webscout 8.2.7__py3-none-any.whl → 8.2.8__py3-none-any.whl

webscout/Provider/AISEARCH/webpilotai_search.py

@@ -0,0 +1,281 @@
+import requests
+import json
+import re
+from typing import Dict, Optional, Generator, Union, Any
+
+from webscout.AIbase import AISearch
+from webscout import exceptions
+from webscout.litagent import LitAgent
+
+
+class Response:
+    """A wrapper class for webpilotai API responses.
+    
+    This class automatically converts response objects to their text representation
+    when printed or converted to string.
+    
+    Attributes:
+        text (str): The text content of the response
+        
+    Example:
+        >>> response = Response("Hello, world!")
+        >>> print(response)
+        Hello, world!
+        >>> str(response)
+        'Hello, world!'
+    """
+    def __init__(self, text: str):
+        self.text = text
+    
+    def __str__(self):
+        return self.text
+    
+    def __repr__(self):
+        return self.text
+
+
+class webpilotai(AISearch):
+    """A class to interact with the webpilotai (WebPilot) AI search API.
+    
+    webpilotai provides a web-based comprehensive search response interface that returns AI-generated 
+    responses with source references and related questions. It supports both streaming and
+    non-streaming responses.
+    
+    Basic Usage:
+        >>> from webscout import webpilotai
+        >>> ai = webpilotai()
+        >>> # Non-streaming example
+        >>> response = ai.search("What is Python?")
+        >>> print(response)
+        Python is a high-level programming language...
+        
+        >>> # Streaming example
+        >>> for chunk in ai.search("Tell me about AI", stream=True):
+        ...     print(chunk, end="", flush=True)
+        Artificial Intelligence is...
+        
+        >>> # Raw response format
+        >>> for chunk in ai.search("Hello", stream=True, raw=True):
+        ...     print(chunk)
+        {'text': 'Hello'}
+        {'text': ' there!'}
+    
+    Args:
+        timeout (int, optional): Request timeout in seconds. Defaults to 90.
+        proxies (dict, optional): Proxy configuration for requests. Defaults to None.
+    """
+
+    def __init__(
+        self,
+        timeout: int = 90,
+        proxies: Optional[dict] = None,
+    ):
+        """Initialize the webpilotai API client.
+        
+        Args:
+            timeout (int, optional): Request timeout in seconds. Defaults to 90.
+            proxies (dict, optional): Proxy configuration for requests. Defaults to None.
+        
+        Example:
+            >>> ai = webpilotai(timeout=120)  # Longer timeout
+            >>> ai = webpilotai(proxies={'http': 'http://proxy.com:8080'})  # With proxy
+        """
+        self.session = requests.Session()
+        self.api_endpoint = "https://api.webpilotai.com/rupee/v1/search"
+        self.timeout = timeout
+        self.last_response = {}
+        
+        # The 'Bearer null' is part of the API's expected headers
+        self.headers = {
+            'Accept': 'application/json, text/plain, */*, text/event-stream',
+            'Content-Type': 'application/json;charset=UTF-8',
+            'Authorization': 'Bearer null',
+            'Origin': 'https://www.webpilot.ai',
+            'Referer': 'https://www.webpilot.ai/',
+            'User-Agent': LitAgent().random(),
+        }
+        
+        self.session.headers.update(self.headers)
+        self.proxies = proxies
+
+    def search(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+    ) -> Union[Response, Generator[Union[Dict[str, str], Response], None, None]]:
+        """Search using the webpilotai API and get AI-generated responses.
+        
+        This method sends a search query to webpilotai and returns the AI-generated response.
+        It supports both streaming and non-streaming modes, as well as raw response format.
+        
+        Args:
+            prompt (str): The search query or prompt to send to the API.
+            stream (bool, optional): If True, yields response chunks as they arrive.
+                                   If False, returns complete response. Defaults to False.
+            raw (bool, optional): If True, returns raw response dictionaries with 'text' key.
+                                If False, returns Response objects that convert to text automatically.
+                                Defaults to False.
+        
+        Returns:
+            Union[Response, Generator[Union[Dict[str, str], Response], None, None]]: 
+                - If stream=False: Returns complete response as Response object
+                - If stream=True: Yields response chunks as either Dict or Response objects
+        
+        Raises:
+            APIConnectionError: If the API request fails
+        
+        Examples:
+            Basic search:
+            >>> ai = webpilotai()
+            >>> response = ai.search("What is Python?")
+            >>> print(response)
+            Python is a programming language...
+            
+            Streaming response:
+            >>> for chunk in ai.search("Tell me about AI", stream=True):
+            ...     print(chunk, end="")
+            Artificial Intelligence...
+            
+            Raw response format:
+            >>> for chunk in ai.search("Hello", stream=True, raw=True):
+            ...     print(chunk)
+            {'text': 'Hello'}
+            {'text': ' there!'}
+        """
+        payload = {
+            "q": prompt,
+            "threadId": ""  # Empty for new search
+        }
+        
+        def for_stream():
+            full_response_content = ""
+            current_event_name = None
+            current_data_buffer = []
+            
+            try:
+                with self.session.post(
+                    self.api_endpoint,
+                    json=payload,
+                    stream=True,
+                    timeout=self.timeout,
+                    proxies=self.proxies
+                ) as response:
+                    if not response.ok:
+                        raise exceptions.APIConnectionError(
+                            f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
+                        )
+                    
+                    # Process the stream line by line
+                    for line in response.iter_lines(decode_unicode=True):
+                        if not line:  # Empty line indicates end of an event
+                            if current_data_buffer:
+                                # Process the completed event
+                                full_data = "\n".join(current_data_buffer)
+                                if current_event_name == "message":
+                                    try:
+                                        data_payload = json.loads(full_data)
+                                        # Check structure based on the API response
+                                        if data_payload.get('type') == 'data':
+                                            content_chunk = data_payload.get('data', {}).get('content', "")
+                                            if content_chunk:
+                                                full_response_content += content_chunk
+                                                
+                                                # Yield the new content chunk
+                                                if raw:
+                                                    yield {"text": content_chunk}
+                                                else:
+                                                    yield Response(content_chunk)
+                                    except json.JSONDecodeError:
+                                        pass
+                                    except Exception as e:
+                                        # Handle exceptions gracefully in stream processing
+                                        pass
+                                
+                                # Reset for the next event
+                                current_event_name = None
+                                current_data_buffer = []
+                            continue
+                        
+                        # Parse SSE fields
+                        if line.startswith('event:'):
+                            current_event_name = line[len('event:'):].strip()
+                        elif line.startswith('data:'):
+                            data_part = line[len('data:'):]
+                            # Remove leading space if present (common in SSE)
+                            if data_part.startswith(' '):
+                                data_part = data_part[1:]
+                            current_data_buffer.append(data_part)
+                    
+                    # Process any remaining data in buffer if stream ended without blank line
+                    if current_data_buffer and current_event_name == "message":
+                        try:
+                            full_data = "\n".join(current_data_buffer)
+                            data_payload = json.loads(full_data)
+                            if data_payload.get('type') == 'data':
+                                content_chunk = data_payload.get('data', {}).get('content', "")
+                                if content_chunk and len(content_chunk) > len(full_response_content):
+                                    delta = content_chunk[len(full_response_content):]
+                                    full_response_content += delta
+                                    
+                                    if raw:
+                                        yield {"text": delta}
+                                    else:
+                                        yield Response(delta)
+                        except (json.JSONDecodeError, Exception):
+                            pass
+                
+            except requests.exceptions.Timeout:
+                raise exceptions.APIConnectionError("Request timed out")
+            except requests.exceptions.RequestException as e:
+                raise exceptions.APIConnectionError(f"Request failed: {e}")
+
+        def for_non_stream():
+            full_response = ""
+            for chunk in for_stream():
+                if raw:
+                    yield chunk
+                else:
+                    full_response += str(chunk)
+            
+            if not raw:
+                # Format the response for better readability
+                formatted_response = self.format_response(full_response)
+                self.last_response = Response(formatted_response)
+                return self.last_response
+
+        return for_stream() if stream else for_non_stream()
+    
+    @staticmethod
+    def format_response(text: str) -> str:
+        """Format the response text for better readability.
+        
+        Args:
+            text (str): The raw response text
+        
+        Returns:
+            str: Formatted text with improved structure
+        """
+        # Clean up formatting
+        # Remove excessive newlines
+        clean_text = re.sub(r'\n{3,}', '\n\n', text)
+        
+        # Ensure consistent spacing around sections
+        clean_text = re.sub(r'([.!?])\s*\n\s*([A-Z])', r'\1\n\n\2', clean_text)
+        
+        # Clean up any leftover HTML or markdown artifacts
+        clean_text = re.sub(r'<[^>]*>', '', clean_text)
+        
+        # Remove trailing whitespace on each line
+        clean_text = '\n'.join(line.rstrip() for line in clean_text.split('\n'))
+        
+        return clean_text.strip()
+
+
+if __name__ == "__main__":
+    from rich import print
+    
+    ai = webpilotai()
+    response = ai.search(input(">>> "), stream=True, raw=False)
+    for chunk in response:
+        print(chunk, end="", flush=True)

webscout/Provider/Aitopia.py

@@ -0,0 +1,316 @@
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
+import json
+import uuid
+import time
+import hashlib
+from typing import Any, Dict, Optional, Generator, Union
+
+from webscout.AIutel import Optimizers
+from webscout.AIutel import Conversation, sanitize_stream # Import sanitize_stream
+from webscout.AIutel import AwesomePrompts
+from webscout.AIbase import Provider, AsyncProvider
+from webscout import exceptions
+from webscout.litagent import LitAgent
+
+class Aitopia(Provider):
+    """
+    A class to interact with the Aitopia API with LitAgent user-agent.
+    """
+
+    AVAILABLE_MODELS = [
+        "Claude 3 Haiku",
+        "GPT-4o Mini",
+        "Gemini 1.5 Flash",
+        "Llama 3.1 70B"
+    ]
+
+    def __init__(
+        self,
+        is_conversation: bool = True,
+        max_tokens: int = 2049,
+        timeout: int = 30,
+        intro: str = None,
+        filepath: str = None,
+        update_file: bool = True,
+        proxies: dict = {},
+        history_offset: int = 10250,
+        act: str = None,
+        model: str = "Claude 3 Haiku",
+        browser: str = "chrome"
+    ):
+        """Initializes the Aitopia API client."""
+        if model not in self.AVAILABLE_MODELS:
+            raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
+            
+        self.url = "https://extensions.aitopia.ai/ai/send"
+        
+        # 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": "text/plain",
+            "accept-language": self.fingerprint["accept_language"],
+            "content-type": "text/plain;charset=UTF-8",
+            "dnt": "1",
+            "origin": "https://chat.aitopia.ai",
+            "priority": "u=1, i",
+            "referer": "https://chat.aitopia.ai/",
+            "sec-ch-ua": self.fingerprint["sec_ch_ua"] or '"Chromium";v="134", "Not:A-Brand";v="24", "Microsoft Edge";v="134"',
+            "sec-ch-ua-mobile": "?0",
+            "sec-ch-ua-platform": f'"{self.fingerprint["platform"]}"',
+            "sec-fetch-dest": "empty",
+            "sec-fetch-mode": "cors",
+            "sec-fetch-site": "same-site",
+            "user-agent": self.fingerprint["user_agent"]
+        }
+        
+        self.session = Session() # Use curl_cffi Session
+        self.session.headers.update(self.headers)
+        self.session.proxies = proxies # Assign proxies directly
+
+        self.is_conversation = is_conversation
+        self.max_tokens_to_sample = max_tokens
+        self.timeout = timeout
+        self.last_response = {}
+        self.model = model
+
+        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 refresh_identity(self, browser: str = None):
+        """
+        Refreshes the browser identity fingerprint.
+        
+        Args:
+            browser: Specific browser to use for the new 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-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
+        
+        return self.fingerprint
+
+    def generate_uuid_search(self):
+        """Generate a UUID and convert to base64-like string."""
+        uuid_str = str(uuid.uuid4())
+        return uuid_str.replace('-', '')
+
+    def generate_hopekey(self):
+        """Generate a random string and hash it."""
+        random_str = str(uuid.uuid4()) + str(time.time())
+        return hashlib.md5(random_str.encode()).hexdigest()
+
+    @staticmethod
+    def _aitopia_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from Aitopia stream JSON objects."""
+        if isinstance(chunk, dict):
+            # Handle Claude 3 Haiku response format
+            if "delta" in chunk and "text" in chunk["delta"]:
+                return chunk["delta"]["text"]
+            # Handle GPT-4o Mini response format
+            elif "choices" in chunk and "0" in chunk["choices"]:
+                return chunk["choices"]["0"]["delta"].get("content")
+            # Add other potential formats here if needed
+        return None
+
+
+    def ask(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+        optimizer: str = None,
+        conversationally: bool = False,
+    ) -> Union[Dict[str, Any], Generator]:
+        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}")
+
+        # Generate hopekey and update headers
+        hopekey = self.generate_hopekey()
+        self.headers["hopekey"] = hopekey
+        
+        # Default history if none provided
+        history = [
+            {
+                "item": "Hello, how can I help you today?",
+                "role": "assistant",
+                # "model": "GPT-4o Mini"
+            }
+        ]
+        
+        # Generate current timestamp for chat_id
+        current_time = int(time.time() * 1000)
+        
+        # Request payload
+        payload = {
+            "history": history,
+            "text": conversation_prompt,
+            "model": self.model,
+            "stream": stream,
+            "uuid_search": self.generate_uuid_search(),
+            "mode": "ai_chat",
+            "prompt_mode": False,
+            "extra_key": "__all",
+            "extra_data": {"prompt_mode": False},
+            "chat_id": current_time,
+            "language_detail": {
+                "lang_code": "en",
+                "name": "English",
+                "title": "English"
+            },
+            "is_continue": False,
+            "lang_code": "en"
+        }
+
+        def for_stream():
+            streaming_text = "" # Initialize outside try block
+            try:
+                response = self.session.post(
+                    self.url, headers=self.headers, json=payload, stream=True, timeout=self.timeout,
+                    impersonate="chrome120" # Add impersonate
+                )
+                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=self._aitopia_extractor, # Use the specific extractor
+                    yield_raw_on_error=False # Skip non-JSON lines or lines where extractor fails
+                )
+
+                for content_chunk in processed_stream:
+                    # content_chunk is the string extracted by _aitopia_extractor
+                    if content_chunk and isinstance(content_chunk, str):
+                        streaming_text += content_chunk
+                        resp = dict(text=content_chunk)
+                        yield resp if not raw else content_chunk
+
+            except CurlError as e:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {str(e)}") from e
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed: {str(e)}")
+            finally:
+                # Update history after stream finishes or fails
+                if streaming_text:
+                    self.last_response = {"text": streaming_text}
+                    self.conversation.update_chat_history(prompt, streaming_text)
+
+        def for_non_stream():
+            try:
+                response = self.session.post(
+                    self.url, headers=self.headers, json=payload, timeout=self.timeout,
+                    impersonate="chrome120" # Add impersonate
+                )
+                response.raise_for_status()
+
+                response_text_raw = response.text # Get raw text
+
+                # Use sanitize_stream to parse the non-streaming JSON response
+                # Assuming non-stream uses the GPT format based on original code
+                processed_stream = sanitize_stream(
+                    data=response_text_raw,
+                    to_json=True, # Parse the whole text as JSON
+                    intro_value=None,
+                    content_extractor=lambda chunk: chunk.get("choices", [{}])[0].get("message", {}).get("content") if isinstance(chunk, dict) else None,
+                    yield_raw_on_error=False
+                )
+                # Extract the single result
+                content = next(processed_stream, None)
+                content = content if isinstance(content, str) else "" # Ensure it's a string
+
+                if content: # Check if content was successfully extracted
+                    self.last_response = {"text": content}
+                    self.conversation.update_chat_history(prompt, content)
+                    return {"text": content}
+                else:
+                    raise exceptions.FailedToGenerateResponseError("No response content found or failed to parse")
+            except CurlError as e:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}") from e
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed: {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]]:
+        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:
+        assert isinstance(response, dict), "Response should be of dict data-type only"
+        return response["text"]
+
+if __name__ == "__main__":
+    print("-" * 80)
+    print(f"{'Model':<50} {'Status':<10} {'Response'}")
+    print("-" * 80)
+
+    for model in Aitopia.AVAILABLE_MODELS:
+        try:
+            test_ai = Aitopia(model=model, timeout=60)
+            response = test_ai.chat("Say 'Hello' in one word", stream=True)
+            response_text = ""
+            for chunk in response:
+                response_text += chunk
+            
+            if response_text and len(response_text.strip()) > 0:
+                status = "✓"
+                # Clean and truncate response
+                clean_text = response_text.strip().encode('utf-8', errors='ignore').decode('utf-8')
+                display_text = clean_text[:50] + "..." if len(clean_text) > 50 else clean_text
+            else:
+                status = "✗"
+                display_text = "Empty or invalid response"
+            print(f"\r{model:<50} {status:<10} {display_text}")
+        except Exception as e:
+            print(f"\r{model:<50} {'✗':<10} {str(e)}")