webscout 8.2.4__py3-none-any.whl → 8.2.6__py3-none-any.whl

webscout/Provider/GithubChat.py

@@ -1,12 +1,13 @@
-import requests
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
 import json
 import time
 from typing import Any, Dict, List, Optional, Union, Generator
 
 from webscout.AIutel import Conversation
 from webscout.AIutel import Optimizers
-from webscout.AIutel import AwesomePrompts
-from webscout.AIbase import Provider
+from webscout.AIutel import AwesomePrompts, sanitize_stream # Import sanitize_stream
+from webscout.AIbase import Provider 
 from webscout import exceptions
 from webscout.litagent import LitAgent
 
@@ -54,7 +55,7 @@ class GithubChat(Provider):
         self.url = "https://github.com/copilot"
         self.api_url = "https://api.individual.githubcopilot.com"
         self.cookie_path = cookie_path
-        self.session = requests.Session()
+        self.session = Session() # Use curl_cffi Session
         self.session.proxies.update(proxies)
         
         # Load cookies for authentication
@@ -158,8 +159,15 @@ class GithubChat(Provider):
                 
             return self._access_token
             
-        except requests.exceptions.RequestException as e:
-            raise exceptions.FailedToGenerateResponseError(f"Failed to get access token: {str(e)}")
+        except:
+            pass
+
+    @staticmethod
+    def _github_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from GitHub Copilot stream JSON objects."""
+        if isinstance(chunk, dict) and chunk.get("type") == "content":
+            return chunk.get("body")
+        return None
 
     def create_conversation(self):
         """Create a new conversation with GitHub Copilot."""
@@ -173,7 +181,10 @@ class GithubChat(Provider):
         headers["Authorization"] = f"GitHub-Bearer {access_token}"
         
         try:
-            response = self.session.post(url, headers=headers)
+            response = self.session.post(
+                url, headers=headers,
+                impersonate="chrome120" # Add impersonate
+            )
             
             if response.status_code == 401:
                 # Token might be expired, try refreshing
@@ -181,7 +192,10 @@ class GithubChat(Provider):
                 access_token = self.get_access_token()
                 headers["Authorization"] = f"GitHub-Bearer {access_token}"
                 response = self.session.post(url, headers=headers)
-                
+            
+            # Check status after potential retry
+            response.raise_for_status() # Raise HTTPError for bad responses (4xx or 5xx)
+
             if response.status_code not in [200, 201]:
                 raise exceptions.FailedToGenerateResponseError(f"Failed to create conversation: {response.status_code}")
                 
@@ -192,40 +206,8 @@ class GithubChat(Provider):
                 raise exceptions.FailedToGenerateResponseError("Failed to extract conversation ID from response")
                 
             return self._conversation_id
-            
-        except requests.exceptions.RequestException as e:
+        except (CurlError, exceptions.FailedToGenerateResponseError, Exception) as e: # Catch CurlError and others
             raise exceptions.FailedToGenerateResponseError(f"Failed to create conversation: {str(e)}")
-
-    def process_response(self, response, prompt: str):
-        """Process streaming response and extract content."""
-        full_text = ""
-        
-        for line in response.iter_lines(decode_unicode=True):
-            if not line or not line.startswith("data: "):
-                continue
-                
-            try:
-                # Parse each line (remove "data: " prefix)
-                json_str = line[6:]
-                if json_str == "[DONE]":
-                    break
-                    
-                data = json.loads(json_str)
-                
-                # Handle different response types
-                if data.get("type") == "content":
-                    token = data.get("body", "")
-                    full_text += token
-                    resp = {"text": token}
-                    yield resp
-                    
-            except json.JSONDecodeError:
-                continue
-        
-        # Update conversation history only for saving to file if needed
-        if full_text:
-            self.last_response = {"text": full_text}
-            self.conversation.update_chat_history(prompt, full_text)
     
     def ask(
         self,
@@ -275,12 +257,14 @@ class GithubChat(Provider):
             "mode": "immersive"
         }
         
+        streaming_text = "" # Initialize for history update
         def for_stream():
+            nonlocal streaming_text # Allow modification of outer scope variable
             try:
                 response = self.session.post(
                     url, 
                     json=request_data,
-                    headers=headers,
+                    headers=headers, # Use updated headers with Authorization
                     stream=True,
                     timeout=self.timeout
                 )
@@ -292,21 +276,35 @@ class GithubChat(Provider):
                     headers["Authorization"] = f"GitHub-Bearer {access_token}"
                     response = self.session.post(
                         url,
-                        json=request_data,
+                        json=request_data, # Use original payload
                         headers=headers,
                         stream=True,
                         timeout=self.timeout
                     )
                 
                 # If still not successful, raise exception
-                if response.status_code != 200:
-                    raise exceptions.FailedToGenerateResponseError(f"Request failed with status code {response.status_code}")
+                response.raise_for_status() # Raise HTTPError for bad responses (4xx or 5xx)
                 
                 # Process the streaming response
-                yield from self.process_response(response, prompt)
+                # 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._github_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 _github_extractor
+                    if content_chunk and isinstance(content_chunk, str):
+                        streaming_text += content_chunk
+                        resp = {"text": content_chunk}
+                        yield resp if not raw else content_chunk
                 
             except Exception as e:
-                if isinstance(e, requests.exceptions.RequestException):
+                if isinstance(e, CurlError): # Check for CurlError
                     if hasattr(e, 'response') and e.response is not None:
                         status_code = e.response.status_code 
                         if status_code == 401:
@@ -314,13 +312,18 @@ class GithubChat(Provider):
                 
                 # If anything else fails
                 raise exceptions.FailedToGenerateResponseError(f"Request failed: {str(e)}")
+            finally:
+                # Update history after stream finishes or fails (if text was generated)
+                if streaming_text:
+                    self.last_response = {"text": streaming_text}
+                    self.conversation.update_chat_history(prompt, streaming_text)
 
         def for_non_stream():
             response_text = ""
             for response in for_stream():
                 if "text" in response:
                     response_text += response["text"]
-            self.last_response = {"text": response_text}
+            # self.last_response and history are updated in for_stream's finally block
             return self.last_response
 
         return for_stream() if stream else for_non_stream()
@@ -358,7 +361,7 @@ if __name__ == "__main__":
     from rich import print
     
     try:
-        ai = GithubChat()
+        ai = GithubChat("cookies.json")
         response = ai.chat("Python code to count r in strawberry", stream=True)
         for chunk in response:
             print(chunk, end="", flush=True)

webscout/Provider/GizAI.py

@@ -0,0 +1,285 @@
+import os
+import base64
+import random
+import json
+from typing import Union, Dict, Any, Optional
+from urllib import response
+
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
+from curl_cffi.const import CurlHttpVersion
+
+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 GizAI(Provider):
+    """
+    A class to interact with the GizAI API.
+    
+    Attributes:
+        system_prompt (str): The system prompt to define the assistant's role.
+        
+    Examples:
+        >>> from webscout.Provider.GizAI import GizAI
+        >>> ai = GizAI()
+        >>> response = ai.chat("What's the weather today?")
+        >>> print(response)
+    """
+    
+    AVAILABLE_MODELS = [
+        "azure-gpt-4-1",
+        "chat-gpt4",
+        "chat-grok-2",
+        "chat-o4-mini",
+        "chat-o4-mini-high",
+        "chat-o4-mini-medium",
+        "claude-haiku",
+        "claude-sonnet",
+        "deepinfra-llama-4-maverick",
+        "deepseek",
+        "deepseek-r1-distill-llama-70b",
+        "gemini-2.0-flash-lite",
+        "gemini-2.5-flash",
+        "gemini-2.5-pro",
+        "gpt-4-1-mini",
+        "gpt-4-1-nano",
+        "gpt-4o-image",
+        "hyperbolic-deepseek-r1",
+        "llama-3-70b",
+        "llama-4-scout",
+        "o3",
+        "phi-4",
+        "qwq-32b"
+    ]
+    
+    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 = "gemini-2.0-flash-lite",
+        system_prompt: str = "You are a helpful assistant."
+    ):
+        """Initializes the GizAI API client."""
+        if model not in self.AVAILABLE_MODELS:
+            raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
+            
+        self.api_url = "https://app.giz.ai/api/data/users/inferenceServer.infer"
+        
+        # Initialize LitAgent for user-agent generation
+        self.agent = LitAgent()
+        
+        # Initialize curl_cffi Session
+        self.session = Session()
+        
+        # Set up the headers
+        self.headers = {
+            "accept": "application/json, text/plain, */*",
+            "content-type": "application/json",
+            "user-agent": self.agent.random(),
+            "origin": "https://app.giz.ai",
+            "referer": "https://app.giz.ai/",
+            "sec-fetch-dest": "empty",
+            "sec-fetch-mode": "cors",
+            "sec-fetch-site": "same-origin"
+        }
+        
+        # Update session headers and proxies
+        self.session.headers.update(self.headers)
+        self.session.proxies = proxies
+        
+        # Store configuration
+        self.system_prompt = system_prompt
+        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 _generate_id(self, length: int = 21) -> str:
+        """Generates a random URL-safe base64 string."""
+        random_bytes = os.urandom(length * 2)  # Generate more bytes initially
+        b64_encoded = base64.urlsafe_b64encode(random_bytes).decode('utf-8')
+        return b64_encoded[:length]
+    
+    def _get_random_ip(self) -> str:
+        """Generates a random IPv4 address string."""
+        return f"{random.randint(0, 255)}.{random.randint(0, 255)}.{random.randint(0, 255)}.{random.randint(0, 255)}"
+    
+    def ask(
+        self,
+        prompt: str,
+        stream: bool = False,  # Parameter kept for compatibility but not used
+        raw: bool = False,
+        optimizer: str = None,
+        conversationally: bool = False,
+    ) -> Dict[str, Any]:
+        """
+        Sends a prompt to the GizAI API and returns the response.
+        
+        Args:
+            prompt (str): The prompt to send to the API.
+            stream (bool): Not supported by GizAI, kept for compatibility.
+            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:
+            Dict[str, Any]: The API response.
+            
+        Examples:
+            >>> ai = GizAI()
+            >>> response = ai.ask("Tell me a joke!")
+        """
+        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 random IDs for request
+        instance_id = self._generate_id()
+        subscribe_id = self._generate_id()
+        x_forwarded_for = self._get_random_ip()
+        
+        # Set up request body - GizAI doesn't support streaming
+        request_body = {
+            "model": "chat",
+            "baseModel": self.model, # Use the specific model ID here
+            "input": {
+                "messages": [{
+                    "type": "human",
+                    "content": conversation_prompt
+                }],
+                "mode": "plan"
+            },
+            "noStream": True,
+            "instanceId": instance_id,
+            "subscribeId": subscribe_id
+        }
+        
+        # Combine default headers with the dynamic x-forwarded-for header
+        request_headers = {**self.headers, "x-forwarded-for": x_forwarded_for}
+        
+        try:
+            # Use curl_cffi session post with impersonate
+            response = self.session.post(
+                self.api_url,
+                headers=request_headers,
+                json=request_body,
+                timeout=self.timeout,
+                impersonate="chrome120",  # Use a common impersonation profile
+                http_version=CurlHttpVersion.V2_0    # Use HTTP/2
+            )
+            response.raise_for_status()  # Check for HTTP errors
+            
+            # Process the response
+            try:
+                response_json = response.json()
+                # GizAI responses have "status" and "output" fields
+                if response_json.get("status") == "completed" and "output" in response_json:
+                    content = response_json["output"]
+                else:
+                    content = ""
+                    # Try to extract content from any available field that might contain the response
+                    for key, value in response_json.items():
+                        if isinstance(value, str) and len(value) > 10:
+                            content = value
+                            break
+            except json.JSONDecodeError:
+                # Handle case where response is not valid JSON
+                content = response.text
+            
+            # Update conversation history
+            self.last_response = {"text": content}
+            self.conversation.update_chat_history(prompt, content)
+            
+            return self.last_response if not raw else content
+        
+        except CurlError as e:
+            raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {str(e)}")
+        except Exception as e:
+            error_text = getattr(e, 'response', None) and getattr(e.response, 'text', '')
+            raise exceptions.FailedToGenerateResponseError(f"Request failed ({type(e).__name__}): {str(e)} - {error_text}")
+    
+    def chat(
+        self,
+        prompt: str,
+        stream: bool = False,  # Parameter kept for compatibility but not used
+        optimizer: str = None,
+        conversationally: bool = False,
+    ) -> str:
+        """
+        Generates a response from the GizAI API.
+        
+        Args:
+            prompt (str): The prompt to send to the API.
+            stream (bool): Not supported by GizAI, kept for compatibility.
+            optimizer (str): Optimizer to use for the prompt.
+            conversationally (bool): Whether to generate the prompt conversationally.
+            
+        Returns:
+            str: The API response text.
+            
+        Examples:
+            >>> ai = GizAI()
+            >>> response = ai.chat("What's the weather today?")
+        """
+        # GizAI doesn't support streaming, so ignore the stream parameter
+        response_data = self.ask(
+            prompt, stream=False, raw=False,
+            optimizer=optimizer, conversationally=conversationally
+        )
+        return self.get_message(response_data)
+    
+    def get_message(self, response: Union[dict, str]) -> str:
+        """
+        Extracts the message from the API response.
+        
+        Args:
+            response (Union[dict, str]): The API response.
+            
+        Returns:
+            str: The message content.
+            
+        Examples:
+            >>> ai = GizAI()
+            >>> response = ai.ask("Tell me a joke!")
+            >>> message = ai.get_message(response)
+        """
+        if isinstance(response, str):
+            return response
+        assert isinstance(response, dict), "Response should be either dict or str"
+        return response.get("text", "")

webscout/Provider/Glider.py

@@ -1,8 +1,9 @@
-import requests
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
 import json
-from typing import Union, Any, Dict, Generator, Optional
+from typing import Union, Any, Dict, Generator, Optional, List
 
-from webscout.AIutel import Optimizers, Conversation, AwesomePrompts
+from webscout.AIutel import Optimizers, Conversation, AwesomePrompts, sanitize_stream # Import sanitize_stream
 from webscout.AIbase import Provider
 from webscout import exceptions
 from webscout.litagent import LitAgent as Lit
@@ -39,7 +40,7 @@ class GliderAI(Provider):
         if model not in self.AVAILABLE_MODELS:
             raise ValueError(f"Invalid model: {model}. Choose from: {', '.join(self.AVAILABLE_MODELS)}")
         
-        self.session = requests.Session()
+        self.session = Session() # Use curl_cffi Session
         self.is_conversation = is_conversation
         self.max_tokens_to_sample = max_tokens
         self.api_endpoint = "https://glider.so/api/chat"
@@ -57,7 +58,7 @@ class GliderAI(Provider):
             "user-agent": Lit().random(),
         }
         self.session.headers.update(self.headers)
-        self.session.proxies = proxies
+        self.session.proxies = proxies # Assign proxies directly
 
         self.__available_optimizers = (
             method for method in dir(Optimizers)
@@ -76,6 +77,14 @@ class GliderAI(Provider):
         )
         self.conversation.history_offset = history_offset
 
+    @staticmethod
+    def _glider_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from Glider stream JSON objects."""
+        if isinstance(chunk, dict):
+            # Handle both standard and DeepSeek response formats within choices
+            return chunk.get("choices", [{}])[0].get("delta", {}).get("content")
+        return None
+
     def ask(
         self,
         prompt: str,
@@ -113,30 +122,32 @@ class GliderAI(Provider):
         }
 
         def for_stream():
-            response = self.session.post(
-                self.api_endpoint, json=payload, stream=True, timeout=self.timeout
-            )
-            if not response.ok:
-                raise exceptions.FailedToGenerateResponseError(
-                    f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
-                )
             streaming_text = ""
-            for value in response.iter_lines(decode_unicode=True):
-                if value:
-                    if value.startswith("data: "):
-                        try:
-                            data = json.loads(value[6:])
-                            # Handle both standard and DeepSeek response formats
-                            if "choices" in data and len(data["choices"]) > 0:
-                                choice = data["choices"][0]
-                                if "delta" in choice and "content" in choice["delta"]:
-                                    content = choice["delta"]["content"]
-                                    if content:
-                                        streaming_text += content
-                                        yield content if raw else {"text": content}
-                        except json.JSONDecodeError:
-                            if "stop" in value:
-                                break
+            try:
+                response = self.session.post(
+                    self.api_endpoint, 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
+                    content_extractor=self._glider_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 _glider_extractor
+                    if content_chunk and isinstance(content_chunk, str):
+                        streaming_text += content_chunk
+                        yield content_chunk if raw else {"text": content_chunk}
+            except CurlError as e:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}") from e
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"An unexpected error occurred ({type(e).__name__}): {e}") from e
             self.last_response.update(dict(text=streaming_text))
             self.conversation.update_chat_history(prompt, self.get_message(self.last_response))
 

webscout/Provider/Groq.py

@@ -9,7 +9,7 @@ from curl_cffi import CurlError
 
 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, AsyncProvider
 from webscout import exceptions
 
@@ -178,6 +178,14 @@ class GROQ(Provider):
         # Set proxies for curl_cffi session
         self.session.proxies = proxies
     
+    @staticmethod
+    def _groq_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[Dict]:
+        """Extracts the 'delta' object from Groq stream JSON chunks."""
+        if isinstance(chunk, dict):
+            # Return the delta object itself, or None if not found
+            return chunk.get("choices", [{}])[0].get("delta")
+        return None
+
     @classmethod
     def update_available_models(cls, api_key=None):
         """Update the available models list from Groq API"""
@@ -262,24 +270,27 @@ class GROQ(Provider):
                         f"Failed to generate response - ({response.status_code}) - {response.text}"
                     )
 
-                message_load = ""
-                for value in response.iter_lines(
-                    decode_unicode=True,
-                    delimiter="" if raw else "data:",
-                    chunk_size=self.stream_chunk_size,
-                ):
-                    try:
-                        resp = json.loads(value)
-                        incomplete_message = self.get_message(resp)
-                        if incomplete_message:
-                            message_load += incomplete_message
-                            resp["choices"][0]["delta"]["content"] = message_load
-                            self.last_response.update(resp)
-                            yield value if raw else resp
-                        elif raw:
-                            yield value
-                    except json.decoder.JSONDecodeError:
-                        pass
+                streaming_text = ""
+                # Use sanitize_stream
+                processed_stream = sanitize_stream(
+                    data=response.iter_content(chunk_size=None), # Pass byte iterator
+                    intro_value="data:",
+                    to_json=True,     # Stream sends JSON
+                    content_extractor=self._groq_extractor, # Use the delta extractor
+                    yield_raw_on_error=False # Skip non-JSON lines or lines where extractor fails
+                )
+
+                for delta in processed_stream:
+                    # delta is the extracted 'delta' object or None
+                    if delta and isinstance(delta, dict):
+                        content = delta.get("content")
+                        if content:
+                            streaming_text += content
+                            resp = {"text": content} # Yield only the new chunk text
+                            self.last_response = {"choices": [{"delta": {"content": streaming_text}}]} # Update last_response structure
+                            yield resp if not raw else content # Yield dict or raw string chunk
+                        # Note: Tool calls in streaming delta are less common in OpenAI format, usually in final message
+
             except CurlError as e:
                 raise exceptions.FailedToGenerateResponseError(f"CurlError: {str(e)}")
             except Exception as e:
@@ -339,7 +350,24 @@ class GROQ(Provider):
                          # Removed response.reason_phrase
                         f"Failed to generate response - ({response.status_code}) - {response.text}"
                     )
-                resp = response.json()
+                
+                response_text = response.text # Get raw text
+
+                # Use sanitize_stream to parse the non-streaming JSON response
+                processed_stream = sanitize_stream(
+                    data=response_text,
+                    to_json=True, # Parse the whole text as JSON
+                    intro_value=None,
+                    # Extractor for non-stream structure (returns the whole parsed dict)
+                    content_extractor=lambda chunk: chunk if isinstance(chunk, dict) else None,
+                    yield_raw_on_error=False
+                )
+                
+                # Extract the single result (the parsed JSON dictionary)
+                resp = next(processed_stream, None)
+                if resp is None:
+                    raise exceptions.FailedToGenerateResponseError("Failed to parse non-stream JSON response")
+
             except CurlError as e:
                 raise exceptions.FailedToGenerateResponseError(f"CurlError: {str(e)}")
             except Exception as e: