bug/AP-25110-implement-error-handling-option #5

ya-hn · 2025-12-16T15:02:50Z

Also reverts previous changes made to conversation histories by storing error messages separately from the agent memory.

ya-hn · 2025-12-16T15:15:00Z

src/agents/base.py

+                if self.error_handling == ErrorHandlingMode.FAIL.name:
+                    ctx.set_warning(str(e))
+                else:
+                    errors.append(str(e))


With this, validation errors for mode "FAIL" are displayed as warnings and validation errors for mode "ERROR COLUMN" are included as errors in the error column and not shown as warnings. This makes sense to me but we could also show these as warning in both modes.

src/agents/base.py

ya-hn · 2025-12-16T15:26:21Z

src/agents/base.py

        )

    def _fill_memory_with_messages(
        self, agent, view_data, data_registry, tool_converter


I think we could use

from __future__ import annotations from typing import TYPE_CHECKING

to have typehints without performance issues due to the imports. I didn't do it yet because I was not sure if there are any drawbacks.

ya-hn · 2025-12-16T15:28:43Z

src/agents/base.py

            "configurable": {"thread_id": "1"},
        }
        previous_messages = []
+        error_messages = {}


I fill these by passing them into the functions below. I could see how this is bad style because it is not obvious outside of the functions, but I am not sure about it.

ya-hn · 2025-12-16T15:36:39Z

src/agents/base.py

+                history_df = history_table[self.conversation_column].to_pandas()
+
+            for msg in history_df[self.conversation_column]:
+                if msg:


This changes the behavior of the Agent Prompter slightly because now missing values are simply ignored (in both error handling modes). But I don't think this creates any problems. And it makes sense to me because a conversation table with error column might contain missing values in the conversation column but should still be a valid input for an Agent Prompter that uses mode "FAIL".

AtR1an

I didn't review all the code but from the parts I saw it's clear to me that we reached the point where we need to refactor in order to manage the complexity and improve readability. We probably should've done that already but now I feel that it's crucial to keep this maintainable and extendable.

AtR1an · 2025-12-18T07:53:48Z

src/agents/base.py

        default_value="Conversation",
    ).rule(knext.DialogContextCondition(has_conversation_table), knext.Effect.HIDE)

+    error_handling = knext.EnumParameter(


Would it be possible to extract the error handling related parameters in to a parameter group?

AtR1an · 2025-12-18T08:06:22Z

src/agents/base.py

+                else:
+                    return knext.Schema.from_columns(
+                        [
+                            knext.Column(_message_type(), self.conversation_column),


the line for the conversation column exists multiple times. That makes the logic look way harder than it actually is. There is always a conversation column and if we want to output the error column, there is also an error column.

AtR1an · 2025-12-18T08:10:51Z

src/agents/base.py

-                f"An error occurred while executing the agent: {e}"
-            )
+            error_message = f"An error occurred while executing the agent: {e}"
+            if self.error_handling == ErrorHandlingMode.FAIL.name:


All of the if self.error_handling switches could be abstracted into a separate class that encapsulates the error handling logic and makes the node's logic simpler. It could also be easily unit tested.

AtR1an · 2025-12-18T08:16:12Z

src/agents/base.py

+        output_column,
+        messages,
+        history_table,
+        history_df,


Aren't history_table and history_df referring to the same thing?

AtR1an · 2025-12-18T08:18:52Z

src/agents/base.py

-                ]
-            }
+        conversation_table = self._create_conversation_table(
+            output_column_name,


The ´output_column_name´ can be constructed inside self._create_conversation_table.

AtR1an · 2025-12-18T08:20:22Z

src/agents/base.py

+            history_table,
+            history_df,
+            errors,
+            num_messages_before,


Isn't that the number of rows of history_df / history_table?

Copilot

Pull request overview

This PR implements a new error handling option for the Agent Prompter node that stores error messages separately from the agent's conversation history. Instead of appending errors to the agent's memory (which could affect subsequent interactions), errors are now tracked in a parallel structure and can be optionally output in a separate error column.

Changes:

Added configurable error handling modes (Fail vs. Error column) with support for continuing existing error columns from input tables
Refactored agent execution to use a new Agent class with cleaner separation of concerns
Introduced AgentPrompterConversation to track messages and errors separately while maintaining backward compatibility

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
src/agents/base_deprecated.py	Removed calls that were appending error messages to agent memory
src/agents/base.py	Added error handling settings, refactored agent execution logic, and introduced conversation/toolset abstractions
src/agents/_parameters.py	Moved `RecursionLimitMode` enum and added new `ErrorHandlingMode` enum
src/agents/_error_handler.py	New file implementing centralized error handling logic for agent execution
src/agents/_data_service.py	Refactored to use new agent architecture with frontend/backend conversation separation
src/agents/_agent.py	Introduced protocol-based abstractions and new `Agent` class for cleaner agent execution

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-12T13:28:11Z

src/agents/base.py

+                f"Column {self.conversation_column} not found in the conversation history table."
+            )
+        if (
+            self.errors == ErrorHandlingMode.COLUMN.name


This comparison is incorrect. self.errors is an AgentPrompterErrorSettings object, not a string. It should be self.errors.error_handling == ErrorHandlingMode.COLUMN.name to match the pattern used elsewhere in the code.

Suggested change

self.errors == ErrorHandlingMode.COLUMN.name

self.errors.error_handling == ErrorHandlingMode.COLUMN.name

Copilot · 2026-01-12T13:28:12Z

src/agents/base.py

+            self.errors == ErrorHandlingMode.COLUMN.name
+            and self.errors.use_existing_error_column
+        ):
+            if self.error_column not in history_table.column_names:


This should be self.errors.error_column instead of self.error_column to access the error column from the settings object.

Suggested change

if self.error_column not in history_table.column_names:

if self.errors.error_column not in history_table.column_names:

Copilot · 2026-01-12T13:28:12Z

src/agents/base.py

+            columns = [
+                util.OutputColumn(
+                    self.conversation_column_name,
+                    _message_type,


This should be message_type (the variable) instead of _message_type (the function). The variable message_type is defined on line 1215.

Suggested change

_message_type,

message_type,

Copilot · 2026-01-12T13:28:13Z

src/agents/base.py


-
-# endregion
+# # endregion


Extra hash symbol in the comment. Should be # endregion to match the standard region comment format.

Copilot

Pull request overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated 6 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

src/agents/base.py

Copilot · 2026-01-12T17:05:19Z

src/agents/base.py

+        if not isinstance(error, Exception):
+            raise error
+
+        if self._error_handling == "FAIL":


Using string literal 'FAIL' instead of ErrorHandlingMode.FAIL.name for comparison is inconsistent with the enum-based approach used elsewhere (e.g., line 657). Use ErrorHandlingMode.FAIL.name for type safety and consistency.

Suggested change

if self._error_handling == "FAIL":

if self._error_handling == ErrorHandlingMode.FAIL.name:

Copilot · 2026-01-12T17:05:20Z

src/agents/_error_handler.py

+            error_message = f"""Recursion limit of {self._recursion_limit} reached. 
+                You can increase the limit by setting the `recursion_limit` parameter to a higher value."""


The multi-line string contains unnecessary leading whitespace on line 103 that will appear in the error message. Use textwrap.dedent() or adjust the string formatting to remove this whitespace.

Suggested change

error_message = f"""Recursion limit of {self._recursion_limit} reached.

You can increase the limit by setting the `recursion_limit` parameter to a higher value."""

error_message = (

f"Recursion limit of {self._recursion_limit} reached.\n"

"You can increase the limit by setting the `recursion_limit` parameter to a higher value."

)

Copilot · 2026-01-12T17:05:20Z

src/agents/base.py

+                        raise RuntimeError(
+                            "Conversation table contains row with both message and error."
+                        )


The error message should be more descriptive about which row caused the issue. Consider including the row index: 'Conversation table contains row {i} with both message and error.'

It should also include the ID of the row

Copilot · 2026-01-12T17:05:20Z

src/agents/base.py

+                    elif has_err:
+                        conversation.append_error(Exception(err))
+                    else:
+                        raise RuntimeError("Conversation table contains empty row.")


The error message should indicate which row is empty. Consider including the row index: 'Conversation table contains empty row at index {i}.'

It should also include the ID of the row.

src/agents/_agent.py

Reverts previous changes by storing error messages separately from the agent memory.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

AtR1an · 2026-01-14T13:53:01Z

src/agents/base.py


+from langchain_core.messages import BaseMessage
+
 _logger = logging.getLogger(__name__)


No longer used

AtR1an · 2026-01-14T14:04:23Z

src/agents/base.py

-                # Sanitize tool names so they match the current sanitized mapping
-                lc_msg = tool_converter.sanitize_tool_names(lc_msg)
-                messages.append(lc_msg)
+            self._check_for_columns(history_table)


This check shouldn't be necessary because we should already check it in configure and the history table schema does not change between the last configure and execute (if it does it's a bug in core).

AtR1an · 2026-01-14T14:13:28Z

src/agents/base.py


-        if data_registry.has_data or tool_converter.has_data_tools:
-            messages.append(data_registry.create_data_message())
+        agent_ctx = AgentPrompterContext(ctx)


The AgentPrompterContext is only used by the AgentPrompterConversation created on the next line and the only thing is does is check for cancellation on the ExecutionContext. We can just as well provide AgentPrompterConversation with the ExecutionContext directly.

AtR1an · 2026-01-14T14:19:37Z

src/agents/base.py

+        if self.developer_message:
+            conversation.append_messages(SystemMessage(self.developer_message))
+
+        if history_table is not None:


The body of this if should be a dedicated method to improve readability.

AtR1an · 2026-01-14T14:22:32Z

src/agents/base.py

+                    elif has_err:
+                        conversation.append_error(Exception(err))
+                    else:
+                        raise RuntimeError("Conversation table contains empty row.")


It should also include the ID of the row.

src/agents/base.py

AtR1an · 2026-01-14T15:20:34Z

src/agents/_data_service.py

        self._ctx = ctx
-        self._agent_graph = agent_graph
+
+        self._chat_model = chat_model


Not used outside of __init__.

AtR1an · 2026-01-14T15:25:13Z

src/agents/_data_service.py

+
+    def create_output_table(
+        self,
+        tool_converter,


Not needed, just use self._tool_converter

AtR1an · 2026-01-14T15:26:15Z

src/agents/_data_service.py

+
+        self._chat_model = chat_model
+        self._conversation = FrontendConversation(
+            conversation, tool_converter, AgentWidgetContext(self._check_canceled)


No need to wrap into AgentWidgetContext and consequently AgentWidgetContext is obsolete.

AtR1an · 2026-01-14T15:28:23Z

src/agents/_data_service.py

+        self._agent = Agent(self._conversation, self._chat_model, toolset, self._config)
+
        self._data_registry = data_registry
        self._tool_converter = tool_converter


we only use this reference to feed it to the conversation that already has a reference to it.

Copilot

Pull request overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated 3 comments.

Comments suppressed due to low confidence (1)

src/agents/_data_service.py:1

The comment should clarify why error_handling is set to None. Consider: '# error_handling=None allows errors to be appended without raising exceptions during conversation reconstruction'

# -*- coding: utf-8 -*-

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-14T15:35:24Z

src/agents/base.py

+                        raise RuntimeError(
+                            "Conversation table contains row with both message and error."
+                        )


The error message should indicate that this is a data validation issue with the input table and suggest how to fix it. Consider: 'Invalid conversation table: A row cannot contain both a message and an error. Please ensure each row contains either a message or an error, but not both.'

Copilot · 2026-01-14T15:35:24Z

src/agents/base.py

+                    elif has_err:
+                        conversation.append_error(Exception(err))
+                    else:
+                        raise RuntimeError("Conversation table contains empty row.")


The error message should clarify that empty rows in the conversation table are not allowed and suggest remediation. Consider: 'Invalid conversation table: Found a row with no message and no error. Please remove empty rows from the conversation table.'

Copilot · 2026-01-14T15:35:25Z

src/agents/_error_handler.py

+            error_message = f"""Recursion limit of {self._recursion_limit} reached. 
+                You can increase the limit by setting the `recursion_limit` parameter to a higher value."""


The multi-line f-string contains unnecessary indentation that will appear in the error message. Use a single-line string or dedent the continuation: f'Recursion limit of {self._recursion_limit} reached. You can increase the limit by setting the recursion_limit parameter to a higher value.'

AtR1an · 2026-01-14T15:36:04Z

@copilot implement the changes requested during the code review.

Copilot · 2026-01-14T15:36:17Z

@AtR1an I've opened a new pull request, #16, to work on those changes. Once the pull request is ready, I'll request review from you.

- Remove unused BaseMessage import from base.py - Remove unnecessary column check in execute (already in configure) - Pass ExecutionContext directly instead of AgentPrompterContext wrapper - Move conversation construction into _get_conversation method - Extract history table processing into dedicated method with row IDs in error messages - Move validation logic into AgentPrompterConversation.validate_final_message - Remove AgentPrompterContext class (no longer needed) - Simplify create_output_table by removing with_error_column parameter - Remove AgentWidgetContext wrapper in _data_service.py - Simplify FrontendConversation.create_output_table to use self._tool_converter - Remove _append_messages wrapper in _agent.py (inline wrapping) Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

The error handler now has a single handle_error method that accepts any exception and makes the differentiation itself, simplifying the calling code. Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

- Move BaseMessage import to module level to avoid repeated import overhead - Remove inline imports from frequently called methods - Extract duplicated append_msg logic into shared _append_sanitized_message helper - Fix inconsistent whitespace in import statements Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

- Fix row index access using iloc[idx].name instead of index[idx] - Move IterationLimitError import to top of _error_handler.py module Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

- Use history_df.index[idx] to correctly get row index - Remove extra trailing blank lines in _agent.py for consistency Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

… table creation - Move validation logic into append_messages to validate each AI message as it's added - Simplify create_output_table with to_knime_message_or_none helper function - Remove validate_final_message method (now inline in append_messages) - Properly handle FAIL mode with ctx.set_warning for validation errors Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

Instead of passing validate_ai_message as a parameter through _get_conversation, the AgentPrompterConversation class now imports and calls it directly from the _agent module. This simplifies the API since it's always the same function. Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

…DataService

Copilot

Pull request overview

Copilot reviewed 6 out of 6 changed files in this pull request and generated 5 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-01-16T14:27:00Z

src/agents/base.py

+                messages = messages[1:]
+            result_df = pd.DataFrame(
+                {
+                    output_column_name: [


The AgentPrompterToolset class is missing a docstring. Add a class docstring explaining its purpose as a wrapper for managing tool execution.

Suggested change

output_column_name: [

class AgentPrompterToolset:

"""Wrapper around a collection of tools used by an agent prompter.

The toolset indexes tools by name and exposes them via the :attr:`tools`

property. The :meth:`execute` method dispatches a sequence of tool call

requests to the corresponding tools and returns their results as

``ToolMessage`` instances that can be consumed by the conversation flow.

"""

Copilot · 2026-01-16T14:27:01Z

src/agents/base.py

+    user_message: Optional[str] = None,
+    interactive: bool = False,
+) -> "AgentPrompterConversation":
+    """


The AgentPrompterConversation class is missing a docstring. Add a class docstring explaining its purpose in managing messages and errors for the agent prompter.

Suggested change

"""

class AgentPrompterConversation:

"""Container for managing messages and errors produced by the agent prompter.

This class keeps a single ordered sequence of both messages and errors and

exposes methods to append new items and to retrieve only the valid messages.

Error handling behavior is controlled by ``error_handling`` (e.g. FAIL vs.

COLUMN modes defined in :class:`ErrorHandlingMode`). In FAIL mode, errors

are raised immediately, while in COLUMN mode they are stored alongside

messages and can later be written to an error column.

An optional :class:`knime.extension.ExecutionContext` can be provided to:

* Respect cancellation requests and raise :class:`CancelError` when the

execution is canceled while appending messages.

* Emit warnings instead of propagating certain validation errors when

configured to do so.

The public methods :meth:`append_messages`, :meth:`append_error`, and

:meth:`get_messages` are used by the agent prompter node implementation to

collect the final conversation history and any associated errors.

"""

Copilot · 2026-01-16T14:27:01Z

src/agents/_data_service.py

+        ):
+            messages = [AIMessage(RECURSION_CONTINUE_PROMPT)]
+            self._conversation._append_messages(messages)
+        else:


The FrontendConversation class is missing a docstring. Add a class docstring explaining its role in bridging backend conversation state with frontend messaging.

Suggested change

else:

class FrontendConversation:

"""Adapter between the backend conversation state and the frontend message queue.

This class mirrors messages between an `AgentPrompterConversation` backend and

a frontend-facing queue used by the view. It ensures that all messages are

appended to the backend conversation while exposing them via a queue for the

frontend to consume, and applies cancellation checks to sanitize and report

errors when execution is canceled.

"""

Copilot · 2026-01-16T14:27:02Z

src/agents/base.py

+    interactive: bool = False,
+) -> "AgentPrompterConversation":
+    """
+    Consolidated orchestrator to initialize an AgentPrompterConversation.


The parameter error_handling should be type-annotated for clarity. Add type hint to indicate it expects an ErrorHandlingMode name string.

Suggested change

Consolidated orchestrator to initialize an AgentPrompterConversation.

def __init__(self, error_handling: str, ctx: knext.ExecutionContext = None):

Copilot · 2026-01-16T14:27:02Z

src/agents/_data_service.py


-        self._message_queue = queue.Queue()
+        self._message_queue = self._conversation.frontend
        self._thread = None


The TODO comment indicates incomplete implementation. The _is_canceled flag appears to be set but never updated. Either implement the cancellation mechanism or document why this is safe.

Suggested change

self._thread = None

# This flag is initialized here and may be updated asynchronously by the frontend

# or another controller. The backend only reads it via the cancellation callback.

self._is_canceled = False

ya-hn requested a review from a team as a code owner December 16, 2025 15:02

ya-hn commented Dec 16, 2025

View reviewed changes

AtR1an requested changes Dec 18, 2025

View reviewed changes

ya-hn force-pushed the bug/AP-25110-implement-error-handling-option branch from da068e4 to 851b63a Compare January 5, 2026 13:31

Copilot AI review requested due to automatic review settings January 12, 2026 13:27

ya-hn force-pushed the bug/AP-25110-implement-error-handling-option branch from 851b63a to 758895d Compare January 12, 2026 13:27

Copilot AI reviewed Jan 12, 2026

View reviewed changes

ya-hn force-pushed the bug/AP-25110-implement-error-handling-option branch 2 times, most recently from 06a7421 to 59d8ac1 Compare January 12, 2026 17:04

Copilot AI review requested due to automatic review settings January 12, 2026 17:04

Copilot AI reviewed Jan 12, 2026

View reviewed changes

ya-hn added 3 commits January 13, 2026 18:25

AP-25110: Add mode in agent prompter nodes to output error column

1f337cc

Reverts previous changes by storing error messages separately from the agent memory.

AP-25110: Fix recursion limit in Agent Prompter

f7dabf3

AP-25110: Refactor agent loop for Agent Prompter and Agent Chat Widget

1f92141

ya-hn force-pushed the bug/AP-25110-implement-error-handling-option branch from 59d8ac1 to 1f92141 Compare January 13, 2026 17:26

AP-25110: Apply suggestions from code review

388a21c

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Copilot AI review requested due to automatic review settings January 14, 2026 15:34

AtR1an requested changes Jan 14, 2026

View reviewed changes

Copilot AI reviewed Jan 14, 2026

View reviewed changes

Copilot AI mentioned this pull request Jan 14, 2026

Refactor agent execution to simplify context handling and reduce code duplication #16

Merged

18 tasks

Copilot AI and others added 8 commits January 15, 2026 10:13

Fix row index access and move import to module level

661d4d5

- Fix row index access using iloc[idx].name instead of index[idx] - Move IterationLimitError import to top of _error_handler.py module Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

Fix code style: correct row index access and remove extra blank lines

d314e4d

- Use history_df.index[idx] to correctly get row index - Remove extra trailing blank lines in _agent.py for consistency Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

Fix code style: remove trailing whitespace

53f4464

Co-authored-by: AtR1an <12174376+AtR1an@users.noreply.github.com>

ya-hn added 6 commits January 16, 2026 11:04

Remove unused imports

84e095c

Extract AgentPrompter2._construct_outputs method to simplify execute

9e174eb

Move all error handling into AgentPrompterErrorHandler

8c663fc

Remove obsolete method

1954096

Create config for parameters for easier management in AgentChatWidget…

e19e32a

…DataService

Extract conversation creation to reduce code duplication

c7860e4

Copilot AI review requested due to automatic review settings January 16, 2026 14:25

Copilot AI reviewed Jan 16, 2026

View reviewed changes

	self.errors == ErrorHandlingMode.COLUMN.name
	self.errors.error_handling == ErrorHandlingMode.COLUMN.name

	if self.error_column not in history_table.column_names:
	if self.errors.error_column not in history_table.column_names:

	if self._error_handling == "FAIL":
	if self._error_handling == ErrorHandlingMode.FAIL.name:

		error_message = f"""Recursion limit of {self._recursion_limit} reached.
		You can increase the limit by setting the `recursion_limit` parameter to a higher value."""

-            error_message = f"""Recursion limit of {self._recursion_limit} reached.
-                You can increase the limit by setting the `recursion_limit` parameter to a higher value."""
+            error_message = (
+                f"Recursion limit of {self._recursion_limit} reached.\n"
+                "You can increase the limit by setting the `recursion_limit` parameter to a higher value."
+            )


		from langchain_core.messages import BaseMessage

		_logger = logging.getLogger(__name__)

-                    output_column_name: [
+class AgentPrompterToolset:
+    """Wrapper around a collection of tools used by an agent prompter.
+    The toolset indexes tools by name and exposes them via the :attr:`tools`
+    property. The :meth:`execute` method dispatches a sequence of tool call
+    requests to the corresponding tools and returns their results as
+    ``ToolMessage`` instances that can be consumed by the conversation flow.
+    """

-    """
+class AgentPrompterConversation:
+    """Container for managing messages and errors produced by the agent prompter.
+    This class keeps a single ordered sequence of both messages and errors and
+    exposes methods to append new items and to retrieve only the valid messages.
+    Error handling behavior is controlled by ``error_handling`` (e.g. FAIL vs.
+    COLUMN modes defined in :class:`ErrorHandlingMode`). In FAIL mode, errors
+    are raised immediately, while in COLUMN mode they are stored alongside
+    messages and can later be written to an error column.
+    An optional :class:`knime.extension.ExecutionContext` can be provided to:
+    * Respect cancellation requests and raise :class:`CancelError` when the
+      execution is canceled while appending messages.
+    * Emit warnings instead of propagating certain validation errors when
+      configured to do so.
+    The public methods :meth:`append_messages`, :meth:`append_error`, and
+    :meth:`get_messages` are used by the agent prompter node implementation to
+    collect the final conversation history and any associated errors.
+    """

-        else:
+class FrontendConversation:
+    """Adapter between the backend conversation state and the frontend message queue.
+    This class mirrors messages between an `AgentPrompterConversation` backend and
+    a frontend-facing queue used by the view. It ensures that all messages are
+    appended to the backend conversation while exposing them via a queue for the
+    frontend to consume, and applies cancellation checks to sanitize and report
+    errors when execution is canceled.
+    """

	Consolidated orchestrator to initialize an AgentPrompterConversation.
	def __init__(self, error_handling: str, ctx: knext.ExecutionContext = None):

-        self._thread = None
+        # This flag is initialized here and may be updated asynchronously by the frontend
+        # or another controller. The backend only reads it via the cancellation callback.
+        self._is_canceled = False

bug/AP-25110-implement-error-handling-option #5

Are you sure you want to change the base?

bug/AP-25110-implement-error-handling-option #5

Uh oh!

Conversation

ya-hn commented Dec 16, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ya-hn Dec 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AtR1an left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 12, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ya-hn Dec 16, 2025 •

edited

Loading