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

webscout/Provider/cerebras.py

@@ -1,10 +1,11 @@
 
 import re
-import requests
+import curl_cffi
+from curl_cffi.requests import Session
 import json
 import os
 from typing import Any, Dict, Optional, Generator, List, Union
-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 UserAgent
@@ -17,7 +18,9 @@ class Cerebras(Provider):
     AVAILABLE_MODELS = [
         "llama3.1-8b",
         "llama-3.3-70b",
-        "deepseek-r1-distill-llama-70b"
+        "deepseek-r1-distill-llama-70b",
+        "llama-4-scout-17b-16e-instruct"
+
     ]
 
     def __init__(
@@ -49,6 +52,8 @@ class Cerebras(Provider):
         self.max_tokens_to_sample = max_tokens
         self.last_response = {}
 
+        self.session = Session() # Initialize curl_cffi session
+
         # Get API key first
         try:
             self.api_key = self.get_demo_api_key(cookie_path)
@@ -74,6 +79,9 @@ class Cerebras(Provider):
             is_conversation, self.max_tokens_to_sample, filepath, update_file
         )
         self.conversation.history_offset = history_offset
+        
+        # Apply proxies to the session
+        self.session.proxies = proxies
 
     # Rest of the class implementation remains the same...
     @staticmethod
@@ -88,7 +96,14 @@ class Cerebras(Provider):
         """Refines the input text by removing surrounding quotes."""
         return text.strip('"')
 
-    def get_demo_api_key(self, cookie_path: str) -> str:
+    @staticmethod
+    def _cerebras_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from Cerebras stream JSON objects."""
+        if isinstance(chunk, dict):
+            return chunk.get("choices", [{}])[0].get("delta", {}).get("content")
+        return None
+
+    def get_demo_api_key(self, cookie_path: str) -> str: # Keep this using requests or switch to curl_cffi
         """Retrieves the demo API key using the provided cookie."""
         try:
             with open(cookie_path, "r") as file:
@@ -114,17 +129,19 @@ class Cerebras(Provider):
         }
 
         try:
-            response = requests.post(
+            # Use the initialized curl_cffi session
+            response = self.session.post(
                 "https://inference.cerebras.ai/api/graphql",
                 cookies=cookies,
                 headers=headers,
                 json=json_data,
                 timeout=self.timeout,
+                impersonate="chrome120" # Add impersonate
             )
             response.raise_for_status()
-            api_key = response.json()["data"]["GetMyDemoApiKey"]
+            api_key = response.json().get("data", {}).get("GetMyDemoApiKey")
             return api_key
-        except requests.exceptions.RequestException as e:
+        except curl_cffi.CurlError as e:
             raise exceptions.APIConnectionError(f"Failed to retrieve API key: {e}")
         except KeyError:
             raise exceptions.InvalidResponseError("API key not found in response.")
@@ -144,41 +161,48 @@ class Cerebras(Provider):
         }
 
         try:
-            response = requests.post(
+            # Use the initialized curl_cffi session
+            response = self.session.post(
                 "https://api.cerebras.ai/v1/chat/completions",
                 headers=headers,
                 json=payload,
                 stream=stream,
-                timeout=self.timeout
+                timeout=self.timeout,
+                impersonate="chrome120" # Add impersonate
             )
             response.raise_for_status()
 
             if stream:
                 def generate_stream():
-                    for line in response.iter_lines():
-                        if line:
-                            line = line.decode('utf-8')
-                            if line.startswith('data:'):
-                                try:
-                                    data = json.loads(line[6:])
-                                    if data.get('choices') and data['choices'][0].get('delta', {}).get('content'):
-                                        content = data['choices'][0]['delta']['content']
-                                        yield content
-                                except json.JSONDecodeError:
-                                    continue
+                    # 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._cerebras_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:
+                        if content_chunk and isinstance(content_chunk, str):
+                            yield content_chunk # Yield the extracted text chunk
 
                 return generate_stream()
             else:
                 response_json = response.json()
-                return response_json['choices'][0]['message']['content']
+                # Extract content for non-streaming response
+                content = response_json.get("choices", [{}])[0].get("message", {}).get("content")
+                return content if content else "" # Return empty string if not found
 
-        except requests.exceptions.RequestException as e:
+        except curl_cffi.CurlError as e:
+            raise exceptions.APIConnectionError(f"Request failed (CurlError): {e}") from e
+        except Exception as e: # Catch other potential errors
             raise exceptions.APIConnectionError(f"Request failed: {e}")
 
     def ask(
         self,
         prompt: str,
         stream: bool = False,
+        raw: bool = False, # Add raw parameter for consistency
         optimizer: str = None,
         conversationally: bool = False,
     ) -> Union[Dict, Generator]:
@@ -199,11 +223,23 @@ class Cerebras(Provider):
 
         try:
             response = self._make_request(messages, stream)
-            if stream:
-                return response
             
-            self.last_response = response
-            return response
+            if stream:
+                # Wrap the generator to yield dicts or raw strings
+                def stream_wrapper():
+                    full_text = ""
+                    for chunk in response:
+                        full_text += chunk
+                        yield chunk if raw else {"text": chunk}
+                    # Update history after stream finishes
+                    self.last_response = {"text": full_text}
+                    self.conversation.update_chat_history(prompt, full_text)
+                return stream_wrapper()
+            else:
+                # Non-streaming response is already the full text string
+                self.last_response = {"text": response}
+                self.conversation.update_chat_history(prompt, response)
+                return self.last_response if not raw else response # Return dict or raw string
 
         except Exception as e:
             raise exceptions.FailedToGenerateResponseError(f"Error during request: {e}")
@@ -216,14 +252,24 @@ class Cerebras(Provider):
         conversationally: bool = False,
     ) -> Union[str, Generator]:
         """Chat with the model."""
-        response = self.ask(prompt, stream, optimizer, conversationally)
+        # Ask returns a generator for stream=True, dict/str for stream=False
+        response_gen_or_dict = self.ask(prompt, stream, raw=False, optimizer=optimizer, conversationally=conversationally)
+        
         if stream:
-            return response
-        return response
+            # Wrap the generator from ask() to get message text
+            def stream_wrapper():
+                for chunk_dict in response_gen_or_dict:
+                    yield self.get_message(chunk_dict)
+            return stream_wrapper()
+        else:
+            # Non-streaming response is already a dict
+            return self.get_message(response_gen_or_dict)
 
     def get_message(self, response: str) -> str:
         """Retrieves message from response."""
-        return response
+        # Updated to handle dict input from ask()
+        assert isinstance(response, dict), "Response should be of dict data-type only for get_message"
+        return response.get("text", "")
 
 
 if __name__ == "__main__":
@@ -231,7 +277,7 @@ if __name__ == "__main__":
     
     # Example usage
     cerebras = Cerebras(
-        cookie_path='cookie.json',
+        cookie_path=r'cookies.json',
         model='llama3.1-8b',
         system_prompt="You are a helpful AI assistant."
     )

webscout/Provider/chatglm.py

@@ -1,11 +1,12 @@
-import requests
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
 import json
 from typing import Any, Dict, Optional, Generator, List, Union
 import uuid
 
 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 exceptions
 from webscout.litagent import LitAgent
@@ -29,7 +30,7 @@ class ChatGLM(Provider):
         plus_model: bool = True,
     ):
         """Initializes the ChatGLM API client."""
-        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://chatglm.cn/chatglm/mainchat-api/guest/stream"
@@ -55,7 +56,7 @@ class ChatGLM(Provider):
             if callable(getattr(Optimizers, method)) and not method.startswith("__")
         )
         self.session.headers.update(self.headers)
-        Conversation.intro = (
+        Conversation.intro = ( # type: ignore
             AwesomePrompts().get_act(
                 act, raise_not_found=True, default=None, case_insensitive=True
             )
@@ -66,7 +67,16 @@ class ChatGLM(Provider):
             is_conversation, self.max_tokens_to_sample, filepath, update_file
         )
         self.conversation.history_offset = history_offset
-        self.session.proxies = proxies
+        self.session.proxies = proxies # Assign proxies directly
+
+    @staticmethod
+    def _chatglm_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from ChatGLM stream JSON objects."""
+        if isinstance(chunk, dict):
+            parts = chunk.get('parts', [])
+            if parts and isinstance(parts[0].get('content'), list) and parts[0]['content']:
+                return parts[0]['content'][0].get('text')
+        return None
 
     def ask(
         self,
@@ -119,45 +129,45 @@ class ChatGLM(Provider):
         }
 
         def for_stream():
+            streaming_text = "" # Initialize outside try block
+            last_processed_content = "" # Track the last processed content
             try:
-                with self.session.post(
-                    self.api_endpoint, json=payload, stream=True, timeout=self.timeout
-                ) as response:
-                    response.raise_for_status()
-                    
-                    streaming_text = ""
-                    last_processed_content = "" # Track the last processed content
-                    for chunk in response.iter_lines():
-                        if chunk:
-                            decoded_chunk = chunk.decode('utf-8')
-                            if decoded_chunk.startswith('data: '):
-                                try:
-                                    json_data = json.loads(decoded_chunk[6:])
-                                    parts = json_data.get('parts', [])
-                                    if parts:
-                                        content = parts[0].get('content', [])
-                                        if content:
-                                            text = content[0].get('text', '')
-                                            new_text = text[len(last_processed_content):]
-                                            if new_text:  # Check for new content
-                                                streaming_text += new_text
-                                                last_processed_content = text
-                                                yield new_text if raw else dict(text=new_text)
-                                except json.JSONDecodeError:
-                                    continue
+                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._chatglm_extractor, # Use the specific extractor
+                    yield_raw_on_error=False # Skip non-JSON lines or lines where extractor fails
+                )
 
+                for current_full_text in processed_stream:
+                    # current_full_text is the full text extracted by _chatglm_extractor
+                    if current_full_text and isinstance(current_full_text, str):
+                        new_text = current_full_text[len(last_processed_content):]
+                        if new_text: # Check for new content
+                            streaming_text += new_text
+                            last_processed_content = current_full_text # Update tracker
+                            yield new_text if raw else dict(text=new_text)
+
+            except CurlError as e:
+                raise exceptions.ProviderConnectionError(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
+            finally:
+                # Update history after stream finishes or fails
+                if streaming_text:
                     self.last_response.update(dict(text=streaming_text))
                     self.conversation.update_chat_history(
                         prompt, self.get_message(self.last_response)
                     )
 
-            except requests.exceptions.RequestException as e:
-                raise exceptions.ProviderConnectionError(f"Request failed: {e}")
-            except json.JSONDecodeError as e:
-                raise exceptions.InvalidResponseError(f"Failed to decode JSON: {e}")
-            except Exception as e:
-                raise exceptions.FailedToGenerateResponseError(f"An unexpected error occurred: {e}")
-
         def for_non_stream():
             for _ in for_stream():
                 pass

webscout/Provider/copilot.py

@@ -2,7 +2,6 @@ import os
 import json
 import base64
 import asyncio
-import requests
 from urllib.parse import quote
 from typing import Optional, Dict, Any, List, Union, Generator
 
@@ -287,8 +286,6 @@ class Copilot(Provider):
                     finally:
                         wss.close()
                         
-            except requests.RequestException as e:
-                raise exceptions.FailedToGenerateResponseError(f"Request failed: {str(e)}")
             except Exception as e:
                 raise exceptions.FailedToGenerateResponseError(f"Error: {str(e)}")
 

webscout/Provider/elmo.py

@@ -1,9 +1,15 @@
-import requests
+from curl_cffi.requests import Session
+from curl_cffi import CurlError
+import json
+from typing import Optional, Union, Any, Dict, Generator
+from webscout import exceptions
 from webscout.AIutel import Optimizers
-from webscout.AIutel import Conversation
+from webscout.AIutel import Conversation, sanitize_stream # Import sanitize_stream
 from webscout.AIutel import AwesomePrompts
 from webscout.AIbase import Provider
 from webscout.litagent import LitAgent
+import re # Import re for the extractor
+
 
 class Elmo(Provider):
     """
@@ -13,7 +19,7 @@ class Elmo(Provider):
     def __init__(
         self,
         is_conversation: bool = True,
-        max_tokens: int = 600,
+        max_tokens: int = 600,  # Note: max_tokens is not used by this API
         timeout: int = 30,
         intro: str = None,
         filepath: str = None,
@@ -22,7 +28,6 @@ class Elmo(Provider):
         history_offset: int = 10250,
         act: str = None,
         system_prompt: str = "You are a helpful AI assistant. Provide clear, concise, and well-structured information. Organize your responses into paragraphs for better readability.",
-
     ) -> None:
         """Instantiates Elmo
 
@@ -37,9 +42,9 @@ class Elmo(Provider):
             history_offset (int, optional): Limit conversation history to this number of last texts. Defaults to 10250.
             act (str|int, optional): Awesome prompt key or index. (Used as intro). Defaults to None.
             system_prompt (str, optional): System prompt for Elmo. Defaults to the provided string.
-            web_search (bool, optional): Enables web search mode when True. Defaults to False.
         """
-        self.session = requests.Session()
+        # Initialize curl_cffi Session
+        self.session = Session()
         self.is_conversation = is_conversation
         self.max_tokens_to_sample = max_tokens
         self.api_endpoint = "https://www.elmo.chat/api/v1/prompt"
@@ -49,20 +54,14 @@ class Elmo(Provider):
         self.system_prompt = system_prompt
         self.headers = {
             "accept": "*/*",
-            "accept-encoding": "gzip, deflate, br, zstd",
             "accept-language": "en-US,en;q=0.9,en-IN;q=0.8",
-            "content-length": "763",
             "content-type": "text/plain;charset=UTF-8",
             "dnt": "1",
             "origin": "chrome-extension://ipnlcfhfdicbfbchfoihipknbaeenenm",
             "priority": "u=1, i",
-            "sec-ch-ua": '"Chromium";v="128", "Not;A=Brand";v="24", "Microsoft Edge";v="128"',
-            "sec-ch-ua-mobile": "?0",
-            "sec-ch-ua-platform": '"Windows"',
             "sec-fetch-dest": "empty",
             "sec-fetch-mode": "cors",
             "sec-fetch-site": "cross-site",
-            "user-agent": LitAgent().random(),
         }
 
         self.__available_optimizers = (
@@ -70,7 +69,10 @@ class Elmo(Provider):
             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  # Assign proxies directly
+
         Conversation.intro = (
             AwesomePrompts().get_act(
                 act, raise_not_found=True, default=None, case_insensitive=True
@@ -82,16 +84,26 @@ class Elmo(Provider):
             is_conversation, self.max_tokens_to_sample, filepath, update_file
         )
         self.conversation.history_offset = history_offset
-        self.session.proxies = proxies
+
+    @staticmethod
+    def _elmo_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from the Elmo stream format '0:"..."'."""
+        if isinstance(chunk, str):
+            match = re.search(r'0:"(.*?)"(?=,|$)', chunk) # Look for 0:"...", possibly followed by comma or end of string
+            if match:
+                # Decode potential unicode escapes like \u00e9 and handle escaped quotes/backslashes
+                content = match.group(1).encode().decode('unicode_escape')
+                return content.replace('\\\\', '\\').replace('\\"', '"')
+        return None
 
     def ask(
         self,
         prompt: str,
-        stream: bool = False,
+        stream: bool = False,  # API supports streaming
         raw: bool = False,
         optimizer: str = None,
         conversationally: bool = False,
-    ) -> dict:
+    ) -> Union[Dict[str, Any], Generator[Any, None, None]]:  # Corrected return type hint
         """Chat with AI
 
         Args:
@@ -106,7 +118,7 @@ class Elmo(Provider):
         {
            "text" : "How may I assist you today?"
         }
-        ```
+        ```json
         """
         conversation_prompt = self.conversation.gen_complete_prompt(prompt)
         if optimizer:
@@ -119,7 +131,6 @@ class Elmo(Provider):
                     f"Optimizer is not one of {self.__available_optimizers}"
                 )
 
-
         payload = {
             "metadata": {
                 "system": {"language": "en-US"},
@@ -145,36 +156,69 @@ class Elmo(Provider):
         }
 
         def for_stream():
-            response = self.session.post(
-                self.api_endpoint,
-                headers=self.headers,
-                json=payload,
-                stream=True,
-                timeout=self.timeout,
-            )
-            if not response.ok:
-                raise Exception(
-                    f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
+            streaming_text = "" # Initialize outside try block
+            try:
+                # Use curl_cffi session post with impersonate
+                # Note: The API expects 'text/plain' but we send JSON.
+                # If this fails, try sending json.dumps(payload) as data with 'Content-Type': 'application/json'
+                response = self.session.post(
+                    self.api_endpoint,
+                    # headers are set on the session, but content-type might need override if sending JSON
+                    json=payload,  # Sending as JSON
+                    stream=True,
+                    timeout=self.timeout,
+                    impersonate="chrome110"  # Use a common impersonation profile
+                )
+                response.raise_for_status()  # Check for HTTP errors
+
+                # 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._elmo_extractor, # Use the specific extractor
+                    yield_raw_on_error=True
+                )
+
+                for content_chunk in processed_stream:
+                    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:  # Catch CurlError
+                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}") from e
+            except Exception as e:  # Catch other potential exceptions (like HTTPError)
+                err_text = getattr(e, 'response', None) and getattr(e.response, 'text', '')
+                raise exceptions.FailedToGenerateResponseError(f"Failed to generate response ({type(e).__name__}): {e} - {err_text}") from e
+            finally:
+                # Update history after stream finishes
+                self.last_response = dict(text=streaming_text)
+                self.conversation.update_chat_history(
+                    prompt, streaming_text
                 )
-            full_response = ""
-            for line in response.iter_lines(decode_unicode=True):
-                if line:
-                    if line.startswith('0:'):
-                        chunk = line.split(':"')[1].strip('"')
-                        formatted_output = (
-                            chunk.replace("\\n", "\n").replace("\\n\\n", "\n\n")
-                        )
-                        full_response += formatted_output
-                        self.last_response.update(dict(text=full_response))
-                        yield formatted_output if raw else dict(text=formatted_output)
-            self.conversation.update_chat_history(
-                prompt, self.get_message(self.last_response)
-            )
 
         def for_non_stream():
-            for _ in for_stream():
-                pass
-            return self.last_response
+            # Aggregate the stream using the updated for_stream logic
+            collected_text = ""
+            try:
+                # Ensure raw=False so for_stream yields dicts
+                for chunk_data in for_stream():
+                    if isinstance(chunk_data, dict) and "text" in chunk_data:
+                        collected_text += chunk_data["text"]
+                    # Handle raw string case if raw=True was passed
+                    elif raw and isinstance(chunk_data, str):
+                         collected_text += chunk_data
+            except Exception as e:
+                 # If aggregation fails but some text was received, use it. Otherwise, re-raise.
+                 if not collected_text:
+                     raise exceptions.FailedToGenerateResponseError(f"Failed to get non-stream response: {str(e)}") from e
+
+            # Update last_response and history *after* aggregation for non-stream
+            self.last_response = {"text": collected_text}
+            self.conversation.update_chat_history(prompt, collected_text)
+            # Return the final aggregated response dict or raw string
+            return collected_text if raw else self.last_response
 
         return for_stream() if stream else for_non_stream()
 
@@ -184,7 +228,7 @@ class Elmo(Provider):
         stream: bool = False,
         optimizer: str = None,
         conversationally: bool = False,
-    ) -> str:
+    ) -> Union[str, Generator[str, None, None]]:  # Corrected return type hint
         """Generate response `str`
         Args:
             prompt (str): Prompt to be send.
@@ -195,23 +239,27 @@ class Elmo(Provider):
             str: Response generated
         """
 
-        def for_stream():
-            for response in self.ask(
-                prompt, True, optimizer=optimizer, conversationally=conversationally
-            ):
-                yield self.get_message(response)
+        def for_stream_chat():  # Renamed inner function
+            # ask() yields dicts or strings when streaming
+            gen = self.ask(
+                prompt, stream=True, raw=False,  # Ensure ask yields dicts
+                optimizer=optimizer, conversationally=conversationally
+            )
+            for response_dict in gen:
+                yield self.get_message(response_dict)  # get_message expects dict
 
-        def for_non_stream():
-            return self.get_message(
-                self.ask(
-                    prompt,
-                    False,
-                    optimizer=optimizer,
-                    conversationally=conversationally,
-                )
+        def for_non_stream_chat():  # Renamed inner function
+            # ask() returns dict or str when not streaming
+            response_data = self.ask(
+                prompt,
+                stream=False,
+                raw=False,  # Ensure ask returns dict
+                optimizer=optimizer,
+                conversationally=conversationally,
             )
+            return self.get_message(response_data)  # get_message expects dict
 
-        return for_stream() if stream else for_non_stream()
+        return for_stream_chat() if stream else for_non_stream_chat()  # Use renamed functions
 
     def get_message(self, response: dict) -> str:
         """Retrieves message only from response
@@ -223,10 +271,11 @@ class Elmo(Provider):
             str: Message extracted
         """
         assert isinstance(response, dict), "Response should be of dict data-type only"
-        return response["text"]
+        return response.get("text", "") # Use .get for safety
 
 
 if __name__ == "__main__":
+    # Ensure curl_cffi is installed
     from rich import print
     ai = Elmo()
     response = ai.chat("write a poem about AI", stream=True)