webscout 7.9__py3-none-any.whl → 8.1__py3-none-any.whl

webscout/Provider/AISEARCH/scira_search.py

@@ -0,0 +1,320 @@
+import requests
+import json
+import re
+import uuid
+import time
+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 SCIRA 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 Scira(AISearch):
+    """A class to interact with the SCIRA AI search API.
+
+    SCIRA provides a powerful search interface that returns AI-generated responses
+    based on web content. It supports both streaming and non-streaming responses.
+
+    Basic Usage:
+        >>> from webscout import Scira
+        >>> ai = Scira()
+        >>> # 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 60.
+        proxies (dict, optional): Proxy configuration for requests. Defaults to None.
+        model (str, optional): Model to use for the search. Defaults to "scira-default".
+        group (str, optional): Group parameter. Defaults to "web".
+    """
+
+    AVAILABLE_MODELS = {
+        "scira-default": "Grok3",
+        "scira-grok-3-mini": "Grok3-mini",  # thinking model
+        "scira-vision": "Grok2-Vision",  # vision model
+        "scira-claude": "Sonnet-3.7",
+        "scira-optimus": "optimus",
+    }
+
+    def __init__(
+        self,
+        timeout: int = 60,
+        proxies: Optional[dict] = None,
+        model: str = "scira-default",
+        deepsearch: bool = False
+    ):
+        """Initialize the SCIRA API client.
+
+        Args:
+            timeout (int, optional): Request timeout in seconds. Defaults to 60.
+            proxies (dict, optional): Proxy configuration for requests. Defaults to None.
+            model (str, optional): Model to use for the search. Defaults to "scira-default" (Grok3).
+            deepsearch (bool, optional): Whether to use deep search mode. If True, uses "extreme" group for more comprehensive results. If False, uses "web" group for faster results. Defaults to False.
+
+        Example:
+            >>> ai = Scira(timeout=120)  # Longer timeout
+            >>> ai = Scira(proxies={'http': 'http://proxy.com:8080'})  # With proxy
+            >>> ai = Scira(model="scira-claude")  # Use Claude model
+            >>> ai = Scira(deepsearch=True)  # Use deep search mode
+        """
+        # Validate model
+        if model not in self.AVAILABLE_MODELS:
+            raise ValueError(
+                f"Invalid model: {model}. Choose from: {list(self.AVAILABLE_MODELS.keys())}"
+            )
+
+        self.session = requests.Session()
+        self.api_endpoint = "https://scira.ai/api/search"
+        self.timeout = timeout
+        self.proxies = proxies
+        self.model = model
+
+        # Set group based on deepsearch parameter
+        self.group = "extreme" if deepsearch else "web"
+        self.last_response = {}
+
+        # Set headers
+        self.headers = {
+            "Content-Type": "application/json",
+            "Accept": "*/*",
+            "User-Agent": LitAgent().random()
+        }
+
+        self.session.headers.update(self.headers)
+
+    def search(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+    ) -> Union[Response, Generator[Union[Dict[str, str], Response], None, None]]:
+        """Search using the SCIRA API and get AI-generated responses.
+
+        This method sends a search query to SCIRA 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
+                - If stream=True: Yields response chunks as they arrive
+
+        Raises:
+            exceptions.APIConnectionError: If there's an issue connecting to the API
+            exceptions.APIResponseError: If the API returns an error response
+
+        Example:
+            >>> ai = Scira()
+            >>> # 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...
+        """
+        # Create a unique message ID
+        message_id = str(uuid.uuid4()).replace("-", "")[:16]
+        self.user_id = str(uuid.uuid4()).replace("-", "")[:16]
+        # Prepare the payload
+        payload = {
+            "id": message_id,
+            "messages": [
+                {
+                    "role": "user",
+                    "content": prompt,
+                    "parts": [
+                        {
+                            "type": "text",
+                            "text": prompt
+                        }
+                    ]
+                }
+            ],
+            "model": self.model,
+            "group": self.group,
+            "user_id": self.user_id,
+            "timezone": "Asia/Calcutta"
+        }
+
+        try:
+            # Send the request
+            response = self.session.post(
+                self.api_endpoint,
+                headers=self.headers,
+                data=json.dumps(payload),
+                stream=True,
+                timeout=self.timeout,
+                proxies=self.proxies
+            )
+
+            # Check for successful response
+            if response.status_code != 200:
+                raise exceptions.APIResponseError(
+                    f"API returned error status: {response.status_code}"
+                )
+
+            # Store the last response
+            self.last_response = {"status_code": response.status_code}
+
+            # Handle streaming response
+            if stream:
+                return self._handle_streaming_response(response, raw)
+
+            # Handle non-streaming response
+            return self._handle_non_streaming_response(response, raw)
+
+        except requests.RequestException as e:
+            raise exceptions.APIConnectionError(f"Error connecting to API: {str(e)}")
+
+    def _handle_streaming_response(
+        self,
+        response: requests.Response,
+        raw: bool
+    ) -> Generator[Union[Dict[str, str], Response], None, None]:
+        """Handle streaming response from the API.
+
+        Args:
+            response (requests.Response): The streaming response from the API
+            raw (bool): Whether to return raw response dictionaries
+
+        Yields:
+            Union[Dict[str, str], Response]: Response chunks as they arrive
+        """
+        for line in response.iter_lines():
+            if line:
+                try:
+                    # Decode the line
+                    decoded_line = line.decode("utf-8")
+
+                    # Check if this is a line starting with "0:" (content)
+                    if re.match(r'^0:', decoded_line):
+                        # Extract the content after "0:"
+                        content = re.sub(r'^0:', '', decoded_line)
+                        # Remove surrounding quotes if present
+                        content = re.sub(r'^"(.*)"$', r'\1', content)
+                        # Replace escaped newlines with actual newlines
+                        content = content.replace('\\n', '\n')
+
+                        if raw:
+                            yield {"text": content}
+                        else:
+                            yield Response(content)
+                except Exception:
+                    # Skip lines that can't be processed
+                    pass
+
+    def _handle_non_streaming_response(
+        self,
+        response: requests.Response,
+        raw: bool
+    ) -> Union[Dict[str, str], Response]:
+        """Handle non-streaming response from the API.
+
+        Args:
+            response (requests.Response): The response from the API
+            raw (bool): Whether to return raw response dictionary
+
+        Returns:
+            Union[Dict[str, str], Response]: Complete response
+        """
+        full_text = ""
+
+        for line in response.iter_lines():
+            if line:
+                try:
+                    # Decode the line
+                    decoded_line = line.decode("utf-8")
+
+                    # Check if this is a line starting with "0:" (content)
+                    if re.match(r'^0:', decoded_line):
+                        # Extract the content after "0:"
+                        content = re.sub(r'^0:', '', decoded_line)
+                        # Remove surrounding quotes if present
+                        content = re.sub(r'^"(.*)"$', r'\1', content)
+                        # Replace escaped newlines with actual newlines
+                        content = content.replace('\\n', '\n')
+                        full_text += content
+                except Exception:
+                    # Skip lines that can't be processed
+                    pass
+
+        if raw:
+            return {"text": full_text}
+        else:
+            return Response(full_text)
+
+    @staticmethod
+    def clean_content(text: str) -> str:
+        """Clean the response content by removing unnecessary formatting.
+
+        Args:
+            text (str): The text to clean
+
+        Returns:
+            str: The cleaned text
+        """
+        # Remove any extra whitespace
+        cleaned_text = re.sub(r'\s+', ' ', text)
+        # Remove any trailing whitespace
+        cleaned_text = cleaned_text.strip()
+
+        return cleaned_text
+
+
+if __name__ == "__main__":
+    from rich import print
+    ai = Scira()
+    user_query = input(">>> ")
+    response = ai.search(user_query, stream=True, raw=False)
+    for chunk in response:
+        print(chunk, end="", flush=True)

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)