webscout 8.2.8__py3-none-any.whl → 8.3__py3-none-any.whl

webscout/Provider/OPENAI/Qwen3.py

@@ -0,0 +1,304 @@
+import requests
+import json
+import time
+import uuid
+from typing import List, Dict, Optional, Union, Generator, Any
+
+from webscout.Provider.OPENAI.base import OpenAICompatibleProvider, BaseChat, BaseCompletions
+from webscout.Provider.OPENAI.utils import (
+    ChatCompletionChunk, ChatCompletion, Choice, ChoiceDelta,
+    ChatCompletionMessage, CompletionUsage,
+    get_last_user_message, get_system_prompt,
+    count_tokens
+)
+
+class Completions(BaseCompletions):
+    def __init__(self, client: 'Qwen3'):
+        self._client = client
+
+    def create(
+        self,
+        *,
+        model: str,
+        messages: List[Dict[str, str]],
+        max_tokens: Optional[int] = 2048,
+        stream: bool = False,
+        temperature: Optional[float] = None,
+        top_p: Optional[float] = None,
+        timeout: Optional[int] = None,
+        proxies: Optional[dict] = None,
+        **kwargs: Any
+    ) -> Union[ChatCompletion, Generator[ChatCompletionChunk, None, None]]:
+        payload = {
+            "data": [
+                get_last_user_message(messages),
+                {
+                    "thinking_budget": kwargs.get("thinking_budget", 38),
+                    "model": self._client.get_model(model),
+                    "sys_prompt": get_system_prompt(messages)
+                },
+                None, None
+            ],
+            "event_data": None,
+            "fn_index": 13,
+            "trigger_id": 31,
+            "session_hash": str(uuid.uuid4()).replace('-', '')
+        }
+
+        request_id = f"chatcmpl-{uuid.uuid4()}"
+        created_time = int(time.time())
+
+        if stream:
+            return self._create_stream(request_id, created_time, model, payload, timeout=timeout, proxies=proxies)
+        else:
+            return self._create_non_stream(request_id, created_time, model, payload, timeout=timeout, proxies=proxies)
+
+    def _create_stream(
+        self, request_id: str, created_time: int, model: str, payload: Dict[str, Any],
+        timeout: Optional[int] = None, proxies: Optional[dict] = None
+    ) -> Generator[ChatCompletionChunk, None, None]:
+        original_proxies = self._client.session.proxies.copy()
+        if proxies is not None:
+            self._client.session.proxies = proxies
+        else:
+            self._client.session.proxies = {}
+        try:
+            session = self._client.session
+            headers = self._client.headers
+            # Step 1: Join the queue
+            join_resp = session.post(self._client.api_endpoint, headers=headers, json=payload, timeout=timeout if timeout is not None else self._client.timeout)
+            join_resp.raise_for_status()
+            event_id = join_resp.json().get('event_id')
+            session_hash = payload["session_hash"]
+
+            # Step 2: Stream data
+            params = {'session_hash': session_hash}
+            stream_resp = session.get(self._client.url + "/gradio_api/queue/data", headers=self._client.stream_headers, params=params, stream=True, timeout=timeout if timeout is not None else self._client.timeout)
+            stream_resp.raise_for_status()
+
+            # --- New logic to yield all content, tool reasoning, and status, similar to Reasoning class ---
+            is_thinking_tag_open = False # True if <think> has been yielded and not yet </think>
+
+            for line in stream_resp.iter_lines():
+                if line:
+                    decoded_line = line.decode('utf-8')
+                    if decoded_line.startswith('data: '):
+                        try:
+                            json_data = json.loads(decoded_line[6:])
+                            if json_data.get('msg') == 'process_generating':
+                                if 'output' in json_data and 'data' in json_data['output'] and len(json_data['output']['data']) > 5:
+                                    updates_list = json_data['output']['data'][5] # This is a list of operations
+                                    for op_details in updates_list:
+                                        action = op_details[0]
+                                        path = op_details[1]
+                                        value = op_details[2]
+
+                                        content_to_yield = None
+                                        is_current_op_tool = False
+                                        is_current_op_text = False
+
+                                        # Case 1: Adding a new content block (tool or text object)
+                                        if action == "add" and isinstance(value, dict) and "type" in value:
+                                            if len(path) == 4 and path[0] == "value" and path[2] == "content":
+                                                block_type = value.get("type")
+                                                content_to_yield = value.get("content")
+                                                if block_type == "tool":
+                                                    is_current_op_tool = True
+                                                elif block_type == "text":
+                                                    is_current_op_text = True
+                                        
+                                        # Case 2: Appending content string to an existing block
+                                        elif action == "append" and isinstance(value, str):
+                                            if len(path) == 5 and path[0] == "value" and path[2] == "content" and path[4] == "content":
+                                                block_index = path[3] # 0 for tool's content, 1 for text's content
+                                                content_to_yield = value
+                                                if block_index == 0: # Appending to tool's content
+                                                    is_current_op_tool = True
+                                                elif block_index == 1: # Appending to text's content
+                                                    is_current_op_text = True
+                                        
+                                        # Case 3: Tool status update (e.g., "End of Thought")
+                                        elif action == "replace" and len(path) == 6 and \
+                                             path[0] == "value" and path[2] == "content" and \
+                                             path[3] == 0 and path[4] == "options" and path[5] == "status": # path[3]==0 ensures it's the tool block
+                                            if value == "done": # Tool block processing is complete
+                                                if is_thinking_tag_open:
+                                                    delta = ChoiceDelta(content="</think>\n\n", role="assistant")
+                                                    yield ChatCompletionChunk(id=request_id, choices=[Choice(index=0, delta=delta)], created=created_time, model=model)
+                                                    is_thinking_tag_open = False
+                                            continue # This operation itself doesn't yield visible content
+
+                                        # Yielding logic
+                                        if is_current_op_tool and content_to_yield:
+                                            if not is_thinking_tag_open:
+                                                delta = ChoiceDelta(content="<think>", role="assistant")
+                                                yield ChatCompletionChunk(id=request_id, choices=[Choice(index=0, delta=delta)], created=created_time, model=model)
+                                                is_thinking_tag_open = True
+                                            
+                                            delta = ChoiceDelta(content=content_to_yield, role="assistant")
+                                            yield ChatCompletionChunk(id=request_id, choices=[Choice(index=0, delta=delta)], created=created_time, model=model)
+
+                                        elif is_current_op_text and content_to_yield:
+                                            if is_thinking_tag_open: # If text starts, close any open thinking tag
+                                                delta = ChoiceDelta(content="</think>", role="assistant")
+                                                yield ChatCompletionChunk(id=request_id, choices=[Choice(index=0, delta=delta)], created=created_time, model=model)
+                                                is_thinking_tag_open = False
+                                            
+                                            delta = ChoiceDelta(content=content_to_yield, role="assistant")
+                                            yield ChatCompletionChunk(id=request_id, choices=[Choice(index=0, delta=delta)], created=created_time, model=model)
+
+                            if json_data.get('msg') == 'process_completed':
+                                if is_thinking_tag_open: # Ensure </think> is yielded if process completes mid-thought
+                                    delta = ChoiceDelta(content="</think>", role="assistant")
+                                    yield ChatCompletionChunk(id=request_id, choices=[Choice(index=0, delta=delta)], created=created_time, model=model)
+                                    is_thinking_tag_open = False
+                                break
+                        except json.JSONDecodeError:
+                            continue
+                        except Exception as e:
+                            # Log or handle other potential exceptions
+                            continue
+            
+            # After the loop, ensure the tag is closed if the stream broke for reasons other than 'process_completed'
+            if is_thinking_tag_open:
+                delta = ChoiceDelta(content="</think>", role="assistant")
+                yield ChatCompletionChunk(id=request_id, choices=[Choice(index=0, delta=delta)], created=created_time, model=model)
+        finally:
+            self._client.session.proxies = original_proxies
+
+    def _create_non_stream(
+        self, request_id: str, created_time: int, model: str, payload: Dict[str, Any],
+        timeout: Optional[int] = None, proxies: Optional[dict] = None
+    ) -> ChatCompletion:
+        original_proxies = self._client.session.proxies.copy()
+        if proxies is not None:
+            self._client.session.proxies = proxies
+        else:
+            self._client.session.proxies = {}
+        try:
+            # For non-streaming, just call the join endpoint and parse the result
+            session = self._client.session
+            headers = self._client.headers
+            resp = session.post(self._client.api_endpoint, headers=headers, json=payload, timeout=timeout if timeout is not None else self._client.timeout)
+            resp.raise_for_status()
+            data = resp.json()
+            # Return the full content as a single message, including all tool and text reasoning if present
+            output = ""
+            if 'output' in data and 'data' in data['output'] and len(data['output']['data']) > 5:
+                updates = data['output']['data'][5]
+                parts = []
+                for update in updates:
+                    if isinstance(update, list) and len(update) > 2 and isinstance(update[2], str):
+                        parts.append(update[2])
+                    elif isinstance(update, list) and isinstance(update[1], list) and len(update[1]) > 4:
+                        if update[1][4] == "content":
+                            parts.append(update[2])
+                        elif update[1][4] == "options" and update[2] != "done":
+                            parts.append(str(update[2]))
+                    elif isinstance(update, dict):
+                        if update.get('type') == 'tool':
+                            parts.append(update.get('content', ''))
+                        elif update.get('type') == 'text':
+                            parts.append(update.get('content', ''))
+                output = "\n".join([str(p) for p in parts if p])
+            else:
+                output = data.get('output', {}).get('data', ["", "", "", "", "", [["", "", ""]]])[5][0][2]
+            message = ChatCompletionMessage(role="assistant", content=output)
+            choice = Choice(index=0, message=message, finish_reason="stop")
+            # Use count_tokens to compute usage
+            prompt_tokens = count_tokens([m.get('content', '') for m in payload['data'] if isinstance(m, dict) and 'content' in m or isinstance(m, str)])
+            completion_tokens = count_tokens(output)
+            usage = CompletionUsage(
+                prompt_tokens=prompt_tokens,
+                completion_tokens=completion_tokens,
+                total_tokens=prompt_tokens + completion_tokens
+            )
+            completion = ChatCompletion(
+                id=request_id,
+                choices=[choice],
+                created=created_time,
+                model=model,
+                usage=usage,
+            )
+            return completion
+        finally:
+            self._client.session.proxies = original_proxies
+
+class Chat(BaseChat):
+    def __init__(self, client: 'Qwen3'):
+        self.completions = Completions(client)
+
+class Qwen3(OpenAICompatibleProvider):
+    url = "https://qwen-qwen3-demo.hf.space"
+    api_endpoint = "https://qwen-qwen3-demo.hf.space/gradio_api/queue/join?__theme=system"
+    AVAILABLE_MODELS = [
+        "qwen3-235b-a22b",
+        "qwen3-32b",
+        "qwen3-30b-a3b",
+        "qwen3-14b",
+        "qwen3-8b",
+        "qwen3-4b",
+        "qwen3-1.7b",
+        "qwen3-0.6b",
+    ]
+    MODEL_ALIASES = {
+        "qwen-3-235b": "qwen3-235b-a22b",
+        "qwen-3-30b": "qwen3-30b-a3b",
+        "qwen-3-32b": "qwen3-32b",
+        "qwen-3-14b": "qwen3-14b",
+        "qwen-3-4b": "qwen3-4b",
+        "qwen-3-1.7b": "qwen3-1.7b",
+        "qwen-3-0.6b": "qwen3-0.6b"
+    }
+
+    def __init__(self):
+        self.timeout = 30
+        self.session = requests.Session()
+        self.session.proxies = {}
+        self.headers = {
+            'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:136.0) Gecko/20100101 Firefox/136.0',
+            'Accept': '*/*',
+            'Accept-Language': 'en-US,en;q=0.5',
+            'Accept-Encoding': 'gzip, deflate, br, zstd',
+            'Referer': f'{self.url}/?__theme=system',
+            'content-type': 'application/json',
+            'Origin': self.url,
+            'Connection': 'keep-alive',
+            'Sec-Fetch-Dest': 'empty',
+            'Sec-Fetch-Mode': 'cors',
+            'Sec-Fetch-Site': 'same-origin',
+            'Pragma': 'no-cache',
+            'Cache-Control': 'no-cache',
+        }
+        self.stream_headers = {
+            'Accept': 'text/event-stream',
+            'Accept-Language': 'en-US,en;q=0.5',
+            'Referer': f'{self.url}/?__theme=system',
+            'User-Agent': self.headers['User-Agent'],
+        }
+        self.session.headers.update(self.headers)
+        self.chat = Chat(self)
+
+    def get_model(self, model):
+        return self.MODEL_ALIASES.get(model, model)
+
+    @property
+    def models(self):
+        class _ModelList:
+            def list(inner_self):
+                return type(self).AVAILABLE_MODELS
+        return _ModelList()
+
+if __name__ == "__main__":
+    client = Qwen3()
+    from rich import print
+    resp = client.chat.completions.create(
+        model="qwen3-14b",
+        messages=[
+            {"role": "system", "content": "You are a helpful assistant."},
+            {"role": "user", "content": "Hello "}
+        ],
+        stream=True
+    )
+    for chunk in resp:
+        print(chunk, end="", flush=True)