CN117033664B

CN117033664B - Service sequence diagram generation method, device, computer equipment and storage medium

Info

Publication number: CN117033664B
Application number: CN202311271665.1A
Authority: CN
Inventors: 张镇鸿
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2023-09-28
Filing date: 2023-09-28
Publication date: 2024-01-09
Anticipated expiration: 2043-09-28
Also published as: CN117033664A

Abstract

The present application relates to a sequence diagram generation method, apparatus, computer device, storage medium and computer program product for a service. The method can be applied to artificial intelligence, such as a scene of generating a sequence diagram of a service through an intelligent question-answer model; the method comprises the following steps: searching knowledge data of the service based on knowledge inquiry information of the service; acquiring entity inquiry information according to the knowledge data, and inquiring entity data corresponding to each business entity based on the entity inquiry information; a business entity is an entity involved in executing a business; acquiring flow inquiry information corresponding to the service according to the data of each entity, and determining interaction data among the service entities according to the flow inquiry information; and generating a sequence chart for representing interaction of the business entities in the business execution process according to the interaction data among the business entities. By adopting the method, the generation efficiency of the sequence diagram can be improved.

Description

Service sequence diagram generation method, device, computer equipment and storage medium

Technical Field

The present invention relates to the field of computer technologies, and in particular, to a method, an apparatus, a computer device, a storage medium, and a computer program product for generating a sequence chart of a service.

Background

The sequence diagram is constructed based on interaction behaviors among the entities in the business process, can represent the dynamic process of completing the business among the entities through interaction information, and can reflect the execution logic of the business process.

In the related art, a business process sequence diagram is manually constructed by an expert in each business field, resulting in low efficiency in constructing the business process sequence diagram.

Disclosure of Invention

In view of the foregoing, it is desirable to provide a method, an apparatus, a computer device, a computer readable storage medium, and a computer program product for generating a sequence chart of a service, which can improve the efficiency of generating the sequence chart.

In a first aspect, the present application provides a method for generating a sequence diagram of a service. The method comprises the following steps:

searching knowledge data of the service based on knowledge inquiry information of the service; acquiring entity inquiry information according to the knowledge data, and inquiring entity data corresponding to each business entity based on the entity inquiry information; a business entity is an entity involved in executing a business; acquiring flow inquiry information corresponding to the service according to the data of each entity, and determining interaction data among the service entities according to the flow inquiry information; and generating a sequence chart for representing interaction of the business entities in the business execution process according to the interaction data among the business entities.

In a second aspect, the present application further provides a service sequence diagram generating device. The device comprises:

the knowledge data searching module is used for searching knowledge data of the service based on knowledge inquiry information of the service;

the entity data query module is used for obtaining entity query information according to the knowledge data and querying entity data corresponding to each business entity based on the entity query information; a business entity is an entity involved in executing a business;

the interactive data determining module is used for obtaining flow inquiry information corresponding to the service according to the data of each entity and determining interactive data among the service entities according to the flow inquiry information;

and the sequence diagram generating module is used for generating a sequence diagram for representing interaction of the business entities in the business execution process according to the interaction data among the business entities.

In some embodiments, the knowledge data searching module is further configured to generate knowledge query information corresponding to the service according to the query template; and processing knowledge inquiry information through a knowledge inquiry and answer model to obtain knowledge data of the business.

In some embodiments, the entity data query module is further configured to process the entity query information through an entity question-answer model to obtain entity data sets belonging to different layers; each entity data set comprises entity data corresponding to business entities belonging to the same layer;

The interactive data determining module is also used for acquiring flow inquiry information belonging to different layers according to each entity data set and determining interactive data among business entities in each layer of entity data set according to each flow inquiry information;

and the sequence diagram generation module is also used for respectively generating sequence diagrams belonging to different layers according to the interactive data among the business entities in the entity data set.

In some embodiments, the entity question-answering model comprises a first question-answering model and a second question-answering model, each entity data set comprises an entity data set belonging to a first level and an entity data set belonging to a second level, and the entity query information is first entity query information; the entity data query module comprises: a first entity data query unit and a second entity data query unit;

the first entity data query unit is used for processing the first entity query information through the first question-answer model to obtain an entity data set belonging to a first layer;

a second entity data query unit, configured to generate second entity query information according to each entity data included in the entity data set belonging to the first layer; and processing the second entity inquiry information through a second question-answer model to obtain an entity data set belonging to a second level.

In some embodiments, the interaction data determination module includes:

a first target entity data set determining unit, configured to traverse each entity data set to obtain a target entity data set belonging to a target hierarchy;

the target flow inquiry information determining unit is used for generating target flow inquiry information corresponding to the service according to the target entity data set;

the target flow data determining unit is used for processing the target flow query information through the flow question-answer model to obtain target flow data of the service;

and the interaction data determining unit is used for determining interaction data among the business entities belonging to the target hierarchy according to the target flow data.

In some embodiments, the interactive data determining unit is further configured to divide the target flow data according to each service entity in the target entity data set, so as to obtain service responsibility data executed by each service entity in the target entity data set and a time sequence identifier of each service responsibility data; and determining interaction data among the business entities belonging to the target hierarchy according to the business responsibility data of the business entities and the time sequence identification of the business responsibility data.

In some embodiments, the sequence diagram generation module comprises:

A second target entity data set determining unit, configured to traverse each entity data set to obtain a target entity data set belonging to a target hierarchy;

a lifeline generating unit for generating lifelines of each business entity in the target entity data set;

the message line generating unit is used for generating message lines corresponding to each interaction data according to the interaction data among the business entities in the target entity data set;

and the sequence diagram generating unit is used for generating a sequence diagram belonging to the target hierarchy according to the lifelines and the message lines corresponding to the interaction data.

In some embodiments, the message line generating unit is further configured to traverse interaction data between each service entity in the target entity data set to obtain target interaction data; the target interaction data is interaction data between the first service entity and the second service entity; and generating a message line corresponding to the target interaction data between the lifeline corresponding to the first service entity and the lifeline corresponding to the second service entity.

In some embodiments, entity data corresponding to each business entity is obtained through an entity question-answer model; the service sequence diagram generating device also comprises an entity question-answer model acquisition module;

The entity question-answering model acquisition module is used for acquiring sample knowledge data and each entity data label; generating sample entity inquiry information according to the sample knowledge data; processing the sample entity inquiry information through a basic question-answer model to obtain the data of each training entity; and adjusting parameters of the basic question-answering model according to the training entity data and the entity data labels to obtain the entity question-answering model.

In some embodiments, the entity question model includes a first question model and a second question model; the basic question-answering model comprises a first basic question-answering model and a second basic question-answering model; the entity data labels comprise a first entity data label belonging to a first layer and a second entity data label belonging to a second layer; the training entity data comprises first training entity data belonging to a first layer and second training entity data belonging to a second layer; the sample entity query is a first sample entity query; the entity question-answering model acquisition module comprises:

the training entity data determining unit is used for processing the first sample entity inquiry information through the first basic question-answer model to obtain first training entity data; generating second sample entity inquiry information according to the first training entity data; processing the second sample entity inquiry information through a second basic question-answer model to obtain second training entity data;

The questioning and answering model acquisition unit is used for adjusting parameters of the first basic questioning and answering model according to the first training entity data and the first entity data labels to obtain a first questioning and answering model; and adjusting parameters of the second basic question-answering model according to the second training entity data and the second entity data labels to obtain a second question-answering model.

In some embodiments, the interaction data between the business entities is determined by the flow data corresponding to the business, and the flow data is obtained by a flow question-answer model; the service sequence diagram generating device also comprises a flow question-answer model acquisition module;

the flow question-answer model acquisition module is used for acquiring entity data of each sample and flow data labels; generating training process inquiry information according to the entity data of each sample; processing the training process inquiry information through a third basic question-answer model to obtain training process data; and adjusting parameters of the third basic question-answering model according to the flow data labels and the training flow data to obtain the flow question-answering model.

In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:

In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:

In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the steps of:

The method, the device, the computer equipment, the storage medium and the computer program product for generating the sequence diagram of the business are characterized in that knowledge data of the business are searched through knowledge inquiry information of the business, entity inquiry information is obtained through the knowledge data, entity inquiry information is obtained according to the knowledge data, entity data corresponding to business entities are inquired according to the entity inquiry information, flow inquiry information is obtained according to the entity data corresponding to the business entities, interaction data among the business entities are determined according to the flow inquiry information, and the sequence diagram for representing interaction of the business entities in the business execution process is generated according to the interaction data among the business entities; by means of the question-answer mode, knowledge data of the service are determined step by step, service entities and entity data for executing the service are extracted, interaction data among the service entities in the process of executing the service are determined, so that a sequence chart for interaction in the process of executing the service is generated according to the interaction data among the service entities.

Drawings

FIG. 1 is an application environment diagram of a method for generating a sequence diagram of a service in one embodiment;

FIG. 2 is a flow diagram of a method for generating a sequence diagram of a service in one embodiment;

FIG. 3 is a flow diagram of a sequence diagram for generating a palm-brushing service in one embodiment;

FIG. 4 is a schematic diagram of a sequence diagram of a palm-brushing service in one embodiment;

FIG. 5 is a schematic diagram of a sequence diagram belonging to a first hierarchy in one embodiment;

FIG. 6 is a schematic diagram of a sequence diagram belonging to a second hierarchy in one embodiment;

FIG. 7 is a flow diagram of one embodiment of obtaining an entity dataset belonging to a second hierarchy;

FIG. 8 is a schematic diagram of obtaining entity data of each business entity belonging to a second hierarchy through a first question-answer model and a second question-answer model in one embodiment;

FIG. 9 is a flow diagram of determining interaction data between business entities, in one embodiment;

FIG. 10 is a flow diagram of a sequence diagram for determining a hierarchy belonging to a target hierarchy in one embodiment;

FIG. 11 is a schematic diagram of a lifeline of a business entity in one embodiment;

fig. 12 is a flowchart of a sequence chart for generating a palm-brushing service in another embodiment;

FIG. 13 is a flowchart of a method for generating a sequence diagram of a service according to another embodiment;

FIG. 14 is a block diagram of a sequence diagram generation apparatus of a service in one embodiment;

fig. 15 is an internal structural view of a computer device in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.

Artificial intelligence (Artificial Intelligence, AI) is the theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use the knowledge to obtain optimal results. In other words, artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence, i.e. research on design principles and implementation methods of various intelligent machines, enables the machines to have functions of sensing, reasoning and decision.

The artificial intelligence technology is a comprehensive subject, and relates to the technology with wide fields, namely the technology with a hardware level and the technology with a software level. Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and other directions.

The method for generating the sequence diagram of the service provided by the embodiment of the application can be applied to an application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104, or may be placed on a cloud or other network server; the sequence diagram generating method of the service may be executed by the terminal 102 or the server 104, or may be executed in cooperation with the terminal 102 and the server 104.

Taking the service sequence diagram generation method as an example, the server 104 can search the knowledge data of the service based on the knowledge inquiry information of the service; the server 104 may obtain entity query information according to the knowledge data, and query entity data corresponding to each service entity based on the entity query information; a business entity is an entity involved in executing a business; the server 104 may obtain flow query information corresponding to the service according to the data of each entity, and determine interaction data between each service entity according to the flow query information; the server 104 may generate a sequence chart for representing interactions of the business entities during the execution of the business according to the interaction data between the business entities.

The terminal 102 may be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, an internet of things device, and a portable wearable device, and the internet of things device may be a smart speaker, a smart television, a smart air conditioner, and a smart vehicle device. The portable wearable device may be a smart watch, smart bracelet, headset, or the like.

The server 104 may be a separate physical server or may be a service node in a blockchain system, where a peer-to-peer network is formed between the service nodes.

The server 104 may be a server cluster formed by a plurality of physical servers, and may be a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

The terminal 102 and the server 104 may be connected by a communication connection manner such as bluetooth, USB (Universal Serial Bus ) or a network, which is not limited herein.

In some embodiments, as shown in fig. 2, a method for generating a sequence chart of a service is provided, where the method is performed by a server or a terminal in fig. 1, and may also be performed by the server and the terminal in fig. 1 cooperatively, where the method is illustrated by way of example by the server in fig. 1, and includes the following steps:

step 202, searching knowledge data of the business based on knowledge inquiry information of the business.

The service is a service executed by an application program, and the service executed by the application program refers to a process involved in executing the service and can be realized by the application program. In practical applications, the service may be a payment service, a search service, a resource transfer service.

Wherein the knowledge inquiry information is a question for inquiring about knowledge related to a service; for example, knowledge inquiry information is used to represent: what a certain service is.

The knowledge inquiry information may be at least one of text, picture, and audio; for example, the knowledge inquiry information may be a question expressed by text, a question expressed by text and pictures together, or a question expressed by audio.

The knowledge data is knowledge content related to the service, and the knowledge data can comprise knowledge content such as definition of the service, key technical points of the service and the like.

In some embodiments, the knowledge inquiry information is used for inquiring about knowledge related to the service, the knowledge data is used for explaining the knowledge related to the service, so that the knowledge data of the service can be searched for based on the knowledge inquiry information of the service, and the knowledge inquiry information can be obtained through inquiry, that is, the server replies the knowledge data of the service when acquiring the knowledge inquiry information of the service, and the knowledge data is used for replying to the knowledge inquiry information.

In practical application, the server can process knowledge inquiry information through a knowledge inquiry model to obtain knowledge data.

Illustratively, when the knowledge inquiry information of the service is a text, the server acquires the text, and processes the text through a knowledge inquiry model to obtain knowledge data of the service; for example, knowledge inquiry information is: "what the payment business is", knowledge data of the business obtained by the question-answer model includes definition of the payment business and key technical points related to the payment business.

When the knowledge inquiry information of the service is a picture and a text, the server extracts the text corresponding to the picture, splices the text and the extracted text to obtain a spliced text, and processes the spliced text through a knowledge inquiry model to obtain knowledge data of the service; for example, the knowledge inquiry information includes text that is: "what the payment business is," the text extracted from the picture included in the knowledge inquiry information is: "Payment by brush palm", splice text: "what the payment business is, the palm payment", the knowledge data of the business that the question-answer model obtains, including the definition of the palm payment and key technical essential that the palm payment involves, the key technical essential that the palm payment involves includes: biometric technology, user registration, payment process, security, application domain, etc.

When the knowledge inquiry information of the service is audio information, the server extracts text corresponding to the audio information, and processes the extracted text by indicating a question-answer model to obtain knowledge data of the service. For example, knowledge inquiry information includes audio, and text extracted from the audio is: "what the payment service is," the knowledge question-answering model determines knowledge data of the payment service from the text.

In some embodiments, the process of obtaining the knowledge question-answering model includes: acquiring sample knowledge inquiry information and a knowledge data tag corresponding to the sample knowledge inquiry information, inputting the sample knowledge inquiry information into an initial question-answer model to obtain training knowledge data, determining a first loss value according to the training knowledge data and the knowledge data tag, adjusting parameters of the initial question-answer model according to the first loss value, and repeating the step of adjusting the parameters of the initial question-answer model until the initial question-answer model is converged under the condition that the initial question-answer model is not converged, wherein the converged initial question-answer model is used as the knowledge question-answer model.

The sample knowledge inquiry information is generated according to the service field, wherein the service field is a service related field, and the knowledge data label is knowledge of the service field; the initial question-answering model has knowledge question-answering capability in the general field, and the initial question-answering model can be a question-answering model obtained by pre-training query information and knowledge in the general field; and carrying out downstream training on the initial question-answer model through sample knowledge query information and knowledge data labels in the service field, wherein the initial question-answer model can learn knowledge in the service field in the training process, so that the converged initial question-answer model has knowledge question-answer capability in the service field, and can process knowledge query information of the service to obtain knowledge data of the service.

Step 204, obtaining entity inquiry information according to the knowledge data, and inquiring entity data corresponding to each business entity based on the entity inquiry information; a business entity is an entity involved in executing a business.

Wherein the entity inquiry information is a question for inquiring about a business entity contained in the knowledge data; for example, entity query information is used to represent: the knowledge data includes which entities are executing the service.

The entity involved in executing the service when the service entity is a palm payment service, and the service entity involved in executing the palm payment includes: a payment object, a palm graph identification class and a payment class; illustratively, the service is a virtual resource transfer, and the entity involved in performing the virtual resource transfer includes: a first resource holding object, a second resource holding object, a biometric class, and a resource transfer class.

The palm image recognition class is a class in which a method for recognizing a palm image is packaged, the payment class is a class in which a method for executing payment is packaged, and the resource transfer class is a class in which a method for executing resource transfer is packaged.

The entity data of the business entity may include descriptive contents of the business entity for explaining the role of the business entity in the execution of the business. Illustratively, the business entity is a payment object, and the entity data is for providing a palm image, performing a determining operation, etc.; the business entity is palm graph identification class, and the entity data is used for identifying the palm graph.

The business entity is biological characteristic identification equipment, and entity data is identity information for identifying living beings.

In some embodiments, the server may obtain an entity query template, embed knowledge data into the entity query template, and obtain entity query information. Illustratively, the entity query template may be: "please extract business entity from < >", entity query template "< >" is used to embed knowledge data.

In some embodiments, the entity query information may also be directly obtained by the server, for example, the generator of the sequence chart generates the entity query information according to the knowledge data, and inputs the entity query information to the server, so that the server obtains the entity query information generated according to the knowledge data.

In some embodiments, the server may process knowledge inquiry information through an entity question-answer model to obtain entity data corresponding to each service entity; illustratively, the server inputs knowledge inquiry information into an entity question-answer model, and each business entity is extracted from the knowledge inquiry information through the entity question-answer model.

In some embodiments, the process of obtaining an entity question-answer model includes: obtaining sample knowledge data and each entity data label corresponding to the sample knowledge data, generating sample entity inquiry information according to the sample knowledge data, processing training entity inquiry information through a basic question-answer model to obtain each training entity data, determining a second loss value according to each training entity data and each entity data label, adjusting parameters of the basic question-answer model according to the second loss value, and repeating the step of adjusting the parameters of the basic question-answer model until the basic question-answer model is converged and taking the converged basic question-answer model as the entity question-answer model under the condition that the basic question-answer model is not converged.

The sample knowledge data is knowledge in the business field, the basic question-answering model has entity extraction capability in the general field, the basic question-answering model can be a question-answering model obtained by pre-training entity query information and entities in the general field, and in practical application, the basic question-answering model can be realized through GPT; and carrying out downstream training on the basic question-answer model through entity inquiry information corresponding to sample knowledge data in the service field and entity data labels, wherein in the training process, the basic question-answer model can learn the related knowledge of an execution entity in the service field, so that the converged basic question-answer model has the capability of extracting the execution entity in the service field, namely, the entity inquiry information corresponding to the service can be processed, and the service entity data is obtained.

Step 206, obtaining the flow inquiry information corresponding to the business according to the data of each entity, and determining the interactive data between each business entity according to the flow inquiry information.

The flow inquiry information is used for inquiring flow data of the execution service, for example, the flow inquiry information is used for indicating: according to the entity data of the business entity, determining what the flow data of the execution business is.

The interaction data between each business entity comprises the data of interaction between different two business entities and also comprises the interaction data between the same business entity; in practical application, the data interacted between two different business entities represents that one business entity transmits the task to the other business entity, and the other business entity continues to execute the task; interaction data between the same business entities, representing one method within a business entity invoking another method to perform a task.

It should be noted that, the execution flow of the service is formed by the data interacted between the service entities; for example, the execution flow of the palm payment service includes: the user provides the palm graph, the recognition device recognizes the palm graph, when the palm graph passes recognition, the recognition device sends a recognition passing message to the payment device, and the payment device executes payment.

In some embodiments, the server may obtain a flow query template, and embed entity data of each business entity into the flow query template to obtain flow query information. Illustratively, the flow query template may be: the entity data of the business entity comprises < >, the flow data of the execution business is what, and the < > "in the flow inquiry template is used for embedding the entity data of each business entity.

In some embodiments, the flow query information may also be directly obtained by the server, for example, the generator of the sequence chart generates the flow query information according to the entity data of each service entity, and inputs the flow query information to the server, so that the server obtains the flow query information generated according to the entity data of each service entity.

In some embodiments, the server determines flow data of the execution service according to the flow query information, extracts service responsibility data executed by each service entity from the flow data, and determines interaction data between the service entities according to the service responsibility data executed by each service entity.

In some embodiments, the server may process the flow query information through a flow query model to obtain flow data of the execution service; the server inputs the flow query information to a flow question-answer model, and outputs flow data of the execution service through the flow question-answer model.

Illustratively, the flow query information is: the "entity data of the service entity includes entity data of the user, entity data of the biological feature identification device and entity data of the payment device, what is the flow data of the payment service", and the flow question-answer model processes the flow query information to obtain flow data: the user provides the biological characteristics to the biological recognition device, the biological recognition device recognizes the biological characteristics, the biological recognition device sends a payment message to the payment device when the biological recognition device passes the recognition, and the payment recognition device executes payment according to the payment message; the determining interaction data between business entities through the flow data comprises: a biometric between the user and the biometric device, a biometric between the biometric device and the biometric device, a payment message between the biometric device and the payment device.

In some embodiments, the process of obtaining a flow question-answer model includes: obtaining entity data of each sample entity, obtaining training process query information according to the entity data of each sample entity, processing the training process query information through a third basic question-answer model to obtain training process data, calculating a third loss value according to a process data label corresponding to the training process query information and the training process data, adjusting parameters of the third basic question-answer model according to the third loss value, and repeatedly executing the process of adjusting the parameters of the third basic question-answer model according to the third loss value under the condition that the third basic question-answer model is not converged until the third basic question-answer model is converged, and taking the converged basic question-answer model as a process question-answer model; in practical applications, the third basic question-answering model may be implemented by GPT.

And step 208, generating a sequence chart for representing interaction of the business entities in the business execution process according to the interaction data among the business entities.

Wherein the sequence diagram is also referred to as a timing diagram, which comprises messages that interact between the business entities.

In some embodiments, the server generates lifelines corresponding to the business entities respectively; the method comprises the steps that interaction data among business entities exist in time sequence, a server traverses the interaction data among the business entities according to the time sequence to obtain target interaction data, a sending business entity and a receiving business entity corresponding to the target interaction data are determined, when the sending business entity and the receiving business entity are different business entities, a message line between a life line of the sending business entity and a life line of the receiving business entity is generated according to a first message line template, and the life line of the business entity points to the life line of the receiving business entity and is marked with the interaction data; when the sending business entity and the receiving business entity are the same business entity, generating a message line which points to the self life line by the life line of the business entity according to a second message line template, wherein the message line is a semi-closed self-association message line and is marked with interactive data.

According to the traversed target interaction data, the message line between the business entities can be determined; and obtaining a sequence chart according to the lifelines corresponding to the service entities and the message lines among the service entities.

In some embodiments, the server may generate a model according to the sequence diagram, process interaction data between each service entity to obtain life line information of the service entity and message line information between the service entities, and generate the sequence diagram according to the life line information of the service entity and the message line information between the service entities.

In some embodiments, the process of obtaining a sequence diagram build model includes: acquiring sample interaction data among sample entities, and lifeline information labels and message line information labels corresponding to the sample interaction data among the sample entities; sample interaction data among sample entities are processed through a base graph construction model, and training life line information and training message line information are obtained; calculating a fourth loss value according to the life line information label and the training life line information, calculating a fifth loss value according to the message line information label and the training message line information, adjusting parameters of the basic diagram building model according to the fourth loss value and the fifth loss value, and repeatedly executing the process of adjusting the parameters of the basic diagram building model according to the fourth loss value and the fifth loss value until the basic diagram building model converges, wherein the converged basic diagram building model is used as a sequence diagram building model under the condition that the basic diagram building model does not converge.

Illustratively, as shown in fig. 3, the service is a palm brushing service, and knowledge inquiry information of the palm brushing service is: the method for obtaining the knowledge data of the palm brushing service by the knowledge question-answering model comprises the following steps: the definition of palm-brushing business and the key technical points that the palm-brushing payment relates to include: biometric technology, user registration, payment process, security, application domain, etc. The server generates entity inquiry information according to knowledge data of the palm brushing service: "please extract service entity from knowledge data of the palm-brushing service", wherein the knowledge data of the palm-brushing service is content such as definition of the palm-brushing service obtained according to a knowledge question-answering model and key technical points related to palm-brushing payment, and the server inputs entity query information into the entity question-answering model, and the entity data of each service entity of the palm-brushing service obtained through the entity question-answering model, for example: a payment object, which is used for providing a palm image; palm image recognition class, which is used for recognizing the palm image; payment class, the role of which is to perform payment. The server generates flow inquiry information of the palm brushing service according to the entity data of each service entity, the flow inquiry information is input into a flow inquiry and answer model, the flow data for executing the palm brushing service is obtained through the flow inquiry and answer model, for example, a payment object executes the palm brushing operation, palm graphs are provided to a palm graph identification class, the palm graph identification class identifies the palm graphs so as to determine the identity information of the payment object, when the identity information obtained by identification is matched with the identity information prestored by the payment object, the palm graph identification class sends an identification passing message to the payment class, and the payment class executes payment according to the identification passing message; the server determines interaction data between the service entities according to the flow data, and comprises the following steps: palm image between the payment object and the palm image recognition class, palm image recognition related data between the palm image recognition class and the palm image recognition class; the recognition between the palm image recognition class and the payment class is through information, and the payment class execute the payment related data. The sequence diagram of the generated palm payment service is shown in fig. 4 according to the interactive data among the service entities. The sequence diagram can reflect interaction among business entities in the palm payment business and can be used for guiding development of application programs corresponding to the palm payment business.

In the method for generating the sequence diagram of the service, knowledge data of the service is searched through knowledge query information of the service, entity query information is obtained through the knowledge data, entity query information is obtained according to the knowledge data, entity data corresponding to service entities is queried according to the entity query information, flow query information is obtained according to the entity data corresponding to the service entities, interaction data among the service entities is determined according to the flow query information, and the sequence diagram for representing interaction of the service entities in the service execution process is generated according to the interaction data among the service entities; by means of the question-answer mode, knowledge data of the service are determined step by step, service entities and entity data for executing the service are extracted, interaction data among the service entities in the process of executing the service are determined, so that a sequence chart for interaction in the process of executing the service is generated according to the interaction data among the service entities.

In some embodiments, looking up knowledge data for a business based on knowledge query information for the business includes: generating knowledge inquiry information corresponding to the service according to the inquiry template; and processing knowledge inquiry information through a knowledge inquiry and answer model to obtain knowledge data of the business.

Wherein, the inquiry template can be: "what" is "< >" in the query template is used to embed the business.

In some embodiments, the server embeds the business into an inquiry template to obtain knowledge inquiry information, for example, the business is face payment, and the knowledge inquiry information is: "what face payment is"; the server inputs knowledge inquiry information into a knowledge question-answer model, and knowledge data of face payment is obtained through the knowledge question-answer model; for example, knowledge data of face payments include: definition of face payment and key technical points related to the face payment.

The sample knowledge inquiry information is generated according to the service field, wherein the service field is a service related field, and the knowledge data label is knowledge of the service field; the initial question-answering model has knowledge question-answering capability in the general field, and the initial question-answering model can be a question-answering model obtained by pre-training query information and knowledge in the general field; and carrying out downstream training on the initial question-answer model through sample knowledge query information and knowledge data labels in the service field, wherein the initial question-answer model can learn knowledge in the service field in the training process, so that the converged initial question-answer model has knowledge question-answer capability in the service field, and can process knowledge query information of the service to obtain knowledge data of the service. In practical applications, the initial question-answering model may be implemented by GPT, which is a pre-trained language model.

In the embodiment, the knowledge data of the service is determined through the knowledge question-answer model, and the knowledge data of the service is not required to be manually searched and screened, so that the efficiency of determining the knowledge data of the service is improved.

In some embodiments, querying entity data corresponding to each business entity based on entity query information includes: processing the entity inquiry information through an entity inquiry and answer model to obtain entity data sets belonging to different layers; each entity data set comprises entity data corresponding to business entities belonging to the same layer; acquiring flow inquiry information corresponding to the service according to the data of each entity, and determining interaction data among the service entities according to the flow inquiry information, wherein the method comprises the following steps: acquiring flow inquiry information belonging to different layers according to each entity data set, and determining interaction data among business entities in each entity data set belonging to each layer according to each flow inquiry information; generating a sequence chart for representing interaction of each business entity in the business execution process according to interaction data among the business entities, wherein the sequence chart comprises the following steps: and respectively generating sequence diagrams belonging to different levels according to the interactive data among the service entities in the entity data sets.

The level may be used to reflect the abstract degree of the service entity, where the higher the level is, the higher the abstract degree of the service entity is, and thus the more abstract the service entity is, the lower the level is, the lower the abstract degree of the service entity is, and thus the more concrete the service entity is.

Illustratively, the biometric features and the face image are business entities of different levels, wherein the degree of abstraction of the biometric features is higher than the degree of abstraction of the face image, obviously the face image is more specific than the biometric features; from another perspective, a business entity that is high in hierarchy is a generalized representation of a business entity that is low in hierarchy.

Entity data sets belonging to different levels, including entity data of each business entity belonging to different levels, for example, two different levels of entity data sets, including an entity data set belonging to a first level including entity data of each business entity belonging to the first level and an entity data set belonging to a second level including entity data of each business entity belonging to the second level.

The process query information belonging to different levels is determined according to entity data sets belonging to different levels, for example, the process query information of a first level is determined according to entity data of each business entity belonging to the first level, and the process query information of a second level is determined according to entity data of each business entity belonging to the second level.

Sequence diagrams belonging to different levels are determined according to the interactive data among the business entities in the entity data sets belonging to different levels, for example, the sequence diagrams belonging to the first level are determined according to the interactive data among the business entities in the entity data sets belonging to the first level, and the sequence diagrams belonging to the second level are determined according to the interactive data among the business entities in the entity data sets belonging to the second level.

In some embodiments, taking a service as a virtual resource transfer service as an example, a server processes knowledge inquiry information corresponding to the virtual resource transfer service through an entity question-answer model to obtain an entity data set belonging to a first level and an entity data set belonging to a second level, where the abstraction degree of the first level is higher than that of the second level, and the entity data set belonging to the first level includes: entity data of the user, entity data of the biometric feature identification device, and entity data of the resource transfer device; the entity data sets belonging to the second hierarchy include: entity data of the user, entity data of the palm image identification class, entity data of the living video identification class, and entity data of the resource transfer class. Wherein the living video recognition class is used for recognizing whether the resource transfer operation is performed by the user himself in an active state, rather than by another person performing the resource transfer operation through the video or picture of the user.

The biological characteristics belonging to the first layer are summarized and expressed by the palm graph and the living video belonging to the second layer, the biological characteristic identification equipment belonging to the first layer is summarized and expressed by the palm graph identification class and the living video identification class belonging to the second layer.

The server generates flow inquiry information belonging to the first layer according to the entity data set belonging to the first layer of the virtual resource transfer service, for example, the flow inquiry information belonging to the first layer of the virtual resource transfer service is: "entity data of a business entity includes: entity data of the user, entity data of the biometric identification device, and entity data of the resource transfer device, what is the flow data of the execution virtual resource transfer service.

The server generates flow inquiry information belonging to the second layer according to the entity data set belonging to the second layer of the virtual resource transfer service, for example, the flow inquiry information belonging to the second layer is: "entity data of a business entity includes: entity data of the user, entity data of the palm image identification class, entity data of the living video identification class, and entity data of the resource transfer class, what is the flow data of executing the virtual resource transfer service.

The server determines interaction data among the business entities belonging to the first hierarchy according to the flow query information belonging to the first hierarchy of the virtual resource transfer business, for example, the interaction data among the business entities belonging to the first hierarchy comprises: the method includes the steps of providing a biometric feature between a user and a biometric feature recognition device, providing data related to the biometric feature recognition between the biometric feature recognition device and the biometric feature recognition device, providing information related to the recognition between the biometric feature recognition device and a resource transfer device, and providing data related to the execution of the resource transfer between the resource transfer device and the resource transfer device.

The server determines interaction data between the service entities belonging to the second hierarchy according to the flow query information belonging to the second hierarchy of the virtual resource transfer service, for example, the interaction data between the service entities belonging to the second hierarchy includes: palm image between user and palm image recognition class, living video between user and living video recognition class, related data for palm image recognition between palm image recognition class and palm image recognition class, related data for living video recognition between living video recognition class and living video recognition class, palm image recognition passing message between palm image recognition class and resource transfer device class, living video recognition passing message between living video recognition class and resource transfer device class, related data for resource transfer between resource transfer device and resource transfer device.

The server obtains a sequence diagram of virtual resource transfer belonging to the first layer according to the interactive data among the business entities belonging to the first layer, as shown in fig. 5; the server obtains a sequence diagram of virtual resource transfer belonging to the second hierarchy according to the interactive data among the service entities belonging to the second hierarchy, as shown in fig. 6.

In practical application, more entity data belonging to different levels of service entities may be extracted according to service requirements, so as to generate more sequence diagrams belonging to different levels, and the embodiment of generating the sequence diagrams belonging to two different levels according to extracting the entity data belonging to two different levels of service entities is merely an example and is not used for limiting the number of levels.

In the above embodiment, entity data of service entities belonging to different levels can be extracted through the entity question-answer model, so as to obtain interaction data among the service entities belonging to different levels, generate sequence diagrams belonging to different levels, and improve efficiency of generating the sequence diagrams belonging to different levels.

In some embodiments, before searching the knowledge data of the business based on the knowledge inquiry information of the business, the method further comprises: acquiring sample knowledge data and each entity data tag; generating sample entity inquiry information according to the sample knowledge data; processing the sample entity inquiry information through a basic question-answer model to obtain the data of each training entity; and adjusting parameters of the basic question-answering model according to the training entity data and the entity data labels to obtain the entity question-answering model.

The sample knowledge data is knowledge in the service field, and comprises service definition and service key points; each entity data tag is entity data of a sample business entity extracted from sample knowledge data.

The basic question-answering model has entity extraction capability in the general field, and can be a question-answering model obtained by pre-training entity inquiry information and entity data in the general field.

The training entity data are entity data of training service entities extracted from the sample knowledge data by the basic question-answer model.

In some embodiments, a server acquires sample knowledge data, embeds the sample knowledge data into an entity query template to obtain sample entity query information, inputs the sample entity query information into a basic question-answer model, extracts feature vectors of the sample entity query information through the basic question-answer model, and processes the feature vectors of the sample entity query information to obtain training entity data of each training service entity; the server acquires entity data labels of all sample service entities corresponding to the sample knowledge data, calculates a second loss value according to the entity data labels of all sample service entities and training entity data of all training service entities, adjusts parameters of a basic question-answering model according to the second loss value, and repeatedly executes the process of adjusting the parameters of the basic question-answering model according to the second loss value until the basic question-answering model converges and takes the basic question-answering model with cooling as the entity question-answering model under the condition that the basic question-answering model does not converge.

The second loss value is calculated according to the entity data label of each sample service entity and the training entity data of each training service entity, and the cross entropy loss value may be calculated according to the entity data label of each sample service entity and the training entity data of each training service entity.

In the above embodiment, the basic question-answering model has the entity extraction capability of the general field, and the downstream training is performed on the basic question-answering model through the sample knowledge data of the service field and the entity data labels of the corresponding sample service entities, and in the training process, the basic question-answering model can learn the relevant knowledge of the execution entities of the service field, so that the converged basic question-answering model has the capability of extracting the execution entities of the service field, namely, the entity query information corresponding to the service can be processed, and the service entity data is obtained.

In some embodiments, the entity question-answering model comprises a first question-answering model and a second question-answering model, each entity data set comprises an entity data set belonging to a first level and an entity data set belonging to a second level, and the entity query information is first entity query information; as shown in fig. 7, processing the entity inquiry information through the entity question-answer model to obtain entity data sets belonging to different layers includes: step 701, processing the first entity inquiry information through a first question-answer model to obtain an entity data set belonging to a first layer; step 702, generating second entity inquiry information according to each entity data included in the entity data set belonging to the first hierarchy; and step 703, processing the second entity inquiry information through the second question-answer model to obtain an entity data set belonging to the second layer.

The first question-answer model is used for extracting entity data of a service entity of a highest level, and in this example, the highest level is a first level, so in this example, the first question-answer model is used for extracting entity data of the service entity belonging to the first level according to first entity inquiry information.

The first entity inquiry information comprises knowledge data of the service and a first layer; generating first entity inquiry information according to knowledge data of the service, the first layer and the first entity inquiry template; illustratively, the first entity query template may be: "please extract <2> business entity from <1>, the" <1> "in the first entity query template is used for embedding knowledge data, and the" <2> "in the first entity query template is used for embedding the first hierarchy; for example, the service is face payment, and the first entity inquiry information is: please extract the business entity of the first level from the knowledge data of < face payment >; in the above example "< knowledge data of face payment >" of the first entity inquiry information is only used to represent the content here, and in practical application, here is specific knowledge data of face payment.

The second question-answering model is used for extracting entity data of business entities belonging to lower levels according to entity data of business entities belonging to higher levels, in the example, the first level is higher than the second level, and the second level is the intersection level compared with the first level, so in the example, the second question-answering model is used for extracting entity data of business entities belonging to the second level according to entity data of business entities belonging to the first level.

The second entity querying information includes: entity data of each business entity belonging to a first level and a second level; generating second entity inquiry information according to entity data of each business entity belonging to the first level, the second level and a second entity inquiry template; illustratively, the second entity query template may be: "please extract <4> business entity according to <3>, the" <3> "in the second entity inquiry template is used for embedding the entity data of each business entity belonging to the first hierarchy, and the" <4> "in the second entity inquiry template is used for embedding the second hierarchy; for example, the second entity query information is: extracting a second-level business entity according to < entity data of each business entity belonging to the first level >; in the above example, "< entity data of each business entity belonging to the first hierarchy >" of the second entity inquiry information is merely used to indicate the content here, and in practical application, the specific content here is the entity data of each business entity belonging to the first hierarchy.

In some embodiments, the server embeds knowledge data of the service and a first level into a first entity inquiry template to obtain first entity inquiry information, inputs the first inquiry information into a first inquiry model, obtains entity data of each service entity belonging to the first level through the first inquiry model, and obtains an entity data set belonging to the first level according to the entity data of each service entity belonging to the first level; the server embeds the entity data set belonging to the first layer and the second layer into a second entity inquiry template to obtain second entity inquiry information, inputs the second inquiry information into a second inquiry model, obtains the entity data of each business entity belonging to the second layer through the second inquiry model, and determines the entity data set belonging to the second layer according to the entity data of each business entity belonging to the second layer.

Illustratively, as shown in fig. 8, the business is a swipe payment, and the server generates first entity inquiry information of the swipe payment: "please extract the business entity of the first level from the knowledge data of the palm-brushing business", input the first entity inquiry information to the first question-answer model, get the entity data of each business entity belonging to the first level: "Payment object, which functions to provide a biometric feature; a biological feature recognition class for recognizing biological features; a payment device operative to perform payment "; the server generates second entity inquiry information according to the entity data of each business entity belonging to the first layer: "please provide the biometric feature according to the payment object; a biological feature recognition class for recognizing biological features; the payment equipment is used for executing payment, extracting a second-level business entity', and inputting second entity inquiry information into the second question-answer model to obtain entity data of each business entity belonging to the second level: "Payment object, function to provide palm image and live video; palm image recognition class, which is used for recognizing the palm image; the living body video identification class is used for identifying living body videos; payment class, the role of which is to perform payment.

After obtaining entity data of each business entity belonging to the first level and entity data of each business entity belonging to the second level according to the first question-answer model and the second question-answer model, obtaining flow query information belonging to the first level according to the entity data of each business entity belonging to the first level, determining interaction data among each business entity belonging to the first level according to the flow query information belonging to the first level, and generating a sequence chart belonging to the first level according to the interaction data among each business entity belonging to the first level; obtaining flow inquiry information belonging to the second level according to the entity data of each business entity belonging to the second level, determining interaction data among the business entities belonging to the second level according to the flow inquiry information belonging to the second level, and generating a sequence chart belonging to the second level according to the interaction data among the business entities belonging to the second level.

It should be noted that, the second question-answer model is used for extracting entity data of a service entity belonging to a lower level according to entity data of a service entity belonging to a higher level, in this embodiment, extracting entity data of a service entity belonging to a second level according to entity data of a service entity belonging to a first level, which is only taken as an example, and in a service requirement situation, the second question-answer model may also extract entity data of a service entity belonging to a third level according to entity data of a service entity belonging to a second level, where the abstraction degree of the third level is lower than that of the second level, and similarly, the second question-answer model may also extract entity data of a service entity belonging to a fourth level according to entity data of a service entity belonging to the third level.

For example, the server generates third entity inquiry information according to entity data of the service entity belonging to the second layer, and processes the third entity inquiry information through the second question-answer model to obtain an entity data set belonging to the third layer; the process query information belonging to the third layer can be obtained according to the entity data of each service entity belonging to the third layer, the interactive data among the service entities belonging to the third layer is determined according to the process query information belonging to the third layer, and the sequence diagram belonging to the third layer is generated according to the interactive data among the service entities belonging to the third layer.

In the above embodiment, the entity data of each business entity belonging to the first level is extracted from the knowledge data of the business through the first question-answer model, and the entity data of each business entity belonging to the second level is extracted through the second question-answer model according to the entity data of each business entity belonging to the first level, so that the sequence diagram belonging to the first level is generated according to the entity data of each business entity belonging to the first level, and the sequence diagram belonging to the second level is generated according to the entity data of each business entity belonging to the second level, thereby improving the efficiency of generating the sequence diagrams belonging to different levels.

In some embodiments, the entity question model includes a first question model and a second question model; the basic question-answering model comprises a first basic question-answering model and a second basic question-answering model; the entity data labels comprise a first entity data label belonging to a first layer and a second entity data label belonging to a second layer; the training entity data comprises first training entity data belonging to a first layer and second training entity data belonging to a second layer; the sample entity query is a first sample entity query; processing the sample entity inquiry information through a basic question-answer model to obtain each training entity data, wherein the method comprises the following steps: processing the first sample entity inquiry information through a first basic question-answer model to obtain first training entity data; generating second sample entity inquiry information according to the first training entity data; processing the second sample entity inquiry information through a second basic question-answer model to obtain second training entity data; according to each training entity data and each entity data label, parameters of a basic question-answering model are adjusted to obtain an entity question-answering model, and the method comprises the following steps: according to the first training entity data and the first entity data labels, parameters of a first basic question-answering model are adjusted to obtain a first question-answering model; and adjusting parameters of the second basic question-answering model according to the second training entity data and the second entity data labels to obtain a second question-answering model.

Wherein the first sample entity query information is generated according to sample knowledge data of the service; the first basic question-answering model has the entity extraction capability of the general field, and the second basic question-answering model has the capability of extracting entity data belonging to a lower-level business entity according to the entity data of a higher-level business entity in the general field.

The first training entity data are entity data of training service entities belonging to a first layer extracted by a first basic question-answer model, and the second training entity data are entity data of training service entities belonging to a second layer extracted by a second basic question-answer model.

The first entity data tag is entity data of each sample service entity belonging to a first layer, which is extracted from a sample knowledge guard of the service according to first sample entity inquiry information; the second entity data tag is determined according to the entity data of each sample service entity belonging to the first layer.

In some embodiments, the server inputs the first sample entity query information to a first basic question-answer model, obtains entity data (each first training entity data) of each first training service entity belonging to the first hierarchy through the first basic question-answer model, generates second sample entity query information according to each first training entity data, inputs the second sample entity query information to a second basic question-answer model, and obtains entity data (each second training entity data) of each second training service entity belonging to the second hierarchy through the second basic question-answer model. The server calculates a first sub-loss value according to each first training entity data and each first entity data label, calculates a second sub-loss value according to each second training entity data and each second entity data label, adjusts parameters of a first basic question-answer model according to the first sub-loss value, adjusts parameters of a second basic question-answer model according to the second sub-loss value, repeatedly executes the process of calculating the second sub-loss value according to each second training entity data and each second entity data label under the condition that the first basic question-answer model or the second basic question-answer model is not converged, adjusts the parameters of the first basic question-answer model according to the first sub-loss value until the first basic question-answer model and the second basic question-answer model are converged, takes the converged first basic question-answer model as the first question-answer model, and takes the converged second basic question-answer model as the second question-answer model.

In practical application, the first basic question-answering model and the second basic question-answering model can be realized through GPT.

The first sub-loss value is calculated according to each first training entity data and each first entity data label, and the cross entropy loss value between each first training entity data and each first entity data label can be used as the first sub-loss value; the second sub-loss value may be calculated from each second training entity data and each second entity data tag, and the cross entropy loss value between each second training entity data and each second entity data tag may be used as the second sub-loss value.

In the above embodiment, the first basic question-answering model has the entity extraction capability of the general field, the second basic question-answering model has the entity data of the general field according to the higher-level business entity, the capability of extracting the entity data of the lower-level business entity, and the first basic question-answering model and the second basic question-answering model are subjected to downstream training through the first sample entity query information, the first entity data tag and the second entity data tag of the business field, so that the converged first basic question-answering model has the capability of extracting the execution entity in the business field, and the converged second basic question-answering model has the capability of extracting the entity data of the business entity of the lower-level according to the higher-level business entity, namely, the entity data of the business entity of different levels can be obtained through the first question-answering model and the second question-answering model.

In some embodiments, as shown in fig. 9, acquiring process query information belonging to different layers according to each entity data set, and determining interaction data between service entities in each layer of entity data set according to each process query information, including: step 901, traversing in each entity data set to obtain a target entity data set belonging to a target hierarchy; step 902, generating target flow inquiry information corresponding to the service according to the target entity data set; step 903, processing the target process query information through a process question-answer model to obtain target process data of the service; step 904, determining interaction data among the business entities belonging to the target hierarchy according to the target flow data.

The target entity data set belonging to the target hierarchy is one of entity data sets belonging to different hierarchies, for example, when each entity data set comprises an entity data set belonging to a first hierarchy and an entity data set belonging to a second hierarchy, the target entity data set can be an entity data set belonging to the first hierarchy obtained by traversing or an entity data set belonging to the second hierarchy obtained by traversing.

In some embodiments, the server traverses in each entity data set belonging to different layers to obtain a target entity data set, embeds the target entity data set in a flow query template to obtain target flow query information, inputs the flow query information into the flow query model, obtains target flow data belonging to the target layers through the flow query model, determines flow data of executing the service according to the flow query information, extracts service responsibility data executed by each service entity in the flow data, and determines interaction data between the service entities according to the service responsibility data executed by each service entity.

And determining interaction data among the business entities in the target entity data set according to the target flow data.

Illustratively, the business is a palm payment, the traversed target entity data set belonging to the target hierarchy is an entity data set belonging to the first hierarchy, comprising: "Payment object, which functions to provide a biometric feature; a biological feature recognition class for recognizing biological features; the payment device, which is used for executing payment, generates target flow inquiry information according to the entity data set belonging to the first layer, and comprises the following steps: a payment object operative to provide a biometric feature; a biological feature recognition class for recognizing biological features; the payment device is used for executing payment and brushing what the payment flow is, and inputting the target flow inquiry information into the flow inquiry and answer model to obtain target flow data: the server extracts interaction data between business entities belonging to a first hierarchy from the target flow data.

When the traversed target entity data set belonging to the target hierarchy is an entity data set belonging to the second hierarchy, interaction data between business entities belonging to the second hierarchy may be determined in the same manner as the above example.

In the above embodiment, the target query information belonging to the target hierarchy is used as the question, and the target flow data of each business entity belonging to the target hierarchy is replied through the flow question-answer model in a question-answer mode, and the interactive data between each business entity belonging to the target hierarchy is determined through the target flow data, so that the efficiency of determining the interactive data between each business entity belonging to the target hierarchy is improved compared with the interactive data between each business entity manually written by a business field expert.

In some embodiments, determining interaction data between business entities belonging to a target hierarchy based on the target flow data includes: dividing target flow data according to each business entity in the target entity data set to obtain business responsibility data executed by each business entity in the target entity data set and time sequence identification of each business responsibility data; and determining interaction data among the business entities belonging to the target hierarchy according to the business responsibility data of the business entities and the time sequence identification of the business responsibility data.

The service responsibility data executed by the service entity can represent responsibility of the service entity in the service execution process, and can be receiving data, sending data, processing data and the like; when the service entity needs to interact with other service entities, the service responsibility data of the service entity comprises other service entities, for example, the data received by the service entity is sent by the other service entities, and further, the interaction data between the service entities can be determined according to the service responsibility data executed by the service entity.

The time sequence identifier of the service responsibility data identifies the time sequence of the service responsibility data in the service execution process, and can limit that the smaller the time sequence identifier of the service responsibility data is, the earlier the time sequence of the service responsibility data is executed, the larger the time sequence identifier of the service responsibility data is, and the later the time sequence of the service responsibility data is executed.

In some embodiments, the server may sequentially traverse the service entities in the target flow data according to the front-to-back order, obtain a sentence segment where the traversed service entity is located, determine service responsibility data executed by the service entity according to the sentence segment where the service entity is located, and determine a time sequence identifier of the service responsibility data executed by the service entity according to the traversed order; for example, according to the sequence from front to back in the target flow data, the first business entity is traversed in the target flow data, and the time sequence identifier of the business responsibility data executed by the business entity can be determined to be 1.

The server determines interaction data among the business entities belonging to the target hierarchy according to the business responsibility data of the business entities and the time sequence identification of the business responsibility data; the interactive data among the business entities also has sequence identifications, such as the interactive data determined according to the business responsibility data with smaller sequence identifications, and is arranged before the interactive data determined according to the business responsibility data with larger sequence identifications; the sequence identifier of the interactive data between each business entity is used for determining the time sequence of the interactive data when generating the sequence chart, and the time sequence of the interactive data arranged in the front in the sequence chart is earlier than the time sequence of the interactive data arranged in the rear in the sequence chart.

Illustratively, the business is face payment and the target hierarchy is the first hierarchy; the target flow data is: the method comprises the steps that a user provides biological characteristics, a biological characteristic recognition device obtains the biological characteristics provided by the user, the biological characteristic recognition device recognizes the biological characteristics to obtain a recognition result, when the recognition result is that the recognition result passes, the biological characteristic recognition device sends a recognition passing message to a payment device, the payment device receives the recognition passing message sent by the biological characteristic recognition device, the payment device executes payment according to the recognition passing message, and a server divides the target flow data according to a service entity to obtain service responsibility data executed by the service entity and a time sequence identifier corresponding to the service responsibility data: "the user provides a biometric feature, 1; the biological characteristic recognition device obtains biological characteristics provided by a user, and 2; the biological feature recognition equipment recognizes biological features to obtain a recognition result, and 3; when the identification result is that the identification passes, the biological characteristic identification device sends an identification passing message to the payment device, and the biological characteristic identification device 4; the payment device receives the identification passing message sent by the biological characteristic identification device, 5; the payment device performs payment by message according to the identification, 6".

According to the business responsibility data belonging to the first layer and the corresponding time sequence identifier, the interactive data among the business entities belonging to the first layer can be determined in turn, which comprises the following steps: the biometric data between the user and the biometric device, the biometric data between the biometric device and the payment device, and the payment data between the payment device and the payment device.

In the above embodiment, the target entity data set belonging to the target hierarchy is divided according to the service entities to obtain the service responsibility data and the corresponding time sequence identifier of each service entity, so that the interactive data between each service entity can be determined according to the service responsibility data of each service entity, and compared with the interactive data manually written by a service domain expert, the efficiency of determining the interactive data between each service entity belonging to the target hierarchy is improved.

In some embodiments, the interaction data between the business entities is determined by the flow data corresponding to the business, and the flow data is obtained by a flow question-answer model; before searching the knowledge data of the business based on the knowledge inquiry information of the business, the method further comprises the following steps: acquiring entity data and flow data labels of all samples; generating training process inquiry information according to the entity data of each sample; processing the training process inquiry information through a third basic question-answer model to obtain training process data; and adjusting parameters of the third basic question-answering model according to the flow data labels and the training flow data to obtain the flow question-answering model.

The sample entity data is entity data of each sample service entity; the flow data label is determined according to the entity data of each sample business entity belonging to the target hierarchy.

Training flow inquiry information is obtained according to entity data of each sample service entity and a flow inquiry template; the third basic question-answering model has the flow generating capability of the general field; the training process data is obtained by processing the training process inquiry information by the third basic question-answering model.

In some embodiments, the server obtains entity data (sample entity data) of each sample service entity, embeds the sample entity data into a flow query template to obtain training flow query information, inputs the training flow query information into a third basic question-answer model, obtains training flow data through the third basic question-answer model, calculates a third loss value according to the training flow data and a flow data label, adjusts parameters of the third basic question-answer model according to the third loss value, and repeatedly executes the process of adjusting parameters of the third basic question-answer model according to the third loss value until the third basic question-answer model converges, and takes the converged basic question-answer model as the flow question-answer model. The server calculates a third loss value according to the training flow data and the flow data label, which may be a cross entropy loss value between the training flow data and the flow data label.

In the above embodiment, the third basic question-answering model has the flow generating capability of the general field, and the downstream training is performed on the third basic question-answering model through the sample entity data and the flow data label of the service field, and in the training process, the third basic question-answering model can learn the relevant knowledge of the service flow in the service field, so that the converged third basic question-answering model has the flow generating capability of the service field, and further can process the flow query information corresponding to the service to obtain the flow data.

In some embodiments, as shown in fig. 10, generating sequence diagrams belonging to different levels according to interaction data between service entities in each entity data set includes: step 1001, traversing in each entity data set to obtain a target entity data set belonging to a target hierarchy; step 1002, generating a life line of each business entity in a target entity data set; step 1003, generating message lines corresponding to each interaction data according to the interaction data among each business entity in the target entity data set; step 1004, generating a sequence diagram belonging to the target hierarchy according to each lifeline and the message line corresponding to each interaction data.

Wherein, the life line of the service entity comprises an entity frame of the service entity and a dotted line perpendicular to the entity frame below the entity frame, as shown in fig. 11; each business entity has a life line, and the life lines of any two business entities are not overlapped; the lifeline of the business entity has a temporal order, and when the business entity performs a first operation earlier than a second operation, the representation of the first operation on the lifeline is above the representation of the second operation on the lifeline.

The message line can represent the interaction data between the same business entity and can also be used for representing the interaction data between two different business entities; for example, the biometric feature recognition device recognizes the biometric feature by calling the biometric feature recognition device to invoke a method of the biometric feature recognition device, and a message line for representing the biometric feature recognition device recognizes the biometric feature exists on a life line of the biometric feature recognition device; the biometric device transmits an identification passing message to the payment device, and a message line indicating that the biometric device transmits the identification passing message to the payment device exists between the lifeline of the biometric device and the lifeline of the payment device.

The message line includes solid lines with directional arrows and interaction data, which can indicate that interaction data is sent by a certain business entity to another business entity or directed to itself by a certain business entity.

When the message line represents the interactive data between the same service entities, the solid line included in the message line is sent by the service entity and points to the service entity, i.e. the solid line is a semi-closed line, such as the message line (identified palm graph) from the palm graph identification class to the palm graph identification class in fig. 6; when the message line represents interaction data between two different business entities, the message line includes a solid line that points from a first business entity of the two different business entities to a second business entity, such as in fig. 6, the user points to a message line of the palm-map identification class (providing a palm-map).

In practical applications, message lines representing interaction data between the same business entities are also referred to as self-associated message lines, or reverse message lines.

In some embodiments, the server traverses in each entity data set to obtain a target entity data set belonging to a target hierarchy, generates an entity frame corresponding to each service entity in the target entity data set, and generates a vertical dashed line under the entity frame to obtain a life line of the service entity; the server traverses according to the sequence to obtain target interaction data in the interaction data among the business entities of the target entity data set, when the target interaction data are the interaction data among the same business entities, the business entity corresponding to the target interaction data is determined, and a self-correlation message line corresponding to the target interaction data is drawn on the life line of the business entity; when the target interaction data is interaction data between two different service entities, determining a first service entity and a second service entity corresponding to the target interaction data, and drawing a message line corresponding to the target interaction data between the first service entity and the second service entity; and obtaining a sequence diagram belonging to the target hierarchy according to the message line corresponding to the target interaction data obtained by each traversal and the life line of each business entity.

The server can also determine the activation block on the life line of the service entity according to the message line related to the service entity; and obtaining a sequence diagram belonging to the target hierarchy according to the message line corresponding to the target interaction data obtained by each traversal, the life line of each service entity and the activation block on the life line of the service entity.

The activation block may be represented by an elongated rectangular box on the lifeline, such as in fig. 6, with a message line from the palm-map identification class to the palm-map identification class corresponding to an activation block.

The determining the activation block on the service entity life line according to the service entity related message line may be determining the earliest message line and the latest message line in the service entity related message line, and generating the activation block of the service entity between the earliest message line and the latest message line on the service entity life line.

After generating the sequence diagram belonging to the target hierarchy, the server performs the next round of traversal in each entity data set to obtain the target hierarchy of the next round, and generates the sequence diagram belonging to the target hierarchy of the next round according to the process, so that each sequence diagram belonging to different hierarchies can be generated.

In the above embodiment, the lifeline of each service entity in the target entity data set and the interactive data between each service entity in the target entity data set are generated, and the message line corresponding to the interactive data is generated, so that the sequence diagram belonging to the target hierarchy can be obtained, and the efficiency of generating the sequence diagrams belonging to different hierarchies is improved.

In some embodiments, generating a message line corresponding to each interaction data according to the interaction data between each business entity in the target entity data set includes: traversing the interactive data among the business entities in the target entity data set to obtain target interactive data; the target interaction data is interaction data between the first service entity and the second service entity; and generating a message line corresponding to the target interaction data between the lifeline corresponding to the first service entity and the lifeline corresponding to the second service entity.

Wherein the target interaction data is one interaction data in the interaction data among the business entities; the first service entity and the second service entity may be the same or different.

In some embodiments, the server traverses each interaction data to obtain the target interaction data according to the time sequence of each interaction data, for example, when the server traverses each interaction data for the first round, the target interaction data obtained by traversing is the interaction data with the earliest time sequence in each interaction data, and when the server traverses each interaction data for the last round, the target interaction data obtained by traversing is the interaction data with the latest time sequence in each interaction data.

And for the traversed target interaction data, the server acquires a first service entity and a second service entity corresponding to the target interaction data, when the first service entity and the second service entity are different, a message line between a life line of the first service entity and a life line of the second service entity is generated, and when the first service entity and the second service entity are identical, a self-correlation message line of the service entity is generated.

In the above embodiment, the message lines between the life lines of the interactive entities are generated according to the interactive data, so as to realize that elements in the sequence diagram are generated by the structured language, realize that the sequence diagram is automatically generated, and improve the efficiency of generating the sequence diagram compared with the process of manually drawing the sequence diagram by an expert.

In some embodiments, the method for generating the sequence diagram of the service can be applied to an application scenario for generating the sequence diagram of the swipe payment.

As shown in fig. 12, the server generates knowledge inquiry information from the swipe payment: what is the palm payment, inputting knowledge inquiry information into a knowledge inquiry and answer model to obtain knowledge data of the palm payment, wherein the knowledge data comprises definition and key technical points of the palm payment; generating first entity inquiry information according to knowledge data of palm payment: "please extract the business entity of the first level from the knowledge data of the palm-brushing business", input the first entity inquiry information into the first question-answer model, get the entity data of each business entity belonging to the first level: "Payment object, which functions to provide a biometric feature; a biological feature recognition class for recognizing biological features; a payment device operative to perform payment ", obtaining second entity inquiry information from entity data of each business entity belonging to the first hierarchy: "please provide the biometric feature according to the payment object; a biological feature recognition class for recognizing biological features; the payment equipment is used for executing payment, extracting a second-level business entity', and inputting second entity inquiry information into the second question-answer model to obtain entity data of each business entity belonging to the second level: "Payment object, function to provide palm image and live video; palm image recognition class, which is used for recognizing the palm image; the living body video identification class is used for identifying living body videos; payment class, the role of which is to perform payment.

Generating flow inquiry information belonging to the first hierarchy according to entity data of each business entity belonging to the first hierarchy, inputting the flow inquiry information belonging to the first hierarchy into a flow inquiry model to obtain flow data of the palm-brushing payment belonging to the first hierarchy, and processing the flow data of the palm-brushing payment belonging to the first hierarchy by a server to obtain interaction data among the business entities belonging to the first hierarchy; and the server generates a sequence diagram belonging to the first hierarchy of the palm-brushing payment according to the interactive data among the business entities belonging to the first hierarchy.

Generating flow inquiry information belonging to the second level according to the entity data of each business entity belonging to the second level, inputting the flow inquiry information belonging to the second level into a flow inquiry model to obtain flow data of the palm-brushing payment belonging to the second level, and processing the flow data of the palm-brushing payment belonging to the second level by a server to obtain interaction data among each business entity belonging to the second level; and the server generates a sequence diagram belonging to the second hierarchy of the palm-brushing payment according to the interactive data among the business entities belonging to the second hierarchy.

In some embodiments, as shown in fig. 13, the sequence diagram generating method of the service includes:

Step 1301, acquiring sample knowledge data and each entity data tag; the entity data labels comprise a first entity data label belonging to a first layer and a second entity data label belonging to a second layer; generating first sample entity inquiry information according to the sample knowledge data; processing the first sample entity inquiry information through a first basic question-answer model to obtain first training entity data; generating second sample entity inquiry information according to the first training entity data; processing the second sample entity inquiry information through a second basic question-answer model to obtain second training entity data;

step 1302, adjusting parameters of a first basic question-answering model according to each first training entity data and each first entity data label to obtain a first question-answering model; according to the second training entity data and the second entity data labels, parameters of a second basic question-answering model are adjusted to obtain a second question-answering model;

step 1303, obtaining entity data and a flow data tag of each sample; generating training process inquiry information according to the entity data of each sample; processing the training process inquiry information through a third basic question-answer model to obtain training process data; parameters of a third basic question-answering model are adjusted according to the flow data labels and the training flow data, and a flow question-answering model is obtained;

Step 1304, generating knowledge inquiry information corresponding to the service according to the inquiry template; knowledge inquiry information is processed through a knowledge inquiry and answer model, so that knowledge data of business are obtained;

step 1305, obtaining entity inquiry information according to the knowledge data, and processing the first entity inquiry information through a first question-answer model to obtain an entity data set belonging to a first level; generating second entity inquiry information according to each entity data included in the entity data set belonging to the first level; processing the second entity inquiry information through a second question-answer model to obtain an entity data set belonging to a second level;

step 1306, traversing in each entity data set to obtain a target entity data set belonging to a target hierarchy; generating target flow inquiry information corresponding to the service according to the target entity data set; processing the target flow inquiry information through a flow inquiry model to obtain target flow data of the service; dividing target flow data according to each business entity in the target entity data set to obtain business responsibility data executed by each business entity in the target entity data set and time sequence identification of each business responsibility data; according to the business responsibility data of each business entity and the time sequence identification of each business responsibility data, determining the interaction data among the business entities belonging to the target hierarchy;

Step 1307, traversing in each entity data set to obtain a target entity data set belonging to a target hierarchy; generating life lines of all business entities in the target entity data set; traversing the interactive data among the business entities in the target entity data set to obtain target interactive data; the target interaction data is interaction data between the first service entity and the second service entity; generating a message line corresponding to the target interaction data between a lifeline corresponding to the first service entity and a lifeline corresponding to the second service entity; and generating a sequence diagram belonging to the target hierarchy according to the lifelines and the message lines corresponding to the interaction data.

It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.

Based on the same inventive concept, the embodiment of the application also provides a service sequence diagram generating device for realizing the service sequence diagram generating method. The implementation scheme of the solution provided by the device is similar to the implementation scheme described in the above method, so the specific limitation in the embodiments of the sequence chart generating device for one or more services provided below may refer to the limitation of the sequence chart generating method for the service hereinabove, and will not be repeated herein.

In some embodiments, as shown in fig. 14, there is provided a sequence diagram generating apparatus for a service, including: a knowledge data lookup module 1401, an entity data query module 1402, an interaction data determination module 1403, and a sequence diagram generation module 1404, wherein:

a knowledge data searching module 1401, configured to search knowledge data of a service based on knowledge inquiry information of the service;

entity data inquiry module 1402, configured to obtain entity inquiry information according to the knowledge data, and inquire entity data corresponding to each service entity based on the entity inquiry information; a business entity is an entity involved in executing a business;

the interactive data determining module 1403 is configured to obtain flow query information corresponding to the service according to the data of each entity, and determine interactive data between each service entity according to the flow query information;

the sequence diagram generating module 1404 is configured to generate a sequence diagram for representing interactions between the service entities in the service execution process according to the interaction data between the service entities.

In some embodiments, the knowledge data searching module 1401 is further configured to generate knowledge query information corresponding to the service according to the query template; and processing knowledge inquiry information through a knowledge inquiry and answer model to obtain knowledge data of the business.

In some embodiments, the entity data query module 1402 is further configured to process the entity query information through an entity question-answer model to obtain entity data sets belonging to different layers; each entity data set comprises entity data corresponding to business entities belonging to the same layer;

the interactive data determining module 1403 is further configured to obtain, according to each entity data set, flow query information pertaining to different levels, and determine, according to each flow query information, interactive data between each business entity in each entity data set pertaining to each level;

the sequence diagram generating module 1404 is further configured to generate sequence diagrams belonging to different levels according to interaction data between service entities in each entity dataset.

In some embodiments, the entity question-answering model comprises a first question-answering model and a second question-answering model, each entity data set comprises an entity data set belonging to a first level and an entity data set belonging to a second level, and the entity query information is first entity query information; the entity data query module 1402 includes: a first entity data query unit and a second entity data query unit;

In some embodiments, the interaction data determination module 1403 includes:

In some embodiments, sequence diagram generation module 1404 includes:

In some embodiments, entity data corresponding to each business entity is obtained through an entity question-answer model; the service sequence diagram generating device further comprises an entity question-answer model obtaining module, which is used for obtaining sample knowledge data and each entity data label; generating sample entity inquiry information according to the sample knowledge data; processing the sample entity inquiry information through a basic question-answer model to obtain the data of each training entity; and adjusting parameters of the basic question-answering model according to the training entity data and the entity data labels to obtain the entity question-answering model.

In some embodiments, the interaction data between the business entities is determined by the flow data corresponding to the business, and the flow data is obtained by a flow question-answer model; the service sequence diagram generating device also comprises a flow question-answer model obtaining module which is used for obtaining each sample entity data and a flow data label; generating training process inquiry information according to the entity data of each sample; processing the training process inquiry information through a third basic question-answer model to obtain training process data; and adjusting parameters of the third basic question-answering model according to the flow data labels and the training flow data to obtain the flow question-answering model.

The modules in the sequence diagram generating device of the service can be realized in whole or in part by software, hardware and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In some embodiments, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 15. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used for storing relevant data of the sequence diagram generating method of the service. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements a sequence diagram generation method for a service.

It will be appreciated by those skilled in the art that the structure shown in fig. 15 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the computer device to which the present application is applied, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.

In some embodiments, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:

In some embodiments, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:

In some embodiments, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:

It should be noted that, the user information (including, but not limited to, user equipment information, user personal information, etc.) and the data (including, but not limited to, data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use, and processing of the related data are required to meet the related regulations.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the various embodiments provided herein may include at least one of relational databases and non-relational databases. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.

Claims

1. A method for generating a sequence diagram of a service, the method comprising:

searching knowledge data of a service based on knowledge inquiry information of the service;

acquiring entity inquiry information according to the knowledge data, and processing the entity inquiry information through an entity inquiry and answer model to acquire entity data sets belonging to different layers; each entity data set comprises entity data corresponding to business entities belonging to the same hierarchy; the business entity is an entity involved in executing the business; the hierarchy is used for reflecting the abstract degree of the business entity;

Traversing each entity data set to obtain a target entity data set belonging to a target hierarchy; generating target flow inquiry information corresponding to the service according to the target entity data set; processing the target flow inquiry information through a flow inquiry model to obtain target flow data of the service; according to the target flow data, determining interaction data among business entities belonging to the target hierarchy;

respectively generating sequence diagrams belonging to different levels according to interaction data among business entities in the entity data sets; the sequence diagrams of different levels are used for guiding the development of software of different levels;

the entity question-answering model comprises a first question-answering model and a second question-answering model, each entity data set comprises an entity data set belonging to a first layer and an entity data set belonging to a second layer, and the entity query information is first entity query information;

the processing the entity inquiry information through the entity question-answering model to obtain entity data sets belonging to different layers comprises the following steps:

processing the first entity inquiry information through the first question-answer model to obtain an entity data set belonging to the first level;

Generating second entity inquiry information according to each entity data included in the entity data set belonging to the first level;

and processing the second entity inquiry information through the second question-answer model to obtain an entity data set belonging to the second level.

2. The method of claim 1, wherein the service-based knowledge inquiry information looks up knowledge data of the service, comprising:

generating knowledge inquiry information corresponding to the service according to the inquiry template;

and processing the knowledge inquiry information through a knowledge question-answering model to obtain knowledge data of the service.

3. The method of claim 1, wherein generating the second entity inquiry information from the entity data included in the entity data set belonging to the first hierarchy includes:

embedding the entity data set belonging to the first layer and the second layer into a second entity inquiry template to obtain second entity inquiry information.

4. The method according to claim 1, wherein determining interaction data between business entities belonging to the target hierarchy based on the target flow data comprises:

dividing the target flow data according to each business entity in the target entity data set to obtain business responsibility data executed by each business entity in the target entity data set and time sequence identification of each business responsibility data;

And determining interaction data among the business entities belonging to the target hierarchy according to the business responsibility data of the business entities and the time sequence identification of the business responsibility data.

5. The method according to claim 1, wherein generating sequence diagrams belonging to different levels according to interaction data between service entities in each of the entity data sets includes:

traversing each entity data set to obtain a target entity data set belonging to a target hierarchy;

generating life lines of all business entities in the target entity data set;

generating message lines corresponding to the interaction data according to the interaction data among the business entities in the target entity data set;

and generating a sequence diagram belonging to the target hierarchy according to the lifelines and the message lines corresponding to the interaction data.

6. The method of claim 5, wherein generating a message line corresponding to each interaction data according to the interaction data between each business entity in the target entity data set comprises:

traversing the interactive data among the business entities in the target entity data set to obtain target interactive data; the target interaction data is interaction data between a first service entity and a second service entity;

And generating a message line corresponding to the target interaction data between the lifeline corresponding to the first service entity and the lifeline corresponding to the second service entity.

7. The method of claim 1, wherein the entity data corresponding to each service entity is obtained through an entity question-answer model; before searching the knowledge data of the service based on the knowledge inquiry information of the service, the method further comprises the following steps:

acquiring sample knowledge data and each entity data tag;

generating sample entity inquiry information according to the sample knowledge data;

processing the sample entity inquiry information through a basic question-answer model to obtain each training entity data;

and adjusting parameters of the basic question-answering model according to the training entity data and the entity data labels to obtain the entity question-answering model.

8. The method of claim 7, wherein the entity question-answering model comprises a first question-answering model and a second question model; the basic question-answering model comprises a first basic question-answering model and a second basic question-answering model; the entity data labels comprise a first entity data label belonging to a first layer and a second entity data label belonging to a second layer; the training entity data comprises first training entity data belonging to the first layer and second training entity data belonging to the second layer; the sample entity query information is first sample entity query information;

The sample entity inquiry information is processed through a basic question-answer model to obtain training entity data, and the method comprises the following steps:

processing the first sample entity inquiry information through the first basic question-answering model to obtain first training entity data;

generating second sample entity inquiry information according to the first training entity data;

processing the second sample entity inquiry information through the second basic question-answering model to obtain second training entity data;

the step of adjusting parameters of the basic question-answering model according to the training entity data and the entity data labels to obtain the entity question-answering model comprises the following steps:

according to the first training entity data and the first entity data labels, adjusting parameters of the first basic question-answering model to obtain the first question-answering model;

and adjusting parameters of the second basic question-answering model according to the second training entity data and the second entity data labels to obtain the second question-answering model.

9. The method according to claim 1, wherein interaction data between service entities is determined by flow data corresponding to the service, the flow data being obtained by a flow question-answer model; before searching the knowledge data of the service based on the knowledge inquiry information of the service, the method further comprises the following steps:

Acquiring entity data and flow data labels of all samples;

generating training process inquiry information according to the sample entity data;

processing the training process inquiry information through a third basic question-answer model to obtain training process data;

and adjusting parameters of the third basic question-answering model according to the flow data label and the training flow data to obtain the flow question-answering model.

10. A traffic sequence diagram generating device, characterized in that the device comprises:

the entity data query module is used for obtaining entity query information according to the knowledge data, and processing the entity query information through an entity question-answer model to obtain entity data sets belonging to different layers; each entity data set comprises entity data corresponding to business entities belonging to the same hierarchy; the business entity is an entity involved in executing the business; the hierarchy is used for reflecting the abstract degree of the business entity;

the interactive data determining module is used for traversing each entity data set to obtain a target entity data set belonging to a target hierarchy; generating target flow inquiry information corresponding to the service according to the target entity data set; processing the target flow inquiry information through a flow inquiry model to obtain target flow data of the service; according to the target flow data, determining interaction data among business entities belonging to the target hierarchy;

The sequence diagram generation module is used for respectively generating sequence diagrams belonging to different levels according to the interactive data among the business entities in the entity data set; the sequence diagrams of different levels are used for guiding the development of software of different levels;

the entity question-answering model comprises a first question-answering model and a second question-answering model, each entity data set comprises an entity data set belonging to a first layer and an entity data set belonging to a second layer, and the entity query information is first entity query information; the entity data query module comprises a first entity data query unit and a second entity data query unit;

the first entity data query unit is configured to process the first entity query information through the first question-answer model to obtain an entity data set belonging to the first hierarchy;

the second entity data query unit is configured to generate second entity query information according to each entity data included in the entity data set belonging to the first layer; and processing the second entity inquiry information through the second question-answer model to obtain an entity data set belonging to the second level.

11. The apparatus of claim 10, wherein the knowledge data lookup module is further configured to generate knowledge query information corresponding to a service according to a query template; and processing the knowledge inquiry information through a knowledge question-answering model to obtain knowledge data of the service.

12. The apparatus of claim 10, wherein the second entity data query unit is further configured to embed the entity data set belonging to the first hierarchy and the second hierarchy into a second entity query template to obtain second entity query information.

13. The apparatus of claim 10, wherein the interactive data determining unit is further configured to divide the target flow data according to each service entity in the target entity data set, and obtain service responsibility data executed by each service entity in the target entity data set and a timing identifier of each service responsibility data; and determining interaction data among the business entities belonging to the target hierarchy according to the business responsibility data of the business entities and the time sequence identification of the business responsibility data.

14. The apparatus of claim 10, wherein the sequence diagram generation module comprises:

15. The apparatus according to claim 14, wherein the message line generating unit is configured to traverse interaction data among the business entities in the target entity dataset to obtain target interaction data; the target interaction data is interaction data between a first service entity and a second service entity; and generating a message line corresponding to the target interaction data between the lifeline corresponding to the first service entity and the lifeline corresponding to the second service entity.

16. The apparatus of claim 10, wherein the entity data corresponding to each service entity is obtained through an entity question-answer model; the apparatus further comprises:

the entity question-answering model acquisition module is used for acquiring sample knowledge data and each entity data label; generating sample entity inquiry information according to the sample knowledge data; processing the sample entity inquiry information through a basic question-answer model to obtain each training entity data; and adjusting parameters of the basic question-answering model according to the training entity data and the entity data labels to obtain the entity question-answering model.

17. The apparatus of claim 16, wherein the entity question-answering model comprises a first question-answering model and a second question model; the basic question-answering model comprises a first basic question-answering model and a second basic question-answering model; the entity data labels comprise a first entity data label belonging to a first layer and a second entity data label belonging to a second layer; the training entity data comprises first training entity data belonging to the first layer and second training entity data belonging to the second layer; the sample entity query information is first sample entity query information;

the entity question-answering model acquisition module comprises:

the training entity data determining unit is used for processing the first sample entity inquiry information through the first basic question-answer model to obtain first training entity data; generating second sample entity inquiry information according to the first training entity data; processing the second sample entity inquiry information through the second basic question-answering model to obtain second training entity data;

the question-answering model obtaining unit is used for adjusting parameters of the first basic question-answering model according to the first training entity data and the first entity data labels to obtain the first question-answering model; and adjusting parameters of the second basic question-answering model according to the second training entity data and the second entity data labels to obtain the second question-answering model.

18. The apparatus of claim 10, wherein the interaction data between the business entities is determined by flow data corresponding to the business, the flow data being obtained by a flow question-answer model; the apparatus further comprises:

the flow question-answer model acquisition module is used for acquiring entity data of each sample and flow data labels; generating training process inquiry information according to the sample entity data; processing the training process inquiry information through a third basic question-answer model to obtain training process data; and adjusting parameters of the third basic question-answering model according to the flow data label and the training flow data to obtain the flow question-answering model.

19. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 9 when the computer program is executed.

20. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 9.

21. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any one of claims 1 to 9.