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

webscout/Provider/IBM.py

@@ -0,0 +1,340 @@
+import json
+import uuid
+from datetime import datetime
+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 (  # Import sanitize_stream
+    AwesomePrompts,
+    Conversation,
+    Optimizers,
+    sanitize_stream,
+)
+from webscout.litagent import LitAgent
+
+
+class IBM(Provider):
+    """
+    A class to interact with the IBM Granite Playground API.
+    """
+    required_auth = False
+    AVAILABLE_MODELS = [
+        "granite-chat",
+        "granite-thinking",
+        "granite-search",
+        "granite-research",
+    ]
+
+    @staticmethod
+    def _ibm_extractor(chunk: Union[str, Dict[str, Any]]) -> Optional[str]:
+        """Extracts content from IBM Granite stream JSON objects."""
+        if isinstance(chunk, dict):
+            if chunk.get("type") == "message.part":
+                part = chunk.get("part", {})
+                return part.get("content")
+        return None
+
+    def get_token(self) -> str:
+        """Fetches a fresh dynamic Bearer token from the IBM UI auth endpoint."""
+        auth_url = "https://www.ibm.com/granite/playground/api/v1/ui/auth"
+        try:
+            # Use the existing session to benefit from cookies/headers
+            response = self.session.get(auth_url, timeout=self.timeout, impersonate="chrome110")
+            if response.ok:
+                data = response.json()
+                token = data.get("token")
+                if token:
+                    self.headers["Authorization"] = f"Bearer {token}"
+                    self.session.headers.update(self.headers)
+                    return token
+            raise exceptions.FailedToGenerateResponseError(f"Failed to fetch auth token: {response.status_code}")
+        except Exception as e:
+            raise exceptions.FailedToGenerateResponseError(f"Error fetching auth token: {str(e)}")
+
+    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 = "granite-chat",
+        system_prompt: str = "You are a helpful assistant.",
+        browser: str = "chrome" # Note: browser fingerprinting might be less effective with impersonate
+    ):
+        """Initializes the IBM Granite API client."""
+        if model not in self.AVAILABLE_MODELS:
+            raise ValueError(f"Invalid model: {model}. Choose from: {self.AVAILABLE_MODELS}")
+
+        self.url = "https://d1eh1ubv87xmm5.cloudfront.net/granite/playground/api/v1/acp/runs"
+
+        # Initialize LitAgent
+        self.agent = LitAgent()
+        self.fingerprint = self.agent.generate_fingerprint(browser)
+        self.headers = {
+            "Accept": "text/event-stream",
+            "Accept-Language": self.fingerprint.get("accept_language", "en-US,en;q=0.9"),
+            "Authorization": "", # Will be filled by get_token
+            "Content-Type": "application/json",
+            "Cache-Control": "no-cache",
+            "Origin": "https://www.ibm.com",
+            "Pragma": "no-cache",
+            "Referer": "https://www.ibm.com/granite/playground",
+            "Sec-Fetch-Dest": "empty",
+            "Sec-Fetch-Mode": "cors",
+            "Sec-Fetch-Site": "cross-site",
+            "User-Agent": self.fingerprint.get("user_agent", ""),
+            "Sec-CH-UA": self.fingerprint.get("sec_ch_ua", ""),
+            "Sec-CH-UA-Mobile": "?0",
+            "Sec-CH-UA-Platform": f'"{self.fingerprint.get("platform", "")}"',
+        }
+
+        # Initialize curl_cffi Session
+        self.session = Session()
+        self.session.headers.update(self.headers)
+        if proxies:
+            self.session.proxies.update(proxies)
+
+        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("__")
+        )
+        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 refresh_identity(self, browser: Optional[str] = None):
+        """
+        Refreshes the browser identity fingerprint.
+
+        Args:
+            browser: Specific browser to use for the new fingerprint
+        """
+        browser = browser or self.fingerprint.get("browser_type", "chrome")
+        self.fingerprint = self.agent.generate_fingerprint(browser)
+
+        # Update headers with new fingerprint
+        self.headers.update({
+            "Accept-Language": self.fingerprint.get("accept_language", "en-US,en;q=0.9"),
+            "User-Agent": self.fingerprint.get("user_agent", ""),
+            "Sec-CH-UA": self.fingerprint.get("sec_ch_ua", ""),
+            "Sec-CH-UA-Platform": f'"{self.fingerprint.get("platform", "")}"',
+        })
+
+        # Update session headers
+        self.session.headers.update(self.headers)
+
+        return self.fingerprint
+
+    def ask(
+        self,
+        prompt: str,
+        stream: bool = False,
+        raw: bool = False,
+        optimizer: Optional[str] = None,
+        conversationally: bool = False,
+        **kwargs: Any,
+    ) -> 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}")
+
+        now = datetime.now().isoformat()
+        payload = {
+            "agent_name": self.model,
+            "input": [
+                {
+                    "role": "user",
+                    "parts": [
+                        {
+                            "content_type": "text/plain",
+                            "content": conversation_prompt,
+                            "content_encoding": "plain",
+                            "role": "user"
+                        }
+                    ],
+                    "created_at": now,
+                    "completed_at": now
+                }
+            ],
+            "mode": "stream",
+            "session_id": str(uuid.uuid4())
+        }
+
+        def for_stream():
+            streaming_text = "" # Initialize outside try block
+            try:
+                # Use curl_cffi session post with impersonate
+                response = self.session.post(
+                    self.url,
+                    data=json.dumps(payload),
+                    stream=True,
+                    timeout=self.timeout,
+                    impersonate="chrome110" # Use a common impersonation profile
+                )
+
+                if response.status_code in [401, 403]:
+                    # Token expired, refresh and retry once
+                    self.get_token()
+                    response = self.session.post(
+                        self.url,
+                        data=json.dumps(payload),
+                        stream=True,
+                        timeout=self.timeout,
+                        impersonate="chrome110"
+                    )
+
+                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="data:",
+                    to_json=True,     # Stream sends JSON
+                    skip_markers=["[DONE]"],
+                    content_extractor=self._ibm_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 _ibm_extractor
+                    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): {str(e)}") from e
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"Request failed ({type(e).__name__}): {str(e)}") from e
+            finally:
+                # Update history after stream finishes or fails
+                if streaming_text:
+                    self.last_response = {"text": streaming_text}
+                    self.conversation.update_chat_history(prompt, streaming_text)
+
+
+        def for_non_stream():
+            try:
+                # For non-stream, we collect from stream
+                final_content = ""
+                for chunk_data in for_stream():
+                    if raw:
+                        if isinstance(chunk_data, str):
+                            final_content += chunk_data
+                        elif isinstance(chunk_data, bytes):
+                            final_content += chunk_data.decode('utf-8', errors='ignore')
+                    else:
+                        if isinstance(chunk_data, dict):
+                            if "text" in chunk_data:
+                                final_content += chunk_data["text"]
+                        elif isinstance(chunk_data, str):
+                            final_content += chunk_data
+
+                if not final_content:
+                    raise exceptions.FailedToGenerateResponseError("Empty response from provider")
+
+                self.last_response = {"text": final_content}
+                return self.last_response if not raw else final_content
+
+            except Exception as e:
+                raise exceptions.FailedToGenerateResponseError(f"Failed to get non-stream response: {str(e)}") from e
+
+
+        return for_stream() if stream else for_non_stream()
+
+    def chat(
+        self,
+        prompt: str,
+        stream: bool = False,
+        optimizer: Optional[str] = None,
+        conversationally: bool = False,
+        **kwargs: Any,
+    ) -> Union[str, Generator[str, None, None]]:
+        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_dict in gen:
+                if raw:
+                    yield cast(str, response_dict)
+                else:
+                    yield self.get_message(cast(Dict[str, Any], response_dict)) # get_message expects dict
+
+        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 cast(str, response_data)
+            else:
+                return self.get_message(cast(Dict[str, Any], response_data)) # get_message expects dict
+
+        return for_stream_chat() if stream else for_non_stream_chat()
+
+    def get_message(self, response: Response) -> str:
+        if not isinstance(response, dict):
+            return str(response)
+        return response["text"]
+
+if __name__ == "__main__":
+    # Ensure curl_cffi is installed
+    print("-" * 80)
+    print(f"{'Model':<50} {'Status':<10} {'Response'}")
+    print("-" * 80)
+
+    for model in IBM.AVAILABLE_MODELS:
+        try:
+            test_ai = IBM(model=model, timeout=60,)
+            response = test_ai.chat("Say 'Hello' in one word", stream=True)
+            response_text = ""
+            if hasattr(response, "__iter__") and not isinstance(response, (str, bytes)):
+                for chunk in response:
+                    response_text += chunk
+            else:
+                response_text = str(response)
+
+            if response_text and len(response_text.strip()) > 0:
+                status = "✓"
+                # Clean and truncate response
+                clean_text = response_text.strip().encode('utf-8', errors='ignore').decode('utf-8')
+                display_text = clean_text[:50] + "..." if len(clean_text) > 50 else clean_text
+            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)}")