Search

CN-121971858-A - Data processing method, device, equipment and readable storage medium

CN121971858ACN 121971858 ACN121971858 ACN 121971858ACN-121971858-A

Abstract

The application discloses a data processing method, a device, equipment and a readable storage medium, wherein the method comprises the steps of obtaining virtual state information, fact information and game operation data, inputting the virtual state information, the fact information and the game operation data as input data into a large language model, generating task connection relations between N task intents and at least two task intents in the N task intents through the large language model, and generating a task graph data set based on the N task intents and one or more task connection relations; generating fact constraint information according to the fact information, generating operation constraint information according to the game operation data, carrying out constraint processing on the task graph data set based on the fact constraint information and the operation constraint information to obtain target task graph data, generating game tasks according to task configuration data associated with the target task graph data, and publishing the game tasks to a game environment. By adopting the method and the device, the development effect and development efficiency aiming at the game task can be improved.

Inventors

  • ZHANG HAOYANG

Assignees

  • 深圳奥拓盖母科技有限公司

Dates

Publication Date
20260505
Application Date
20260327

Claims (15)

  1. 1. A method of data processing, comprising: Obtaining virtual state information corresponding to a target virtual role controlled by an object in a game world, obtaining fact information and game operation data associated with the virtual state information in a knowledge base of the game world, and determining the virtual state information, the fact information and the game operation data as input data; Inputting the input data into a large language model, generating N task intents associated with the virtual state information through the large language model, and generating task connection relations between at least two task intents in the N task intents, and generating task graph data sets based on the N task intents and one or more task connection relations, wherein M task intents included in one task graph data in the task graph data sets are respectively used for generating associated task configuration data, N is a positive integer greater than 1, M is a positive integer less than or equal to N, and one task configuration data refers to task recommendation configuration matched with the virtual state information; generating fact constraint information according to the fact information, generating operation constraint information according to the game operation data, and performing constraint processing on the task graph data set based on the fact constraint information and the operation constraint information to obtain target task graph data; generating a game task for pushing the game progress corresponding to the target virtual character according to task configuration data associated with the target task graph data, and publishing the game task to a game environment where the target virtual character is located.
  2. 2. The method of claim 1, wherein the obtaining virtual state information corresponding to a target virtual character in a game world, and obtaining fact information and game operation data associated with the virtual state information in a knowledge base of the game world, comprises: Acquiring scenario node information of the target virtual character in a game world, game preference data corresponding to the target virtual character and entity interaction information aiming at the game world, and determining the scenario node information, the game preference data and the entity interaction information as virtual state information corresponding to the target virtual character; Obtaining fact information in a knowledge base of the game world according to the virtual state information, obtaining game operation targets and game budget information aiming at the target virtual roles from the knowledge base, and determining the game operation targets and the game budget information as game operation data.
  3. 3. The method of claim 1, wherein the inputting the input data into a large language model, generating N task intents associated with the virtual state information by the large language model, comprises: Generating a task prompt word for indicating a large language model to generate data according to the virtual state information, the fact information and the game operation data in the input data; Inputting the task prompt word into the large language model, and extracting virtual state characteristics corresponding to the virtual state information in the task prompt word through the large language model; Acquiring one or more candidate task intentions associated with the input data from the knowledge base through the large language model, and extracting features of the one or more candidate task intentions to obtain candidate intention features respectively corresponding to the one or more candidate task intentions; And performing feature matching on the virtual state feature and one or more candidate intention features to obtain intention similarity corresponding to the one or more candidate intention features, and determining candidate task intentions with the intention similarity being greater than or equal to an intention similarity threshold value in the one or more candidate task intentions as N task intentions associated with the virtual state information.
  4. 4. The method of claim 1, wherein the N task intents include a task intent N i and a task intent N j , N is a positive integer greater than 1, i, j are both positive integers less than or equal to N, the task intent N i and the task intent N j are different, the task graph data set includes S task graph data, the S task graph data includes task graph data S r , S is a positive integer, r is a positive integer less than or equal to S, and the generating the task graph data set based on the N task intents and one or more task connection relationships includes: If the task connection relationship between the task intention N i and the task intention N j is a task connectable relationship, performing intention connection on the task intention N i and the task intention N j to obtain initial task graph data; And adding the task intention with the task connectable relation with the task intention N i and the task intention with the task connectable relation with the task intention N j in the N task intents to the initial task graph data to obtain task graph data S r .
  5. 5. The method of claim 4, wherein the generating a task connection relationship between at least two task intents of the N task intents comprises: Acquiring a first task type of the task intention N i and a second task type of the task intention N j , and acquiring a task execution requirement of the task intention N i and a task execution reward of the task intention N j ; If the first task type and the second task type meet a sequence relation, determining that a task connection relation between the task intention N i and the task intention N j is a task connectable relation, wherein the sequence relation is used for indicating that a task sequence execution sequence exists between a game task corresponding to the task intention N i and a game task corresponding to the task intention N j ; If the task execution requirement and the task execution reward satisfy a condition dependency relationship, determining that a task connection relationship between the task intention N i and the task intention N j is a task connectable relationship, wherein the condition dependency relationship is used for indicating that a task execution process corresponding to the task execution requirement depends on the task execution reward; If the precedence relationship is not satisfied between the first task type and the second task type, and the condition dependency relationship is not satisfied between the task execution requirement and the task execution reward, determining that the task connection relationship between the task intention N i and the task intention N j is a task non-connectable relationship.
  6. 6. The method of claim 1, wherein the task graph data set includes S task graph data, wherein the S task graph data includes task graph data S r , wherein S is a positive integer, wherein r is a positive integer less than or equal to S, wherein the task intents include task types and task execution requirements, wherein M task intents in the task graph data S r include task intents M h , wherein M is a positive integer, wherein h is a positive integer less than or equal to M, wherein the method further comprises: Acquiring historical task configuration data associated with the task intent M h ; Extracting features of the task intention M h through the large language model to obtain task intention features corresponding to task intention M h , extracting features of the historical task configuration data to obtain task configuration semantic features and task configuration range features corresponding to the historical task configuration data, wherein the task intention features comprise task type features corresponding to task types of the task intention M h and execution requirement features corresponding to task execution requirements of the task intention M h , the task configuration semantic features are used for representing association relations among one or more data fields of the historical task configuration data, and the task configuration range features are used for reflecting data ranges of the one or more data fields; Performing feature fusion on the task configuration semantic features and the task type features to obtain type fusion features, and generating a type configuration data set corresponding to the task type according to the type fusion features; Feature fusion is carried out on the task configuration range features and the execution requirement features to obtain requirement fusion features, and a requirement configuration data set corresponding to the task execution requirements is generated based on the requirement fusion features; Acquiring subtask configuration data associated with the task intention M h from the type configuration data set and the requirement configuration data set; and determining the subtask configuration data respectively associated with the M task intents as task configuration data corresponding to the task graph data S r .
  7. 7. The method of claim 1, wherein generating fact constraint information from the fact information comprises: acquiring historical role information and historical event information from the knowledge base according to scenario node information in the virtual state information, wherein the historical role information refers to a fixed virtual role which appears in the scenario node information, the fixed virtual role refers to a virtual role which cannot be controlled by the object in the game world, and the historical event information refers to an event which occurs in the scenario node information; Task role information and task event information respectively associated with N task intents in the task graph data set are acquired, role constraint information is generated based on the task role information and the historical role information, and event constraint information is generated based on the task event information and the historical event information; And determining the role constraint information and the event constraint information as fact constraint information.
  8. 8. The method of claim 1, wherein the generating operation constraint information from the game operation data comprises: Acquiring task rewards included in task configuration data respectively associated with the M task intents, acquiring game budget information in the game operation data, and generating rewards constraint information based on the task rewards and the game budget information; acquiring task estimated completion time corresponding to the M task intents respectively, acquiring task completion interval data indicated by a game operation target in the game operation data, and generating time constraint information based on the task estimated completion time and the task completion interval data; And determining the rewards constraint information and the time constraint information as operation constraint information.
  9. 9. The method of claim 1, wherein the operation constraint information comprises rewards constraint information and time constraint information, the fact constraint information comprises role constraint information and event constraint information, the task graph data set comprises S task graph data, the S task graph data comprises task graph data S r , S is a positive integer, and r is a positive integer less than or equal to S, constraint processing is performed on the task graph data set based on the fact constraint information and the operation constraint information to obtain target task graph data, and the method comprises: Acquiring task estimated completion time and task configuration data corresponding to M task intents in the task graph data S r respectively, and acquiring unit rewards values included in the M task configuration data respectively; Summing the M unit rewards to obtain a task graph rewards corresponding to the task graph data S r , and if the task graph rewards are smaller than or equal to the total rewards threshold indicated by the rewards constraint information and the M unit rewards are smaller than or equal to the unit rewards threshold indicated by the rewards constraint information, and M estimated task completion times in the task graph data S r are all within the task completion interval data indicated by the time constraint information, determining the task graph data S r as intermediate task graph data; Acquiring task role information and task event information respectively included in the M task configuration data of the intermediate task graph data, and if abnormal role information different from historical role information indicated by the role constraint information exists in the M task role information, replacing the abnormal role information with the historical role information to obtain updated task role information; if abnormal event information different from the historical event information indicated by the event constraint information exists in the M pieces of task event information, updating the information form of the abnormal event information to a fixed information form included in the historical event information to obtain updated task event information; And determining intermediate task graph data comprising the updated task role information and the updated task event information as target task graph data.
  10. 10. The method of claim 1, wherein the task configuration data associated with the target task graph data includes a task interface identifier, a task document identifier, and a prize value, wherein the generating a game task for advancing a game progress corresponding to the target virtual character from the task configuration data associated with the target task graph data includes: Acquiring task interface rendering data of a game task corresponding to task configuration data associated with the target task graph data from the knowledge base according to the task interface identification; acquiring task file contents corresponding to task configuration data associated with the target task graph data from the knowledge base according to the task file identification, wherein the task file contents comprise field contents respectively corresponding to one or more data fields in the task configuration data associated with the target task graph data; The rewarding data are used for being issued to a game account of the target virtual character when the target virtual character completes the game task; And generating the game task for pushing the game progress corresponding to the target virtual character according to the task interface rendering data, the task file content and the rewarding data.
  11. 11. The method of claim 1, wherein the number of target avatars is at least two, the method further comprising: Dividing at least two target virtual roles into target test roles and target comparison roles according to the role identifications corresponding to the at least two target virtual roles respectively, wherein the number of the target comparison roles is larger than that of the target test roles, the game tasks are distributed to the game environment where the target test roles are located, and the game environment where the target comparison roles are located is distributed with the fixed tasks in the knowledge base; Acquiring a task completion rate, a task feedback rate and a task completion time of the target test character for the game task, and generating a task evaluation score for the game task based on the task completion rate, the task feedback rate and the task completion time; if the task evaluation score is greater than or equal to a score threshold, stopping issuing the fixed task, and issuing the game task to a game environment where the target comparison role is located; And if the task evaluation score is smaller than the score threshold, stopping issuing the game task, and issuing the fixed task to the game environment where the target test role is located.
  12. 12. A data processing apparatus, comprising: the receiving and transmitting module is used for acquiring virtual state information corresponding to a target virtual role controlled by an object in the game world, acquiring fact information and game operation data associated with the virtual state information in a knowledge base of the game world, and determining the virtual state information, the fact information and the game operation data as input data; The task map generation module is used for inputting the input data into a large language model, generating N task intents related to the virtual state information through the large language model, generating task connection relations between at least two task intents in the N task intents, and generating task map data sets based on the N task intents and one or more task connection relations, wherein M task intents included in one task map data in the task map data sets are respectively used for generating related task configuration data, N is a positive integer greater than 1, M is a positive integer less than or equal to N, and one task configuration data refers to task recommendation configuration matched with the virtual state information; The task graph constraint module is used for generating fact constraint information according to the fact information, generating operation constraint information according to the game operation data, and carrying out constraint processing on the task graph data set based on the fact constraint information and the operation constraint information to obtain target task graph data; And the task issuing module is used for generating a game task for advancing the game progress corresponding to the target virtual role according to the task configuration data associated with the target task graph data, and issuing the game task to the game environment where the target virtual role is located.
  13. 13. A computer device comprises a processor, a memory, and a network interface; The processor is connected to the memory and the network interface, wherein the network interface is configured to provide a data communication function, the memory is configured to store a computer program, and the processor is configured to invoke the computer program to cause the computer device to perform the method of any of claims 1-11.
  14. 14. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein a computer program adapted to be loaded and executed by a processor to cause a computer device having the processor to perform the method of any of claims 1-11.
  15. 15. A computer program product, characterized in that the computer program product comprises a computer program stored in a computer readable storage medium and adapted to be read and executed by a processor to cause a computer device with the processor to perform the method of any of claims 1-11.

Description

Data processing method, device, equipment and readable storage medium Technical Field The present application relates to the field of computer technologies, and in particular, to a data processing method, apparatus, device, and readable storage medium. Background An operator can develop a game task in the game world that is performed by a target virtual character controlled by a user and then distribute the game task to a game environment in which the target virtual character is located. However, since the game tasks acquired by all users are the same, the game experience of the users is easy to be monotonous, and if the users feel dissatisfied with the game tasks, operators are required to develop the game patches including new game tasks, and then issue the game patches to the users, so that the users can update the game tasks by installing the game patches, and users which are not matched with the new game tasks can still exist, therefore, the whole development and updating process is not only tedious, but also the development effect on the game tasks is difficult to be effectively improved. Disclosure of Invention The embodiment of the application provides a data processing method, a device, equipment and a readable storage medium, which can improve the development effect and development efficiency aiming at game tasks. In one aspect, an embodiment of the present application provides a data processing method, including: Obtaining virtual state information corresponding to a target virtual role controlled by an object in the game world, obtaining fact information and game operation data associated with the virtual state information in a knowledge base of the game world, and determining the virtual state information, the fact information and the game operation data as input data; Inputting input data into a large language model, generating N task intents related to virtual state information through the large language model, generating task connection relations between at least two task intents in the N task intents, and generating task graph data sets based on the N task intents and one or more task connection relations, wherein M task intents included in one task graph data in the task graph data sets are respectively used for generating related task configuration data, N is a positive integer greater than 1, M is a positive integer less than or equal to N, and one task configuration data refers to task recommendation configuration matched with the virtual state information; generating fact constraint information according to the fact information, generating operation constraint information according to the game operation data, and performing constraint processing on the task graph data set based on the fact constraint information and the operation constraint information to obtain target task graph data; Generating a game task for pushing the game progress corresponding to the target virtual character according to task configuration data associated with the target task graph data, and releasing the game task to a game environment where the target virtual character is located. Wherein, obtaining virtual state information corresponding to a target virtual character in the game world, and obtaining fact information and game operation data associated with the virtual state information in a knowledge base of the game world, comprises: acquiring scenario node information of a target virtual character in a game world, game preference data corresponding to the target virtual character and entity interaction information aiming at the game world, and determining the scenario node information, the game preference data and the entity interaction information as virtual state information corresponding to the target virtual character; and acquiring fact information in a knowledge base of the game world according to the virtual state information, acquiring game operation targets and game budget information aiming at target virtual roles from the knowledge base, and determining the game operation targets and the game budget information as game operation data. Wherein inputting the input data to the large language model, generating N task intents associated with the virtual state information through the large language model, comprises: generating a task prompt word for indicating the large language model to generate data according to the virtual state information, the fact information and the game operation data in the input data; inputting the task prompt word into a large language model, and extracting virtual state characteristics corresponding to virtual state information in the task prompt word through the large language model; acquiring one or more candidate task intentions associated with input data from a knowledge base through a large language model, and extracting features of the one or more candidate task intentions to obtain candidate intention features respectively corresponding to the one or more candidate t