webscout 8.2.7__py3-none-any.whl → 8.2.9__py3-none-any.whl

webscout/Provider/OPENAI/venice.py

@@ -0,0 +1,431 @@
+import time
+import uuid
+import requests
+import json
+from typing import List, Dict, Optional, Union, Generator, Any
+
+# Import base classes and utility structures
+from .base import OpenAICompatibleProvider, BaseChat, BaseCompletions
+from .utils import (
+    ChatCompletionChunk, ChatCompletion, Choice, ChoiceDelta,
+    ChatCompletionMessage, CompletionUsage, count_tokens
+)
+
+# Attempt to import LitAgent, fallback if not available
+try:
+    from webscout.litagent import LitAgent
+except ImportError:
+    print("Warning: LitAgent not found. Some functionality may be limited.")
+
+# --- Venice Client ---
+
+class Completions(BaseCompletions):
+    def __init__(self, client: 'Venice'):
+        self._client = client
+
+    def create(
+        self,
+        *,
+        model: str,
+        messages: List[Dict[str, str]],
+        max_tokens: Optional[int] = 2049,
+        stream: bool = False,
+        temperature: Optional[float] = 0.8,
+        top_p: Optional[float] = 0.9,
+        **kwargs: Any
+    ) -> Union[ChatCompletion, Generator[ChatCompletionChunk, None, None]]:
+        """
+        Creates a model response for the given chat conversation.
+        Mimics openai.chat.completions.create
+        """
+        # Extract system message if present for systemPrompt parameter
+        system_prompt = self._client.system_prompt
+        for msg in messages:
+            if msg["role"] == "system":
+                system_prompt = msg["content"]
+                break
+
+        # Prepare the payload for Venice API
+        payload = {
+            "requestId": str(uuid.uuid4())[:7],
+            "modelId": self._client.convert_model_name(model),
+            "prompt": messages,
+            "systemPrompt": system_prompt,
+            "conversationType": "text",
+            "temperature": temperature if temperature is not None else self._client.temperature,
+            "webEnabled": True,
+            "topP": top_p if top_p is not None else self._client.top_p,
+            "includeVeniceSystemPrompt": False,
+            "isCharacter": False,
+            "clientProcessingTime": 2000
+        }
+
+        # Add optional parameters if provided
+        if max_tokens is not None and max_tokens > 0:
+            payload["max_tokens"] = max_tokens
+
+        # Add any additional parameters
+        for key, value in kwargs.items():
+            if key not in payload:
+                payload[key] = value
+
+        request_id = f"chatcmpl-{uuid.uuid4()}"
+        created_time = int(time.time())
+
+        if stream:
+            return self._create_stream(request_id, created_time, model, payload)
+        else:
+            return self._create_non_stream(request_id, created_time, model, payload)
+
+    def _create_stream(
+        self, request_id: str, created_time: int, model: str, payload: Dict[str, Any]
+    ) -> Generator[ChatCompletionChunk, None, None]:
+        try:
+            response = self._client.session.post(
+                self._client.api_endpoint,
+                json=payload,
+                stream=True,
+                timeout=self._client.timeout
+            )
+
+            # Handle non-200 responses
+            if response.status_code != 200:
+                raise IOError(
+                    f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
+                )
+
+            # Track token usage across chunks
+            prompt_tokens = 0
+            completion_tokens = 0
+            total_tokens = 0
+
+            # Estimate prompt tokens based on message length
+            prompt_tokens = 0
+            for msg in payload.get("prompt", []):
+                prompt_tokens += count_tokens(msg.get("content", ""))
+            prompt_tokens += count_tokens(payload.get("systemPrompt", ""))
+
+            for line in response.iter_lines():
+                if not line:
+                    continue
+
+                try:
+                    # Decode bytes to string
+                    line_data = line.decode('utf-8').strip()
+                    if '"kind":"content"' in line_data:
+                        data = json.loads(line_data)
+                        if 'content' in data:
+                            content = data['content']
+
+                            # Format the content (replace escaped newlines)
+                            content = self._client.format_text(content)
+
+                            # Update token counts
+                            completion_tokens += 1
+                            total_tokens = prompt_tokens + completion_tokens
+
+                            # Create the delta object
+                            delta = ChoiceDelta(
+                                content=content,
+                                role="assistant",
+                                tool_calls=None
+                            )
+
+                            # Create the choice object
+                            choice = Choice(
+                                index=0,
+                                delta=delta,
+                                finish_reason=None,
+                                logprobs=None
+                            )
+
+                            # Create the chunk object
+                            chunk = ChatCompletionChunk(
+                                id=request_id,
+                                choices=[choice],
+                                created=created_time,
+                                model=model,
+                                system_fingerprint=None
+                            )
+
+                            # Convert chunk to dict using Pydantic's API
+                            if hasattr(chunk, "model_dump"):
+                                chunk_dict = chunk.model_dump(exclude_none=True)
+                            else:
+                                chunk_dict = chunk.dict(exclude_none=True)
+
+                            # Add usage information to match OpenAI format
+                            usage_dict = {
+                                "prompt_tokens": prompt_tokens,
+                                "completion_tokens": completion_tokens,
+                                "total_tokens": total_tokens,
+                                "estimated_cost": None
+                            }
+
+                            chunk_dict["usage"] = usage_dict
+
+                            # Return the chunk object for internal processing
+                            yield chunk
+                except json.JSONDecodeError:
+                    continue
+                except UnicodeDecodeError:
+                    continue
+
+            # Final chunk with finish_reason="stop"
+            delta = ChoiceDelta(
+                content=None,
+                role=None,
+                tool_calls=None
+            )
+
+            choice = Choice(
+                index=0,
+                delta=delta,
+                finish_reason="stop",
+                logprobs=None
+            )
+
+            chunk = ChatCompletionChunk(
+                id=request_id,
+                choices=[choice],
+                created=created_time,
+                model=model,
+                system_fingerprint=None
+            )
+
+            if hasattr(chunk, "model_dump"):
+                chunk_dict = chunk.model_dump(exclude_none=True)
+            else:
+                chunk_dict = chunk.dict(exclude_none=True)
+            chunk_dict["usage"] = {
+                "prompt_tokens": prompt_tokens,
+                "completion_tokens": completion_tokens,
+                "total_tokens": total_tokens,
+                "estimated_cost": None
+            }
+
+            yield chunk
+
+        except Exception as e:
+            print(f"Error during Venice stream request: {e}")
+            raise IOError(f"Venice request failed: {e}") from e
+
+    def _create_non_stream(
+        self, request_id: str, created_time: int, model: str, payload: Dict[str, Any]
+    ) -> ChatCompletion:
+        try:
+            # For non-streaming, we still use streaming internally to collect the full response
+            response = self._client.session.post(
+                self._client.api_endpoint,
+                json=payload,
+                stream=True,
+                timeout=self._client.timeout
+            )
+
+            # Handle non-200 responses
+            if response.status_code != 200:
+                raise IOError(
+                    f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
+                )
+
+            # Collect the full response
+            full_text = ""
+            for line in response.iter_lines():
+                if not line:
+                    continue
+
+                try:
+                    # Decode bytes to string
+                    line_data = line.decode('utf-8').strip()
+                    if '"kind":"content"' in line_data:
+                        data = json.loads(line_data)
+                        if 'content' in data:
+                            content = data['content']
+                            full_text += content
+                except json.JSONDecodeError:
+                    continue
+                except UnicodeDecodeError:
+                    continue
+
+            # Format the text (replace escaped newlines)
+            full_text = self._client.format_text(full_text)
+
+            # Estimate token counts
+            prompt_tokens = 0
+            for msg in payload.get("prompt", []):
+                prompt_tokens += count_tokens(msg.get("content", ""))
+            prompt_tokens += count_tokens(payload.get("systemPrompt", ""))
+            completion_tokens = count_tokens(full_text)
+            total_tokens = prompt_tokens + completion_tokens
+
+            # Create the message object
+            message = ChatCompletionMessage(
+                role="assistant",
+                content=full_text
+            )
+
+            # Create the choice object
+            choice = Choice(
+                index=0,
+                message=message,
+                finish_reason="stop"
+            )
+
+            # Create the usage object
+            usage = CompletionUsage(
+                prompt_tokens=prompt_tokens,
+                completion_tokens=completion_tokens,
+                total_tokens=total_tokens
+            )
+
+            # Create the completion object
+            completion = ChatCompletion(
+                id=request_id,
+                choices=[choice],
+                created=created_time,
+                model=model,
+                usage=usage,
+            )
+
+            return completion
+
+        except Exception as e:
+            print(f"Error during Venice non-stream request: {e}")
+            raise IOError(f"Venice request failed: {e}") from e
+
+class Chat(BaseChat):
+    def __init__(self, client: 'Venice'):
+        self.completions = Completions(client)
+
+class Venice(OpenAICompatibleProvider):
+    """
+    OpenAI-compatible client for Venice AI API.
+
+    Usage:
+        client = Venice()
+        response = client.chat.completions.create(
+            model="mistral-31-24b",
+            messages=[{"role": "user", "content": "Hello!"}]
+        )
+    """
+
+    AVAILABLE_MODELS = [
+        "mistral-31-24b",
+        "llama-3.2-3b-akash",
+        "qwen2dot5-coder-32b",
+        "deepseek-coder-v2-lite",
+    ]
+
+    # No model mapping needed as we use the model names directly
+
+    def __init__(
+        self,
+        timeout: Optional[int] = None,
+        browser: str = "chrome"
+    ):
+        """
+        Initialize the Venice client.
+
+        Args:
+            timeout: Request timeout in seconds (None for no timeout)
+            browser: Browser to emulate in user agent
+        """
+        self.timeout = timeout
+        self.temperature = 0.8  # Default temperature
+        self.top_p = 0.9  # Default top_p
+        self.system_prompt = "You are a helpful AI assistant."  # Default system prompt
+        self.api_endpoint = "https://venice.ai/api/inference/chat"
+        self.session = requests.Session()
+
+        # Initialize LitAgent for user agent generation
+        agent = LitAgent()
+        self.fingerprint = agent.generate_fingerprint(browser)
+
+        # Headers for the request
+        self.headers = {
+            "User-Agent": self.fingerprint["user_agent"],
+            "accept": self.fingerprint["accept"],
+            "accept-language": self.fingerprint["accept_language"],
+            "content-type": "application/json",
+            "origin": "https://venice.ai",
+            "referer": "https://venice.ai/chat/",
+            "sec-ch-ua": self.fingerprint["sec_ch_ua"] or '"Google Chrome";v="133", "Chromium";v="133", "Not?A_Brand";v="24"',
+            "sec-ch-ua-mobile": "?0",
+            "sec-ch-ua-platform": f'"{self.fingerprint["platform"]}"',
+            "sec-fetch-dest": "empty",
+            "sec-fetch-mode": "cors",
+            "sec-fetch-site": "same-origin"
+        }
+
+        self.session.headers.update(self.headers)
+
+        # Initialize the chat interface
+        self.chat = Chat(self)
+
+    def format_text(self, text: str) -> str:
+        """
+        Format text by replacing escaped newlines with actual newlines.
+
+        Args:
+            text: Text to format
+
+        Returns:
+            Formatted text
+        """
+        # Use a more comprehensive approach to handle all escape sequences
+        try:
+            # First handle double backslashes to avoid issues
+            text = text.replace('\\\\', '\\')
+
+            # Handle common escape sequences
+            text = text.replace('\\n', '\n')
+            text = text.replace('\\r', '\r')
+            text = text.replace('\\t', '\t')
+            text = text.replace('\\"', '"')
+            text = text.replace("\\'", "'")
+
+            # Handle any remaining escape sequences using JSON decoding
+            # This is a fallback in case there are other escape sequences
+            try:
+                # Add quotes to make it a valid JSON string
+                json_str = f'"{text}"'
+                # Use json module to decode all escape sequences
+                decoded = json.loads(json_str)
+                return decoded
+            except json.JSONDecodeError:
+                # If JSON decoding fails, return the text with the replacements we've already done
+                return text
+        except Exception as e:
+            # If any error occurs, return the original text
+            print(f"Warning: Error formatting text: {e}")
+            return text
+
+    def convert_model_name(self, model: str) -> str:
+        """
+        Convert model names to ones supported by Venice.
+
+        Args:
+            model: Model name to convert
+
+        Returns:
+            Venice model name
+        """
+        # If the model is already a valid Venice model, return it
+        if model in self.AVAILABLE_MODELS:
+            return model
+
+        # Default to the most capable model
+        print(f"Warning: Unknown model '{model}'. Using 'mistral-31-24b' instead.")
+        return "mistral-31-24b"
+
+    @property
+    def models(self):
+        class _ModelList:
+            def list(inner_self):
+                return type(self).AVAILABLE_MODELS
+        return _ModelList()
+
+    @classmethod
+    def models(cls):
+        """Return the list of available models for Venice."""
+        return cls.AVAILABLE_MODELS