Logging and Chat History

The Dify client provides comprehensive logging functionality for both workflow and chat applications.

Quick Examples

# Get workflow logs
workflow_app = client.get_app(app_id="your-workflow-app-id")

# Get paginated logs
logs = workflow_app.get_logs(page=1, limit=10)
for entry in logs.data:
    print(f"Workflow run {entry.workflow_run.id}: {entry.workflow_run.status}")

# Iterate through all logs
for entry in workflow_app.iter_logs(limit=10):
    print(f"Log entry {entry.id} created at {entry.created_at}")

# Get workflow node executions
workflow_run_id = logs.data[0].workflow_run.id
executions = workflow_app.get_node_executions(workflow_run_id)
for node in executions.data:
    print(f"Node {node.title} ({node.node_type}): {node.status}")
    if node.error:
        print(f"Error: {node.error}")

# Get agent/chat conversation logs
agent_app = client.get_app(app_id="your-agent-app-id")
conversations = agent_app.get_logs(page=1, limit=10)
for conversation in conversations.data:
    print(f"Conversation {conversation.id}: {conversation.name}")

# Get messages for a conversation
conversation_id = conversations.data[0].id
messages = agent_app.get_messages(conversation_id=conversation_id, limit=10)
for message in messages.data:
    print(f"Message: {message.query}")
    if message.answer:
        print(f"Answer: {message.answer}")

Chat Session

class Chat

Represents a chat conversation session. Used for tracking and retrieving message history.

__init__(client: DifyBaseClient, app: App, id: str, info: dict = None)

Initialize a new chat session.

Parameters:

client – The Dify client instance
app – The parent application instance
id – Unique identifier for the chat session
info – Optional additional information about the chat session

property messages(max_pages: int = 10) → list

Retrieves chat messages for the conversation, with pagination support. Returns a list of message dictionaries.

Parameters:: max_pages – Maximum number of pages to retrieve (default: 10)
Returns:: List of message dictionaries containing conversation history

Workflow Logging

class WorkflowLogEntry

Represents a single workflow execution log entry.

id: str: Unique identifier for the log entry.

workflow_run: WorkflowRun: Details about the workflow run.

created_from: str: Source from which the workflow was created.

created_by_role: str: Role of the user who created the workflow run.

created_by_account: Dict | None: Account information of the creator.

created_by_end_user: Dict: End user information.

created_at: int: Timestamp when the log entry was created.

class WorkflowRun

Contains information about a specific workflow run.

id: str: Unique identifier for the workflow run.

version: str: Version of the workflow.

status: str: Current status of the workflow run.

error: str | None: Error message if the workflow run failed.

elapsed_time: float: Time taken to execute the workflow.

total_tokens: int: Total number of tokens used.

total_steps: int: Total number of steps executed.

created_at: int: Timestamp when the workflow run was created.

finished_at: int: Timestamp when the workflow run finished.

exceptions_count: int: Number of exceptions encountered during execution.

Workflow Node Executions

class WorkflowNodeExecution

Represents the execution of a single node in a workflow.

id: str: Unique identifier for the node execution.

node_type: str: Type of the workflow node.

title: str: Title of the node.

status: str: Execution status of the node.

error: str | None: Error message if the node execution failed.

elapsed_time: float: Time taken to execute the node.

inputs: Dict[str, Any] | None: Input parameters provided to the node.

outputs: Dict[str, Any] | None: Output values produced by the node.

process_data: Dict[str, Any] | None: Process data during node execution.

index: int: Index of the node execution.

predecessor_node_id: str | None: ID of the predecessor node.

node_id: str: ID of the node being executed.

execution_metadata: WorkflowNodeExecutionMetadata | None: Metadata about the node execution.

extras: Dict[str, Any]: Additional extra data.

created_at: int: Timestamp when the node execution was created.

created_by_role: str: Role of the user who triggered the execution.

created_by_account: Dict[str, Any] | None: Account information of the creator.

created_by_end_user: Dict[str, Any]: End user information.

finished_at: int: Timestamp when the node execution finished.

class WorkflowNodeExecutionMetadata

Metadata about workflow node execution.

parallel_id: str | None: ID for parallel execution.

parallel_start_node_id: str | None: ID of the starting node in parallel execution.

tool_info: Dict[str, Any] | None: Information about tools used.

total_tokens: int | None: Total tokens used.

total_price: str | None: Total price of execution.

currency: str | None: Currency for the price.

class WorkflowNodeExecutions

Container for a list of node executions.

data: List[WorkflowNodeExecution]: List of node executions.

Agent and Chat Logging

class AgentConversation

Represents a single agent or chat conversation.

id: str: Unique identifier for the conversation.

status: str: Current status of the conversation.

from_source: str: Source from which the conversation originated.

from_end_user_id: str | None: ID of the end user who started the conversation.

from_end_user_session_id: str | None: Session ID of the end user.

from_account_id: str | None: Account ID of the creator.

from_account_name: str | None: Name of the account creator.

name: str: Name of the conversation.

summary: str: Summary of the conversation.

read_at: int | None: Timestamp when the conversation was last read.

created_at: int: Timestamp when the conversation was created.

updated_at: int: Timestamp when the conversation was last updated.

annotated: bool: Whether the conversation has been annotated.

agent_model_config: Dict: Model configuration for the agent (aliased as “model_config” in API).

message_count: int: Number of messages in the conversation.

user_feedback_stats: Dict: Statistics about user feedback.

admin_feedback_stats: Dict: Statistics about admin feedback.

status_count: Dict: Count of different statuses.

Chat Messages

class ChatMessage

Represents a single chat message in a conversation.

id: str: Unique identifier for the message.

conversation_id: str: ID of the conversation this message belongs to.

inputs: Dict[str, Any] | None: Input parameters provided with the message.

query: str | None: The user’s query or question.

answer: str | None: The assistant’s answer.

message: Dict[str, Any] | None: Full message object with additional metadata.

feedback: Dict[str, Any] | None: User feedback on the message.

created_at: int: Timestamp when the message was created.

created_by_role: str | None: Role of the user who created the message.

created_by_account: Dict[str, Any] | None: Account information of the creator.

created_by_end_user: Dict[str, Any] | None: End user information.

from_source: str | None: Source from which the message originated.

from_end_user_id: str | None: ID of the end user.

from_end_user_session_id: str | None: Session ID of the end user.

from_account_id: str | None: Account ID of the creator.

from_account_name: str | None: Name of the account creator.

class PaginatedChatMessages

Paginated response containing chat messages.

data: List[ChatMessage]: List of chat messages in the current page.

page: int: Current page number.

limit: int: Number of items per page.

total: int: Total number of items.

has_more: bool: Whether there are more pages available.

class PaginatedAgentLogs

Paginated response containing agent conversation logs.

data: List[AgentConversation]: List of conversations in the current page.

page: int: Current page number.

limit: int: Number of items per page.

total: int: Total number of items.

has_more: bool: Whether there are more pages available.