich import print
+
+    cookies_file_path = "cookies.json"
+
+    print("-" * 80)
+    print(f"{'Model':<50} {'Status':<10} {'Response'}")
+    print("-" * 80)
+
+    for model in MCPCore.AVAILABLE_MODELS:
+        try:
+            test_ai = MCPCore(cookies_path=cookies_file_path, model=model, timeout=60)
+            response = test_ai.chat("Say 'Hello' in one word", stream=True)
+            response_text = ""
+            # Accumulate the response text without printing in the loop
+            for chunk in response:
+                response_text += chunk
+
+            if response_text and len(response_text.strip()) > 0:
+                status = "✓"
+                # Truncate response if too long
+                display_text = response_text.strip()[:50] + "..." if len(response_text.strip()) > 50 else response_text.strip()
+            else:
+                status = "✗"
+                display_text = "Empty or invalid response"
+            # Print the final status and response, overwriting the "Testing..." line
+            print(f"\r{model:<50} {status:<10} {display_text}")
+        except Exception as e:
+            # Print error, overwriting the "Testing..." line
+            print(f"\r{model:<50} {'✗':<10} {str(e)}")

webscout/Provider/Marcus.py

@@ -0,0 +1,198 @@
+from curl_cffi.requests import Session
+from curl_cffi import CurlError
+import json
+from typing import Union, Any, Dict, Optional, Generator
+
+from webscout.AIutel import Optimizers
+from webscout.AIutel import Conversation
+from webscout.AIutel import AwesomePrompts, sanitize_stream # Import sanitize_stream
+from webscout.AIbase import Provider
+from webscout import exceptions
+
+class Marcus(Provider):
+    """
+    This class provides methods for interacting with the AskMarcus API.
+    Improved to match webscout provider standards.
+    """
+
+    def __init__(
+        self,
+        is_conversation: bool = True,
+        max_tokens: int = 2048, # 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
+    ):
+        """Initializes the Marcus API."""
+        # Initialize curl_cffi Session
+        self.session = Session()
+        self.is_conversation = is_conversation
+        self.max_tokens_to_sample = max_tokens
+        self.api_endpoint = "https://www.askmarcus.app/api/response"
+        self.timeout = timeout
+        self.last_response = {}
+        
+        self.headers = {
+            'content-type': 'application/json',
+            'accept': '*/*',
+            'origin': 'https://www.askmarcus.app',
+            'referer': 'https://www.askmarcus.app/chat',
+        }
+        
+        # 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,
+    ) -> Union[Dict[str, Any], Generator[Any, None, None]]:
+        """Sends a prompt to the AskMarcus API and returns the response."""
+        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 exceptions.FailedToGenerateResponseError(
+                    f"Optimizer is not one of {self.__available_optimizers}"
+                )
+
+        data = {"message": conversation_prompt}
+
+        def for_stream():
+            streaming_text = "" # Initialize outside try block
+            try:
+                # Use curl_cffi session post with impersonate
+                response = self.session.post(
+                    self.api_endpoint,
+                    # headers are set on the session
+                    json=data,
+                    stream=True,
+                    timeout=self.timeout,
+                    # proxies are set on the session
+                    impersonate="chrome110" # Use a common impersonation profile
+                )
+                response.raise_for_status() # Check for HTTP errors
+                
+                # Use sanitize_stream to decode bytes and yield text chunks
+                processed_stream = sanitize_stream(
+                    data=response.iter_content(chunk_size=None), # Pass byte iterator
+                    intro_value=None, # No prefix
+                    to_json=False,    # It's plain text
+                    yield_raw_on_error=True
+                )
+
+                for content_chunk in processed_stream:
+                    if content_chunk and isinstance(content_chunk, str):
+                        streaming_text += content_chunk # Aggregate text
+                        yield {"text": content_chunk} if not raw else content_chunk
+                # Update history after stream finishes
+                self.last_response = {"text": streaming_text} # Store aggregated text
+                self.conversation.update_chat_history(
+                    prompt, streaming_text
+                )
+
+            except CurlError as e: # Catch CurlError
+                raise exceptions.ProviderConnectionError(f"Error connecting to Marcus (CurlError): {str(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.ProviderConnectionError(f"Error connecting to Marcus ({type(e).__name__}): {str(e)} - {err_text}") from e
+
+        def for_non_stream():
+            try:
+                # Use curl_cffi session post with impersonate
+                response = self.session.post(
+                    self.api_endpoint,
+                    # headers are set on the session
+                    json=data,
+                    timeout=self.timeout,
+                    # proxies are set on the session
+                    impersonate="chrome110" # Use a common impersonation profile
+                )
+                response.raise_for_status() # Check for HTTP errors
+                
+                response_text_raw = response.text # Get raw text
+
+                # Process the text using sanitize_stream (even though it's not streaming)
+                processed_stream = sanitize_stream(
+                    data=response_text_raw,
+                    intro_value=None, # No prefix
+                    to_json=False     # It's plain text
+                )
+                # Aggregate the single result
+                full_response = "".join(list(processed_stream))
+
+                self.last_response = {"text": full_response}
+                self.conversation.update_chat_history(prompt, full_response)
+                # Return dict or raw string
+                return full_response if raw else self.last_response
+
+            except CurlError as e: # Catch CurlError
+                 raise exceptions.ProviderConnectionError(f"Error connecting to Marcus (CurlError): {str(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.ProviderConnectionError(f"Error connecting to Marcus ({type(e).__name__}): {str(e)} - {err_text}") from e
+
+
+        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, Generator[str, None, None]]:
+        """Generates a response from the AskMarcus API."""
+        response_data = self.ask(
+            prompt, stream=False, raw=False, # Always get the full response
+            optimizer=optimizer, conversationally=conversationally
+        )
+        if stream:
+            def stream_wrapper():
+                yield self.get_message(response_data)
+            return stream_wrapper()
+        else:
+            return self.get_message(response_data)
+
+    def get_message(self, response: Dict[str, Any]) -> str:
+        """Extracts the message from the API response."""
+        assert isinstance(response, dict), "Response should be of dict data-type only"
+        return response.get("text", "")
+
+if __name__ == "__main__":
+    # Ensure curl_cffi is installed
+    from rich import print
+    ai = Marcus()
+    response = ai.chat("hi", stream=True)
+    for chunk in response:
+        print(chunk, end="", flush=True)