Task execution method, system, electronic device, storage medium and program product

Abstract

The application provides a task execution method, system, electronic device, storage medium and program product, and relates to the fields of task processing and system scheduling. The method is applied to a heterogeneous virtual environment, and includes: decomposing a to-be-executed task to obtain a subtask graph, at least one subtask in the subtask graph being executed by at least one heterogeneous node in the heterogeneous virtual environment; generating an execution instruction corresponding to each subtask based on a preset communication protocol, the execution instructions corresponding to different subtasks having the same format; sending the execution instructions corresponding to the at least one subtask to the at least one heterogeneous node based on the subtask graph, and performing data transmission between the at least one heterogeneous node through a preset bus; and determining a target execution result corresponding to the to-be-executed task based on a subexecution result returned by the at least one heterogeneous node. The application solves the technical problem of low accuracy and continuity in executing a task in the related art.

Description

Technical Field

[0001] This application relates to the fields of task processing and artificial intelligence, and more specifically, to a task execution method, system, electronic device, storage medium, and program product. Background Technology

[0002] With the development of artificial intelligence and automation technologies, intelligent systems are now capable of understanding and executing simple task instructions. However, these intelligent systems encounter bottlenecks when handling complex enterprise-level application scenarios, such as task workflows that span web, desktop, and mobile devices. Automation tools in these technologies are limited to single environments; for example, web automation is limited to browser operations, desktop automation is limited to specific operating systems, and mobile device automation can only control mobile devices. When task flows cross these different technological domains, the problem of fragmented collaboration among automation tools becomes prominent, resulting in lower accuracy and consistency in task execution.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This application provides a task execution method, system, electronic device, storage medium, and program product to at least solve the technical problems of low accuracy and consistency in task execution in related technologies.

[0005] According to one aspect of the embodiments of this application, a task execution method is provided, comprising: decomposing a task to be executed to obtain a subtask graph, wherein at least one subtask in the subtask graph is executed by at least one heterogeneous node in a heterogeneous virtual environment; generating execution instructions corresponding to at least one subtask based on a preset communication protocol, wherein the execution instructions corresponding to different subtasks have the same format; sending the execution instructions corresponding to at least one subtask to at least one heterogeneous node based on the subtask graph, wherein the at least one heterogeneous node transmits data through a preset bus; and determining the target execution result corresponding to the task to be executed based on the sub-execution results returned by at least one heterogeneous node.

[0006] According to another aspect of the embodiments of this application, a task execution system is also provided, applied to a heterogeneous virtual environment. The system includes: a task planning component, used to decompose the task to be executed to obtain a subtask graph, wherein at least one subtask in the subtask graph is executed by at least one heterogeneous node in the heterogeneous virtual environment; and an instruction processing component, used to generate execution instructions corresponding to at least one subtask sequentially based on a preset communication protocol, send the execution instructions corresponding to at least one subtask to at least one heterogeneous node based on the subtask graph, and determine the target execution result corresponding to the task to be executed based on the sub-execution results returned by at least one heterogeneous node, wherein at least one heterogeneous node transmits data through a preset bus.

[0007] According to one aspect of the embodiments of this application, a task execution apparatus is provided, applied to a heterogeneous virtual environment. The apparatus includes: a decomposition module, used to decompose the task to be executed to obtain a subtask graph, wherein at least one subtask in the subtask graph is executed by at least one heterogeneous node in the heterogeneous virtual environment; a generation module, used to generate execution instructions corresponding to at least one subtask based on a preset communication protocol, wherein the execution instructions corresponding to different subtasks have the same format; a sending module, used to send the execution instructions corresponding to at least one subtask to at least one heterogeneous node based on the subtask graph, wherein the at least one heterogeneous node transmits data through a preset bus; and a determination module, used to determine the target execution result corresponding to the task to be executed based on the sub-execution results returned by at least one heterogeneous node.

[0008] According to another aspect of the embodiments of this application, a computing device is also provided, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of this application when it runs.

[0009] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory storing an executable program; and a processor connected to the memory via a bus for running the program, wherein the program executes the methods in various embodiments of this application when it runs.

[0010] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of this application.

[0011] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of this application.

[0012] According to another aspect of the embodiments of this application, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the methods in various embodiments of this application.

[0013] According to another aspect of the embodiments of this application, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of this application.

[0014] In this embodiment, firstly, the task to be executed is decomposed to obtain a subtask graph, where at least one subtask in the subtask graph is executed by at least one heterogeneous node in a heterogeneous virtual environment. Next, based on a preset communication protocol, execution instructions corresponding to at least one subtask are generated, with the execution instructions for different subtasks having the same format. Then, based on the subtask graph, the execution instructions corresponding to at least one subtask are sent to at least one heterogeneous node, with data transmission between the at least one heterogeneous node via a preset bus. Finally, based on the sub-execution results returned by at least one heterogeneous node, the target execution result corresponding to the task to be executed is determined. This application first decomposes the task to be executed, breaking down the complex task flow into a series of independently executable subtask graphs, and generates execution instructions based on a preset communication protocol. This ensures that subtasks located on different heterogeneous nodes all have execution instructions following the same format, simplifying instruction transmission and execution across environments and ensuring the identifiability of task instructions. This application establishes a data sharing channel between heterogeneous nodes through a preset bus, enabling the result of the previous subtask to be transmitted to the next subtask in a timely and accurate manner, ensuring the continuity of the task flow and the integrity of the data. By aggregating the sub-execution results returned by heterogeneous nodes, the target execution result can be comprehensively evaluated and determined, ensuring the accuracy and consistency of the overall task execution, thereby solving the technical problem of low accuracy and consistency in task execution in related technologies.

[0015] The above general description and the following detailed description are for illustrative and explanatory purposes only and do not constitute a limitation thereof. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0017] Figure 1 This is a schematic diagram illustrating an application scenario of a task execution method according to an embodiment of this application;

[0018] Figure 2This is a flowchart of a task execution method according to an embodiment of this application;

[0019] Figure 3 This is a schematic diagram of an optional task execution system architecture according to an embodiment of this application;

[0020] Figure 4 This is a schematic diagram of a task execution system according to an embodiment of this application;

[0021] Figure 5 This is a schematic diagram of a task execution device according to an embodiment of this application;

[0022] Figure 6 This is a structural block diagram of a computing device according to an embodiment of this application;

[0023] Figure 7 This is a structural block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some, not all, of the embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative effort should fall within the scope of protection of the present application.

[0025] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in other orders. Other orders herein refer to orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed, or inherent to such processes, methods, products, or apparatus.

[0026] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0027] First, some nouns or terms that appear in the description of the embodiments of this application shall be interpreted as follows:

[0028] Heterogeneous Virtual Environment: This can refer to a unified execution space built using virtualization technology, which includes various heterogeneous computing units such as web page containers, desktop operating system sandboxes, and mobile emulators.

[0029] Global Context Bus: This can refer to a data persistence and sharing mechanism across heterogeneous containers, used to pass variables, files, and clipboard content between physically isolated execution units such as web pages, desktops, and mobile devices, ensuring the context continuity of cross-platform task flows.

[0030] Unified Action Description Language (UADL): This can be a cross-scenario intermediate state collaboration protocol proposed in this application, which shields the differences of underlying heterogeneous systems and defines a unified semantics across four dimensions of capabilities: code, browser, mobile device, and user computer, to guide collaborative work across various ends.

[0031] According to an embodiment of this application, a task execution method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0032] The task execution method provided in this application embodiment can be applied to, for example, Figure 1 The application scenarios shown are not limited to these. In, for example... Figure 1In the application scenario shown, server 10 can connect to one or more user terminal devices 20 via a local area network (LAN), wide area network (WAN), internet connection, or other types of data network. These user terminal devices 20 may include, but are not limited to, smartphones, tablets, laptops, PDAs, personal computers, smart home devices, and in-vehicle devices. The user terminal devices 20 can interact with the user through a graphical user interface to implement the methods provided in this application embodiment.

[0033] In this embodiment, the system consisting of a user terminal device and a server can perform the following steps: The user terminal device can interact with the server and send a task to be executed to the server. The server can break down the task to be executed to obtain a subtask graph; generate execution instructions corresponding to at least one subtask based on a preset communication protocol; send the execution instructions corresponding to at least one subtask to at least one heterogeneous node based on the subtask graph; and determine the target execution result corresponding to the task to be executed based on the sub-execution results returned by at least one heterogeneous node. The server can return the target execution result to the user terminal device.

[0034] With the rapid development of high-performance computing units, the methods provided in this application embodiment can also be applied to model-in-machine systems in other application scenarios. In one optional embodiment, the model-in-machine system has multiple built-in models. Users can select one model to adjust as needed to obtain their own model. The high-performance computing unit built into the model-in-machine system can then directly call the adjusted model to execute the methods provided in this application embodiment. In another optional embodiment, the model-in-machine system has a pre-trained model built-in. Therefore, the high-performance computing unit built into the model-in-machine system can directly call this model to execute the methods provided in this application embodiment.

[0035] Furthermore, when users need to train their own models, they can upload their own datasets via the client. These datasets are then sent to the server, allowing the server to adjust the pre-trained model using the dataset to obtain the user's customized model, which can then be deployed to the production environment. To facilitate users' model adjustment needs, the server provides complete adjustment tools, development frameworks, and processes, supporting multiple adjustment strategies. This allows the adjusted model to better adapt to different application domains and achieve a high degree of customization.

[0036] Under the aforementioned operating environment, this application provides the following: Figure 2 The task execution method shown is applied to heterogeneous virtual environments. Figure 2 This is a flowchart of a task execution method according to an embodiment of this application. For example... Figure 2 As shown, the method may include the following steps:

[0037] Step S202: Decompose the task to be executed to obtain a subtask map.

[0038] In this subtask, at least one subtask in the subtask graph is executed by at least one heterogeneous node in the heterogeneous virtual environment.

[0039] The aforementioned tasks to be executed can refer to comprehensive tasks initiated by users or external systems that require coordination of various sub-tasks in a heterogeneous virtual environment to complete together. These tasks can be multi-step, cross-platform processes, such as financial auditing, document processing, or data aggregation in enterprise scenarios, which may involve web application operations, desktop software usage, and mobile device interaction.

[0040] The aforementioned subtask graph refers to a directed network of task nodes that decomposes a task to be executed into a network with dependencies, temporal constraints, and execution paths through semantic parsing and logical reasoning. Each node can represent an operation that can be independently executed by heterogeneous nodes. Nodes can form directed edges through data dependencies or control dependencies, creating a graph structure. For example, in the process of "querying exchange rates—filling a table → sending an application notification," the subtask graph can clearly show that filling the table depends on the output of the exchange rate query, and sending the application notification must be triggered after the table is filled, which is a temporal constraint. It can also label the execution environment and input sources required for each subtask. This subtask graph structure explicitly expresses the execution order of tasks and supports parallel branching, conditional jumps, and exception rollbacks, enabling intelligent scheduling of heterogeneous nodes to execute collaboratively and avoiding execution failures due to missing dependencies or incorrect ordering.

[0041] The aforementioned heterogeneous virtual environment can refer to a unified execution framework that includes various different computing environments, such as web page containers, desktop operating system sandboxes, and mobile device emulators. These environments are integrated together through software technology to form a system that can be uniformly scheduled and managed.

[0042] The aforementioned heterogeneous nodes can refer to the basic units that constitute a heterogeneous virtual environment. These can be web page containers, desktop operating system sandboxes, or mobile device emulators. Each heterogeneous node can be responsible for executing specific types of subtasks, such as web browsing, desktop file operations, or mobile application interaction. Heterogeneous nodes can collaborate through preset communication mechanisms to jointly complete the subtask graph corresponding to the tasks to be executed.

[0043] In one optional embodiment, upon receiving a task to be executed, the task can be decomposed into a series of executable subtasks. Then, based on the characteristics and requirements of each subtask, a suitable execution environment can be intelligently determined, such as web page operation, desktop application interaction, or mobile application control. These subtasks can then be assigned to corresponding heterogeneous nodes in a heterogeneous virtual environment for execution. For example, a web page data scraping subtask can be assigned to a web page container, and a table data processing subtask can be handled by a desktop operating system sandbox.

[0044] In the above process, by decomposing the task to be executed into an independently executable subtask graph, multiple subtasks can be processed in parallel, reducing waiting time and accelerating the overall execution speed. Each subtask is executed by a suitable heterogeneous node, making full use of the advantages of different environments, ensuring that different subtasks are executed by suitable heterogeneous nodes, enhancing cross-platform collaboration capabilities, and breaking down complex task flows into smaller subtasks, which can simplify the execution logic and code implementation of tasks.

[0045] Step S204: Based on a preset communication protocol, generate execution instructions corresponding to at least one subtask.

[0046] The execution instructions for different subtasks have the same format.

[0047] The aforementioned pre-defined communication protocol can refer to a unified communication standard, such as the unified action description language proposed in this application, used to transmit execution instructions and data between heterogeneous nodes. The pre-defined communication protocol helps ensure that the format of instructions remains consistent across different environments, thereby simplifying cross-platform collaboration. The pre-defined communication protocol can be implemented by defining standardized message formats, transmission mechanisms, and data encoding rules.

[0048] The aforementioned execution instructions can refer to specific action commands generated under a preset communication protocol, used to instruct heterogeneous nodes how to execute corresponding subtasks. Execution instructions can encapsulate the operational details of the subtask, including but not limited to operation type, target element, input data, and expected result. Execution instructions for different subtasks can point to different operations, but can maintain a consistent format for unified management and scheduling.

[0049] In one optional embodiment, execution instructions corresponding to at least one subtask can be generated based on a preset communication protocol. For example, a preset communication protocol can be loaded, which can define a fixed format for the execution instructions. Then, the required operation parameters can be extracted from the subtask, such as understanding the target URL, the locator type and value of the data location when crawling data in a web environment. The extracted parameters can be filled into the corresponding preset communication protocol to generate standardized execution instructions. The preset communication protocol defines a unified syntax and semantics, and the generated instruction format can remain consistent, which helps to improve the universality and portability of the execution instructions.

[0050] In the above process, the execution instructions for different subtasks have the same format, ensuring that the execution instructions for different subtasks have the same format standard. This promotes interoperability between heterogeneous nodes and enables tasks to be executed across environments. The unified format of execution instructions reduces the complexity of the adapter parsing instructions and speeds up the response and execution speed of heterogeneous nodes.

[0051] Step S206: Based on the subtask graph, send the execution instructions corresponding to at least one subtask to at least one heterogeneous node.

[0052] In this system, at least one heterogeneous node transmits data through a preset bus.

[0053] The aforementioned preset bus can refer to a virtual data sharing and transmission mechanism, such as a global context bus. The preset bus can operate independently of individual heterogeneous nodes, enabling coordination and data transfer between heterogeneous environments. It can synchronize various results and status information generated during subtask execution, including variable values, file paths, and clipboard contents, ensuring data can flow between different heterogeneous nodes.

[0054] In one optional embodiment, execution instructions can be directed to heterogeneous nodes of the corresponding environment type based on the subtask graph and the environment type of each subtask. Upon receiving the execution instructions, the driver adaptation components on each heterogeneous node can convert them into actual operation commands suitable for the local environment. For subtasks that require data transfer across nodes, a preset bus can be used as an intermediary. When a subtask completes, the execution result can be encoded and transmitted via the preset bus to the next heterogeneous node containing the subtask, thereby achieving continuous data flow.

[0055] In the above process, the pre-defined bus ensures that heterogeneous nodes can share and understand unified context information, which helps maintain the integrity of the task flow. Accurate distribution of execution instructions avoids data retransmission and redundant operations, improving overall execution efficiency. Distributing execution instructions to heterogeneous nodes and transmitting data via the pre-defined bus improves the task execution process and enhances cross-platform task flow processing capabilities and collaborative efficiency in heterogeneous virtual environments.

[0056] Step S208: Based on the sub-execution results returned by at least one heterogeneous node, determine the target execution result corresponding to the task to be executed.

[0057] The aforementioned sub-execution results can refer to the execution status and data returned by the heterogeneous nodes after each subtask is completed. Sub-execution results may include success flags, error messages, and output data generated during the operation.

[0058] The aforementioned target execution results can refer to the completion status and outcome of the task to be executed, or they can be a summary and integration of the execution results of each subtask in the subtask graph.

[0059] In one optional embodiment, a preset bus can be monitored to receive sub-execution results from various heterogeneous nodes. These sub-execution results can include data, status information, completion flags, etc. The collected sub-execution results can also be verified to check whether each subtask was completed correctly as expected and to identify possible execution anomalies or failures. Next, the verified sub-execution results can be integrated according to the logical order of the task flow to ensure the integrity and consistency of the target execution result. For example, if a task includes retrieving data from a webpage, processing data in a desktop application, and sending results via a mobile application, then the data retrieved from the webpage, the results of desktop data processing, and the notifications from the mobile application can be combined in an orderly manner. The overall execution status of the tasks to be executed can also be determined based on the integrated results, and feedback on task completion can be provided to the user or external systems, while also providing execution logs or reports.

[0060] In the above process, collecting and integrating the results of each sub-execution helps ensure the consistency of each part of the complex task flow and the reliability of the results. For users, they receive complete task flow results that have been completed as expected, simplifying the understanding of the results and the difficulty of subsequent operations.

[0061] In this embodiment, firstly, the task to be executed is decomposed to obtain a subtask graph, where at least one subtask in the subtask graph is executed by at least one heterogeneous node in a heterogeneous virtual environment. Next, based on a preset communication protocol, execution instructions corresponding to at least one subtask are generated, with the execution instructions for different subtasks having the same format. Then, based on the subtask graph, the execution instructions corresponding to at least one subtask are sent to at least one heterogeneous node, with data transmission between the at least one heterogeneous node via a preset bus. Finally, based on the sub-execution results returned by at least one heterogeneous node, the target execution result corresponding to the task to be executed is determined. This application first decomposes the task to be executed, breaking down the complex task flow into a series of independently executable subtask graphs, and generates execution instructions based on a preset communication protocol. This ensures that subtasks located on different heterogeneous nodes all have execution instructions following the same format, simplifying instruction transmission and execution across environments and ensuring the identifiability of task instructions. This application establishes a data sharing channel between heterogeneous nodes through a preset bus, enabling the result of the previous subtask to be transmitted to the next subtask in a timely and accurate manner, ensuring the continuity of the task flow and the integrity of the data. By aggregating the sub-execution results returned by heterogeneous nodes, the target execution result can be comprehensively evaluated and determined, ensuring the accuracy and consistency of the overall task execution, thereby solving the technical problem of low accuracy and consistency in task execution in related technologies.

[0062] In the above embodiments of this application, generating execution instructions corresponding to at least one subtask based on a preset communication protocol includes: converting the task information corresponding to at least one subtask according to the preset communication protocol to obtain execution instructions corresponding to at least one subtask.

[0063] The task information mentioned above refers to a description of the specific subtask to be executed, which may include, but is not limited to, the following aspects: operation type, input data, expected result, environment configuration, security, and permissions. Operation type describes the basic nature of the subtask, such as clicking, entering text, navigating to a specific URL, or other types of operations. Input data refers to the specific information required to complete the subtask, such as the text to be entered, the file path to be uploaded, and the data parameters to be queried. In a subtask like filling out a report, the task information may include the specific values ​​of each data item in the report. Expected result defines the desired state or result after the subtask is successfully executed, as well as the output data to be captured. For example, a screenshot subtask may have an expected screenshot range and output filename. Environment configuration may include specific configuration information for the target heterogeneous nodes, such as browser version, operating system settings, and mobile device model, ensuring that the instructions are executed in the correct context. Security and permissions involve security restrictions and required permissions during subtask execution, ensuring that operations do not violate system policies.

[0064] In one optional embodiment, for each subtask in the subtask graph, specific task information corresponding to each subtask can be determined, which may include, but is not limited to, the purpose, operation object, required parameters, and expected results of the subtask. Subsequently, this task information can be formatted using a preset communication protocol to generate standardized execution instructions, preparing for execution by heterogeneous nodes. The subtask description can be parsed using a task planning model to identify operation details, such as operation type (e.g., browsing a webpage, filling out a form, sending a message); operation object (e.g., specific webpage elements, desktop application windows, mobile application controls); and related parameters. Execution instructions can be constructed according to a preset format based on the identified task information. The execution instructions can contain sufficient information to guide heterogeneous nodes on how to perform operations, such as node type, specific commands, and input / output variable definitions.

[0065] In the above process, protocol formatting helps simplify instruction management in heterogeneous environments, reducing development and maintenance costs. Standardized instruction formats make task information clearer, facilitating user understanding of the meaning and purpose of each operation, and improving debuggability and traceability. The use of pre-defined communication protocols enables heterogeneous nodes to understand and execute unified instructions, ensuring correct execution flows across web environments, desktop operating systems, and mobile devices, thus enhancing cross-platform execution capabilities.

[0066] In the above embodiments of this application, the task information corresponding to at least one subtask is format-converted based on a preset communication protocol to obtain the execution instruction corresponding to at least one subtask, including: obtaining the field values ​​corresponding to multiple preset fields from the task information corresponding to any one subtask based on multiple preset fields contained in the preset format template corresponding to the preset communication protocol; filling the field values ​​corresponding to the multiple preset fields into the preset format template to obtain the execution instruction corresponding to any one subtask.

[0067] The aforementioned preset format template refers to a standardized template used to generate execution instructions. It can be designed based on a preset communication protocol and defines the basic structure and required fields of the instructions. The preset format template ensures that heterogeneous nodes can recognize and process execution instructions in a unified format from the upper layer, thereby enabling collaborative work across environments.

[0068] The aforementioned field values ​​refer to the fill items for preset fields in a pre-defined template, derived from the parsing of specific task information. Each field value can represent actual parameters or states during task execution, such as the Uniform Resource Locator (URL) of the webpage to be accessed, the form fields to be filled in, the coordinates of the button to be clicked, and the name of the application to be operated on. Field values ​​can transform abstract task information into specific execution instructions. By obtaining and correctly filling in these field values, highly targeted and clearly defined instructions can be generated for various nodes in a heterogeneous environment to execute specific tasks.

[0069] In one optional embodiment, a preset communication protocol can be used to convert the extracted task information into execution instructions. Specifically, a preset format template defined by the preset communication protocol can be loaded. This preset format template predefines multiple fields, including node type, operation type, locator type, etc., which constitute the basic framework of the execution instructions. Subsequently, values ​​corresponding to the template fields can be extracted from the parsed subtask information, such as determining that the node to be operated is a browser, the action to be performed is a click, or the object to be operated is a button located by a Cascading Style Sheets selector. Then, by filling these field values ​​into the preset format template, execution instructions suitable for specific subtasks can be generated. This process involves deep parsing of task information and precise matching of preset communication protocol fields, ensuring the accuracy and standardization of the instructions, and providing a foundation for subsequent execution in heterogeneous environments.

[0070] In the above process, the preset communication protocol template ensures that the execution instructions have a unified format, simplifying the understanding and execution of instructions by heterogeneous nodes and reducing ambiguity and errors in task execution. Through flexible matching of preset fields, different types of tasks can be handled, generating corresponding execution instructions for web interface operations, desktop software interactions, or mobile application control, thus improving the execution capability of cross-platform task flows. The use of preset format templates reduces the computational cost of generating instructions. Simultaneously, standardized instructions facilitate rapid parsing and execution in heterogeneous environments, accelerating the overall task flow scheduling speed. The preset fields and format templates of the preset communication protocol are easy to maintain and update. When new functions need to be added or new environments need to be adapted, only the template needs to be modified or fields added to quickly support new execution requirements.

[0071] In the above embodiments of this application, the preset communication protocol encapsulates task operation instructions corresponding to multiple heterogeneous nodes in a heterogeneous virtual environment; multiple preset fields include: node type, operation type, operation range, locator type, locator value, and output variable type.

[0072] The node types mentioned above can specify which heterogeneous node the execution instructions will run on, and each node type can correspond to a specific computing environment.

[0073] The operation types mentioned above can refer to specific actions performed at a certain node, such as clicking a button, filling out a form, running code, taking a screenshot, etc. Different operation types correspond to different underlying calls or application programming interfaces.

[0074] The aforementioned scope of operation refers to the defined context or environment in which the operation occurs. This could be a specified webpage area, desktop window, or mobile application interface, ensuring that the operation takes place in the correct object or scenario. By limiting the scope of operation, the target of the operation can be accurately located in complex and dynamic computing environments, preventing accidental operations.

[0075] The locator type field described above specifies how to locate the target element or data within a given operational scope. This field ensures accurate identification and location of the target element, even across heterogeneous nodes, using appropriate location strategies for the specific environment.

[0076] The locator values ​​mentioned above can refer to actual values ​​associated with the locator type, used to specifically identify target elements or data.

[0077] The output variable types mentioned above can refer to the storage format of the results or data that may be generated after the instruction is executed. This can be text, images, file paths, or other types of data. By specifying the output variable types, the results of instruction execution can be stored and transmitted in the correct format, ensuring that subsequent subtasks can correctly read and use this data.

[0078] In one optional embodiment, the design of the preset communication protocol takes into account the task instruction transmission requirements in heterogeneous virtual environments, encapsulating a series of preset fields to describe task operation instructions in a standardized form. These fields may include node type, operation type, operation range, locator type, locator value, and output variable type. When constructing execution instructions, appropriate field combinations can be selected according to the needs of the subtask, and specific task information can be filled into these preset fields to form a complete preset communication protocol instruction. For example, a subtask of "locating and clicking a specific element in a webpage environment using path language" would explicitly indicate in the instruction that the node type is webpage, the operation type is click, the locator type is path language positioning, and the specific value, etc.

[0079] In the above settings, understanding the meaning of the preset communication protocol fields in different heterogeneous environments enables the correct parsing and execution of instructions, ensuring the smooth operation of cross-platform tasks. Detailed field definitions allow for accurate location and manipulation of target objects, reducing execution errors, especially for complex and dynamic user interfaces. The use of preset communication protocols simplifies task description, enabling even complex application scenarios to be described and executed through concise instruction sets, reducing development and maintenance workload. The definitions of preset fields provide flexibility, allowing for the easy addition of more node types and operation types in the future to adapt to the evolving needs of heterogeneous virtual environments and applications.

[0080] In the above embodiments of this application, when sending the first execution instruction corresponding to the first subtask, the first execution instruction is converted from the first heterogeneous node corresponding to the first subtask into a first control instruction matching the first heterogeneous node. The first control instruction is used to control the first heterogeneous node to execute the first subtask. When sending the second execution instruction corresponding to other subtasks, the second execution instruction is converted from the second heterogeneous node corresponding to the other subtasks into a second control instruction matching the second heterogeneous node. The second control instruction is used to control the second heterogeneous node to obtain the historical sub-execution result corresponding to the previous subtask through a preset bus, and to execute other subtasks based on the historical sub-execution result. The other subtasks are subtasks in the subtask graph other than the first subtask.

[0081] The aforementioned historical sub-execution results refer to situations in a cross-task flow where the execution of the current subtask depends on the results or states of previous subtasks. These historical sub-execution results can be data values, file paths, images, text content, or other forms of output, and can be shared and transferred between heterogeneous nodes via a global context bus. Historical sub-execution results enable the making of next operational decisions based on the results of previous steps in a continuous task flow, thereby achieving coherent execution of cross-tasks.

[0082] In one optional embodiment, when execution instructions are sent to heterogeneous nodes, the driver adaptation components on each heterogeneous node can parse the received execution instructions in a preset communication protocol format to understand task information, such as navigating to a website in a web container, launching a specific application in a desktop operating system, or sending a message in a mobile emulator. Then, based on the characteristics of the local environment, the preset communication protocol instructions can be converted into matching control instructions. For subsequent subtasks in the sequence, additional attention can be paid to the preset bus to retrieve the required historical sub-execution results from the previously completed subtask, such as specific data, files, or variable states, ensuring that the current subtask can execute correctly based on previous results.

[0083] In the above process, by converting common preset communication protocol commands into control commands that match the heterogeneous environment, the execution behavior of each heterogeneous node can be accurately controlled, avoiding execution errors caused by environmental differences. The use of a preset bus ensures the continuity and consistency of data during task execution, enabling subtasks to be linked together to form a complete task flow loop and improving the smoothness of automated execution. The control command conversion mechanism allows each node to understand and execute unified task commands, realizing cross-end task flow collaboration.

[0084] In the above embodiments of this application, the operation interfaces of multiple heterogeneous nodes in the heterogeneous virtual environment are mapped to virtual atomic objects; multiple heterogeneous nodes are equipped with virtual data channels corresponding to a preset bus.

[0085] The aforementioned virtual atomic objects can refer to an abstract representation of various operation interfaces of heterogeneous nodes used to describe and encapsulate them at a unified level. The design of virtual atomic objects enables the expression of various specific actions in cross-end task flows in a standardized manner.

[0086] The aforementioned virtual data channel can refer to a component of a pre-defined bus, providing a virtual space for data transmission between different heterogeneous nodes. The virtual data channel is independent of specific physical devices and computing environments, enabling data and status information to flow between different platforms.

[0087] In one alternative embodiment, the operation interfaces of different heterogeneous nodes can be mapped to unified virtual atomic objects. These virtual atomic objects represent basic operable units, such as clicks, input, and file uploads. Whether on a browser, desktop application, or mobile device, the same operation will be mapped to atomic objects with the same name, although the internal implementation details can differ. Furthermore, virtual data channels can be mounted within each heterogeneous node. These virtual data channels can be directly connected to a preset bus for real-time transmission of data and status change information generated during task execution, ensuring data flow and sharing across heterogeneous environments.

[0088] In the above process, the introduction of virtual atomic objects eliminates heterogeneity at the operational level, allowing the same type of operation to have the same name and parameters in different environments, simplifying the complexity of task scheduling and execution. By mounting virtual data channels within each node, secure data isolation can be ensured, while real-time sharing can be achieved through a preset bus, improving the continuity and efficiency of cross-device task flows. This mapping mechanism is designed to be flexible, facilitating the addition of more heterogeneous nodes or operation types in the future; only the mapping rules for atomic objects and data channels need to be added or adjusted accordingly.

[0089] In the above embodiments of this application, the method further includes: in response to receiving task description text, performing semantic recognition on the task description text using a text processing model to obtain the task to be executed; and decomposing the task to be executed to obtain a subtask graph, including: decomposing the task to be executed using a task planning model to obtain a subtask graph.

[0090] The task description text mentioned above can refer to natural language expressions provided by users or external systems, describing the details of the task to be performed and the expected results. It can be detailed instructions or a more vague concept or requirement description. For example, a task description like "Please help me complete this month's financial statements, including collecting new exchange rate information, updating the data in the tables, and sending an email to the finance team" provides the overall direction and components of the task.

[0091] The aforementioned text processing model can refer to a model based on natural language understanding technology such as machine learning or deep learning, which can parse task description text, extract important information and intent, and convert it into a format or task definition that a computer can understand.

[0092] The aforementioned task planning model refers to further analyzing the structure and dependencies of tasks based on the processed task description, decomposing high-order tasks into a series of executable subtasks, each subtask representing a specific step in the task flow. The task planning model considers the temporality of tasks, operational feasibility, and resource availability, organizing these subtasks in an optimal way to ensure they are executed in the correct order and within the appropriate environment.

[0093] In one optional embodiment, upon receiving a natural language task description text provided by a user, semantic recognition can be performed using a text processing model. This text processing model can be based on deep learning technology and capable of understanding complex language expressions and application logic. The text processing model can parse the text, extract important information and operational details, and obtain a machine-understandable task description. Next, a task planning model can be used to decompose the overall task into a series of smaller, more specific sub-task graphs based on the context and objectives of the task description. This decomposition process can take into account factors such as the execution order of tasks, required resources, and dependencies between operations, ensuring that the execution of each sub-task together constitutes the complete process of the task to be executed.

[0094] In the above process, the use of text processing models enables the understanding of natural language instructions, lowering the barrier to user interaction and allowing users to submit task requests without requiring programming knowledge. Task planning models automatically decompose tasks and autonomously determine the optimal execution path, improving execution efficiency and automation. The establishment of subtask graphs allows for dynamic adjustments to execution strategies based on actual conditions; even in complex cross-platform task flows, optimal solutions can be found by decomposing and reorganizing subtasks.

[0095] The technical solution proposed in this application is described below with reference to an optional embodiment. This application proposes a cross-platform task flow collaborative execution method and system for heterogeneous virtual environments. In actual scenarios such as enterprise office, financial auditing, and new media operation, users often require complex cross-platform task flows. This application proposes how to construct a heterogeneous virtual environment, enabling intelligent models to schedule and coordinate different environments such as web pages, desktops, and mobile devices in a continuous task flow without manual intervention in underlying data transmission and environment switching. This application can be applied to cloud-native automated execution platforms, whose functional modules may include an intelligent task flow scheduling engine, a cross-platform collaborative controller, and a cloud-based heterogeneous sandbox cluster. This application constructs a cross-platform collaborative execution system for heterogeneous virtual environments. It unifies upper-layer task intents through a unified action description language and connects lower-layer heterogeneous data through a global context bus, achieving continuous execution of the task flow.

[0096] The system architecture design of this application can be set as follows: The system can adopt a layered collaborative architecture, which can be divided into the following layers from top to bottom: L1: Intent planning and collaboration layer (Brain), which is responsible for understanding the user's natural language and decomposing complex task flows into a cross-platform collaborative task graph. L2: Unified protocol and data layer (Protocol & Data), which can include a unified action description language generation engine and a global context bus. L3: Driver adaptation layer (Adapter), which can convert unified action description language instructions into specific low-level calls, shielding heterogeneous differences. L4: Heterogeneous virtual environment layer (Infrastructure), which can dynamically orchestrate a cloud sandbox cluster, including web page containers, desktop operating system sandboxes, mobile device emulators, etc.

[0097] Figure 3 This is a schematic diagram of an optional task execution system architecture according to an embodiment of this application, such as... Figure 3 As shown, the task execution system includes: L1: Intent Planning and Collaboration Layer, L2: Unified Protocol and Data Layer, L3: Driver Adaptation Layer, and L4: Heterogeneous Virtual Environment Layer. In the L1: Intent Planning and Collaboration Layer, the task to be executed is decomposed using a task planning model to obtain a subtask graph. The subtask graph generates corresponding execution instructions using a preset communication protocol generation engine and distributes the instructions. In the L3: Driver Adaptation Layer, the web adapter, desktop adapter, mobile device adapter, and code executor respectively call the heterogeneous nodes in the L4: Heterogeneous Virtual Environment Layer—the web container, desktop operating system sandbox, mobile emulator, and Python runtime—to execute instructions. The L2: Unified Protocol and Data Layer can read and write data with the heterogeneous nodes in the L4: Heterogeneous Virtual Environment Layer through a preset bus and perform unified data persistent storage.

[0098] This application proposes a cross-platform capability mapping based on a unified action description language (UAL). UAL serves as a universal language connecting the model and the underlying environment, abstracting heterogeneous operations into standardized structures (such as JSON). It can define capability dimensions including: Browser capabilities, encapsulating browser-specific operations (such as navigate, navigate_back, snapshot, fill_upload, click, etc.); Mobile device capabilities, encapsulating mobile operations (such as start_application, stop_application, swipe, input_text, tap, etc.); User computer capabilities, encapsulating desktop system-level operations (such as start_application, stop_application, shell, press_key, full_screen, active_window, close_window, etc.); and Code space capabilities, allowing the embedding of code blocks (such as Python, Rust, JavaScript, etc.) for data cleaning or logical judgment, and tools (including run_code, etc.) when the model deems user interaction and complex computation necessary.

[0099] The collaborative workflow of the global context bus in this application can be as follows: To address the problem that data retrieved from web pages is difficult to directly paste into operating system software, this application introduces a bus mechanism independent of the execution environment: unified persistence, where heterogeneous nodes, including browsers, user computers, and mobile devices, are all mounted on the same virtual data channel. Context flow: In the previous step, such as the output of a browser node, the result is marked as a Context_Variable and written to the bus; in the next step, such as when a mobile device node is executing, the adapter automatically pulls the variable value from the bus and injects it into the current operation.

[0100] A specific implementation scenario example: cross-platform financial reimbursement verification. To more clearly illustrate the advantages of this application, the following example of automated corporate financial verification is used. User instruction: "Download today's exchange rate for currency A from the bank's website, then find the exchange rate column in the financial table on the desktop and enter it. Finally, send a message to me on mobile device A to notify me of completion." Collaborative execution process: Task flow decomposition. Using the model, the instruction is broken down into Task 1: Web page retrieval; Task 2: Desktop entry; Task 3: Mobile device notification. Web environment execution: Start the web page container, access the bank's official website; identify the exchange rate value, such as 7.25, and store the exchange rate value in the global context bus (Key: usd_rate). Desktop environment collaboration: Start the desktop operating system and open the spreadsheet application. The adapter reads the value (usd_rate=7.25) from the bus and fills it into the cell using the Unified Action Description Language instruction (keyboard_input). Mobile device environment collaboration: Start the mobile device emulator, open application A, and execute the message sending operation to complete the task flow loop.

[0101] This application can also employ a vision-based model, bypassing the underlying application programming interface (API) and instead recognizing screenshots through a visual model, sending only mouse and keyboard coordinate events. This application overcomes collaborative fragmentation by breaking down the boundaries of physical devices through a heterogeneous virtual environment layer and a global context bus. The model no longer operates on a single device but rather on a unified virtual space, allowing data to flow between web pages, applications, and desktops. It enhances task flow execution capabilities; compared to single-point instruction generation, this application emphasizes the continuity of the flow. Through the unified semantics of a unified action description language, it can understand and execute cross-platform tasks with complex dependencies. It also enhances stability and security by decoupling task logic from the underlying driver; instructions are filtered and executed within a sandbox via an adapter, avoiding the risks of directly running code and shielding against instability caused by differences in the underlying environment.

[0102] This application presents a cross-platform task flow collaborative architecture for heterogeneous environments. It innovatively constructs a heterogeneous virtual environment layer encompassing web pages, mobile devices, and desktops. Through driver adaptation technology, it uniformly maps the operation interfaces of different terminals to virtual atomic objects, enabling the model to have a perspective that transcends physical boundaries. Based on a global context bus-based data flow mechanism, it proposes a data sharing architecture independent of the execution environment, supporting data transfer and state synchronization between different underlying entities, thus solving the problem of data silos between heterogeneous applications. Based on a unified action description language (MAL), it defines a global collaborative instruction planning mechanism with multi-dimensional capabilities, enabling the model to generate cross-platform collaborative task graphs with temporal dependencies, achieving unified semantic description and scheduling of complex task flows.

[0103] According to another aspect of the embodiments of this application, a task execution system is also provided, applied to a heterogeneous virtual environment. This system can execute the task execution methods of the above embodiments, and the specific implementation methods and preferred application scenarios are the same as those of the above embodiments, and will not be repeated here.

[0104] Figure 4 This is a schematic diagram of a task execution system according to an embodiment of this application. Figure 4 As shown, the system includes the following: a task planning component 402 and an instruction processing component 404.

[0105] The task planning component 402 is used to decompose the task to be executed to obtain a subtask graph, wherein at least one subtask in the subtask graph is executed by at least one heterogeneous node in the heterogeneous virtual environment; the instruction processing component 404 is used to generate execution instructions corresponding to at least one subtask in sequence based on a preset communication protocol, send the execution instructions corresponding to at least one subtask to at least one heterogeneous node based on the subtask graph, and determine the target execution result corresponding to the task to be executed based on the sub-execution results returned by at least one heterogeneous node, wherein at least one heterogeneous node transmits data through a preset bus.

[0106] The instruction processing component includes: an instruction generation engine, used to generate execution instructions corresponding to at least one subtask sequentially based on a preset communication protocol, and to send the execution instructions corresponding to at least one subtask sequentially to at least one heterogeneous node based on the subtask graph; and a preset bus, used to transmit the sub-execution results corresponding to at least one subtask.

[0107] The system also includes: multiple driver adaptation components, deployed on multiple heterogeneous nodes in a heterogeneous virtual environment, used to convert received execution instructions into control instructions that match the multiple heterogeneous nodes; and multiple task orchestration components, deployed on multiple heterogeneous nodes, used to execute control instructions when the received control instructions correspond to the first subtask in the subtask graph, and to obtain the historical sub-execution results corresponding to the previous subtask from a preset bus when the control instructions correspond to other subtasks in the subtask graph, and to execute control instructions based on the historical sub-execution results. The other subtasks are subtasks in the subtask graph other than the first subtask.

[0108] According to another aspect of the embodiments of this application, a task execution apparatus is also provided, applied to a heterogeneous virtual environment. This apparatus can execute the task execution method of the above embodiments; the specific implementation method and preferred application scenarios are the same as those of the above embodiments, and will not be repeated here.

[0109] Figure 5 This is a schematic diagram of a task execution device according to an embodiment of this application, such as... Figure 5 As shown, the device includes the following: a disassembly module 502, a generation module 504, a sending module 506, and a determination module 508.

[0110] The system includes a decomposition module 502 for decomposing the task to be executed into a subtask graph, wherein at least one subtask in the subtask graph is executed by at least one heterogeneous node in a heterogeneous virtual environment; a generation module 504 for generating execution instructions corresponding to at least one subtask based on a preset communication protocol, wherein the execution instructions corresponding to different subtasks have the same format; a sending module 506 for sending the execution instructions corresponding to at least one subtask to at least one heterogeneous node based on the subtask graph, wherein the at least one heterogeneous node transmits data through a preset bus; and a determination module 508 for determining the target execution result corresponding to the task to be executed based on the sub-execution results returned by at least one heterogeneous node.

[0111] The generation module is also used to convert the format of the task information corresponding to at least one subtask based on a preset communication protocol to obtain the execution instructions corresponding to at least one subtask.

[0112] The generation module is also used to obtain the field values ​​corresponding to multiple preset fields from the task information corresponding to any subtask based on multiple preset fields contained in the preset format template corresponding to the preset communication protocol; and to fill the field values ​​corresponding to multiple preset fields into the preset format template to obtain the execution instruction corresponding to any subtask.

[0113] The preset communication protocol encapsulates task operation instructions corresponding to multiple heterogeneous nodes in a heterogeneous virtual environment; multiple preset fields include: node type, operation type, operation range, locator type, locator value, and output variable type.

[0114] Specifically, when sending the first execution instruction corresponding to the first subtask, the first execution instruction is converted from the first heterogeneous node corresponding to the first subtask into a first control instruction matching the first heterogeneous node. The first control instruction is used to control the first heterogeneous node to execute the first subtask. When sending the second execution instruction corresponding to other subtasks, the second execution instruction is converted from the second heterogeneous node corresponding to the other subtasks into a second control instruction matching the second heterogeneous node. The second control instruction is used to control the second heterogeneous node to obtain the historical sub-execution result corresponding to the previous subtask through a preset bus, and to execute other subtasks based on the historical sub-execution result. The other subtasks are subtasks in the subtask graph other than the first subtask.

[0115] In this heterogeneous virtual environment, the operation interfaces of multiple heterogeneous nodes are mapped to virtual atomic objects; multiple heterogeneous nodes are equipped with virtual data channels corresponding to preset buses.

[0116] The decomposition module is also used to respond to the received task description text, perform semantic recognition on the task description text using a text processing model to obtain the task to be executed, and decompose the task to be executed using a task planning model to obtain a subtask map.

[0117] The disassembly module 502, generation module 504, sending module 506, and determining module 508 described above correspond to steps S202 to S208 in the above embodiments. Each module and its corresponding step implements the same instances and application scenarios, but are not limited to the content disclosed in the above embodiments. The modules or units described above may be hardware or software components stored in memory and processed by one or more processors. The modules may also be part of a device and may run in the server 10 provided in the above embodiments.

[0118] The preferred embodiments involved in the above embodiments of this application are the same as the solutions, application scenarios and implementation processes provided in the above embodiments, and will not be repeated here.

[0119] Embodiments of this application may provide a computing device. Figure 6 This is a structural block diagram of a computing device according to an embodiment of this application. Figure 6 As shown, the computing device 60 may include one or more (one shown in the figure) processors 602, memory 604, memory controller, and peripheral interfaces.

[0120] The aforementioned computing device can be understood as an integrated intelligent terminal, including but not limited to servers, desktop computers, PCs (Personal Computers), and all-in-one model machines. Furthermore, the computing device may pre-install the model described in the above embodiments of this application.

[0121] Specifically, this computing device can pre-install various types of models, including but not limited to models in fields such as natural language processing, visual processing, speech processing, code processing, and multimodal task processing, thus providing diverse model choices. In different product forms, this computing device can support one or more model usage methods, including but not limited to model training, model invocation, model fine-tuning, model deployment, model inference, and application. In some product forms, this computing device also supports model management, including but not limited to multi-type model management (supporting the management of discriminative, generative, and other model types), model version control (supporting the control of different model versions), and model evaluation (evaluating model performance and effectiveness based on model evaluation tools). In other product forms, this computing device can also create applications based on models, providing API calling capabilities. Models can be called into created applications through API interfaces, and application management tools are provided to control and manage applications.

[0122] Furthermore, this computing device can also include data management (supporting the creation and management of model tuning datasets), a training center (providing abundant training resources to help users learn and master AI technologies), and basic control capabilities (providing enterprise-level basic control capabilities to ensure system security and efficient operation). Through these functions, it provides a comprehensive, integrated device for AI development, training, deployment, and application.

[0123] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and apparatus in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby implementing the methods in the above embodiments. The memory may include high-speed random access memory (RAM) and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memories. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to terminal A via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks (LANs), mobile communication networks, and combinations thereof.

[0124] The processor can invoke an executable program stored in memory via a transmission device to execute the method described in any of the above embodiments.

[0125] Embodiments of this application may provide an electronic device. Figure 7 This is a structural block diagram of an electronic device according to an embodiment of this application. Figure 7 As shown, the electronic device may include: an input / output device 702; a memory 704; and a processor 706, wherein the processor 706 is connected to the input / output device 702 and the memory 704 via a bus 708.

[0126] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and apparatus in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby implementing the methods in the above embodiments. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to terminal A via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0127] The processor can invoke an executable program stored in memory via a transmission device to perform the following methods: decomposing the task to be executed to obtain a subtask map, wherein at least one subtask in the subtask map is executed by at least one heterogeneous node in a heterogeneous virtual environment; generating execution instructions corresponding to at least one subtask based on a preset communication protocol, wherein the execution instructions corresponding to different subtasks have the same format; sending the execution instructions corresponding to at least one subtask to at least one heterogeneous node based on the subtask map, wherein at least one heterogeneous node transmits data through a preset bus; determining the target execution result corresponding to the task to be executed based on the sub-execution results returned by at least one heterogeneous node; and executing the methods in the various embodiments of this application.

[0128] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0129] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0130] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause the processing unit to execute the methods in the various embodiments of this application.

[0131] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0132] Embodiments of this application also provide a computer-readable storage medium. Optionally, in this embodiment, the computer-readable storage medium can be used to store program code executed by the method provided in the above embodiments.

[0133] Optionally, in this embodiment, the storage medium may be located in a computing device.

[0134] Optionally, in this embodiment, the computer-readable storage medium is configured to store an executable program, which, when the executable program is running, controls the device where the computer-readable storage medium is located to execute the method described in any of the above embodiments.

[0135] Embodiments of this application also provide a computer program product. Optionally, in this embodiment, the computer program product may include a computer program that, when executed by a processor, implements the methods provided in the embodiments described above.

[0136] The aforementioned computer program products can refer to software programs that have been written, tested, and released, and can run on computers or other devices. Computer program products can include application programs, operating systems, utility software, etc., used to achieve specific functions or solve specific problems.

[0137] Embodiments of this application also provide a computer program product. Optionally, the computer program product may include a non-volatile computer-readable storage medium, which can be used to store a computer program that, when executed by a processor, implements the method provided in the above embodiments.

[0138] The aforementioned non-volatile computer-readable storage medium can refer to a medium for storing data. Non-volatile computer-readable storage media can retain data without loss when power is off and can be used to store long-term data, such as operating systems, applications, and user files. Non-volatile storage media can include hard disk drives, solid-state drives, optical disks, and flash memory storage devices, etc.

[0139] Embodiments of this application also provide a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it implements the method provided in the above embodiments.

[0140] The aforementioned computer program can refer to a set of instructions used to tell the computer to perform specific tasks or operations. Computer programs can be written by programmers using specific programming languages ​​and can include algorithms, data structures, logic, and control flow. Computer programs can be used for a variety of purposes, including application software, operating systems, etc.

[0141] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0142] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0143] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0144] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0145] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0146] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A task execution method, characterized in that, include: The task to be executed is broken down to obtain a subtask graph, wherein at least one subtask in the subtask graph is executed by at least one heterogeneous node in a heterogeneous virtual environment. Based on a preset communication protocol, execution instructions corresponding to the at least one subtask are generated, wherein the execution instructions corresponding to different subtasks have the same format; Based on the subtask graph, the execution instructions corresponding to the at least one subtask are sent to the at least one heterogeneous node, wherein the at least one heterogeneous node transmits data with each other through a preset bus; Based on the sub-execution results returned by the at least one heterogeneous node, the target execution result corresponding to the task to be executed is determined.

2. The method according to claim 1, characterized in that, The step of generating execution instructions corresponding to the at least one subtask based on a preset communication protocol includes: Based on the preset communication protocol, the task information corresponding to the at least one subtask is converted into a format to obtain the execution instructions corresponding to the at least one subtask.

3. The method according to claim 2, characterized in that, The step of converting the format of the task information corresponding to the at least one subtask based on the preset communication protocol to obtain the execution instructions corresponding to the at least one subtask includes: Based on the multiple preset fields contained in the preset format template corresponding to the preset communication protocol, the field values ​​corresponding to the multiple preset fields are obtained from the task information corresponding to any subtask; The field values ​​corresponding to the multiple preset fields are filled into the preset format template to obtain the execution instruction corresponding to any subtask.

4. The method according to claim 3, characterized in that, The preset communication protocol encapsulates task operation instructions corresponding to multiple heterogeneous nodes in the heterogeneous virtual environment. The preset fields include: node type, operation type, operation range, locator type, locator value, and output variable type.

5. The method according to any one of claims 1 to 4, characterized in that, When sending the first execution instruction corresponding to the first subtask, the first execution instruction is converted by the first heterogeneous node corresponding to the first subtask into a first control instruction matching the first heterogeneous node. The first control instruction is used to control the first heterogeneous node to execute the first subtask. When sending a second execution instruction corresponding to another subtask, the second execution instruction is converted by the second heterogeneous node corresponding to the other subtask into a second control instruction matching the second heterogeneous node. The second control instruction is used to control the second heterogeneous node to obtain the historical sub-execution result corresponding to the previous subtask through the preset bus, and to execute the other subtask based on the historical sub-execution result. The other subtask is a subtask in the subtask graph other than the first subtask.

6. The method according to any one of claims 1 to 4, characterized in that, The operation interfaces of multiple heterogeneous nodes in the heterogeneous virtual environment are mapped to virtual atomic objects; the multiple heterogeneous nodes are equipped with virtual data channels corresponding to the preset bus.

7. The method according to any one of claims 1 to 4, characterized in that, The method further includes: In response to receiving task description text, the text processing model is used to perform semantic recognition on the task description text to obtain the task to be executed; The process of breaking down the task to be performed to obtain a sub-task graph includes: The task to be executed is decomposed using a task planning model to obtain the subtask map.

8. A task execution system, characterized in that, include: A task planning component is used to decompose the task to be executed to obtain a subtask graph, wherein at least one subtask in the subtask graph is executed by at least one heterogeneous node in a heterogeneous virtual environment. An instruction processing component is used to sequentially generate execution instructions corresponding to the at least one subtask based on a preset communication protocol, send the execution instructions corresponding to the at least one subtask to the at least one heterogeneous node based on the subtask graph, and determine the target execution result corresponding to the task to be executed based on the sub-execution results returned by the at least one heterogeneous node, wherein the at least one heterogeneous node transmits data through a preset bus.

9. The system according to claim 8, characterized in that, The instruction processing component includes: An instruction generation engine is used to generate execution instructions corresponding to the at least one subtask sequentially based on a preset communication protocol, and to send the execution instructions corresponding to the at least one subtask sequentially to the at least one heterogeneous node based on the subtask graph. The preset bus is used to transmit the sub-execution results corresponding to the at least one sub-task.

10. The system according to claim 9, characterized in that, The system also includes: Multiple driver adaptation components are deployed on multiple heterogeneous nodes in the heterogeneous virtual environment to convert received execution instructions into control instructions that match the multiple heterogeneous nodes. Multiple task orchestration components are deployed on the multiple heterogeneous nodes respectively. When the received control instruction corresponds to the first subtask in the subtask graph, the component executes the control instruction. When the control instruction corresponds to other subtasks in the subtask graph, the component obtains the historical sub-execution result corresponding to the previous subtask from the preset bus and executes the control instruction based on the historical sub-execution result. The other subtasks are subtasks in the subtask graph other than the first subtask.

11. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor, connected to the memory via a bus, is used to run the program, wherein the program executes the method according to any one of claims 1 to 7.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the computer-readable storage medium is located to perform the method of any one of claims 1 to 7.

13. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1 to 7.

Images