webscout 8.3.7__py3-none-any.whl → 2025.10.13__py3-none-any.whl

webscout/Provider/Apriel.py

@@ -0,0 +1,306 @@
+"""
+A class to interact with the Apriel Gradio chat API (servicenow-ai-apriel-chat.hf.space).
+
+This provider integrates the Apriel chat model into the Webscout framework.
+"""
+from typing import Generator, Optional, Union, Any, Dict
+import json
+import time
+from curl_cffi import CurlError
+from curl_cffi.requests import Session
+
+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
+from webscout.sanitize import sanitize_stream
+
+
+class Apriel(Provider):
+    """
+    A class to interact with the Apriel Gradio chat API.
+
+    Attributes:
+        system_prompt (str): The system prompt to define the assistant's role.
+
+    Examples:
+        >>> from webscout.Provider.apriel import Apriel
+        >>> ai = Apriel()
+        >>> response = ai.chat("What's the weather today?")
+        >>> print(response)
+        'The weather today is sunny with a high of 75°F.'
+    """
+    required_auth = False
+    AVAILABLE_MODELS = ["UNKNOWN"]
+
+    def __init__(
+        self,
+        is_conversation: bool = True,
+        max_tokens: int = 600,
+        timeout: int = 30,
+        intro: str = None,
+        filepath: str = None,
+        update_file: bool = True,
+        proxies: dict = {},
+        history_offset: int = 10250,
+        act: str = None,
+        system_prompt: str = "You are a helpful assistant.",
+        model: str = "UNKNOWN"
+    ):
+        """
+        Initializes the Apriel API with given parameters.
+
+        Args:
+            is_conversation (bool): Whether the provider is in conversation mode.
+            max_tokens (int): Maximum number of tokens to sample.
+            timeout (int): Timeout for API requests.
+            intro (str): Introduction message for the conversation.
+            filepath (str): Filepath for storing conversation history.
+            update_file (bool): Whether to update the conversation history file.
+            proxies (dict): Proxies for the API requests.
+            history_offset (int): Offset for conversation history.
+            act (str): Act for the conversation.
+            system_prompt (str): The system prompt to define the assistant's role.
+        """
+        self.session = Session()
+        self.is_conversation = is_conversation
+        self.max_tokens_to_sample = max_tokens
+        self.api_endpoint = "https://servicenow-ai-apriel-chat.hf.space"
+        self.timeout = timeout
+        self.last_response = {}
+        self.system_prompt = system_prompt
+
+        # Initialize LitAgent for user agent generation
+        self.agent = LitAgent()
+
+        self.headers = {
+            "Content-Type": "application/json",
+            "User-Agent": self.agent.random(),
+            "Accept": "text/event-stream",
+            "Cache-Control": "no-cache",
+        }
+
+        self.__available_optimizers = (
+            method
+            for method in dir(Optimizers)
+            if callable(getattr(Optimizers, method)) and not method.startswith("__")
+        )
+        self.session.headers.update(self.headers)
+        self.session.proxies = proxies
+
+        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 _get_session_hash(self) -> str:
+        """Generate or get a session hash for the Gradio API."""
+        try:
+            url = f"{self.api_endpoint}/gradio_api/heartbeat"
+            response = self.session.get(url, timeout=self.timeout)
+            response.raise_for_status()
+            return str(int(time.time()))
+        except Exception:
+            return str(int(time.time()))
+
+    def _join_queue(self, session_hash: str, message: str, fn_index: int = 1, trigger_id: int = 16) -> Optional[str]:
+        """Send the user message to /gradio_api/queue/join and return event_id if available."""
+        url = f"{self.api_endpoint}/gradio_api/queue/join"
+        payload = {
+            "data": [[], {"text": message, "files": []}, None],
+            "event_data": None,
+            "fn_index": fn_index,
+            "trigger_id": trigger_id,
+            "session_hash": session_hash,
+        }
+        response = self.session.post(url, json=payload, timeout=self.timeout)
+        response.raise_for_status()
+        try:
+            return response.json().get("event_id")
+        except Exception:
+            return None
+
+    def _run_predict(self, session_hash: str, fn_index: int = 3, trigger_id: int = 16) -> None:
+        """Call /gradio_api/run/predict to start processing the queued request."""
+        url = f"{self.api_endpoint}/gradio_api/run/predict"
+        payload = {
+            "data": [],
+            "event_data": None,
+            "fn_index": fn_index,
+            "trigger_id": trigger_id,
+            "session_hash": session_hash
+        }
+        response = self.session.post(url, json=payload, timeout=self.timeout)
+        response.raise_for_status()
+
+    @staticmethod
+    def _apriel_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from Apriel Gradio stream JSON objects."""
+        if isinstance(chunk, dict):
+            msg = chunk.get("msg")
+            if msg == "process_generating":
+                output = chunk.get("output", {})
+                data = output.get("data")
+                if data and isinstance(data, list) and len(data) > 0:
+                    ops = data[0]
+                    for op in ops:
+                        if isinstance(op, list) and len(op) > 2 and op[0] == "append":
+                            return op[2]
+        return None
+
+    def ask(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+        optimizer: str = None,
+        conversationally: bool = False,
+    ) -> Union[Dict[str, Any], Generator]:
+        """
+        Sends a prompt to the Apriel Gradio API and returns the response.
+
+        Args:
+            prompt (str): The prompt to send to the API.
+            stream (bool): Whether to stream the response.
+            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.
+        """
+        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}"
+                )
+
+        session_hash = self._get_session_hash()
+        event_id = self._join_queue(session_hash, conversation_prompt)
+        self._run_predict(session_hash)
+
+        def for_stream():
+            streaming_text = ""
+            try:
+                url = f"{self.api_endpoint}/gradio_api/queue/data?session_hash={session_hash}"
+                response = self.session.get(
+                    url,
+                    stream=True,
+                    timeout=self.timeout,
+                    impersonate="chrome110"
+                )
+                if not response.ok:
+                    raise exceptions.FailedToGenerateResponseError(
+                        f"Failed to generate response - ({response.status_code}, {response.reason}) - {response.text}"
+                    )
+
+                # Use sanitize_stream
+                processed_stream = sanitize_stream(
+                    data=response.iter_content(chunk_size=None),
+                    intro_value="data:",
+                    to_json=True,
+                    content_extractor=self._apriel_extractor,
+                    yield_raw_on_error=False
+                )
+
+                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:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed (CurlError): {e}")
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"An unexpected error occurred ({type(e).__name__}): {e}")
+            finally:
+                if streaming_text:
+                    self.last_response = {"text": streaming_text}
+                    self.conversation.update_chat_history(prompt, streaming_text)
+
+        def for_non_stream():
+            for _ in for_stream():
+                pass
+            return self.last_response if not raw else self.last_response.get("text", "")
+
+        return for_stream() if stream else for_non_stream()
+
+    def chat(
+        self,
+        prompt: str,
+        stream: bool = False,
+        optimizer: str = None,
+        conversationally: bool = False,
+        raw: bool = False,
+    ) -> Union[str, Generator[str, None, None]]:
+        """
+        Generates a response from the Apriel API.
+
+        Args:
+            prompt (str): The prompt to send to the API.
+            stream (bool): Whether to stream the response.
+            optimizer (str): Optimizer to use for the prompt.
+            conversationally (bool): Whether to generate the prompt conversationally.
+            raw (bool): Whether to return raw response chunks.
+
+        Returns:
+            str: The API response.
+        """
+
+        def for_stream():
+            for response in self.ask(
+                prompt, True, raw=raw, optimizer=optimizer, conversationally=conversationally
+            ):
+                if raw:
+                    yield response
+                else:
+                    yield self.get_message(response)
+
+        def for_non_stream():
+            result = self.ask(
+                prompt,
+                False,
+                raw=raw,
+                optimizer=optimizer,
+                conversationally=conversationally,
+            )
+            if raw:
+                return result
+            else:
+                return self.get_message(result)
+
+        return for_stream() if stream else for_non_stream()
+
+    def get_message(self, response: dict) -> str:
+        """
+        Extracts the message from the API response.
+
+        Args:
+            response (dict): The API response.
+
+        Returns:
+            str: The message content.
+        """
+        assert isinstance(response, dict), "Response should be of dict data-type only"
+        return response.get("text", "")
+
+if __name__ == "__main__":
+    from rich import print
+    ai = Apriel(timeout=60)
+    response = ai.chat("write a poem about AI", stream=True, raw=False)
+    for chunk in response:
+        print(chunk, end="", flush=True)