webscout 8.2.2__py3-none-any.whl → 2026.1.19__py3-none-any.whl

webscout/Provider/HeckAI.py

@@ -1,233 +1,392 @@
-import requests
-import json
-import uuid
-import sys
-from typing import Any, Dict, Optional, Generator, Union
-
-from webscout.AIutel import Optimizers
-from webscout.AIutel import Conversation
-from webscout.AIutel import AwesomePrompts, sanitize_stream
-from webscout.AIbase import Provider, AsyncProvider
-from webscout import exceptions
-from webscout.litagent import LitAgent
-
-class HeckAI(Provider):
-    """
-    A class to interact with the HeckAI API with LitAgent user-agent.
-    """
-
-    AVAILABLE_MODELS = [
-        "deepseek/deepseek-chat",
-        "openai/gpt-4o-mini",
-        "deepseek/deepseek-r1",
-        "google/gemini-2.0-flash-001"
-    ]
-
-    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 = "google/gemini-2.0-flash-001",
-        language: str = "English"
-    ):
-        """Initializes the HeckAI API client."""
-        if model not in self.AVAILABLE_MODELS:
-            raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
-            
-        self.url = "https://api.heckai.weight-wave.com/api/ha/v1/chat"
-        self.session_id = str(uuid.uuid4())
-        self.language = language
-        
-        # Use LitAgent for user-agent
-        self.headers = {
-            'User-Agent': LitAgent().random(),
-            'Content-Type': 'application/json',
-            'Origin': 'https://heck.ai',
-            'Referer': 'https://heck.ai/',
-            'Connection': 'keep-alive'
-        }
-        
-        self.session = requests.Session()
-        self.session.headers.update(self.headers)
-        self.session.proxies.update(proxies)
-
-        self.is_conversation = is_conversation
-        self.max_tokens_to_sample = max_tokens
-        self.timeout = timeout
-        self.last_response = {}
-        self.model = model
-        self.previous_question = None
-        self.previous_answer = None
-
-        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 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}")
-
-        # Payload construction
-        payload = {
-            "model": self.model,
-            "question": conversation_prompt,
-            "language": self.language,
-            "sessionId": self.session_id,
-            "previousQuestion": self.previous_question,
-            "previousAnswer": self.previous_answer,
-            "imgUrls": []
-        }
-        
-        # Store this message as previous for next request
-        self.previous_question = conversation_prompt
-
-        def for_stream():
-            try:
-                with requests.post(self.url, headers=self.headers, data=json.dumps(payload), stream=True, timeout=self.timeout) as response:
-                    if response.status_code != 200:
-                        raise exceptions.FailedToGenerateResponseError(
-                            f"Request failed with status code {response.status_code}"
-                        )
-                    
-                    streaming_text = ""
-                    in_answer = False
-                    
-                    for line in response.iter_lines(decode_unicode=True):
-                        if not line:
-                            continue
-                            
-                        # Remove "data: " prefix
-                        if line.startswith("data: "):
-                            data = line[6:]
-                        else:
-                            continue
-                        
-                        # Check for control markers
-                        if data == "[ANSWER_START]":
-                            in_answer = True
-                            continue
-                            
-                        if data == "[ANSWER_DONE]":
-                            in_answer = False
-                            continue
-                            
-                        if data == "[RELATE_Q_START]" or data == "[RELATE_Q_DONE]":
-                            continue
-                            
-                        # Process content if we're in an answer section
-                        if in_answer:
-                            streaming_text += data
-                            resp = dict(text=data)
-                            yield resp if raw else resp
-                    
-                    self.previous_answer = streaming_text
-                    self.conversation.update_chat_history(prompt, streaming_text)
-                    
-            except requests.RequestException as e:
-                raise exceptions.FailedToGenerateResponseError(f"Request failed: {str(e)}")
-
-        def for_non_stream():
-            full_text = ""
-            for chunk in for_stream():
-                if isinstance(chunk, dict) and "text" in chunk:
-                    full_text += chunk["text"]
-            self.last_response = {"text": full_text}
-            return self.last_response
-
-        return for_stream() if stream else for_non_stream()
-
-    @staticmethod
-    def fix_encoding(text):
-        if isinstance(text, dict) and "text" in text:
-            try:
-                text["text"] = text["text"].encode("latin1").decode("utf-8")
-                return text
-            except (UnicodeError, AttributeError) as e:
-                return text
-        elif isinstance(text, str):
-            try:
-                return text.encode("latin1").decode("utf-8")
-            except (UnicodeError, AttributeError) as e:
-                return text
-        return text
-
-    def chat(
-        self,
-        prompt: str,
-        stream: bool = False,
-        optimizer: str = None,
-        conversationally: bool = False,
-    ) -> str:
-        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 HeckAI.AVAILABLE_MODELS:
-        try:
-            test_ai = HeckAI(model=model, timeout=60)
-            response = test_ai.chat("Say 'Hello' in one word", stream=True)
-            response_text = ""
-            for chunk in response:
-                response_text += chunk
-                print(f"\r{model:<50} {'Testing...':<10}", end="", flush=True)
-            
-            if response_text and len(response_text.strip()) > 0:
-                status = "✓"
-                # Truncate response if too long
-                display_text = response_text.strip()[:50] + "..." if len(response_text.strip()) > 50 else response_text.strip()
-            else:
-                status = "✗"
-                display_text = "Empty or invalid response"
-            print(f"\r{model:<50} {status:<10} {display_text}")
-        except Exception as e:
-            print(f"\r{model:<50} {'✗':<10} {str(e)}")
+import json
+import uuid
+from typing import Any, Dict, Generator, Optional, Union, cast
+
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
+
+from webscout import exceptions
+from webscout.AIbase import Provider, Response
+from webscout.AIutel import AwesomePrompts, Conversation, Optimizers, sanitize_stream
+from webscout.litagent import LitAgent
+
+
+class HeckAI(Provider):
+    """
+    Provides an interface to interact with the HeckAI API using a LitAgent user-agent.
+
+    This class supports conversational AI interactions with multiple available models,
+    manages session state, handles streaming and non-streaming responses, and integrates
+    with conversation history and prompt optimizers.
+
+    Attributes:
+        AVAILABLE_MODELS (list): List of supported model identifiers.
+        url (str): API endpoint URL.
+        session_id (str): Unique session identifier for the conversation.
+        language (str): Language for the conversation.
+        headers (dict): HTTP headers used for API requests.
+        session (Session): curl_cffi session for HTTP requests.
+        is_conversation (bool): Whether to maintain conversation history.
+        max_tokens_to_sample (int): Maximum tokens to sample (not used by API).
+        timeout (int): Request timeout in seconds.
+        last_response (dict): Stores the last API response.
+        model (str): Model identifier in use.
+        previous_question (str): Last question sent to the API.
+        previous_answer (str): Last answer received from the API.
+        conversation (Conversation): Conversation history manager.
+    """
+
+    required_auth = False
+    AVAILABLE_MODELS = [
+        "google/gemini-2.5-flash-preview",
+        "deepseek/deepseek-chat",
+        "deepseek/deepseek-r1",
+        "openai/gpt-4o-mini",
+        "openai/gpt-4.1-mini",
+        "x-ai/grok-3-mini-beta",
+        "meta-llama/llama-4-scout",
+        "openai/gpt-5-mini",
+        "openai/gpt-5-nano",
+    ]
+
+    def __init__(
+        self,
+        is_conversation: bool = True,
+        max_tokens: int = 2049,
+        timeout: int = 30,
+        intro: Optional[str] = None,
+        filepath: Optional[str] = None,
+        update_file: bool = True,
+        proxies: dict = {},
+        history_offset: int = 10250,
+        act: Optional[str] = None,
+        model: str = "google/gemini-2.5-flash-preview",
+        language: str = "English",
+    ):
+        """
+        Initializes the HeckAI API client.
+
+        Args:
+            is_conversation (bool): Whether to maintain conversation history.
+            max_tokens (int): Maximum tokens to sample (not used by this API).
+            timeout (int): Timeout for API requests in seconds.
+            intro (str, optional): Introductory prompt for the conversation.
+            filepath (str, optional): File path for storing conversation history.
+            update_file (bool): Whether to update the conversation file.
+            proxies (dict): Proxy settings for HTTP requests.
+            history_offset (int): Offset for conversation history truncation.
+            act (str, optional): Role or act for the conversation.
+            model (str): Model identifier to use.
+            language (str): Language for the conversation.
+
+        Raises:
+            ValueError: If the provided model is not in AVAILABLE_MODELS.
+        """
+        if model not in self.AVAILABLE_MODELS:
+            raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
+
+        self.url = "https://api.heckai.weight-wave.com/api/ha/v1/chat"
+        self.session_id = str(uuid.uuid4())
+        self.language = language
+
+        # Use LitAgent (keep if needed for other headers or logic)
+        self.headers = {
+            "Content-Type": "application/json",
+            "Origin": "https://heck.ai",  # Keep Origin
+            "Referer": "https://heck.ai/",  # Keep Referer
+            "User-Agent": LitAgent().random(),  # Use random user agent
+        }
+
+        # Initialize curl_cffi Session
+        self.session = Session()
+        # Update curl_cffi session headers and proxies
+        self.session.headers.update(self.headers)
+        if proxies:
+            self.session.proxies.update(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.previous_question = None
+        self.previous_answer = None
+
+        self.__available_optimizers = (
+            method
+            for method in dir(Optimizers)
+            if callable(getattr(Optimizers, method)) and not method.startswith("__")
+        )
+        self.conversation = Conversation(
+            is_conversation, self.max_tokens_to_sample, filepath, update_file
+        )
+        self.conversation.history_offset = history_offset
+
+        if act:
+            self.conversation.intro = (
+                AwesomePrompts().get_act(
+                    cast(Union[str, int], act),
+                    default=self.conversation.intro,
+                    case_insensitive=True,
+                )
+                or self.conversation.intro
+            )
+        elif intro:
+            self.conversation.intro = intro
+
+    def ask(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+        optimizer: Optional[str] = None,
+        conversationally: bool = False,
+        **kwargs: Any,
+    ) -> Response:
+        """
+        Sends a prompt to the HeckAI API and returns the response.
+
+        Args:
+            prompt (str): The prompt or question to send to the API.
+            stream (bool): If True, yields streaming responses as they arrive.
+            raw (bool): If True, yields raw string chunks instead of dicts.
+            optimizer (str, optional): Name of the optimizer to apply to the prompt.
+            conversationally (bool): If True, optimizer is applied to the full conversation prompt.
+
+        Returns:
+            Union[Dict[str, Any], Generator]: If stream is False, returns a dict with the response text.
+            If stream is True, yields response chunks as dicts or strings.
+
+        Raises:
+            Exception: If the optimizer is not available.
+            exceptions.FailedToGenerateResponseError: On API or network errors.
+        """
+        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}")
+
+        # Payload construction
+        payload = {
+            "model": self.model,
+            "question": conversation_prompt,
+            "language": self.language,
+            "sessionId": self.session_id,
+            "previousQuestion": self.previous_question,
+            "previousAnswer": self.previous_answer,
+            "imgUrls": [],
+            "superSmartMode": False,  # Added based on API request data
+        }
+
+        # Store this message as previous for next request
+        self.previous_question = conversation_prompt
+
+        def for_stream():
+            streaming_text = ""  # Initialize outside try block
+            try:
+                response = self.session.post(
+                    self.url,
+                    data=json.dumps(payload),
+                    stream=True,
+                    timeout=self.timeout,
+                    impersonate="chrome110",
+                )
+                response.raise_for_status()
+
+                processed_stream = sanitize_stream(
+                    data=response.iter_content(chunk_size=1024),
+                    intro_value="data: ",
+                    to_json=False,
+                    start_marker="data: [ANSWER_START]",
+                    end_marker="data: [ANSWER_DONE]",
+                    skip_markers=[
+                        "data: [RELATE_Q_START]",
+                        "data: [RELATE_Q_DONE]",
+                        "data: [REASON_START]",
+                        "data: [REASON_DONE]",
+                    ],
+                    yield_raw_on_error=True,
+                    strip_chars="",
+                    raw=raw,
+                )
+
+                for content_chunk in processed_stream:
+                    if content_chunk and isinstance(content_chunk, str):
+                        content_chunk = content_chunk.replace("\\\\", "\\").replace('\\"', '"')
+                    if raw:
+                        if content_chunk and isinstance(content_chunk, str):
+                            streaming_text += content_chunk
+                        yield content_chunk
+                    else:
+                        if content_chunk and isinstance(content_chunk, str):
+                            streaming_text += content_chunk
+                            yield dict(text=content_chunk)
+
+                # Only update history if we received a valid response
+                if streaming_text:
+                    self.previous_answer = streaming_text
+                    try:
+                        if streaming_text and isinstance(streaming_text, str):
+                            sanitized_text = streaming_text.strip()
+                            if sanitized_text:
+                                self.conversation.update_chat_history(prompt, sanitized_text)
+                    except Exception as e:
+                        print(f"Warning: Failed to update conversation history: {str(e)}")
+            except CurlError as e:
+                raise exceptions.FailedToGenerateResponseError(
+                    f"Request failed (CurlError): {str(e)}"
+                ) from e
+            except Exception as e:
+                err_text = ""
+                if hasattr(e, "response"):
+                    response_obj = getattr(e, "response")
+                    if hasattr(response_obj, "text"):
+                        err_text = getattr(response_obj, "text")
+                raise exceptions.FailedToGenerateResponseError(
+                    f"Request failed ({type(e).__name__}): {str(e)} - {err_text}"
+                ) from e
+
+        def for_non_stream():
+            full_text = ""
+            try:
+                for chunk_data in for_stream():
+                    if raw:
+                        if isinstance(chunk_data, str):
+                            chunk_data = chunk_data.replace("\\\\", "\\").replace('\\"', '"')
+                            full_text += chunk_data
+                    else:
+                        if isinstance(chunk_data, dict) and "text" in chunk_data:
+                            text = chunk_data["text"].replace("\\\\", "\\").replace('\\"', '"')
+                            full_text += text
+            except Exception as e:
+                if not full_text:
+                    raise exceptions.FailedToGenerateResponseError(
+                        f"Failed to get non-stream response: {str(e)}"
+                    ) from e
+            self.last_response = {"text": full_text}
+            return full_text if raw else self.last_response
+
+        return for_stream() if stream else for_non_stream()
+
+    @staticmethod
+    def fix_encoding(text):
+        """
+        Fixes encoding issues in the response text.
+
+        Args:
+            text (Union[str, dict]): The text or response dict to fix encoding for.
+
+        Returns:
+            Union[str, dict]: The text or dict with encoding corrected if possible.
+        """
+        if isinstance(text, dict) and "text" in text:
+            try:
+                val = str(text["text"]).encode("latin1").decode("utf-8")
+                text["text"] = val.replace("\\\\", "\\").replace(
+                    '\\"', '"'
+                )  # Handle escaped backslashes
+                return text
+            except (UnicodeError, AttributeError):
+                val = str(text["text"])
+                text["text"] = val.replace("\\\\", "\\").replace(
+                    '\\"', '"'
+                )  # Handle escaped backslashes
+                return text
+        elif isinstance(text, str):
+            try:
+                return (
+                    text.encode("latin1").decode("utf-8").replace("\\\\", "\\").replace('\\"', '"')
+                )
+            except (UnicodeError, AttributeError):
+                return text.replace("\\\\", "\\").replace('\\"', '"')  # Handle escaped backslashes
+        return text
+
+    def chat(
+        self,
+        prompt: str,
+        stream: bool = False,
+        optimizer: Optional[str] = None,
+        conversationally: bool = False,
+        **kwargs: Any,
+    ) -> Union[str, Generator[str, None, None]]:
+        """
+        Sends a prompt to the HeckAI API and returns only the message text.
+
+        Args:
+            prompt (str): The prompt or question to send to the API.
+            stream (bool): If True, yields streaming response text.
+            optimizer (str, optional): Name of the optimizer to apply to the prompt.
+            conversationally (bool): If True, optimizer is applied to the full conversation prompt.
+            **kwargs: Additional parameters including raw.
+
+        Returns:
+            Union[str, Generator[str, None, None]]: The response text, or a generator yielding text chunks.
+        """
+        raw = kwargs.get("raw", False)
+
+        def for_stream_chat():
+            # ask() yields dicts or strings when streaming
+            gen = self.ask(
+                prompt, stream=True, raw=raw, optimizer=optimizer, conversationally=conversationally
+            )
+            for response in gen:
+                if raw:
+                    yield cast(str, response)
+                else:
+                    yield self.get_message(cast(Response, response))
+
+        def for_non_stream_chat():
+            # ask() returns dict or str when not streaming
+            response_data = self.ask(
+                prompt,
+                stream=False,
+                raw=raw,
+                optimizer=optimizer,
+                conversationally=conversationally,
+            )
+            if raw:
+                return response_data if isinstance(response_data, str) else str(response_data)
+            return self.get_message(response_data)  # get_message expects dict
+
+        return for_stream_chat() if stream else for_non_stream_chat()
+
+    def get_message(self, response: Response) -> str:
+        """
+        Extracts the message text from the API response.
+
+        Args:
+            response (Response): The API response dictionary.
+
+        Returns:
+            str: The extracted message text. Returns an empty string if not found.
+        """
+        # Validate response format
+        if not isinstance(response, dict):
+            return str(response)
+
+        # Handle missing text key gracefully
+        if "text" not in response:
+            return ""
+
+        # Ensure text is a string
+        text = response["text"]
+        if not isinstance(text, str):
+            text = str(text)
+
+        return text.replace("\\\\", "\\").replace('\\"', '"')
+
+
+if __name__ == "__main__":
+    from rich import print
+
+    ai = HeckAI(model="openai/gpt-5-nano")
+    response = ai.chat("tell me about humans", stream=True, raw=False)
+    if hasattr(response, "__iter__") and not isinstance(response, (str, bytes)):
+        for chunk in response:
+            print(chunk, end="", flush=True)
+    else:
+        print(response)