CN112394922A

CN112394922A - Decision configuration method, business decision method and decision engine system

Info

Publication number: CN112394922A
Application number: CN202011413107.0A
Authority: CN
Inventors: 周智杰; 王小东; 廖浩
Original assignee: Chengdu New Hope Finance Information Co Ltd
Current assignee: Chengdu New Hope Finance Information Co Ltd
Priority date: 2020-12-02
Filing date: 2020-12-02
Publication date: 2021-02-23
Anticipated expiration: 2040-12-02
Also published as: CN112394922B

Abstract

The embodiment of the application provides a decision configuration method, a service decision method and a decision engine system, wherein the decision configuration method comprises the following steps: responding to a decision flow configuration operation for the target service, and creating a decision flow containing at least one decision branch for the target service, wherein each branch in the at least one decision branch comprises at least one decision event; responding to a first configuration operation on a target event in the decision flow, and determining a target parameter item and a target rule operator according to the first configuration operation, wherein the target event is any one decision event in the decision flow, and the target rule operator comprises at least one of a numerical operator, a logic operator or a text operator; responding to drag configuration operation of the target parameter item and the target rule operator, and generating a conditional decision rule for the target event; and the decision flow is used for outputting a decision result according to the conditional decision rule of each decision event configured in the decision flow when triggered to execute.

Description

Decision configuration method, business decision method and decision engine system

Technical Field

The present application relates to the technical field of decision engines, and in particular, to a decision configuration method, a business decision method, and a decision engine system.

Background

Under the background of the internet era, the information quantity and the business quantity of a plurality of industries are rapidly increased, and scenes needing rapid judgment and rapid decision making are more and more, so that an intelligent decision engine is necessary to be provided.

The decision engine is a general tool, is suitable for all scenes needing quick decision making, and can comb the contents of each field into a series of rules, thereby providing a processing strategy and a processing result for the decision event judgment process of the corresponding field. The user can configure specific events according to actual requirements, and then the decision engine judges corresponding events according to actually obtained data and actually configured contents.

However, the current decision engine requires a large amount of learning cost for the user to learn the script knowledge corresponding to the decision engine, and has a high requirement for the user and a complicated configuration process.

Disclosure of Invention

The present application aims to provide a decision configuration method, a business decision method and a decision engine system, which can solve the problems in the prior art that the decision engine has high requirements for users and the configuration process is complicated.

In a first aspect, the present application provides a method for decision configuration, the method comprising:

responding to a decision flow configuration operation for a target service, and creating a decision flow containing at least one decision branch for the target service, wherein each branch of the at least one decision branch comprises at least one decision event;

in response to a first configuration operation on a target event in the decision flow, determining a target parameter item and a target rule operator according to the first configuration operation, wherein the target event is any decision event in the decision flow, and the target rule operator comprises at least one of a numerical operator, a logic operator or a text operator;

generating a conditional decision rule for the target event in response to a drag configuration operation on the target parameter item and the target rule operator;

and the decision flow is used for outputting a decision result according to the conditional decision rule of each decision event configured in the decision flow when triggered to execute.

In the method, a user can rapidly configure the decision flow through a dragging and pulling operation mode, the method can rapidly generate corresponding conditional decision rules for each decision event in the decision flow based on the dragging and pulling configuration mode, when the method is applied to the decision configuration process of the decision engine, the requirement on the user can be reduced, the user does not need to learn a large amount of programming knowledge and script knowledge intentionally for the configuration process, and the configuration difficulty is reduced. When a slightly complex rule configuration scene is faced, the method can be used for configuration by using the idea of meta programming, the conditional decision rule can be generated for the decision event by dragging configuration operation of the target parameter item and the target rule operator, and compared with a mode that a user firstly splits a plurality of sub-expressions according to the service content and then carries out nested assembly configuration on each sub-expression, the method can reduce the use number of intermediate variables in the configuration process by the user and reduce the configuration difficulty.

In an alternative embodiment, the method further comprises:

responding to the visual test operation of the decision flow to obtain test parameters;

inputting the test parameters into the decision flow, and performing decision simulation on the test parameters through each conditional decision rule in the decision flow;

in the process of decision simulation, according to the simulation sequence or the simulation state of each decision event in the decision flow, simulation marking is carried out, and the simulation marking result is displayed on the decision flow.

By the implementation mode, the test simulation process of the decision flow can be visualized, a user can conveniently and quickly know the test state, and the tracking of the simulation execution process of the decision flow is facilitated.

In an alternative embodiment, the method further comprises:

displaying an online programming interface in response to a second configuration operation for the target event;

and generating a conditional decision rule for the target event according to the code content acquired from the online programming interface.

Through the implementation mode, a user can perform configuration operation in an online programming mode under the scene of complex business logic, such as when a target rule operator or a target parameter item is lacked, or when the condition decision rule configuration process of a decision event is difficult to complete only through a row of expressions, or when the business logic is complex, the implementation mode can generate the condition decision rule for the corresponding decision event based on the content obtained by online programming, and the user does not need to additionally download and install a code editor, so that the operation difficulty of the user can be reduced.

In an alternative embodiment, after displaying the online programming interface, the method further comprises:

generating an event file for the target event;

and carrying out online detection on the code content in the online programming interface according to the type of the event file or the code content acquired from the online programming interface.

Through the implementation mode, online detection can be performed on online programming content, and compared with a processing mode that a user performs self-checking after writing all codes or an offline code processing tool is adopted to compile a complete code document, the implementation mode can support code writing and detection without the need of executing compiling triggering operation additionally by the user.

In an optional embodiment, the performing, according to the type of the event file or according to the code content acquired from the online programming interface, online detection on the code content in the online programming interface includes:

responding to code writing operation of the online programming interface, and calling a preset language service middleware according to the type of the event file or code content acquired from the online programming interface in the code writing process;

calling a language service module through the language service middleware, and detecting codes of code contents in the online programming interface through the language service module;

and displaying a code detection result in the online programming interface.

Through the implementation mode, the corresponding language service middleware and the language service module can be automatically called according to the type of the event file or the code content obtained by online programming, so that the online code detection is carried out according to the called middleware and module, the code detection result can be timely displayed in an online programming interface, for a user, multiple programming languages do not need to be learnt deliberately, and the code correction can be carried out by the user through the code detection result displayed in the online programming interface timely.

In an alternative embodiment, the code detection result includes: at least one of a keyword hint flag, a code completion hint, and an exception code hint.

By the implementation mode, the online programming keyword prompt, the automatic code completion or the abnormal code detection can be supported.

In an alternative embodiment, the language service module comprises a lexical analyzer and a syntax analyzer, and the language service module supports at least one of Java language, Python language, Jcc language, C + + language and Groovy language.

The implementation mode can improve compatibility, and the implementation mode can automatically call the corresponding language service module according to the programming language type used by the user during online programming, so that the code language self-adaptive detection is carried out on the online coding process, and for the user, the user does not need to learn multiple programming languages, and the requirement on the user can be reduced.

In an alternative embodiment, in response to a first configuration operation on a target event in the decision stream, determining a target parameter item and a target rule operator according to the first configuration operation comprises:

acquiring the target parameter item from a parameter pool according to the parameter type selected by the first configuration operation;

and acquiring the target rule operator from an operation rule pool according to the expression type selected by the first configuration operation.

Through the implementation mode, the target parameter item and the target rule operator can be obtained based on the content selected by the first configuration operation, the data definition workload required to be completed by a user in the traditional configuration process can be reduced, the code function learning cost of the user for the complex operation function can be reduced, and the implementation mode can support the user to carry out dragging configuration based on the self-defined domain modeling language so as to carry out decision configuration.

In a second aspect, the present application provides a business decision method, including:

acquiring a to-be-processed parameter corresponding to a target service through a preset data interface;

preprocessing the parameters to be processed to obtain parameters to be decided, wherein the preprocessing comprises data cleaning;

inputting the parameter to be decided into a decision flow configured by the method of the first aspect, and performing decision operation on the parameter to be decided according to a conditional decision rule of each decision event configured in the decision flow to obtain a decision result corresponding to the parameter to be decided.

By the method, the decision flow obtained by the method configuration of the first aspect can be based on, and a decision result is rapidly provided for the to-be-processed parameter corresponding to the target service.

In a third aspect, the present application provides a decision engine system, comprising:

a memory;

a processor;

the memory has stored thereon a computer program which, when executed by the processor, performs the method of the first or second aspect.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments of the present application will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and that those skilled in the art can also obtain other related drawings based on the drawings without inventive efforts.

Fig. 1 is a schematic architecture diagram of a decision engine system according to an embodiment of the present disclosure.

Fig. 2 is a flowchart of a decision configuration method according to an embodiment of the present application.

Fig. 3 is a schematic diagram of a decision flow in an example provided by an embodiment of the present application.

Fig. 4 is a schematic configuration diagram of a decision event in an example provided by the embodiment of the present application.

Fig. 5 is a partial flowchart of a decision configuration method according to an embodiment of the present application.

Fig. 6 is another partial flowchart of a decision configuration method according to an embodiment of the present disclosure.

Fig. 7 is a flowchart of a service decision method according to an embodiment of the present application.

Fig. 8 is a functional implementation structure block diagram of a decision engine system according to an embodiment of the present disclosure.

Detailed Description

The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.

In the prior art, there are many scenarios that a decision engine needs to provide a quick decision, for example, for an internet financial institution, it is difficult to obtain credit investigation reports of a specific professional institution (e.g., a people bank), and a large amount of other credit investigation data with complex types are often needed to develop a wind control system, so as to counter various fraud risks and credit risks, and a decision scheme needs to be quickly provided for a plurality of complex business scenarios when the wind control system is built. At present, some risk management and control schemes of enterprises need to use external data of dozens of enterprises or even more, and the external data needs to be subjected to butt joint, cleaning, derivation and conversion, and then is sent to a specific decision model in a decision engine for analysis, and finally a decision result is obtained. The decision result is typically: qualified, unqualified, passing, failing, waiting, and the like.

In fact, how to configure the core decision model or rule in the decision engine and how to implement the model and rule to the actual product user is troublesome for both the product user and the background research and development staff. Taking a wind control product in the financial field as an example, a product usually requires a product designer to deploy dozens or even hundreds of rules, and needs to use several, dozens or even hundreds of models, if a rule is to be deployed or a model is to be used each time, corresponding background research and development personnel are temporarily inquired or research and development data are temporarily inquired, so that the communication cost and the learning cost are high. Especially, in some scenes, if a constant rule is adopted and is easily known, optimization iteration needs to be continuously performed on the wind control system, and continuous switching of the rule, the model and the decision logic brings great difficulty to development departments.

Therefore, how to provide a tool convenient for practical use for business or product users, reduce the learning cost of users, and reduce the difficulty in configuring and using a decision engine is necessary for research and development departments or business departments. It will be appreciated that such scenarios requiring efficient decision making are not only common in the financial industry, but also in other industries than finance. If a decision scenario that a plurality of conditions are required to determine whether to execute a certain function is needed for efficient decision making, a decision engine can be adopted to provide a decision result.

At present, decision engine systems such as JBoss Drools, Mandarax, OpenRules, JEOPS, infosafe, roole, Apache CamelIBM, URule, rad, Drools and the like which are common in the market have the defects of high product learning cost and complex use mode, and may not support users to write codes by themselves when facing complex business processing logic, and the configuration process and the use process are difficult for product users. However, the configuration process of commercial decision engine products such as ODM, Oracle Business Rules, flag Rules engine, Jess, topriles, policy intelligence decision engine, Yiborui decision engine, China's large data wind control platform, Sparkling Logic, ILOG, etc. in the market is also tedious, a user is required to finish the complicated rule splitting and recombining process by himself, the user is required to design a large number of intermediate variables for the split small Rules respectively, and some decision engines do not support Java and Python languages, and the user of the product is required to learn a small and numerous programming language used by a specific decision engine additionally. Therefore, the current decision engine system has high requirements for users, and the configuration process of the decision flow is tedious.

In view of the above, the present invention provides an improved cloud-based general decision engine system implemented based on an Integrated Development Environment (IDE), where the system can execute the method provided in the embodiments of the present application, and for a simpler service, a configuration process can be quickly completed in a simple configuration manner by dragging, and when configuration is performed in a dragging configuration manner, a meta-programming concept is adopted to simplify the configuration operation difficulty of a user. For slightly complex services, besides a dragging configuration mode, the configuration is realized in a cloud online programming mode, a code editor required by programming is not required to be additionally downloaded, installed and learned by a user, online code editing, debugging and testing can be carried out on a browser, code compiling and executing are finished at the cloud (at one side of a server), code writing and code detection can be supported, and a code detection result can be displayed in the code compiling process of the user. The configuration mode of the whole decision flow is simple and convenient, a user can freely create or update the decision flow based on the operations of graphic editing and flow dragging and arranging, the simulation/execution result of the decision flow is visual, and the simulation execution process can be tracked.

Referring to fig. 1, fig. 1 is a schematic diagram illustrating an architecture of a decision engine system according to an embodiment of the present disclosure. The decision engine system is a cloud-end general decision engine system implemented based on an Integrated Development Environment (IDE).

As shown in fig. 1, the decision engine system includes a data access layer, a data computation layer, a decision engine layer, and an application layer.

The data access layer may provide some data interaction interfaces and APIs (Application Programming interfaces), so as to call data that needs to be called or accessed from the outside, where the data may be credit investigation data, consumption data, three-party data, and the like that can be obtained after an authorization is granted.

And the data calculation layer is used for processing and analyzing various data obtained from the data access layer. The data calculation layer can support distributed offline data calculation and real-time data calculation, massive offline data can be calculated and extracted through technologies such as Spark, Impala and Hive (for example, ETL extraction, the ETL is called Extract-Transform-Load as a whole, namely a data extraction-transformation-loading process, and is a process of loading (Load) data of a business system to a data warehouse after extraction, cleaning and transformation, so as to integrate scattered, disordered and standard non-uniform data together), feature processing is performed, and tool such as Flink and Storm can be used for performing ETL, feature processing and index calculation on massive real-time data. The analyzed data or features may be input into some business models or decision logic, which may be a decision stream in the embodiment of the present application.

The decision engine layer is used for providing editing configuration, simulation test and execution functions of the decision flow, visualizing simulation and execution results of the decision flow, and can also be used for realizing flow arrangement of business decision logic, wherein the flow arrangement can be regarded as macro configuration of the decision flow, and the configuration of specific nodes and specific events in the decision flow can be regarded as micro configuration of the decision flow.

Optionally, the decision engine layer may also be used for monitoring and early warning, and may be used for monitoring and early warning on a plurality of modules in the decision engine system, for example, the decision execution condition of the decision flow, the decision result, the data quality of the data source input into the decision flow (the data quality may be evaluated in terms of the number of times of interruption of the decision flow, the processing and analysis time of the data by the data computation layer, the stability of the data source, and the like), the execution time of the decision flow, and the like may be monitored. The data which needs to be monitored can be cleaned through a Flink tool, the data are cleaned to be indexes meeting monitoring requirements, monitoring rules and logic expressions needed by monitoring services are configured through a web page which can be dragged and configured, the logic expressions are analyzed and calculated through a domain modeling language, and the index data meeting monitoring and early warning conditions are pushed in a picture and text combination mode (the index data can be pushed in a telephone mode, an enterprise WeChat mode, a nail mode and the like).

And the application layer is used for executing specific service functions according to the decision result of the decision engine layer. What interface is called, what data is generated, what operation is performed, and the like when the decision result indicates "pass", and the like can be set in the application layer. Different functions can be set in the application layer according to different service requirements, for example, different service functions can be set in the application layer for pre-loan, mid-loan, post-loan, admission or other non-financial scenarios. Business functions in the application layer include, but are not limited to: information identification, compliance verification, qualification review, automatic approval, and the like.

In the embodiment of the present application, data between different users are isolated from each other, that is, each user can only see the self-configured decision flow and the self-configured conditional decision rule for the user of the decision engine.

It is understood that the architecture shown in fig. 1 is only a software functional architecture diagram of the decision engine system, and in practical applications, there may be more or less functional layers or an architecture different from that shown in fig. 1. In some application scenarios, the functionality of portions of the functional layers may be combined.

Referring to fig. 2, fig. 2 is a schematic diagram of a decision configuration method according to an embodiment of the present application. The method can be applied to the decision engine system provided by the embodiment of the application.

As shown in FIG. 2, the method may include steps S21-S23.

S21: and responding to a decision flow configuration operation of the target service, and creating a decision flow containing at least one decision branch for the target service, wherein each branch of the at least one decision branch comprises at least one decision event.

The business in the embodiment of the present application may be a business in a financial industry or a non-financial industry. For example, the service applicable to the decision scenario of the embodiment of the present application may be a service scenario of other industries, in addition to a service scenario of anti-fraud, credit wind control, insurance underwriting, automatic claim settlement, and the like in the financial industry, for example, a service scenario of an automatic driving scenario (for example, it is determined whether mode switching is currently appropriate according to various environmental data and self-checking data), an equipment maintenance scenario (for example, it is determined whether it is necessary to immediately maintain a certain equipment according to various parameters), a monitoring early warning scenario (for example, it is necessary to perform an early warning under what conditions is determined when facing various complex parameters), and the like.

Regarding the decision flow configuration operation in S21, there may be, but not limited to: a decision flow creation operation, an event creation operation, a branch creation operation, an event configuration operation, a branch configuration operation, and the like. Each decision branch in this implementation may be used to point to a node. A decision flow comprising a plurality of nodes and a plurality of decision branches is obtained through S21.

Any node in the embodiments of the present application may correspond to a decision event (referred to as an event for short). The decision rule corresponding to the event can be configured for the target event corresponding to the event configuration operation through the event configuration operation (in the simulation or execution stage, if the decision rule corresponding to the target event is executed, the decision rule of the next event can be executed according to the branch corresponding to the target event). The relation of each decision event can be configured according to the branch configuration operation, and the processing sequence among the decision events can be indicated through the branch configuration operation.

FIG. 3 shows a schematic diagram of a decision flow in one example. Nodes corresponding to "a", "B", "C", "D", "E", "F", "G" in fig. 3 represent decision events. The user can configure the execution priority and weight among the events. Each event may correspond to a conditional decision rule.

Optionally, the nodes newly built for the decision flow or the nodes already built and configured may be the target nodes in the embodiment of the present application when selected. The event corresponding to the target node is marked as a target event, and S22 may be executed when the target event is determined. In some embodiments, node configuration may be performed first, and then a corresponding decision flow may be generated based on the node configuration result (i.e., in some embodiments, S22 and S23 may be performed first, and then S21 may be performed).

S22: in response to a first configuration operation on a target event in the decision stream, determining a target parameter item and a target rule operator according to the first configuration operation, wherein the target event is any one decision event in the decision stream, and the target rule operator comprises at least one of a numerical operator, a logical operator or a text operator.

Wherein the first configuration operation may be, but is not limited to: double-click, selection and dragging operation. The target parameter item may be obtained from a parameter pool according to a parameter type selected by the first configuration operation. A target rule operator may be obtained from the operation rule pool based on the expression type selected by the first configuration operation.

Therefore, the target parameter item and the target rule operator can be obtained based on the content selected by the first configuration operation, the data definition workload required to be completed by a user in the traditional configuration process can be reduced, the code function learning cost of the user for complex operation functions can be reduced, and the implementation mode can support the user to carry out dragging configuration based on the self-defined domain modeling language so as to carry out decision configuration. The operator supported by the operation rule pool and the parameter items supported by the parameter pool provided by the embodiment of the application are rich.

Upon selection of the target parameter item or target rule operator, S23 may be executed.

S23: conditional decision rules are generated for the target event in response to a drag configuration operation on the target parameter items and the target rule operator.

Wherein, the rule operator whose expression type in the operation rule pool is "numerical operator" includes but is not limited to: "addition (+)", "subtraction (-)", "multiplication (", "division (/)", "square root (sqrt)", "absolute value (abs)", "base-2 logarithm (log)", "base-e logarithm (ln)", "base-10 logarithm (lnTen)", "factorial (nPower)", "base-e exponential function (xExp)", "floor (floor)", "ceiling (ceil)", "maximum value (max)", "minimum value (min)", "average value (avg)", "sum (sum)", "x raised power (xPowerY)", "remainder (mod)", and the like.

Rule operators in the operation rule pool whose expression type is "logical operator" include, but are not limited to: "and", "or", "not", "true", "false", "greater than", "(>)," (<) "," (+) ")," (-), "(- ═) or less", "(| -),") and the like.

Rule operators in the operation rule pool whose expression type is "text operator" include, but are not limited to: "in … … (functin)", "not in … … (functin)", "start with … … (functionstart)", "not start with … … (functionstart)", "end with … … (functionsent)", "not end with … … (functionstrintinend)", "contain … … (functicontain)", "not contain … … (functionstate)", "in set (functinset)", "not in set (functinset)", "regularly contain (functigex)", "not contain (functinotregex)", "empty (funcIsEmpty)", "not empty (funcIsEmpty)" and the like. Where the ellipses in these rule operators represent what needs to be configured by the user.

In this embodiment of the present application, each rule operator in the operation rule pool may be referred to as a computation operator or a meta operator, and each rule operator is implemented by DSL (Domain specific Language), and may be regarded as a function.

As shown in fig. 4, a drag configuration diagram of one decision event in one example is shown.

Taking an "age judgment" decision event as an object event (the object event may be any event in fig. 3) as an example, when the object parameter item with the parameter type of "user information" and the "age" is selected from the parameter pool, the object parameter item may be added to the configuration area of the rule expression based on a drag operation, and similarly, when the object rule operator "greater than" in the operation rule pool and with the expression type of "logical operator" is selected, the object rule operator may be added to the configuration area of the rule expression based on the drag operation, and the user may still configure the content added to the configuration area of the rule expression by dragging. Based on this principle, a conditional decision rule (as shown in fig. 4) generated from a row of expression conditions, which is obtained based on a drag configuration operation in the configuration area of a regular expression, means that if the input data type includes "age" in the user information and the age is greater than 22, and the input data type includes "name" in the user information and the name is one of "a 1" and "a 2", when this condition is established, the "valuation: 0 ", and when the condition is not satisfied, outputs" assign: the result of 1 "is used for determining whether the user age is greater than 22 and determining whether the user name is one of a1 and a2, if the results of the determinations are yes, one assignment result is output for the target event, otherwise another assignment result is output, and the different assignment result is used for indicating different processing modes (for example, assignment 0 may indicate rejection, and assignment 1 may indicate pass), so as to obtain the conditional decision rule.

In configuring for the target decision event, an interrupt condition may also be configured, and the interrupt condition may be used to indicate under what conditions the decision flow will stop simulation, stop execution.

In the method, for a user of a decision engine, the user does not need to pay attention to a bottom code, only needs to select and drag a rule operator in an operation rule pool and parameter items in a parameter pool, and equipment or a system executing the method can automatically perform nested assembly of expressions based on dragging configuration operation of the user, so that the expression of a conditional decision rule group can be written in one row theoretically, and calculation is performed. In the method, the conditional decision rule is generated based on the dragged rule operator, so that the DSL-based expression calculation engine is realized, and for a user, the user does not need to split the rule by himself or herself and then respectively configure and assemble the split small rules, namely, the user does not need to split a complex expression into a plurality of atomic expressions to participate in calculation.

For the decision flow which is already established, each decision branch in the decision flow can be updated in response to the updating operation for the decision flow, and the updating content includes but is not limited to adding, deleting and modifying the decision branch or the decision event in the decision branch. The update process may also be implemented in a drag configuration.

In the method of S21-S23, the user may perform fast configuration on the decision stream through a drag and pull operation manner, and the method may quickly generate corresponding conditional decision rules for each decision event in the decision stream based on a drag and pull configuration manner. When a slightly complex rule configuration scene is faced, the method can be used for configuration by using the idea of meta programming, the conditional decision rule can be generated for a decision event by dragging configuration operation of a target parameter item and a target rule operator, a user does not need to split a plurality of sub-expressions according to business content and then nest, assemble and configure each sub-expression, the method can reduce the use amount of intermediate variables in the configuration process by the user, and the configuration difficulty can be reduced.

Optionally, when a complex business requirement is encountered, for example, when a rule operator in the operation rule pool and a parameter item in the parameter pool cannot meet a user configuration requirement, or it is difficult to meet the business requirement through simple drag-and-drop configuration, the method provided by the embodiment of the present application may support configuring the target event in an online programming manner. The implementation may include steps S24-S25 (see FIG. 5). Of course, the method of the embodiment of the present application also supports an offline programming mode of the import script to write codes, so as to configure the target event.

S24: and displaying the online programming interface in response to the second configuration operation for the target event.

The second configuration operation may be an online programming trigger operation for a target event, and the specific trigger manner should not be construed as a limitation to the present application.

The user may enter code content and annotation content based on the online programming interface.

And when the target event needs to be configured in an online programming mode, responding to the second configuration operation and generating an event file for the target event. The user can select the file saving type for the event file according to the actually grasped programming language, and the code content in the current event file can be distinguished by the file saving type according to which programming language is written.

Illustratively, the suffix name of the event file may be, but is not limited to, ". java", ". py", and the type of the event file is distinguishable by the file suffix name of the event file.

S25: and generating a conditional decision rule for the target event according to the code content acquired from the online programming interface.

When the code content in the online programming interface is detected and compiled, a conditional decision rule can be generated for the target event.

Through the implementation manner of the S24-S25, a user can perform configuration operation in an online programming manner in a scenario facing complex business logic, for example, when a target rule operator or a target parameter item is lacked, or when it is difficult to complete a conditional decision rule configuration process of a decision event only through an expression, and the implementation manner can generate a conditional decision rule for a corresponding decision event based on content obtained through online programming, and does not need to additionally download and install a code editor, so that the difficulty in user operation can be reduced.

Optionally, the event file may be generated for the target event after the online programming interface is displayed, or the event file may be generated for the target event first and then the online programming interface is displayed.

When the code content exists in the online programming interface, the code content in the online programming interface can be detected online according to the type of the event file or the code content acquired from the online programming interface.

The method comprises the steps that code content in the online programming interface can be detected in real time in the process that a user inputs codes into the online programming interface, when the content in the online programming interface is detected to be changed, the code content in the online programming interface can be packaged into a file to be detected (the type of the file to be detected is the same as that of an event file, and the event file can be a file to be detected), and the code content in the file to be detected is detected online.

As shown in fig. 6, the process of online detection may include the steps of: S26-S28.

S26: and responding to the code writing operation of the online programming interface, and calling a preset language service middleware according to the type of the event file or the code content acquired from the online programming interface in the code writing process.

S27: and calling the language service module through the language service middleware, and detecting the code content in the online programming interface through the language service module.

S28: and displaying the code detection result in the online programming interface.

As one implementation mode, the type of the file can be determined according to the suffix name of the file, so that the language type corresponding to the code content in the file is determined, and after the language type is determined, the preset language service middleware can be called.

As another implementation manner, the code content acquired in the online programming interface may be identified according to some keywords in the code content, for example, the keywords may be function names used in some specific programming languages, and the language type corresponding to the code content may also be determined by a keyword identification manner.

Illustratively, the preset language service middleware may include an API (Application Programming Interface), an http protocol, and a WebSocket communication tool.

The language service module comprises a lexical analyzer and a syntactic analyzer. The lexical analyzer can be used for carrying out lexical detection on the code content in the current online programming interface, and the grammar analyzer can be used for carrying out grammar detection on the code content in the current online programming interface.

The language service module supports at least one of Java language, Python language, Jcc language, C + + language and Groovy language.

For the determined target event, because the same file requires the same language to be written in codes, only the language service module of one language is called in the same online programming process.

The lexical analyzer and the syntax analyzer can be regarded as encapsulated sub-modules in the existing code editor, and in the prior art, the sub-modules are not generally split for use, but relevant components of the whole code editor are downloaded and installed to a local terminal for use. However, in the embodiment of the present application, a lexical analyzer and a parser corresponding to different languages are provided in advance for different programming languages in the decision engine system, and an integrated language environment capable of supporting multiple languages is provided (i.e., the embodiment of the present application provides an online IDE). For a user, a complex code editor does not need to be downloaded and installed by self, only the code content in the online programming interface needs to be concerned, and the method can automatically call the lexical analyzer and the syntactic analyzer of the corresponding programming language according to the code content in the online programming interface.

The process of calling the lexical analyzer and the syntax analyzer can be regarded as that the code content in the online editing interface is distributed to the execution environment of the system background for compiling and detecting.

Optionally, the code detection result may include: at least one of a keyword hint flag, a code completion hint, and an exception code hint.

The keyword hint flag may be highlighted in a set color (e.g., yellow, red, green, etc.), and the keyword may be an undefined parameter that appears for the first time, for example.

The prompting content of the code completion prompt can be determined based on the detection results of the lexical analyzer and the grammar analyzer, selectable code fields or grammars can be recommended through the code completion function according to the code content input by the user, and the recommended code fields or grammars are usually high-frequency use words obtained through statistics, or special expression content under some specific grammars, or built-in functions appointed by programming grammars. In some embodiments, the code completion function may be omitted if efficiency of interaction is a concern.

The prompt content of the abnormal code prompt can be determined based on the detection results of the lexical analyzer and the syntactic analyzer, and can be the content of character missing, character errors, grammar errors and the like determined by the lexical analyzer and the syntactic analyzer.

Therefore, the method can support keyword prompt, code automatic completion or abnormal code detection of online programming.

In an application scene, a user can perform online programming only by opening a browser, does not need to download and install a complex code editor, can write codes at any time and any place, and directly enters a development state. Illustratively, a user can open an online editor of a web page version, the online editor is used for providing an online programming interface, a file to be detected can be generated in real time based on code content in the online editor, and the file to be detected can also be generated in real time for the code content in the online editor according to a set detection period. And then calling a language service middleware at the rear end based on the file type or the code content of the file to be detected, calling a corresponding language service module by the language service middleware according to the identified programming language and the identified programming language, carrying out online compilation, language inspection, lexical analysis and grammatical analysis on the code content of the file to be detected by the language service module to obtain an online detection result, and generating highlight keyword prompt content, code completion prompt content and error code prompt content based on the online detection result. And the online detection result obtained by the language service module is fed back to the front-end online editor through the language service middleware, and the front-end online editor displays the online detection result. In other application scenarios, the interaction process among the online editor, the language service middleware and the language service module can be regarded as the interaction process between the local terminal and the cloud host, and the user can connect the local terminal for inputting the code content to the cloud host and perform online programming and online code detection by managing resources in the cloud host.

Through the implementation mode, the corresponding language service middleware and the language service module can be automatically called according to the type of the event file or the code content obtained by online programming, so that online code detection is performed according to the called middleware and module, the code detection result can be timely displayed in an online programming interface, and for a user, various programming languages do not need to be learned deliberately, so that the user can timely correct the code through the code detection result displayed in the online programming interface. The method has strong compatibility, and can automatically call the corresponding language service module according to the programming language type used by the user during the online programming, thereby carrying out code language self-adaptive detection on the online coding process, and for the user, the method does not need to learn a plurality of programming languages, and can reduce the requirements for the user.

Optionally, the method may further include: S301-S303.

S301: and responding to the visual test operation of the decision flow to obtain test parameters.

And when a test button in the configuration operation interface of the decision flow is triggered, the visual test operation on the decision flow is considered to be initiated. The test parameters may be pre-imported parameters or temporarily recorded parameters. The test parameters may include parameters of various data types, and the test parameters may include, but are not limited to, user age value, user name content, partial field in the identification number, native place, number of repayment, etc. The specific test parameters can be set according to the actual service.

S302: and inputting the test parameters into the decision flow, and performing decision simulation on the test parameters through each conditional decision rule in the decision flow.

S303: in the process of decision simulation, according to the simulation sequence or the simulation state of each decision event in the decision flow, simulation marking is carried out, and the simulation marking result is displayed on the decision flow.

Optionally, a decision processing log may be generated for the decision simulation process.

For example, the simulation execution order of each decision event may be displayed on the decision flow, and the input parameters and the output parameters corresponding to each decision event may be displayed on the decision flow. The simulated, non-simulated and simulated decision events may be marked in different ways (e.g., in different colors).

By the implementation mode, the test simulation process of the decision flow can be visualized, a user can conveniently and quickly know the test state, the tracking of the simulation execution process of the decision flow is facilitated, and the test process can be transparent.

Optionally, the method may further include: and displaying the rule summary of each decision event on the configured decision flow. The rule summary content may be part of the fields, variables in the conditional decision rule.

Based on the same inventive concept, the embodiment of the application also provides a service decision method. The method can be applied to the decision engine system of the embodiment of the application.

As shown in fig. 7, the method may include: steps S31-S33.

S31: and acquiring a to-be-processed parameter corresponding to the target service through a preset data interface.

The parameters to be processed can be obtained through a data interface arranged in a data access layer of the decision engine system. The data type of the parameter to be processed is related to the actual service type, and may be the same as the data type of the test parameter. The parameters to be processed may include data obtained from an internal data interface of the enterprise (e.g., financial assets, user names, etc.), and may also include data provided by external systems external to the enterprise (e.g., financial data for third party credit, transaction data provided by a union bank, administrative data provided by a business institution or tax institution, etc.).

S32: and preprocessing the parameters to be processed to obtain the parameters to be decided, wherein the preprocessing comprises data cleaning.

The parameters to be processed can be preprocessed through a data calculation layer in the decision engine system, for example, offline calculation or real-time calculation can be performed on various data sources from a heterogeneous system, so that feature extraction, data cleaning and index conversion calculation are performed on the parameters to be processed. Illustratively, the metric conversion calculation may be to convert the data content into a parameter to be decided that meets the parameter type requirement of the decision flow.

S33: and inputting the parameters to be decided into the decision flow configured by the decision configuration method, and performing decision operation on the parameters to be decided according to the conditional decision rules of each decision event configured in the decision flow to obtain a decision result corresponding to the parameters to be decided.

Wherein S33 may be performed by a decision engine layer in the decision engine system.

By the method, the obtained decision flow can be configured based on the decision configuration method, and a decision result is rapidly provided for the to-be-processed parameters corresponding to the target service.

In an application scenario, a user can perform editing configuration, simulation or execution of a decision flow based on a decision engine layer in a decision engine system, can configure a conditional decision rule for a single decision event in the decision flow, and can implement various configurations for the decision flow based on a configuration mode of graph editing, wherein configuration contents include, but are not limited to, adding events, adding variables, assigning variables, adding branches (adding edges, implementing node connection, the connection refers to unidirectional connection, that is, connection can only be from front to back, but cannot be from back to front), configuring input data and output data types, formats and the like of each event.

When a user newly establishes a decision flow in a configuration operation interface of the decision flow, the initial position of the decision flow generates a starting node by default. Based on the starting node, the user can create a new node (corresponding to a new event, each node can be regarded as a decision event), create a new edge (corresponding to a new branch), and construct a regular expression of each node in a drag configuration manner on the operation interface according to the actual service decision logic.

Each regular expression can be regarded as a conditional decision rule, and the expressions support numerical calculation, logic calculation and text calculation. The user can configure a plurality of nodes on the operation interface, and the user can configure a plurality of nodes at will, for example, can select one node to configure all rules, or can select rules for configuring the rules of one node into a plurality of nodes in a split manner, for complex business logic, if the configuration mode of dragging and pulling is not enough to meet the configuration requirement, the code can be written by using an online IDE to realize flexible configuration, or a simple text editing mode (visible as offline programming) for importing scripts can be adopted to configure.

Each decision flow has an independent output result, branches in each decision flow can have independent output results, and the output result of the whole decision flow is the decision result. The input parameters of the decision flow are derived from a data access layer and a data calculation layer.

After the configuration of the decision flow is completed, simulation and execution can be performed based on a decision engine layer, and the simulation process and the execution process are the same in principle and different only in the adopted data source.

The simulation and execution process can comprise the following steps: taking a snapshot of the decision flow (the snapshot is generated after the decision flow configuration is completed), and restoring configuration data of each node in the decision flow based on the snapshot of the decision flow (the configuration data is used for describing conditional decision rules of each decision event in the decision flow). And assembling to obtain a complete decision flow according to the conditional decision rules configured for each node in the decision flow. Then, the input data (as a data source) actually used in the decision flow can be read, the child nodes are sequentially added into the node execution queue from the starting node of the decision flow, and whether all the parent nodes of each node in the queue are executed or not is judged. Taking a target node in a decision flow as an example (the target node corresponds to a target event in the decision flow), if all father nodes of the target node are executed, executing operations such as rules, expressions, variable assignments and the like in the target node (i.e., executing a conditional decision rule corresponding to the target event), sequentially adding all child nodes of the target node into an execution node queue, and sequentially performing recursion until all father nodes and child nodes associated with the target node are executed (i.e., until the unexecuted node queue is empty). The target node may be any node in the decision flow.

The final decision engine output result may be assembled and output according to the output content (e.g., field format, field value) of the terminal child node of each branch of the decision flow, and the final decision engine output result may be used as a decision result, that is, the overall decision result of the entire decision flow may be obtained according to the combination of the decision results of the plurality of child nodes.

When the simulation is carried out, parameters for carrying out simulation test can be input or imported into the selected or called decision flow, a user can click a test button on an operation interface, so that a test starting process is triggered, and whether each node passes or not is calculated according to the conditional decision rule actually configured for each node in the decision flow. And judging that the corresponding input data is executed even if the result corresponding to the corresponding input data is determined to be 'failed' according to the actual conditional decision rule.

The calculation result (whether executed or passed), the calculation rule, and the intermediate variable of each node may be stored, and a decision processing log may be generated based on the calculation result, the calculation rule, and the intermediate variable. The input data and the output result corresponding to each node can be displayed in the simulation test process, so that a user can know the simulation condition of each node in time.

Based on the same inventive concept, the application also provides a decision engine system.

The decision engine system has an operation processing capability, and as shown in fig. 8, an implementation structure of the decision engine system may include a memory 801, a processor 802 and a communication bus 803.

The decision engine system can be used to implement the aforementioned decision configuration method or business decision method. The decision engine system can be deployed with the functional modules in the decision engine layer.

The communication bus 803 is used to enable direct or indirect connection between the various components in the decision engine system.

The memory 801 is a storage medium, and may be a high-speed RAM memory or a non-volatile memory (non-volatile memory).

The Processor 802 has an arithmetic Processing capability, and may be, but is not limited to, a general-purpose Processor such as a Central Processing Unit (CPU) or a Network Processor (NP); but may also be a dedicated processor or a processor built from other programmable logic devices. The processor 802 may implement the methods, steps, and logic blocks provided by embodiments of the present application.

The memory 801 stores a computer program executable by the processor 802, and the processor 802 is configured to execute the computer program stored in the memory 801, so as to implement some or all of the steps of the method provided by the foregoing embodiments.

It should be noted that the structure shown in fig. 8 is only an illustration, and there may be more components or other configurations different from those shown in fig. 8 in specific applications.

In summary, by the method and the system provided by the embodiment of the application, the rule configuration of the decision engine can be realized at the cloud, a decision engine solution with high universality and compatibility is provided, the configuration difficulty is reduced in a drag-and-pull configuration mode, the configuration process of the decision flow is more flexible, the calculation rules are rich, the rapid compiling of a complex logic expression is supported in the decision flow configuration process, the calculation process of the decision flow can be executed at the cloud, and the complex environment deployment of a user at a local terminal is not needed. The scheme provided by the embodiment of the application supports cloud function computing, can realize cross-platform and cross-equipment online programming, and can detect the written code. The processing process and the processing result of the whole decision flow can be visualized, the decision result, the intermediate result and the calculation rule of the whole decision flow can be dynamically displayed on the flow node corresponding to the corresponding event, and the nodes which do not meet the decision rule (namely, the events which do not meet the conditional decision rule) can be specially marked, so that the user can conveniently locate the problem in time. For complex service logic, online code writing according to services is supported, a set of online version of IDE is realized, a user can open a browser to write codes and scripts, code detection results are provided for code contents written by the user, and functions of code marking prompt, code automatic completion and the like can be realized.

In the decision configuration method provided by the embodiment of the application, the DSL-based expression calculation engine function can be realized, common numerical calculation, logic calculation, text calculation and logic operation used in combination are supported, and the nested combination of expressions is supported, so that a user can write the expressions one line and perform rule calculation, and the processing difficulty of the user is reduced.

When an online IDE function is embedded in a decision engine system, online editing and compiling of an execution code can be realized, so that not only is high grammar brightness supported, but also real-time error prompt and automatic completion prompt of the code are supported, and a plurality of programming languages are supported, cloud function execution can be supported for a code file which is actually written, the online written code is processed by calling a language service module, and a user only needs to write the code by adopting a language familiar to the user.

The embodiment of the application also provides a cloud function, which can automatically identify the programming language of the complex logic code programmed by a user, package the content of the programmed code into files, distribute the files to a cloud function executor (the cloud function executor comprises language service modules and functions supporting multiple programming languages), and automatically and dynamically execute corresponding functions according to the programming language type corresponding to the code, so that the code is executed and compiled online. Variables, characteristics, models, algorithms, indexes and the like which are actually needed can be called in the cloud function.

In the embodiments provided in the present application, it should be understood that the disclosed embodiments may be implemented in other ways. The above-described system embodiments are merely illustrative, and for example, a division of modules is merely a logical division, and an actual implementation may have another division, and for example, a plurality of modules or components may be combined or integrated into another system. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment. In addition, the functional modules in the embodiments of the present application may be integrated according to actual needs.

It should be noted that, if the functions corresponding to the above methods are implemented in the form of software functional modules and sold or used as independent products, the functions may be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, or portions thereof, which substantially or substantially contribute to the prior art, may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device to perform all or part of the steps of the methods of the embodiments of the present application.

In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions.

The above embodiments are merely examples of the present application and are not intended to limit the scope of the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method of decision configuration, the method comprising:

2. The method of claim 1, further comprising:

inputting the test parameters into the decision flow, and performing decision simulation on the test parameters through a conditional decision rule of each decision event in the decision flow;

3. The method of claim 1, further comprising:

4. The method of claim 3, wherein after displaying the online programming interface, the method further comprises:

generating an event file for the target event;

5. The method according to claim 4, wherein the online detection of the code content in the online programming interface according to the type of the event file or the code content obtained from the online programming interface comprises:

and displaying a code detection result in the online programming interface.

6. The method of claim 5, wherein the code detection result comprises: at least one of a keyword hint flag, a code completion hint, and an exception code hint.

7. The method of claim 5, wherein the language service module comprises a lexical analyzer and a parser, and wherein the language service module supports at least one of Java, Python, Jcc, C + + and Groovy languages.

8. The method of claim 1, wherein determining, in response to a first configuration operation on a target event in the decision stream, a target parameter term and a target rule operator according to the first configuration operation comprises:

9. A method for business decision, the method comprising:

inputting the parameter to be decided into a decision flow configured by the method of any one of claims 1 to 8, and performing decision operation on the parameter to be decided according to a conditional decision rule of each decision event configured in the decision flow to obtain a decision result corresponding to the parameter to be decided.

10. A decision engine system, comprising:

a memory;

a processor;

the memory has stored thereon a computer program which, when executed by the processor, performs the method of any of claims 1-9.