Merge branch 'main' into feat/rag-2

api/core/workflow/graph_engine/entities/graph_runtime_state.py

@@ -22,7 +22,7 @@ class GraphRuntimeState(BaseModel):
     #
     # Note: Since the type of this field is `dict[str, Any]`, its values may not remain consistent
     # after a serialization and deserialization round trip.
-    outputs: dict[str, Any] = {}
+    outputs: dict[str, Any] = Field(default_factory=dict)
 
     node_run_steps: int = 0
     """node run steps"""

api/core/workflow/nodes/knowledge_retrieval/knowledge_retrieval_node.py

@@ -8,7 +8,7 @@ from typing import TYPE_CHECKING, Any, Optional, cast
 
 from sqlalchemy import Float, and_, func, or_, text
 from sqlalchemy import cast as sqlalchemy_cast
-from sqlalchemy.orm import Session
+from sqlalchemy.orm import sessionmaker
 
 from core.app.app_config.entities import DatasetRetrieveConfigEntity
 from core.app.entities.app_invoke_entities import ModelConfigWithCredentialsEntity
@@ -175,7 +175,7 @@ class KnowledgeRetrievalNode(BaseNode):
             redis_client.zremrangebyscore(key, 0, current_time - 60000)
             request_count = redis_client.zcard(key)
             if request_count > knowledge_rate_limit.limit:
-                with Session(db.engine) as session:
+                with sessionmaker(db.engine).begin() as session:
                     # add ratelimit record
                     rate_limit_log = RateLimitLog(
                         tenant_id=self.tenant_id,
@@ -183,7 +183,6 @@ class KnowledgeRetrievalNode(BaseNode):
                         operation="knowledge",
                     )
                     session.add(rate_limit_log)
-                    session.commit()
                 return NodeRunResult(
                     status=WorkflowNodeExecutionStatus.FAILED,
                     inputs=variables,
@@ -389,6 +388,15 @@ class KnowledgeRetrievalNode(BaseNode):
                                 "segment_id": segment.id,
                                 "retriever_from": "workflow",
                                 "score": record.score or 0.0,
+                                "child_chunks": [
+                                    {
+                                        "id": str(getattr(chunk, "id", "")),
+                                        "content": str(getattr(chunk, "content", "")),
+                                        "position": int(getattr(chunk, "position", 0)),
+                                        "score": float(getattr(chunk, "score", 0.0)),
+                                    }
+                                    for chunk in (record.child_chunks or [])
+                                ],
                                 "segment_hit_count": segment.hit_count,
                                 "segment_word_count": segment.word_count,
                                 "segment_position": segment.position,
@@ -572,7 +580,7 @@ class KnowledgeRetrievalNode(BaseNode):
     def _process_metadata_filter_func(
         self, sequence: int, condition: str, metadata_name: str, value: Optional[Any], filters: list
     ):
-        if value is None:
+        if value is None and condition not in ("empty", "not empty"):
             return
 
         key = f"{metadata_name}_{sequence}"

api/core/workflow/nodes/llm/entities.py

@@ -13,7 +13,7 @@ class ModelConfig(BaseModel):
     provider: str
     name: str
     mode: LLMMode
-    completion_params: dict[str, Any] = {}
+    completion_params: dict[str, Any] = Field(default_factory=dict)
 
 
 class ContextConfig(BaseModel):

api/core/workflow/nodes/loop/loop_node.py

@@ -313,30 +313,31 @@ class LoopNode(BaseNode):
                 and event.node_type == NodeType.LOOP_END
                 and not isinstance(event, NodeRunStreamChunkEvent)
             ):
-                check_break_result = True
+                # Check if variables in break conditions exist and process conditions
+                # Allow loop internal variables to be used in break conditions
+                available_conditions = []
+                for condition in break_conditions:
+                    variable = self.graph_runtime_state.variable_pool.get(condition.variable_selector)
+                    if variable:
+                        available_conditions.append(condition)
+
+                # Process conditions if at least one variable is available
+                if available_conditions:
+                    input_conditions, group_result, check_break_result = condition_processor.process_conditions(
+                        variable_pool=self.graph_runtime_state.variable_pool,
+                        conditions=available_conditions,
+                        operator=logical_operator,
+                    )
+                    if check_break_result:
+                        break
+                else:
+                    check_break_result = True
                 yield self._handle_event_metadata(event=event, iter_run_index=current_index)
                 break
 
             if isinstance(event, NodeRunSucceededEvent):
                 yield self._handle_event_metadata(event=event, iter_run_index=current_index)
 
-                # Check if all variables in break conditions exist
-                exists_variable = False
-                for condition in break_conditions:
-                    if not self.graph_runtime_state.variable_pool.get(condition.variable_selector):
-                        exists_variable = False
-                        break
-                    else:
-                        exists_variable = True
-                if exists_variable:
-                    input_conditions, group_result, check_break_result = condition_processor.process_conditions(
-                        variable_pool=self.graph_runtime_state.variable_pool,
-                        conditions=break_conditions,
-                        operator=logical_operator,
-                    )
-                    if check_break_result:
-                        break
-
             elif isinstance(event, BaseGraphEvent):
                 if isinstance(event, GraphRunFailedEvent):
                     # Loop run failed

api/core/workflow/nodes/parameter_extractor/parameter_extractor_node.py

@@ -1,3 +1,4 @@
+import contextlib
 import json
 import logging
 import uuid
@@ -666,10 +667,8 @@ class ParameterExtractorNode(BaseNode):
             if result[idx] == "{" or result[idx] == "[":
                 json_str = extract_json(result[idx:])
                 if json_str:
-                    try:
+                    with contextlib.suppress(Exception):
                         return cast(dict, json.loads(json_str))
-                    except Exception:
-                        pass
         logger.info("extra error: %s", result)
         return None
 
@@ -686,10 +685,9 @@ class ParameterExtractorNode(BaseNode):
             if result[idx] == "{" or result[idx] == "[":
                 json_str = extract_json(result[idx:])
                 if json_str:
-                    try:
+                    with contextlib.suppress(Exception):
                         return cast(dict, json.loads(json_str))
-                    except Exception:
-                        pass
+
         logger.info("extra error: %s", result)
         return None