unique_toolkit 1.62.2__tar.gz → 1.62.4__tar.gz

{unique_toolkit-1.62.2 → unique_toolkit-1.62.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: unique-toolkit
-Version: 1.62.2
+Version: 1.62.4
 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>

{unique_toolkit-1.62.2 → unique_toolkit-1.62.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "unique_toolkit"
-version = "1.62.2"
+version = "1.62.4"
 description = ""
 readme = "README.md"
 requires-python = ">=3.12"

{unique_toolkit-1.62.2 → unique_toolkit-1.62.4}/unique_toolkit/agentic/history_manager/history_manager.py

@@ -10,7 +10,10 @@ from unique_toolkit._common.feature_flags.schema import (
 )
 from unique_toolkit._common.validators import LMI
 from unique_toolkit.agentic.history_manager.loop_token_reducer import LoopTokenReducer
-from unique_toolkit.agentic.history_manager.utils import transform_chunks_to_string
+from unique_toolkit.agentic.history_manager.utils import (
+    serialize_tool_content_json,
+    transform_chunks_to_string,
+)
 from unique_toolkit.agentic.reference_manager.reference_manager import ReferenceManager
 from unique_toolkit.agentic.tools.config import get_configuration_dict
 from unique_toolkit.agentic.tools.schemas import ToolCallResponse
@@ -347,10 +350,10 @@ class HistoryManager:
 def _strip_uncited_sources_from_content(content: str, cited: set[int]) -> str:
     """Filter a tool response content string to only keep cited source items.
 
-    The content is expected to be a JSON array of
-    ``{"source_number": N, "content": "..."}`` dicts.  Items whose
-    ``source_number`` is not in *cited* are removed.  If the content is
-    not in the expected JSON format it is returned unchanged.
+    The content is expected to be a JSON array of dicts containing at least
+    a "source_number" key.  Items whose "source_number" is not in *cited*
+    are removed.  If the content is not valid JSON or not in the expected
+    format it is returned unchanged.
     """
     try:
         data = json.loads(content)
@@ -366,4 +369,4 @@ def _strip_uncited_sources_from_content(content: str, cited: set[int]) -> str:
         return content
 
     filtered = [item for item in data if item.get("source_number") in cited]
-    return json.dumps(filtered)
+    return serialize_tool_content_json(filtered)

{unique_toolkit-1.62.2 → unique_toolkit-1.62.4}/unique_toolkit/agentic/history_manager/loop_token_reducer.py

@@ -12,6 +12,7 @@ from unique_toolkit.agentic.history_manager.history_construction_with_contents i
     FileContentSerialization,
     get_full_history_with_contents_and_tool_calls,
 )
+from unique_toolkit.agentic.history_manager.utils import serialize_tool_content_json
 from unique_toolkit.agentic.reference_manager.reference_manager import ReferenceManager
 from unique_toolkit.app.schemas import ChatEvent
 from unique_toolkit.app.unique_settings import UniqueSettings
@@ -651,7 +652,7 @@ class LoopTokenReducer:
             else:
                 raise ValueError(f"Unexpected content type: {type(message.content)}")
 
-            content = json.dumps(
+            content = serialize_tool_content_json(
                 {
                     "source_number": source_offset,
                     "content": content_dict.get("content"),
@@ -685,13 +686,14 @@ class LoopTokenReducer:
         sources = [
             {
                 "source_number": source_offset + i,
+                "content_id": chunk.id,
                 "content": chunk.text,
             }
             for i, chunk in enumerate(content_chunks)
         ]
 
         return LanguageModelToolMessage(
-            content=json.dumps(sources),
+            content=serialize_tool_content_json(sources),
             tool_call_id=message.tool_call_id,
             name=message.name,
         )

{unique_toolkit-1.62.2 → unique_toolkit-1.62.4}/unique_toolkit/agentic/history_manager/utils.py

@@ -14,6 +14,11 @@ from unique_toolkit.language_model.schemas import (
 logger = logging.getLogger(__name__)
 
 
+def serialize_tool_content_json(value: Any) -> str:
+    """Serialize tool-role JSON text while preserving readable Unicode."""
+    return json.dumps(value, ensure_ascii=False)
+
+
 def convert_tool_interactions_to_content_messages(
     loop_history: list[LanguageModelMessage],
 ) -> list[LanguageModelMessage]:
@@ -75,11 +80,12 @@ def transform_chunks_to_string(
     sources: list[dict[str, Any]] = [
         {
             "source_number": max_source_number + i,
+            "content_id": chunk.id,
             "content": chunk.text,
         }
         for i, chunk in enumerate(content_chunks)
     ]
-    return json.dumps(sources), sources
+    return serialize_tool_content_json(sources), sources
 
 
 def load_sources_from_string(