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

webscout/Provider/ChatGPTClone.py

@@ -1,6 +1,7 @@
 import time
 import uuid
-import cloudscraper
+# import cloudscraper
+from curl_cffi.requests import Session, RequestsError
 import json
 import re
 from typing import Any, Dict, Optional, Generator, Union
@@ -9,10 +10,10 @@ from datetime import date
 
 from webscout.AIutel import Optimizers
 from webscout.AIutel import Conversation
-from webscout.AIutel import AwesomePrompts
+from webscout.AIutel import AwesomePrompts, sanitize_stream # Import sanitize_stream
 from webscout.AIbase import Provider
 from webscout import WEBS, exceptions
-from webscout.litagent import LitAgent
+# from webscout.litagent import LitAgent
 
 class ChatGPTClone(Provider):
     """
@@ -22,6 +23,11 @@ class ChatGPTClone(Provider):
     
     url = "https://chatgpt-clone-ten-nu.vercel.app"
     AVAILABLE_MODELS = ["gpt-4", "gpt-3.5-turbo"]
+    SUPPORTED_IMPERSONATION = [
+        "chrome110", "chrome116", "chrome119", "chrome120",
+        "chrome99_android", "edge99", "edge101",
+        "safari15_3", "safari15_6_1", "safari17_0", "safari17_2_1"
+    ]
     
     def __init__(
         self,
@@ -37,15 +43,18 @@ class ChatGPTClone(Provider):
         model: str = "gpt-4",
         temperature: float = 0.6,
         top_p: float = 0.7,
-        browser: str = "chrome",
+        impersonate: str = "chrome120",
         system_prompt: str = "You are a helpful assistant."
     ):
-        """Initialize the ChatGPT Clone client."""
+        """Initialize the ChatGPT Clone client using curl_cffi."""
         if model not in self.AVAILABLE_MODELS:
             raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
+        if impersonate not in self.SUPPORTED_IMPERSONATION:
+            raise ValueError(f"Invalid impersonate browser: {impersonate}. Choose from: {self.SUPPORTED_IMPERSONATION}")
             
         self.model = model
-        self.session = cloudscraper.create_scraper()
+        self.impersonate = impersonate
+        self.session = Session(impersonate=self.impersonate, proxies=proxies, timeout=timeout)
         self.is_conversation = is_conversation
         self.max_tokens_to_sample = max_tokens
         self.timeout = timeout
@@ -54,28 +63,17 @@ class ChatGPTClone(Provider):
         self.top_p = top_p
         self.system_prompt = system_prompt
 
-        # Initialize LitAgent for user agent generation
-        self.agent = LitAgent()
-        # Use fingerprinting to create a consistent browser identity
-        self.fingerprint = self.agent.generate_fingerprint(browser)
-
-        # Use the fingerprint for headers
         self.headers = {
-            "Accept": self.fingerprint["accept"],
-            "Accept-Encoding": "gzip, deflate, br, zstd",
-            "Accept-Language": self.fingerprint["accept_language"],
             "Content-Type": "application/json",
-            "DNT": "1",
             "Origin": self.url,
             "Referer": f"{self.url}/",
-            "Sec-CH-UA": self.fingerprint["sec_ch_ua"] or '"Not)A;Brand";v="99", "Microsoft Edge";v="127", "Chromium";v="127"',
-            "Sec-CH-UA-Mobile": "?0",
-            "Sec-CH-UA-Platform": f'"{self.fingerprint["platform"]}"',
-            "User-Agent": self.fingerprint["user_agent"],
+            "DNT": "1",
+            "Sec-Fetch-Dest": "empty",
+            "Sec-Fetch-Mode": "cors",
+            "Sec-Fetch-Site": "same-origin",
+            "TE": "trailers"
         }
-        
-        # Create session cookies with unique identifiers
-        self.cookies = {"__Host-session": uuid.uuid4().hex, '__cf_bm': uuid.uuid4().hex}
+        self.session.headers.update(self.headers)
 
         self.__available_optimizers = (
             method
@@ -92,35 +90,32 @@ class ChatGPTClone(Provider):
             is_conversation, self.max_tokens_to_sample, filepath, update_file
         )
         self.conversation.history_offset = history_offset
-        self.session.proxies = proxies
-        
-        # Set consistent headers for the scraper session
-        for header, value in self.headers.items():
-            self.session.headers[header] = value
-
-    def refresh_identity(self, browser: str = None):
-        """Refreshes the browser identity fingerprint."""
-        browser = browser or self.fingerprint.get("browser_type", "chrome")
-        self.fingerprint = self.agent.generate_fingerprint(browser)
-        
-        # Update headers with new fingerprint
-        self.headers.update({
-            "Accept": self.fingerprint["accept"],
-            "Accept-Language": self.fingerprint["accept_language"],
-            "Sec-CH-UA": self.fingerprint["sec_ch_ua"] or self.headers["Sec-CH-UA"],
-            "Sec-CH-UA-Platform": f'"{self.fingerprint["platform"]}"',
-            "User-Agent": self.fingerprint["user_agent"],
-        })
-        
-        # Update session headers
-        for header, value in self.headers.items():
-            self.session.headers[header] = value
-        
-        # Generate new cookies
-        self.cookies = {"__Host-session": uuid.uuid4().hex, '__cf_bm': uuid.uuid4().hex}
-        
-        return self.fingerprint
 
+    def refresh_identity(self, impersonate: str = None):
+        """Re-initializes the curl_cffi session with a new impersonation target."""
+        impersonate = impersonate or self.impersonate
+        if impersonate not in self.SUPPORTED_IMPERSONATION:
+            raise ValueError(f"Invalid impersonate browser: {impersonate}. Choose from: {self.SUPPORTED_IMPERSONATION}")
+        self.impersonate = impersonate
+        self.session = Session(
+            impersonate=self.impersonate,
+            proxies=self.session.proxies,
+            timeout=self.timeout
+        )
+        self.session.headers.update(self.headers)
+        return self.impersonate
+
+    @staticmethod
+    def _chatgptclone_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from the ChatGPTClone stream format '0:"..."'."""
+        if isinstance(chunk, str):
+            match = re.search(r'0:"((?:[^\\"]|\\.)*)"', chunk) # Use the existing regex
+            if match:
+                content = match.group(1)
+                # Decode JSON string escapes and then unicode escapes
+                decoded_content = json.loads(f'"{content}"').encode().decode('unicode_escape')
+                return decoded_content
+        return None
     def ask(
         self,
         prompt: str,
@@ -129,7 +124,7 @@ class ChatGPTClone(Provider):
         optimizer: str = None,
         conversationally: bool = False,
     ) -> Union[Dict[str, Any], Generator]:
-        """Send a message to the ChatGPT Clone API"""
+        """Send a message to the ChatGPT Clone API using curl_cffi"""
         conversation_prompt = self.conversation.gen_complete_prompt(prompt)
         if optimizer:
             if optimizer in self.__available_optimizers:
@@ -149,38 +144,53 @@ class ChatGPTClone(Provider):
             "model": self.model
         }
 
+        api_url = f"{self.url}/api/chat"
+
+        def _make_request(attempt_refresh=True):
+            try:
+                response = self.session.post(api_url, json=payload, stream=True)
+                response.raise_for_status()
+                return response
+            except RequestsError as e:
+                if attempt_refresh and e.response and e.response.status_code in [403, 429]:
+                    self.refresh_identity()
+                    return _make_request(attempt_refresh=False)
+                else:
+                    err_msg = f"Request failed: {e}"
+                    if e.response is not None:
+                        err_msg = f"Failed to generate response - ({e.response.status_code}, {e.response.reason}) - {e.response.text}"
+                    raise exceptions.FailedToGenerateResponseError(err_msg) from e
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"An unexpected error occurred: {e}") from e
+
         def for_stream():
+            response = _make_request()
+            streaming_text = "" # Initialize outside try block
             try:
-                with self.session.post(f"{self.url}/api/chat", headers=self.headers, cookies=self.cookies, json=payload, stream=True, timeout=self.timeout) as response:
-                    if not response.ok:
-                        # If we get a non-200 response, try refreshing our identity once
-                        if response.status_code in [403, 429]:
-                            self.refresh_identity()
-                            # Retry with new identity
-                            with self.session.post(f"{self.url}/api/chat", headers=self.headers, cookies=self.cookies, json=payload, stream=True, timeout=self.timeout) as retry_response:
-                                if not retry_response.ok:
-                                    raise exceptions.FailedToGenerateResponseError(
-                                        f"Failed to generate response after identity refresh - ({retry_response.status_code}, {retry_response.reason}) - {retry_response.text}"
-                                    )
-                                response = retry_response
-                        else:
-                            raise exceptions.FailedToGenerateResponseError(
-                                f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
-                            )
-
-                    streaming_text = ""
-                    for line in response.iter_lines(decode_unicode=True):
-                        if line:
-                            match = re.search(r'0:"(.*?)"', line)
-                            if match:
-                                content = match.group(1)
-                                streaming_text += content
-                                yield content if raw else dict(text=content)
-                                
-                    self.last_response.update(dict(text=streaming_text))
-                    self.conversation.update_chat_history(prompt, streaming_text)
+                # Use sanitize_stream
+                processed_stream = sanitize_stream(
+                    data=response.iter_content(chunk_size=None), # Pass byte iterator
+                    intro_value=None, # No simple prefix
+                    to_json=False,    # Content is text after extraction
+                    content_extractor=self._chatgptclone_extractor, # Use the specific extractor
+                    yield_raw_on_error=True # Yield even if extractor fails (might get metadata lines)
+                )
+
+                for content_chunk in processed_stream:
+                    # content_chunk is the string extracted by _chatgptclone_extractor
+                    if content_chunk and isinstance(content_chunk, str):
+                        streaming_text += content_chunk
+                        yield content_chunk if raw else dict(text=content_chunk)
+
+            except RequestsError as e:
+                raise exceptions.FailedToGenerateResponseError(f"Stream interrupted by request error: {e}") from e
             except Exception as e:
-                raise exceptions.FailedToGenerateResponseError(f"Request failed: {e}")
+                raise exceptions.FailedToGenerateResponseError(f"Error processing stream: {e}") from e
+            finally:
+                # Update history after stream finishes or fails
+                self.last_response.update(dict(text=streaming_text))
+                self.conversation.update_chat_history(prompt, streaming_text)
+                response.close()
 
         def for_non_stream():
             for _ in for_stream():
@@ -202,25 +212,26 @@ class ChatGPTClone(Provider):
                 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:
         """Extract message text from response"""
         assert isinstance(response, dict)
-        formatted_text = response["text"].replace('\\n', '\n').replace('\\n\\n', '\n\n')
+        if not isinstance(response, dict) or "text" not in response:
+            return str(response)
+        # Extractor handles formatting
+        formatted_text = response.get("text", "")
         return formatted_text
 
 if __name__ == "__main__":
     from rich import print
-    ai = ChatGPTClone(timeout=5000)
+    ai = ChatGPTClone(timeout=120, impersonate="chrome120")
     response = ai.chat("write a poem about AI", stream=True)
     for chunk in response:
-        print(chunk, end="", flush=True) 
+        print(chunk, end="", flush=True)

webscout/Provider/ChatSandbox.py

@@ -0,0 +1,342 @@
+from typing import Optional, Union, Any, Dict, Generator, List
+from uuid import uuid4
+import json
+import re
+import random
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
+from curl_cffi.const import CurlHttpVersion
+
+from webscout.AIutel import sanitize_stream
+from webscout.AIutel import Optimizers
+from webscout.AIutel import Conversation
+from webscout.AIutel import AwesomePrompts
+from webscout.AIbase import Provider
+from webscout import exceptions
+from webscout.litagent import LitAgent
+
+class ChatSandbox(Provider):
+    """
+    Sends a chat message to the specified model via the chatsandbox API.
+
+    This provider allows you to interact with various AI models through the chatsandbox.com
+    interface, supporting different models/models like OpenAI, DeepSeek, Llama, etc.
+
+    Attributes:
+        model (str): The model to chat with (e.g., "openai", "deepseek", "llama").
+
+    Examples:
+        >>> from webscout.Provider.chatsandbox import ChatSandbox
+        >>> ai = ChatSandbox(model="openai")
+        >>> response = ai.chat("Hello, how are you?")
+        >>> print(response)
+        'I'm doing well, thank you for asking! How can I assist you today?'
+    """
+    AVAILABLE_MODELS = ["openai", "deepseek", "llama", "gemini", "mistral-large"]
+
+    def __init__(
+        self,
+        model: str = "openai",
+        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,
+    ):
+        """
+        Initializes the ChatSandbox API with given parameters.
+
+        Args:
+            model (str): The model to chat with (e.g., "openai", "deepseek", "llama").
+            is_conversation (bool): Whether the provider is in conversation mode.
+            max_tokens (int): Maximum number of tokens to sample.
+            timeout (int): Timeout for API requests.
+            intro (str): Introduction message for the conversation.
+            filepath (str): Filepath for storing conversation history.
+            update_file (bool): Whether to update the conversation history file.
+            proxies (dict): Proxies for the API requests.
+            history_offset (int): Offset for conversation history.
+            act (str): Act for the conversation.
+
+        Examples:
+            >>> ai = ChatSandbox(model="openai", system_prompt="You are a friendly assistant.")
+            >>> print(ai.model)
+            'openai'
+        """
+        if model not in self.AVAILABLE_MODELS:
+            raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
+
+        # Initialize curl_cffi Session
+        self.session = Session()
+        self.model = model
+        self.is_conversation = is_conversation
+        self.max_tokens_to_sample = max_tokens
+        self.api_endpoint = "https://chatsandbox.com/api/chat"
+        self.timeout = timeout
+        self.last_response = {}
+
+        # Initialize LitAgent for user agent generation
+        self.agent = LitAgent()
+
+        # Set up headers
+        self.headers = {
+            'authority': 'chatsandbox.com',
+            'accept': '*/*',
+            'accept-encoding': 'gzip, deflate, br',
+            'accept-language': 'en-US,en;q=0.9',
+            'content-type': 'application/json',
+            'origin': 'https://chatsandbox.com',
+            'referer': f'https://chatsandbox.com/chat/{self.model}',
+            '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',
+            'user-agent': self.agent.random(),
+            'dnt': '1',
+            'sec-gpc': '1',
+        }
+
+        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
+
+        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
+
+    @staticmethod
+    def _chatsandbox_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from the chatsandbox stream format."""
+        if isinstance(chunk, str):
+            try:
+                data = json.loads(chunk)
+                if isinstance(data, dict) and "reasoning_content" in data:
+                    return data["reasoning_content"]
+                return chunk
+            except json.JSONDecodeError:
+                return chunk
+        return None
+
+    def ask(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+        optimizer: str = None,
+        conversationally: bool = False,
+    ) -> Union[Dict[str, Any], Generator]:
+        """
+        Sends a prompt to the ChatSandbox API and returns the response.
+
+        Args:
+            prompt (str): The prompt to send to the API.
+            stream (bool): Whether to stream the response.
+            raw (bool): Whether to return the raw response.
+            optimizer (str): Optimizer to use for the prompt.
+            conversationally (bool): Whether to generate the prompt conversationally.
+
+        Returns:
+            Union[Dict[str, Any], Generator]: The API response.
+
+        Examples:
+            >>> ai = ChatSandbox()
+            >>> response = ai.ask("Tell me a joke!")
+            >>> print(response)
+            {'text': 'Why did the scarecrow win an award? Because he was outstanding in his field!'}
+        """
+        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}"
+                )
+
+        # Prepare the payload
+        payload = {
+            "messages": [conversation_prompt],
+            "character": self.model
+        }
+
+        def for_stream():
+            try:
+                # Use curl_cffi session post with updated impersonate and http_version
+                response = self.session.post(
+                    self.api_endpoint,
+                    headers=self.headers,
+                    json=payload,
+                    stream=True,
+                    timeout=self.timeout,
+                    impersonate="chrome120",  # Try a different impersonation profile
+                    http_version=CurlHttpVersion.V1_1  # Force HTTP/1.1
+                )
+                if not response.ok:
+                    raise exceptions.FailedToGenerateResponseError(
+                        f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
+                    )
+                
+                streaming_response = ""
+                # Use sanitize_stream with the custom extractor
+                processed_stream = sanitize_stream(
+                    data=response.iter_content(chunk_size=None),  # Pass byte iterator
+                    intro_value=None,  # No simple prefix to remove here
+                    to_json=False,     # Content is not JSON
+                    content_extractor=self._chatsandbox_extractor  # Use the specific extractor
+                )
+
+                for content_chunk in processed_stream:
+                    if content_chunk and isinstance(content_chunk, str):
+                        streaming_response += content_chunk
+                        yield content_chunk if raw else dict(text=content_chunk)
+
+                self.last_response.update(dict(text=streaming_response))
+                self.conversation.update_chat_history(
+                    prompt, self.get_message(self.last_response)
+                )
+            except CurlError as e:  # Catch CurlError
+                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}")
+            except Exception as e:  # Catch other potential exceptions
+                # Include the original exception type in the message for clarity
+                raise exceptions.FailedToGenerateResponseError(f"An unexpected error occurred ({type(e).__name__}): {e}")
+
+        def for_non_stream():
+            # This function implicitly uses the updated for_stream
+            for _ in for_stream():
+                pass
+            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,
+    ) -> str:
+        """
+        Generates a response from the ChatSandbox API.
+
+        Args:
+            prompt (str): The prompt to send to the API.
+            stream (bool): Whether to stream the response.
+            optimizer (str): Optimizer to use for the prompt.
+            conversationally (bool): Whether to generate the prompt conversationally.
+
+        Returns:
+            str: The API response.
+
+        Examples:
+            >>> ai = ChatSandbox()
+            >>> response = ai.chat("What's the weather today?")
+            >>> print(response)
+            'I don't have real-time weather data, but I can help you find weather information online.'
+        """
+        def for_stream():
+            for response in self.ask(
+                prompt,
+                stream=True,
+                raw=False,
+                optimizer=optimizer,
+                conversationally=conversationally,
+            ):
+                yield response.get("text", "")
+
+        if stream:
+            return for_stream()
+        else:
+            return self.get_message(
+                self.ask(
+                    prompt,
+                    stream=False,
+                    raw=False,
+                    optimizer=optimizer,
+                    conversationally=conversationally,
+                )
+            )
+
+    def get_message(self, response: Dict[str, Any]) -> str:
+        """
+        Extract the message from the API response.
+        
+        Args:
+            response (Dict[str, Any]): The API response.
+            
+        Returns:
+            str: The extracted message.
+        """
+        if not isinstance(response, dict):
+            return str(response)
+            
+        raw_text = response.get("text", "")
+
+        # Try to parse as JSON
+        try:
+            data = json.loads(raw_text)
+            if isinstance(data, dict):
+                # Check for different response formats
+                if "reasoning_content" in data:
+                    return data["reasoning_content"]
+                elif "content" in data:
+                    return data["content"]
+                elif "message" in data:
+                    return data["message"]
+                elif "response" in data:
+                    return data["response"]
+                elif "text" in data:
+                    return data["text"]
+                # Return the whole JSON if no specific field is found
+                return json.dumps(data, ensure_ascii=False)
+        except json.JSONDecodeError:
+            # If it's not JSON, return the raw text
+            pass
+
+        return raw_text.strip()
+
+# --- Example Usage ---
+if __name__ == "__main__":
+    from rich import print
+    # Ensure curl_cffi is installed
+    print("-" * 80)
+    print(f"{'Model':<50} {'Status':<10} {'Response'}")
+    print("-" * 80)
+
+    for model in ChatSandbox.AVAILABLE_MODELS:
+        try:
+            test_ai = ChatSandbox(model=model, timeout=60)
+            response = test_ai.chat("Say 'Hello' in one word")
+            response_text = response
+            
+            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(f"{model:<50} {status:<10} {display_text}")
+        except Exception as e:
+            print(f"{model:<50} {'✗':<10} {str(e)}")