Assistant Events

Assistant Events are specialized event handling for AI Assistant interactions. They are responsible for monitoring AI Assistant execution processes, state changes, and user interactions, providing processing capabilities for core interaction events such as user intent recognition, assistant switching, and session management.

The Assistant Events element hierarchical structure is Meta (events.Meta) → Type (events.AIAssistantType) → Instance. Developers can quickly create Assistant Events instance elements through JitAi's visual development tools.

Of course, developers can also create their own Type elements, or override the official events.AIAssistantType element provided by JitAi in their own App to implement their own encapsulation.

Quick Start

Creating Instance Elements

Directory Structure

Recommended Directory Structure

events/
└── testAssistantEvents/
    ├── __init__.py
    ├── e.json
    └── inner.py

e.json File

e.json Configuration File

{
  "asyncType": false,
  "title": "Test Assistant Events",
  "sender": "aiassistants.toolTest",
  "operate": "afterRun",
  "func": "services.standardServicesTested.sayHello",
  "backendBundleEntry": ".",
  "type": "events.AIAssistantType"
}

Business Logic Code

inner.py File

# -*-coding:utf-8-*-
from datatypes.Meta import datatypes

def customFunc(eventOutData):
    """
    Custom event handler function
    :param eventOutData: Event output data
    """
    # Business logic for processing event data
    pass

Usage Example

Getting and Using Assistant Events

# Get assistant event instance
event = app.getElement("events.testAssistantEvents")

# Event will automatically trigger when AI Assistant executes afterRun
# Calls the specified target function: services.standardServicesTested.sayHello

Element Configuration

e.json Configuration

Parameter	Type	Native Type	Required	Description
type	str	str	Yes	Fixed value: events.AIAssistantType
title	str	str	Yes	Event display name
sender	str	str	Yes	Event sender, usually the fullName of AI Assistant
operate	str	str	Yes	Event trigger timing: basic timing (beforeRun/afterRun), node timing (nodeId.beforeNodeRun etc.), manual operation timing (nodeId.approved etc.)
func	str	str	Yes	Target function called when event is triggered
asyncType	bool	bool	No	Whether to execute asynchronously, default false
backendBundleEntry	str	str	Yes	Fixed value: "."

Business Logic Code Configuration

The inner.py file is used to define custom event processing logic:

Custom Handler Function Example

def customFunc(eventOutData):
    """
    Custom event handler function
    :param eventOutData: Event output data
    """
    # Business logic for processing event data
    print(f"Assistant event triggered: {eventOutData}")
    return eventOutData

Methods

getSender

Get the complete event sender identifier in the format "sender.operate".

Parameter Details

No parameters

Return Value

Type	Description
str	Complete identifier of the event sender

Usage Example

Get Event Sender

event = app.getElement("events.testAssistantEvents")
sender = event.getSender()
print(sender)  # Output: aiassistants.toolTest.afterRun

isValid

Validate whether the event is valid, used to determine if the event should be triggered under current conditions.

Parameter Details

Parameter	Type	Native Type	Required	Description
rowObj	object	object	No	Data row object
args	tuple	tuple	No	Additional parameters
kwargs	dict	dict	No	Keyword parameters

Return Value

Type	Description
bool	Whether the event is valid

Usage Example

Validate Event Validity

event = app.getElement("events.testAssistantEvents")
is_valid = event.isValid()
print(is_valid)  # Output: True

handleNode

Process event nodes, parse event parameters and prepare to call target functions.

Parameter Details

Parameter	Type	Native Type	Required	Description
node	object	object	Yes	Event node object
args	tuple	tuple	No	Parameters passed to target function
kwargs	dict	dict	No	Keyword parameters passed to target function

Return Value

Type	Description
tuple	Returns processed (node, args, kwargs)

Usage Example

Process Event Node

event = app.getElement("events.testAssistantEvents")
# Usually called automatically by the event system, developers generally don't need to use it directly

属性

sender

事件发送方的标识，通常为触发事件的AI助理的fullName。

operate

事件触发的操作时机，支持多种触发场景：

• 基础时机: beforeRun(助理运行前)、afterRun(助理运行后)
• 节点时机: ${nodeId}.beforeNodeRun(节点到达时)、${nodeId}.afterNodeRun(节点执行后)
• 人工操作时机: ${nodeId}.approved(同意后)、${nodeId}.rejected(拒绝后)、${nodeId}.replied(回复后)、${nodeId}.edited(编辑后)

func

事件触发时调用的目标函数，格式：{元素名}.{函数名}。

asyncType

事件是否异步执行，为true时事件处理不会阻塞主流程。

高级特性

事件参数解析机制

助理事件具有智能的参数解析能力，会根据目标函数的参数定义自动调整传递的参数。如果目标函数没有设置参数，事件系统会自动传递空参数。

参数解析示例

# 目标函数定义
def targetFunction(param1, param2):
    return f"处理参数：{param1}, {param2}"

# 事件会自动解析并传递正确的参数数量和类型

节点级事件配置

助理事件支持对AI助理流程中的特定节点进行事件监听，实现更精细的流程控制。

节点级事件配置示例

# 监听特定节点的执行
{
  "title": "数据处理节点事件",
  "sender": "aiassistants.dataProcessor",
  "operate": "node001.beforeNodeRun",  # 监听节点node001的到达
  "func": "services.validationService.validateData"
}

# 监听人工审核节点
{
  "title": "审核通过事件",
  "sender": "aiassistants.approvalAssistant",
  "operate": "approvalNode.approved",  # 监听审核节点的通过操作
  "func": "services.workflowService.processApproval"
}

# 自定义事件处理逻辑
def customFunc(eventOutData):
    node_id = eventOutData.get("nodeId", "")
    node_type = eventOutData.get("nodeType", "")
    operation = eventOutData.get("operation", "")

    if operation == "beforeNodeRun":
        # 节点执行前的预处理
        print(f"节点 {node_id} 即将执行")
        # 可以进行数据验证、权限检查等

    elif operation == "approved":
        # 审核通过后的处理
        approval_data = eventOutData.get("approvalData", {})
        print(f"审核节点 {node_id} 审核通过")
        # 可以触发后续流程、发送通知等

    return eventOutData

与AI助理的协同工作

助理事件与AI助理组件深度集成，支持多种协同场景：

协同工作示例

# 监听助理执行完成事件进行日志记录
{
  "sender": "aiassistants.customerService",
  "operate": "afterRun",
  "func": "services.logService.recordInteraction"
}

# 监听关键业务节点进行状态同步
{
  "sender": "aiassistants.businessProcessor",
  "operate": "paymentNode.afterNodeRun",
  "func": "services.paymentService.confirmPayment"
}

# 自定义协同处理逻辑
def customFunc(eventOutData):
    operation = eventOutData.get("operation", "")

    if operation == "afterRun":
        # 助理执行完成后的协同处理
        result = eventOutData.get("result", {})
        # 可以触发其他助理、发送通知、更新状态等

    return eventOutData

通过合理配置助理事件，可以实现AI助理行为的自动化监控、日志记录、状态同步等功能，为企业级AI应用提供完整的事件驱动支持。