unique-toolkit 2026.20.0.dev24__tar.gz → 2026.20.0.dev26__tar.gz

{unique_toolkit-2026.20.0.dev24 → unique_toolkit-2026.20.0.dev26}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: unique-toolkit
-Version: 2026.20.0.dev24
+Version: 2026.20.0.dev26
 Summary: 
 Author: Cedric Klinkert, Martin Fadler, Sadique Sheik, Fabian Schläpfer, Pascal Hauri, Ejiroghene Oki-Peter, Rami Azouz, Ricardo Visini, Sanzio Monti
 Author-email: Cedric Klinkert <cedric.klinkert@unique.ch>, Martin Fadler <martin.fadler@unique.ch>, Sadique Sheik <sadique@unique.ch>, Fabian Schläpfer <fabian@unique.ch>, Pascal Hauri <pascal@unique.ch>, Ejiroghene Oki-Peter <oki-peter.ejiroghene@unique.ai>, Rami Azouz <rami.ext@unique.ch>, Ricardo Visini <ricardo.visini@unique.ch>, Sanzio Monti <sanzio.monti@unique.ai>
@@ -16,7 +16,7 @@ Requires-Dist: sseclient>=0.0.27,<0.0.28
 Requires-Dist: openai>=1.105.0,<3
 Requires-Dist: platformdirs>=4.0.0,<5
 Requires-Dist: pillow>=12.1.1
-Requires-Dist: unique-sdk>=2026.20.0.dev21,<2026.20.0rc0
+Requires-Dist: unique-sdk>=2026.20.0.dev22,<2026.20.0rc0
 Requires-Dist: jambo>=0.1.2,<0.2
 Requires-Dist: jinja2>=3.1.6,<4
 Requires-Dist: docxtpl>=0.20.1,<0.21

{unique_toolkit-2026.20.0.dev24 → unique_toolkit-2026.20.0.dev26}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "unique_toolkit"
-version = "2026.20.0.dev24"
+version = "2026.20.0.dev26"
 description = ""
 readme = "README.md"
 requires-python = ">=3.12"
@@ -28,7 +28,7 @@ dependencies = [
     "openai>=1.105.0,<3",
     "platformdirs>=4.0.0,<5",
     "pillow>=12.1.1",
-    "unique-sdk>=2026.20.0.dev21,<2026.20.0rc0",
+    "unique-sdk>=2026.20.0.dev22,<2026.20.0rc0",
     "jambo>=0.1.2,<0.2",
     "jinja2>=3.1.6,<4",
     "docxtpl>=0.20.1,<0.21",

{unique_toolkit-2026.20.0.dev24 → unique_toolkit-2026.20.0.dev26}/unique_toolkit/_common/config_checker/validator.py

@@ -288,20 +288,20 @@ class ConfigValidator:
             # Get the old value from original JSON
             old_value = None
             try:
-                current: Any = original_json
+                node: Any = original_json
                 for key in loc:
-                    if isinstance(current, dict) and isinstance(key, str):
-                        current = current.get(key)
-                    elif isinstance(current, list) and isinstance(key, int):
+                    if isinstance(node, dict) and isinstance(key, str):
+                        node = node.get(key)
+                    elif isinstance(node, list) and isinstance(key, int):
                         try:
-                            current = current[key]
+                            node = node[key]
                         except (IndexError, TypeError):
-                            current = None
+                            node = None
                             break
                     else:
-                        current = None
+                        node = None
                         break
-                old_value = current
+                old_value = node
             except Exception:
                 pass
 

{unique_toolkit-2026.20.0.dev24 → unique_toolkit-2026.20.0.dev26}/unique_toolkit/agentic/message_log_manager/service.py

@@ -4,9 +4,9 @@ Logger Service for Unique Agentic tools.
 Target of the method is to extend the step tracking on all levels of the tool.
 """
 
-from collections import defaultdict
 from logging import getLogger
 
+from unique_toolkit.agentic.message_log_order import next_message_log_order
 from unique_toolkit.chat.schemas import (
     MessageLog,
     MessageLogDetails,
@@ -18,10 +18,6 @@ from unique_toolkit.content.schemas import ContentReference
 
 _LOGGER = getLogger(__name__)
 
-# Per-request counters for message log ordering - keyed by message_id
-# This is a mandatory global variable since we have in the system a bug which makes it impossible to use it as a proper class variable.
-_request_counters = defaultdict(int)
-
 
 class MessageStepLogger:
     """
@@ -55,8 +51,7 @@ class MessageStepLogger:
             Next order number for this specific request
         """
 
-        _request_counters[message_id] += 1
-        return _request_counters[message_id]
+        return next_message_log_order(message_id=message_id)
 
     def create_message_log_entry(
         self,

unique_toolkit-2026.20.0.dev26/unique_toolkit/agentic/message_log_order.py

@@ -0,0 +1,26 @@
+"""Per-request message log order counters (public API).
+
+``MessageStepLogger`` and :meth:`~unique_toolkit.services.chat_service.ChatService.complete_responses_with_references_async`
+share the same sequence so user-visible steps and rate-limit banners interleave with
+correct ``order`` values.
+
+This module lives outside the ``message_log_manager`` package intentionally: importing
+``message_log_manager`` triggers its ``__init__.py`` which pulls in ``ChatService``,
+creating a circular import. Placing this module one level up breaks the cycle while
+keeping the counters accessible to both callers.
+"""
+
+from collections import defaultdict
+
+_counters: defaultdict[str, int] = defaultdict(int)
+
+
+def next_message_log_order(*, message_id: str) -> int:
+    """Increment and return the next order index for logs belonging to ``message_id``."""
+    _counters[message_id] += 1
+    return _counters[message_id]
+
+
+def reset_message_log_order_counters() -> None:
+    """Clear all order counters (intended for test fixtures)."""
+    _counters.clear()

{unique_toolkit-2026.20.0.dev24 → unique_toolkit-2026.20.0.dev26}/unique_toolkit/services/chat_service.py

@@ -1,4 +1,6 @@
+import asyncio
 import logging
+from contextlib import suppress
 from typing import TYPE_CHECKING, Any, Sequence, overload
 
 import unique_sdk
@@ -18,6 +20,7 @@ from typing_extensions import Self, deprecated
 
 from unique_toolkit._common.utils.files import is_file_content, is_image_content
 from unique_toolkit.agentic.feature_flags import feature_flags
+from unique_toolkit.agentic.message_log_order import next_message_log_order
 from unique_toolkit.app.unique_settings import UniqueContext, UniqueSettings
 from unique_toolkit.chat.cancellation import CancellationWatcher
 from unique_toolkit.chat.constants import (
@@ -1740,59 +1743,157 @@ class ChatService(ChatServiceDeprecated):
         reasoning: Reasoning | None = None,
         other_options: dict[str, Any] | None = None,
     ) -> ResponsesLanguageModelStreamResponse:
+        # UN-19878: rate-limit retry UX via message logs. After the streaming refactor
+        # (UN-20053) merges and is adopted, this should be integrated with the shared
+        # streaming layer instead of living on ChatService.
+        # Single-element box so nested callbacks and finally see assignments (pyright).
+        rate_limit_log_box: list[MessageLog | None] = [None]
+        _ticker_task: asyncio.Task[None] | None = None
+
+        async def _cancel_ticker() -> None:
+            nonlocal _ticker_task
+            if _ticker_task is not None and not _ticker_task.done():
+                _ = _ticker_task.cancel()
+                with suppress(asyncio.CancelledError):
+                    await _ticker_task
+            _ticker_task = None
+
         async def _on_rate_limit_retry(attempt: int, wait_secs: float) -> None:
+            nonlocal _ticker_task
+
             if not rate_limit_retry_config.log_message_on_retry:
                 return
             if not feature_flags.enable_new_answers_ui_un_14411.is_enabled(
                 self._company_id
             ):
                 return
-            from unique_toolkit.agentic.message_log_manager.service import (
-                _request_counters,
-            )
+
+            await _cancel_ticker()
 
             msg_id = self._assistant_message_id
-            _request_counters[msg_id] += 1
-            order = _request_counters[msg_id]
+            # Use round(), not int() and not :.0f in isolation: tenacity adds jitter, so
+            # wait_secs is often non-integer. Rounding once keeps the first banner and the
+            # countdown start value identical (e.g. avoid "31s" then 29, 28…).
+            wait_display_seconds = max(0, round(float(wait_secs)))
+            initial = (
+                f"Rate limit reached; retrying in {wait_display_seconds}s "
+                f"(attempt {attempt})"
+            )
+
             try:
-                await self.create_message_log_async(
-                    message_id=msg_id,
-                    text=f"Rate limit reached; retrying in {wait_secs:.0f}s (attempt {attempt})",
-                    status=MessageLogStatus.RUNNING,
-                    order=order,
-                )
+                if rate_limit_log_box[0] is None:
+                    order = next_message_log_order(message_id=msg_id)
+                    rate_limit_log_box[0] = await self.create_message_log_async(
+                        message_id=msg_id,
+                        text=initial,
+                        status=MessageLogStatus.RUNNING,
+                        order=order,
+                    )
+                else:
+                    existing = rate_limit_log_box[0]
+                    if existing.message_log_id is None:
+                        return
+                    rate_limit_log_box[0] = await self.update_message_log_async(
+                        message_log_id=existing.message_log_id,
+                        order=existing.order,
+                        text=initial,
+                        status=MessageLogStatus.RUNNING,
+                    )
             except Exception:
                 _LOGGER.warning(
                     "Failed to write rate-limit retry message log",
                     exc_info=True,
                 )
+                return
 
-        return await stream_responses_with_references_async(
-            company_id=self._company_id,
-            user_id=self._user_id,
-            assistant_message_id=self._assistant_message_id,
-            user_message_id=self._user_message_id,
-            chat_id=self._chat_id,
-            assistant_id=self._assistant_id,
-            model_name=model_name,
-            messages=messages,
-            content_chunks=content_chunks,
-            tools=tools,
-            temperature=temperature,
-            debug_info=debug_info,
-            start_text=start_text,
-            include=include,
-            instructions=instructions,
-            max_output_tokens=max_output_tokens,
-            metadata=metadata,
-            parallel_tool_calls=parallel_tool_calls,
-            text=text,
-            tool_choice=tool_choice,
-            top_p=top_p,
-            reasoning=reasoning,
-            other_options=other_options,
-            on_rate_limit_retry=_on_rate_limit_retry,
-        )
+            active = rate_limit_log_box[0]
+            if active is None or active.message_log_id is None:  # pyright: ignore[reportUnnecessaryComparison]
+                return
+
+            log_id = active.message_log_id
+            log_order = active.order
+
+            async def _countdown() -> None:
+                remaining = wait_display_seconds
+                while remaining > 0:
+                    await asyncio.sleep(1)
+                    remaining -= 1
+                    if remaining <= 0:
+                        break
+                    try:
+                        _ = await self.update_message_log_async(
+                            message_log_id=log_id,
+                            order=log_order,
+                            text=(
+                                f"Rate limit reached; retrying in {remaining}s "
+                                f"(attempt {attempt})"
+                            ),
+                            status=MessageLogStatus.RUNNING,
+                        )
+                    except Exception:
+                        _LOGGER.debug(
+                            "Rate limit countdown message log update failed",
+                            exc_info=True,
+                        )
+                        break
+
+            _ticker_task = asyncio.create_task(_countdown())
+
+        stream_error: BaseException | None = None
+        try:
+            return await stream_responses_with_references_async(
+                company_id=self._company_id,
+                user_id=self._user_id,
+                assistant_message_id=self._assistant_message_id,
+                user_message_id=self._user_message_id,
+                chat_id=self._chat_id,
+                assistant_id=self._assistant_id,
+                model_name=model_name,
+                messages=messages,
+                content_chunks=content_chunks,
+                tools=tools,
+                temperature=temperature,
+                debug_info=debug_info,
+                start_text=start_text,
+                include=include,
+                instructions=instructions,
+                max_output_tokens=max_output_tokens,
+                metadata=metadata,
+                parallel_tool_calls=parallel_tool_calls,
+                text=text,
+                tool_choice=tool_choice,
+                top_p=top_p,
+                reasoning=reasoning,
+                other_options=other_options,
+                on_rate_limit_retry=_on_rate_limit_retry,
+            )
+        except BaseException as exc:
+            stream_error = exc
+            raise
+        finally:
+            await _cancel_ticker()
+            fin = rate_limit_log_box[0]
+            if fin is not None and fin.message_log_id is not None:
+                try:
+                    if stream_error is None:
+                        _ = await self.update_message_log_async(
+                            message_log_id=fin.message_log_id,
+                            order=fin.order,
+                            text="Rate limit resolved; resuming.",
+                            status=MessageLogStatus.COMPLETED,
+                        )
+                    else:
+                        _ = await self.update_message_log_async(
+                            message_log_id=fin.message_log_id,
+                            order=fin.order,
+                            text="Rate limit retries exhausted.",
+                            status=MessageLogStatus.FAILED,
+                        )
+                except Exception:
+                    _LOGGER.warning(
+                        "Failed to finalize rate-limit message log",
+                        exc_info=True,
+                    )
 
     # Chat Content Methods
     ############################################################################