logging fix, add thinking mode

README.md

@@ -86,16 +86,13 @@ https://ollama.com/download
 ### 5. **Sťahovanie modelov**
 
 ```bash
-ollama pull qwen3-next:80b-cloud
+ollama pull qwen3.5:cloud
 ```
 
 ```bash
 ollama pull gpt-oss:20b-cloud
 ```
 
-```bash
-ollama pull deepseek-v3.1:671b-cloud
-```
 
 ### **Inštalácia cez DOCKER**
 

api/fetch_api_data.py

@@ -1,11 +1,32 @@
 import httpx
 import logging
+import json
 from pydantic import BaseModel
 from cachetools import TTLCache
+from typing import Callable
 from tenacity import retry, stop_after_attempt, wait_exponential
 
 from api.config import HTTP_TIMEOUT, HTTP_MAX_CONNECTIONS, HTTP_MAX_KEEPALIVE, CACHE_TTL, CACHE_MAX_SIZE
 
+logger = logging.getLogger(__name__)
+
+logger.handlers.clear()
+logger.setLevel(logging.INFO)
+
+handler = logging.StreamHandler()
+handler.setFormatter(logging.Formatter(
+    fmt='%(asctime)s - %(levelname)s - %(message)s',
+    datefmt='%H:%M:%S'
+))
+logger.addHandler(handler)
+logger.propagate = False
+
+httpx_logger = logging.getLogger("httpx")
+httpx_logger.handlers.clear()
+httpx_logger.setLevel(logging.INFO)
+httpx_logger.addHandler(handler)
+httpx_logger.propagate = False
+
 _cache = TTLCache(maxsize=CACHE_MAX_SIZE, ttl=CACHE_TTL)
 
 _client = httpx.AsyncClient(
@@ -14,12 +35,6 @@ _client = httpx.AsyncClient(
                         max_keepalive_connections=HTTP_MAX_KEEPALIVE),
 )
 
-logging.basicConfig(
-    level=logging.INFO,
-    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
-)
-logger = logging.getLogger(__name__)
-
 def docstring_from_model(model: type[BaseModel]):
     def decorator(func):
         if func.__doc__:
@@ -32,34 +47,45 @@ def docstring_from_model(model: type[BaseModel]):
         return func
     return decorator
 
+_log_callback: Callable[[str], None] | None = None
+
+def set_log_callback(cb: Callable[[str], None] | None):
+    global _log_callback
+    _log_callback = cb
+
+def _log(msg: str):
+    logger.info(msg)
+    if _log_callback is not None:
+        _log_callback(msg)
+
 @retry(stop=stop_after_attempt(3), wait=wait_exponential(multiplier=1, min=1, max=5))
 async def fetch_api_data(icon: str, url: str, params: dict, remove_keys: list = None) -> dict:
     try:
         cache_key = f"{url}:{sorted(params.items())}"
         if cache_key in _cache:
-            logger.info(f"\n💾 {icon}Cache hit: {cache_key}\n")
+            _log(f"💾 {icon}  Cache hit")
             return _cache[cache_key]
 
-        logger.info(f"\n🔨 {icon} Input parameters: {params}\n")
+        _log(f"🔨 {icon} Input parameters:\n{json.dumps(params, ensure_ascii=False, indent=2)}")
         response = await _client.get(url, params=params)
         response.raise_for_status()
-        logger.debug(f"\n🖇️ {icon} Request URL: {response.url}\n")
+        _log(f"🖇️ {icon}  Request URL: {response.url}")
 
         data = response.json()
         if remove_keys and isinstance(data, dict):
             for key in remove_keys:
                 data.pop(key, None)
-        logger.info(f"\n✅ {icon} Success: {url}")
 
         _cache[cache_key] = data
+        _log(f"✅ {icon}  Success: {url}")
         return data
 
     except httpx.HTTPStatusError as e:
-        logger.error(f"\n{icon}HTTP error: {e.response.status_code} - {e.response.text}\n")
+        _log(f"❌  {icon}HTTP error: {e.response.status_code} - {e.response.text}")
         return {"error": "http_error", "status_code": e.response.status_code}
     except httpx.RequestError as e:
-        logger.error(f"\n{icon}Request error: {str(e)}\n")
+        _log(f"❌  {icon}Request error: {str(e)}")
         return {"error": "request_error", "status_code": str(e)}
     except Exception as e:
-        logger.critical(f"\n{icon}Unexpected error: {str(e)}\n", exc_info=True)
+        _log(f"❌  {icon}Unexpected error: {str(e)}")
         return {"error": "unexpected_error", "status_code": str(e)}

app.py

@@ -6,6 +6,8 @@ from core.config import DEFAULT_MODEL, MAX_HISTORY
 from core.init_agent import assistant_agent
 from core.stream_response import stream_response
 
+from api.fetch_api_data import set_log_callback
+
 STARTERS = [
     ("What legal data can the agent find?","magnifying_glass"),
     ("What is the agent not allowed to do or use?","ban"),
@@ -60,10 +62,22 @@ async def main(message: cl.Message):
     if len(history) > MAX_HISTORY:
         history = history[-MAX_HISTORY:]
 
-    msg = cl.Message(content="")
+    async with cl.Step(name="🔍 Fetching data...") as step:
-    async for token in stream_response(agent, history):
+        log_lines = []
-        await msg.stream_token(token)
+
-    await msg.update()
+        def on_log(line: str):
+            log_lines.append(line)
+            step.output = "\n".join(log_lines)
+
+        set_log_callback(on_log)
+
+        msg = cl.Message(content="")
+        async for token in stream_response(agent, history):
+            await msg.stream_token(token)
+        await msg.update()
+
+        set_log_callback(None)
+        await step.update()
 
     if msg.content:
         history.append({"role": "assistant", "content": msg.content})

core/stream_response.py

@@ -10,4 +10,5 @@ async def stream_response(agent: Agent, prompt: list[dict] | str) -> AsyncGenera
             if event.type == "raw_response_event" and isinstance(event.data, ResponseTextDeltaEvent):
                 yield event.data.delta # <-- sends the next piece of response text
     except Exception as e:
-        yield f"⚠️🖨️ Error: {e}"
+        yield f"⚠️🖨 Error: {e}"
+

public/styles.css

@@ -30,7 +30,6 @@
   text-align: left;
 }
 
-
 #starters button img {
   margin-bottom: 6px;
 }
@@ -52,4 +51,4 @@
     align-items: center !important;
     gap: 8px !important;
     min-height: 1rem;
-}
+}