English| 中文
RuleGo
is a lightweight, high-performance, embedded orchestration rule engine based on Go
language. It is also a flexible and highly customizable event processing framework. It can aggregate, distribute, filter, transform, enrich and execute various actions on input messages.
This project is largely inspired by thingsboard .
RuleGo documentation is hosted on: rulego.cc .
- Development language: Go 1.18
- Lightweight: No external middleware dependencies, can efficiently process and link data on low-cost devices, suitable for IoT edge computing.
- High performance: Thanks to the high-performance characteristics of
Go
, in addition,RuleGo
adopts technologies such as coroutine pool and object pool. For 10W data processingJS script filtering->JS script data processing->HTTP push
, the average time is 9 seconds. - Embedded: Support embedding
RuleGo
into existing projects, non-intrusively utilizing its features. - Componentized: All business logic is componentized and can be flexibly configured and reused.
- Rule chain: You can flexibly combine and reuse different components to achieve highly customizable and scalable business processes.
- Process orchestration: Support dynamic orchestration of rule chains, you can encapsulate your business into
RuleGo
components, and then achieve your highly changing business needs by building blocks. - Easy to extend: Provide rich and flexible extension interfaces and hooks, such as: custom components, component registration management, rule chain DSL parser, coroutine pool, rule node message inflow/outflow callback, rule chain processing end callback.
- Dynamic loading: Support dynamic loading of components and extension components through
Go plugin
. - Built-in common components:
Message type Switch
,JavaScript Switch
,JavaScript filter
,JavaScript converter
,HTTP push
,MQTT push
,Send email
,Log record
and other components. You can extend other components by yourself. - Context isolation mechanism: Reliable context isolation mechanism, no need to worry about data streaming in high concurrency situations.
RuleGo
is a rule engine based on orchestration, which is best at decoupling your system.
- If your system is complex and bloated with code
- If your business scenario is highly customized or frequently changed
- If your system needs to interface with a large number of third-party systems or protocols
- Or you need an end-to-end IoT solution
- Or you need to process data from heterogeneous systems centrally
- Or you want to try hot deployment in
Go
language... ThenRuleGo
framework will be a very good solution.
- Edge computing: For example: You can deploy
RuleGo
on the edge server, preprocess, filter, aggregate or calculate the data before reporting it to the cloud. The data processing rules and distribution rules can be dynamically configured and modified through the rule chain without restarting the system. - Internet of Things: For example: Collect device data reporting, and after the rule judgment of the rule chain, trigger one or more actions, such as: send email, send alarm, and link with other devices or systems.
- Data distribution: For example: You can distribute data to different systems according to different message types, such as HTTP, MQTT or gRPC.
- Application integration: Use
RuleGo
as a glue to various different systems or protocols, such as: kafka, message queue, database, chatGPT, third-party systems. - Data processing from heterogeneous systems: For example: Receive data from different data sources (such as MQTT, HTTP, etc.), and then filter, format conversion, and then distribute to databases, business systems or dashboards.
- Highly customized business: For example: Decouple highly customized or frequently changed business and hand it over to
RuleGo
rule chain for management. Business requirements change without restarting the main program. - Complex business orchestration: For example: Encapsulate the business into custom components, and use
RuleGo
to orchestrate and drive these custom components, and support dynamic adjustment. - Microservice orchestration: For example: Use
RuleGo
to orchestrate and drive microservices, or dynamically call third-party services to process business and return results. - Business code and business logic decoupling: For example: User points calculation system, risk control system.
- Flexible configuration and highly customized event processing framework: For example: Asynchronously or synchronously process different message types.
- Automation: For example, process automation systems, marketing automation systems.
Use the go get
command to install RuleGo
:
go get github.com/rulego/rulego
First, define the rule chain in Json format. The rule chain definition does not require learning a specific rule syntax or DSL, just configure the components and connect them with certain relationships, and you can achieve your functional requirements. Rule chain definition: Reference rule chain
RuleGo is extremely simple and lightweight. Just follow these 2 steps:
- Import the
RuleGo
package and create a rule engine instance:
import "github.com/rulego/rulego"
//Create a rule engine instance, each rule engine instance has only one root rule chain
ruleEngine, err := rulego.New("rule01", []byte(ruleFile))
- Pass the message, message type, message metadata to the rule engine instance for processing, and then the message will be processed according to the configured rule chain logic:
//Define message metadata
metaData := types.NewMetadata()
metaData.PutValue("productType", "test01")
//Define message content and message type
msg := types.NewMsg(0, "TELEMETRY_MSG", types.JSON, metaData, "{\"temperature\":35}")
//Pass the message to the rule engine for processing
ruleEngine.OnMsg(msg)
Dynamically update the rule chain
//Update the root rule chain
err := ruleEngine.ReloadSelf([]byte(ruleFile))
//Update a node under the rule chain
ruleEngine.ReloadChild("rule_chain_test", nodeFile)
//Get the rule chain definition
ruleEngine.DSL()
Rule engine instance management:
//Load all rule chain definitions in the folder to the rule engine pool
rulego.Load("/rules", rulego.WithConfig(config))
//Get a created rule engine instance by ID
ruleEngine, ok := rulego.Get("rule01")
//Delete a created rule engine instance
rulego.Del("rule01")
Configuration:
See documentation for details
//Create a default configuration
config := rulego.NewConfig()
//Debug node callback, node configuration must be configured debugMode:true to trigger call
//Node entry and exit information will call this callback function
config.OnDebug = func (chainId,flowType string, nodeId string, msg types.RuleMsg, relationType string, err error) {
}
//Global rule chain end callback
//If you just want to call for a single message, use the ruleEngine.OnMsgWithOptions method
//Note: If the rule chain has multiple branch endpoints, it will be called multiple times
config.OnEnd = func (msg types.RuleMsg, err error) {
}
//Use configuration
ruleEngine, err := rulego.New("rule01", []byte(ruleFile), rulego.WithConfig(config))
Rule nodes are the basic components of the rule chain, they are functions that implement specific business logic. Rule nodes can filter, transform, enrich or perform some actions on the incoming messages. Rule nodes can adjust their behavior and output by configuring parameters.
You can easily encapsulate your business into RuleGo
node components, and flexibly configure and reuse them, like building blocks to achieve your business requirements.
- Custom node components: examples/custom_component or documentation
- Provide custom components in
go plugin
way: examples/plugin or documentation RuleGo
provides a lot of standard components , as well as extended components
Rule chains are the core concept of RuleGo, they are directed acyclic graphs composed of multiple rule nodes, each rule node is a component that can implement different business logic, nodes are connected by relationship types (relation type). Rule chains can be dynamically configured and modified, support nesting and orchestration, and implement complex business processes.
The following example defines 3 rule nodes, which are to filter->transform->push data, the rule chain logic is as follows:
Rule chain definition:
{
"ruleChain": {
"name": "Test rule chain",
"root": true
},
"metadata": {
"nodes": [
{
"id": "s1",
"type": "jsFilter",
"name": "Filter",
"debugMode": true,
"configuration": {
"jsScript": "return msg!='bb';"
}
},
{
"id": "s2",
"type": "jsTransform",
"name": "Transform",
"debugMode": true,
"configuration": {
"jsScript": "metadata['test']='test02';\n metadata['index']=50;\n msgType='TEST_MSG_TYPE2';\n var msg2=JSON.parse(msg);\n msg2['aa']=66;\n return {'msg':msg2,'metadata':metadata,'msgType':msgType};"
}
},
{
"id": "s3",
"type": "restApiCall",
"name": "Push data",
"debugMode": true,
"configuration": {
"restEndpointUrlPattern": "http://192.168.216.21:9099/api/socket/msg",
"requestMethod": "POST",
"maxParallelRequestsCount": 200
}
}
],
"connections": [
{
"fromId": "s1",
"toId": "s2",
"type": "True"
},
{
"fromId": "s2",
"toId": "s3",
"type": "Success"
}
]
}
}
Other rule chain examples:
RuleGo
provides Endpoint
module for unified data integration and processing of heterogeneous systems.
For more details, please refer to: Endpoint
RuleGo
almost does not increase system overhead, resource consumption is extremely low, because it uses object coroutine pool and object pool, even higher performance than directly calling business methods, especially suitable for running on edge servers.
Machine: Raspberry Pi 2 (900MHz Cortex-A7*4,1GB LPDDR2)
Data size: 260B
Rule chain: JS script filtering->JS complex transformation->HTTP push
Test results: 100 concurrent and 500 concurrent, memory consumption does not change much around 19M
Any form of contribution is welcome, including submitting issues, suggestions, documentation, tests or code. Please follow these steps:
- Clone the project repository to your local machine
- Create a new branch and make modifications
- Submit a merge request to the main branch
- Wait for review and feedback
RuleGo
uses Apache 2.0 license, please refer to LICENSE file for details.