Data processing method, task processing method and service development platform

Abstract

The embodiment of the invention provides a data processing method, a task processing method and a service development platform, and the data processing method comprises the steps: in response to a service construction request for a target service task, extracting service information corresponding to the target service task from the service construction request; according to the service information, a target service component is determined from multiple candidate service components, the candidate service components are pre-compiled task execution tools, and the multiple candidate service components are used for providing different service capabilities; determining a service link corresponding to the target service component according to the service information; and constructing a target service application according to the target service component and the service link, so as to execute the target service task according to the target service application.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of artificial intelligence technology, and in particular to data processing methods, task processing methods, and service development platforms. Background Technology

[0002] In practical applications, with the development of artificial intelligence technology, AI applications have gradually become core tools, typically capable of handling massive amounts of data, generating intelligent content, and making complex decisions. For example, large language models can be used to achieve various capabilities such as copywriting generation, intelligent customer service, virtual assistants, and data analysis. However, although AI applications are powerful, their development is hampered by a complex technology stack, requiring mastery of programming languages, prompt word engineering, fine-tuning, and other techniques. This complexity makes efficient management difficult, leading to high development costs and low R&D efficiency. Consequently, the developed AI applications often fail to meet actual application needs. Therefore, an effective technical solution is urgently needed to address these issues. Summary of the Invention

[0003] In view of the above, embodiments of this specification provide a data processing method. One or more embodiments of this specification also relate to a data processing apparatus, a task processing method, a task processing device, a service development platform, a computing device, a computer-readable storage medium, and a computer program product, to address the technical deficiencies existing in the prior art.

[0004] According to a first aspect of the embodiments of this specification, a data processing method is provided, comprising: In response to a service construction request for a target service task, service information corresponding to the target service task is extracted from the service construction request. Based on the service information, a target service component is determined from multiple candidate service components, wherein the candidate service components are pre-written task execution tools, and the multiple candidate service components are used to provide different service capabilities; Based on the service information, determine the service link corresponding to the target service component; Based on the target service component and the service link, a target service application is constructed to execute the target service task.

[0005] According to a second aspect of the embodiments of this specification, a data processing apparatus is provided, comprising: The first determining module is configured to extract service information corresponding to the target service task from the service construction request in response to the service construction request for the target service task. The second determining module is configured to determine a target service component from multiple candidate service components based on the service information, wherein the candidate service components are pre-written task execution tools, and the multiple candidate service components are used to provide different service capabilities; The third determining module is configured to determine the service link corresponding to the target service component based on the service information. The building module is configured to build a target service application based on the target service component and the service link, so as to execute the target service task based on the target service application.

[0006] According to a third aspect of the embodiments of this specification, a task processing method is provided, including: In response to a task processing request, identify the tasks to be processed; Based on the target service application, execute the task to be processed and obtain the task processing result corresponding to the task to be processed; The target service application is constructed according to the data processing method provided in the embodiments of this specification.

[0007] According to a fourth aspect of the embodiments of this specification, a task processing apparatus is provided, comprising: The determination module is configured to determine the tasks to be processed in response to task processing requests; The execution module is configured to execute the task to be processed according to the target service application and obtain the task processing result corresponding to the task to be processed. The target service application is constructed according to the data processing method provided in the embodiments of this specification.

[0008] According to a fifth aspect of the embodiments of this specification, a service development platform is provided for executing the data processing method provided in the embodiments of this specification, and for managing multiple candidate service components, wherein the candidate service components are pre-written task execution tools, and the multiple candidate service components are used to provide different service capabilities.

[0009] According to a sixth aspect of the embodiments of this specification, a computing device is provided, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the above method.

[0010] According to a seventh aspect of the embodiments of this specification, a computer-readable storage medium is provided that stores a computer program / instructions that, when executed by a processor, implement the steps of the above-described method.

[0011] According to an eighth aspect of the embodiments of this specification, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described method.

[0012] One embodiment of this specification provides a data processing method, comprising: in response to a service construction request for a target service task, extracting service information corresponding to the target service task from the service construction request; determining a target service component from a plurality of candidate service components based on the service information, wherein the candidate service components are pre-written task execution tools, and the plurality of candidate service components are used to provide different service capabilities; determining a service link corresponding to the target service component based on the service information; and constructing a target service application based on the target service component and the service link, so as to execute the target service task based on the target service application.

[0013] In the above method, in response to a service construction request for a target service task, service information corresponding to the target service task is extracted from the service construction request. Based on the service information, a target service component is determined from multiple candidate service components. Since a task execution tool is pre-written as a candidate service component, it is not necessary to write one when selecting the target service component. The task execution tool corresponding to the service information can be directly selected as the target service component. Based on the service information, the service link corresponding to the target service component is determined. This allows the target service application to be built directly based on the target service component and service link, further reducing development costs while meeting the application requirements corresponding to the target service task. Attached Figure Description

[0014] Figure 1 This is a schematic diagram illustrating an application scenario of a data processing method provided in one embodiment of this specification; Figure 2 This is a flowchart illustrating a data processing method provided in one embodiment of this specification; Figure 3 This is a schematic diagram of a service link in a data processing method provided in one embodiment of this specification; Figure 4 This is a flowchart illustrating the processing procedure of a data processing method provided in one embodiment of this specification. Figure 5 This is a schematic diagram of the structure of a data processing device provided in one embodiment of this specification; Figure 6 This is a flowchart illustrating a task processing method provided in one embodiment of this specification; Figure 7 This is a structural block diagram of a computing device provided in one embodiment of this specification. Detailed Implementation

[0015] Many specific details are set forth in the following description to provide a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0016] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of this specification. The singular forms “a,” “described,” and “the” as used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0017] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this specification, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0018] Furthermore, 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, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties. Moreover, 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 entry points are provided for users to choose to authorize or refuse.

[0019] In one or more embodiments of this specification, a large model refers to a deep learning model with a large number of model parameters, typically containing hundreds of millions, tens of billions, hundreds of billions, trillions, or even tens of trillions of model parameters. A large model can also be called a foundation model. It is pre-trained using large-scale unlabeled corpora to produce a pre-trained model with hundreds of millions of parameters. Such models can adapt to a wide range of downstream tasks and have good generalization ability. Examples include Large Language Models (LLMs) and multi-modal pre-training models.

[0020] In practical applications, large models only require a small number of samples to fine-tune the pre-trained model before they can be applied to different tasks. Large models can be widely used in fields such as Natural Language Processing (NLP) and Computer Vision. Specifically, they can be applied to computer vision tasks such as Visual Question Answering (VQA), Image Captioning (IC), and Image Generation, as well as NLP tasks such as text-based sentiment classification, text summarization, and machine translation. The main application scenarios for large models include digital assistants, intelligent robots, search, online education, office software, e-commerce, and intelligent design.

[0021] First, the terms and concepts used in one or more embodiments of this specification will be explained.

[0022] AI: Artificial Intelligence refers to the human-like intelligent behaviors exhibited by computer systems, including the ability to understand language, recognize images, reason and make decisions, and generate content.

[0023] API: Application Programming Interface, is a set of predefined rules and protocols that allow different software systems to communicate and exchange data.

[0024] RAG: Retrieval-Augmented Generation, is a technical architecture that combines external knowledge retrieval with large model generation to improve the accuracy and credibility of AI responses.

[0025] Agent: An intelligent agent is an AI system that can autonomously perceive its environment, make plans and decisions, invoke tools, execute tasks, and reflect on itself, rather than simply passively answering questions.

[0026] C language: a programming language.

[0027] C++: A programming language.

[0028] This specification provides a data processing method, and also relates to a data processing apparatus, a task processing method, a task processing apparatus, a service development platform, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail in the following embodiments.

[0029] See Figure 1, Figure 1 A schematic diagram illustrating an application scenario of a data processing method provided according to an embodiment of this specification is shown.

[0030] Figure 1 It includes end-side device 102 and cloud-side device 104, wherein cloud-side device 104 can be a service development platform used to build service applications.

[0031] In practice, the user sends a service building request to the cloud device 104 through the terminal device 102. The cloud device 104 responds to the service building request for the target service task, determines the service information corresponding to the target service task, and determines the target service component from multiple candidate service components based on the service information. It also determines the service link corresponding to the target service component based on the service information. This enables the target service application to be built directly based on the target service component and service link. The service interface of the target service application is then sent to the terminal device 102 so that the terminal device 102 can call the target service application through the service interface.

[0032] The edge device 102 may include a browser, an app (application), or a web application such as an H5 (Hypertext Markup Language 5) application, a lightweight application (also known as a mini-program), or a cloud application. The edge device can be developed based on a software development kit (SDK) provided by the server, such as a real-time communication (RTC) SDK. The edge device can be deployed in an electronic device and depends on the device's operation or certain apps within the device to run. The electronic device may have a display screen and support information browsing, such as a personal mobile terminal like a mobile phone, tablet, or personal computer. Various other types of applications can also be configured in the electronic device, such as human-computer interaction applications, model training applications, text processing applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, and social media platform software.

[0033] Cloud-side device 104 can be understood as a server providing various services, including physical servers and cloud servers. Examples include servers providing communication services to multiple clients, servers supporting backend training of models used on clients, and servers processing data sent by clients. It should be noted that cloud-side device 104 can be implemented as a distributed server cluster composed of multiple servers, or as a single server. Cloud-side device 104 can also be a server for a distributed system, or a server integrated with blockchain. Cloud-side device 104 can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology.

[0034] It is worth noting that the data processing method provided in the embodiments of this specification can be executed by the cloud-side device 104 or by the end-side device 102; in other embodiments, the data processing method provided in the embodiments of this specification can also be executed jointly by the end-side device 102 and the cloud-side device 104.

[0035] See Figure 2 , Figure 2 A flowchart of a data processing method according to an embodiment of this specification is shown, which specifically includes the following steps.

[0036] Step 202: In response to a service construction request for a target service task, extract the service information corresponding to the target service task from the service construction request.

[0037] Specifically, the data processing methods provided in the embodiments of this specification can be applied to a service development platform for the component-based development of service applications. Service applications may include AI applications. The AI ​​application can be a software system or product function that uses artificial intelligence technology as its core capability and is able to perceive, understand, generate, or make decisions. For example, it could be a contract review application, image retrieval application, text generation application, video generation application, etc., which are not limited in this specification. The service development platform provided in the embodiments of this specification can realize the construction of service applications based on an efficient development paradigm. Specifically, the efficient development paradigm can be understood as a systematic, modular, and engineering-oriented methodology and practice, which can build AI applications as software systems and achieve modularity, configurability, orchestration, observability, and visualization. Efficient development paradigms can include, for example, the RAG paradigm and template-based approach. Based on this paradigm, a unified RAG engine can be established, which can be reused by all question-and-answer applications. Enhanced retrieval generation can be achieved by configuring knowledge base paths and prompt word templates. The intelligent agent and workflow orchestration development paradigm can enable AI to autonomously complete multi-step tasks by defining task flowcharts. The prompt word engineering and centralized management development paradigm can enable direct modification of prompt words by establishing a prompt word version library and providing a visual editor.

[0038] The target service task can be understood as the service task that the service application to be built needs to perform. For example, if the target service task is a text generation task, then the service application to be built could be a service application for text generation. The service build request can be a request sent by a client to build the target service application. This service build request can carry service information of the service application to be built. This service build request can be sent by either the service provider or the developer through a client. The service information can include the service requirements and service content of the service application to be built. For example, if the service application to be built is a text generation application, then the service information can include a text message, such as "A service application needs to be built for text generation."

[0039] Based on this, it is possible to receive service construction requests for target service tasks and obtain the service information carried in the service construction request.

[0040] In one embodiment of this specification, the service construction request may carry service information corresponding to the target service task. Based on this, the service information corresponding to the target service task can be directly extracted from the service construction request. In another embodiment of this specification, the service construction request may carry the user's service requirements for the target service task. These service requirements may also be a piece of text information. Based on this, the service requirements can be obtained from the service construction request, and intent recognition can be performed on the service requirements to obtain the service information corresponding to the target service task. This embodiment of the specification does not limit this aspect.

[0041] Step 204: Based on the service information, determine the target service component from multiple candidate service components, wherein the candidate service components are pre-written task execution tools, and the multiple candidate service components are used to provide different service capabilities.

[0042] Specifically, after obtaining service information, the target service component corresponding to the service information can be selected from multiple candidate service components managed in the service development platform based on the service information.

[0043] In this context, candidate service components can be understood as service components with different service capabilities provided in the service development platform, while target service components can be understood as target service components selected from candidate service components that correspond to the service information. That is to say, the service capabilities provided by the target service component correspond to the service information.

[0044] Based on this, the service capabilities corresponding to the service information can be determined, and the target service component corresponding to the service capability can be selected from multiple candidate components. For example, if the service information is "a service application needs to be built for text generation", then the service capabilities corresponding to this service information may include text generation capabilities, intent recognition capabilities, knowledge retrieval capabilities, etc., and the corresponding target service component can be determined based on these service capabilities.

[0045] In practical implementation, when determining the service capabilities corresponding to the service information, semantic parsing can be performed on the service information. For example, keyword extraction and intent recognition can be performed on the service information. The service capabilities corresponding to the service information can be determined based on the keyword extraction results and intent recognition results. Continuing with the previous example, if the service information is "a service application needs to be built for text generation", then the keyword extraction results obtained from the keyword extraction of the service information are "build", "perform", "service application", and "text generation". The intent recognition results obtained from the intent recognition of the service information are "the user wants to develop a service application with text generation as its core function".

[0046] Alternatively, based on the historical building experience of the service application, after determining the explicit service capabilities corresponding to the service information from the service information, the implicit service capabilities related to the explicit service capabilities can be extracted from the historical building experience of the service application, and then the explicit service capabilities and implicit service capabilities can be used as the service capabilities corresponding to the service information. Continuing with the previous example, if the service information is "a service application needs to be built for text generation," then the explicit service capability can be extracted from the service information as "text generation capability" using keyword extraction or intent recognition. This text generation capability can be to automatically generate coherent, relevant, and stylistically appropriate text based on the input. Based on this explicit service capability, the implicit service capabilities related to this explicit service capability can be extracted from the historical construction experience of text generation applications. These implicit service capabilities include intent recognition capability, knowledge retrieval capability, context understanding capability, content security and compliance check capability, and format / style control capability. It is understandable that, based on the historical construction experience of text generation applications, when generating text, it is usually necessary to understand the user's input intent to determine the generation strategy. During the generation process, external knowledge can be introduced to improve the accuracy of the generated content. In multi-turn dialogues with users, it is necessary to maintain consistency in multi-turn dialogues or long text generation, and the security and compliance of the generated content need to be checked. During the generation process, text needs to be generated according to a specified template, tone, and length. Based on this, the implicit service capabilities related to the explicit service capability can be determined based on existing historical construction experience, thereby determining the service capability corresponding to the service information, and further determining the target service component based on the service capability.

[0047] Alternatively, a mapping rule template between service information and service capabilities can be pre-built. Specifically, a mapping rule template between service information keywords and service capabilities can be built. For example, if the service information keyword is text generation, content generation, or automatic writing, the service capability corresponding to the service information keyword is text generation capability. Mapping rules are created for these service information keywords and service capabilities so that when it is determined that the service information includes a specific service information keyword, the service capability corresponding to the specific service information keyword can be determined according to the pre-built mapping rule template, and the service capability can be identified as the service capability corresponding to the service information.

[0048] It is understood that intent recognition and keyword extraction can be implemented using large models, or other trained machine learning models and deep learning models, etc., and the embodiments in this specification do not limit this.

[0049] In practical applications, the service development platform, as a component-based development framework, can provide multiple candidate service components for connecting and organizing context information. This enables the application (i.e., the service application to be built) to have task planning and tool execution capabilities. Candidate service components can include data loaders, text segmenters, vector storage and retrieval units, etc., and can support input / output, prompt templates, and model inference capabilities for various task processing models. By pre-writing service components in the service development platform, for example, using C (or C++), subsequent calls to these components significantly improve code execution efficiency. Furthermore, the service development platform also allows for reverse engineering. That is, without source code or documentation, by analyzing the compiled program, the design logic, algorithm structure, data format, or working principle can be reconstructed, thereby achieving reverse development and reconstruction of the candidate service components within the service development platform.

[0050] In specific implementation, determining the target service component from multiple candidate service components based on the service information includes: Based on the service information, at least one sub-service task corresponding to the target service task is determined; Based on the at least one sub-service task, determine the target service component corresponding to the at least one sub-service task from the plurality of candidate service components.

[0051] The at least one sub-service task corresponding to the target service task can be understood as the target service task being broken down into at least one sub-service task. It can be understood that one sub-service task can correspond to one service capability. For example, if the target service task is a text generation task, the text generation task can be broken down into an intent recognition sub-task, a knowledge retrieval sub-task, and a text generation sub-task.

[0052] Specifically, based on the service capabilities corresponding to the service information, the target service task can be broken down into sub-service tasks corresponding to each service capability, and based on each sub-service task, the target service component corresponding to each sub-service task can be determined from multiple candidate service components.

[0053] In summary, by breaking down the target service task based on service information, the selection of candidate service components can be achieved, which facilitates the subsequent construction of the target service application.

[0054] Optionally, determining the target service component from multiple candidate service components based on the service information includes: The multiple candidate service components and / or the service information are displayed through a display interface; In response to a component selection instruction for the plurality of candidate service components, the candidate service component corresponding to the component selection instruction is determined as the target service component.

[0055] The display interface can be understood as the graphical interface (i.e., visual interface) provided by the service development platform. This interface can present multiple candidate service components to developers, allowing them to manually select the desired target service component. The component selection command can be understood as a click operation on the candidate service components on the display interface.

[0056] Specifically, multiple candidate service components and / or service information can be displayed through the interface. Developers can make selections on the interface, and the service development platform responds to these selections by choosing the target service component from among the multiple candidate service components, thereby further constructing the service link corresponding to the target service component.

[0057] Furthermore, the service development platform can be deployed on multiple clients, allowing the graphical interface to be displayed across multiple clients. Developers can adjust the candidate service components and service chains displayed in the graphical interface of the service development platform through each client to achieve collaborative development among multiple clients.

[0058] In summary, by setting up a graphical interface, developers can manually select candidate service components, thus achieving visualization of the target service application construction process.

[0059] Step 206: Determine the service link corresponding to the target service component based on the service information.

[0060] The service chain corresponding to the target service component includes information about the target service component and the execution order between them. The service information may include the service requirements and content of the target service application to be built. These requirements and content can be text information described in natural language, and the corresponding service capabilities can be extracted from them.

[0061] Specifically, the service link corresponding to the target service component can be determined based on information such as the service requirements and service content of the target service application to be built, which are included in the service information.

[0062] In practical applications, the intent recognition and service capability decomposition of service information can be realized based on the large model. The service information input by the developer or service provider can be input into the large model to obtain the service capabilities obtained by the large model through intent recognition of the service information, as well as at least one sub-service task obtained by decomposing the target service task. This facilitates the subsequent construction of service links and further realizes the construction of the target service application.

[0063] In practical applications, there are multiple target service components; The step of determining the service link corresponding to the target service component based on the service information includes: Based on the service information, determine the execution order among multiple target service components; Based on the execution order, construct the service links corresponding to the multiple target service components.

[0064] Specifically, the execution order of multiple target service components can be determined based on information such as service requirements and service content in the service information. In practical applications, a large model can be used to identify the intent of service information and decompose service capabilities. Service information input by the developer or service provider can be fed into the large model to obtain the execution order of multiple target service components output by the large model, thereby facilitating the subsequent construction of service links and further realizing the construction of the target service application.

[0065] In practice, the essence of a service chain is the sequential flow and gradual processing of data among multiple capability modules. Therefore, multiple service capabilities can be sorted. The key to sorting lies in determining which service capability's output is another service capability's input, and which service capability needs to be completed first for subsequent service capabilities to execute correctly. The execution order obtained by sorting multiple service capabilities can serve as the execution order among the multiple target service components corresponding to those service capabilities.

[0066] Optionally, when constructing the service link corresponding to the target service component based on the service information, multiple service capabilities corresponding to the service information can be determined, and the multiple service capabilities can be sorted to obtain the sorting result between the multiple service capabilities. The sorting result is used as the execution order between the multiple target service components corresponding to the multiple service capabilities. The multiple target service components are sorted according to the execution order to construct the service link. The specific implementation methods for constructing service links can include: 1. Extracting the final expected output from the service information and clarifying the final output goal. For example, if the service information is "build a service application for text generation", then the final output is a piece of natural language text that meets the requirements; 2. Determining all service capabilities corresponding to the service information; 3. Analyzing the data dependencies between multiple service capabilities. This can involve labeling the inputs and outputs of each service capability and analyzing whether each service capability depends on the outputs of other service capabilities. For example, the input of the intent recognition capability can be a user request, and the output can be a structured intent label. The input of the knowledge retrieval capability can be the intent recognition result and user input information, and the output can be relevant knowledge fragments. This knowledge retrieval capability can depend on the output of the intent recognition capability; 4. Based on the inputs, outputs, and dependencies of multiple service capabilities, converting them into a directed acyclic graph and performing topological sorting to construct the service links.

[0067] For example, see Figure 3 , Figure 3 The diagram illustrates a service link in a data processing method according to an embodiment of this specification, as shown below. Figure 3 As shown, in the graphical interface, target service component 1, target service component 2, target service component 3, target service component 4, and target service component 5 together constitute the service chain. Taking the construction of an event copywriting generation application as an example, target service component 1 can be an input component, target service component 2 can be an intent recognition component, target service component 3 can be a knowledge retrieval component, target service component 4 can be a tool invocation component, and target service component 5 can be an output component. Based on these five target service components and the service chain constructed according to the execution order among them, the event copywriting generation application is constituted. Based on this, the service development platform can realize the visual construction of the service chain of the target service application based on this graphical interface.

[0068] In practical applications, the service development platform includes an intelligent link management component, which is used to organize and use various candidate service components to achieve service goals and implement project logic. The Agent (i.e., the target service application) can achieve more complex task planning and execution capabilities based on this intelligent link management component.

[0069] In summary, the service chain is constructed by analyzing the execution order of the target service components. This service chain then calls each target service component in the execution order, thereby further constructing the target service application.

[0070] Furthermore, after determining the target service component from multiple candidate service components based on the service information, the data processing method further includes: If the task processing model corresponding to the target service task is determined based on the service information, the model call interface corresponding to the task processing model is determined, wherein the model call interface is used by the target service component to call the task processing model.

[0071] The task processing model can include, but is not limited to, various large models, agents, machine learning models, deep learning models, and neural network models. The target service task corresponds to a task processing model, indicating that to execute the target service task, the corresponding task processing model needs to be called. In other words, the task processing model is the model used to execute the target service task. For example, for a text generation task, a text generation model might be called; for a text retrieval task, a text retrieval model might be called. This task processing model can be a model capable of executing the target service task, obtained by training on the training data corresponding to the target service task. For example, for a text generation task, a text generation model might be called, and this text generation model could be a model trained on the training data (including sample text and labeled text) corresponding to the text generation task.

[0072] Specifically, when it is determined from service information that executing a target service task requires calling a corresponding task processing model, the model call interface corresponding to that task processing model can be obtained from the service development platform. In the service chain, each target service component can call the task processing model through this model call interface. That is to say, the service chain can include each target service component, the model call interface, and the execution order between each target service component and the task processing model. In practical applications, the service development platform can store pre-developed model call interfaces corresponding to different task processing models. When the task processing model corresponding to the target service task is determined, the model call interface corresponding to the pre-developed task processing model can be directly obtained from the service development platform.

[0073] Furthermore, if the execution of the target service task requires calling the corresponding task processing model based on the service information, a model call interface corresponding to that task processing model can be generated. In specific implementation, the interface target corresponding to the task processing model can be determined. This interface target may include interface caller information, interface call frequency and concurrency, interface input / output format, and other interface information (such as whether account verification or rate limiting is required). This interface target can be determined by user requirements. Then, a model call interface specification can be designed based on the interface target. This may include encoding the interface path, request body, response body, communication protocol, etc., to obtain the encoded program as the model call interface.

[0074] In practical applications, the service development platform can provide the ability to interact with various LLM models, including prompt word formatting, prediction, and result parsing capabilities. In other words, the service development platform can manage the model call interfaces corresponding to various LLM models.

[0075] In summary, by managing the model call interfaces corresponding to different task processing models, the system provides the ability to interact with various task processing models, making it easy to call the task processing models at any time in the target service application, thereby improving the accuracy and efficiency of the target service task execution.

[0076] Step 208: Construct a target service application based on the target service component and the service link, so as to execute the target service task based on the target service application.

[0077] Specifically, after determining the target service component and the corresponding service link, a target service application can be built based on the target service component and service link. When executing the target service task based on the target service application, the target service component can be called through the service link to execute the target service task.

[0078] Furthermore, after constructing the target service application based on the target service component and the service link, the data processing method further includes: The target service application is monitored based on the application management component.

[0079] In practical applications, the application management component can be an Observability module provided by the service development platform. This component can be understood as an integrated system function module used to collect, analyze, and display the runtime behavior of the target service application. It can be used for debugging, testing, logging, monitoring, and evaluation to ensure the stability and maintainability of the target service application.

[0080] Specifically, when a target service application malfunctions (e.g., provides an incorrect answer or gets stuck), the application management component can view the complete execution chain of the target service application, replay the user dialogue context, and check the input of prompt words, search results, and model output to quickly locate the cause of the error. Before the target service application goes live, the application management component can also verify whether the AI ​​behavior of the target service application meets expectations. It can provide a test sandbox environment for the target service application, and batch input test cases, automatically comparing the output for accuracy and compliance, thereby achieving testing of the target service application. The application management component can record every step of the target service application's operation and status, such as user input, the target service application's output, called tools, API returns, error messages, and time consumption statistics. The application management component can also monitor the health status of the target service application in real time, monitoring the request volume, response time, error rate, LLM call cost, and Agent decision steps. The application management component can also evaluate the quality and effectiveness of the target service application, assessing the accuracy of the target service application's answers, obtaining human feedback information such as user ratings or error markings, and evaluating service metrics such as customer service resolution rate and user dwell time.

[0081] For example, if the target service application is an intelligent recruitment assistant, and a user inputs a question like "What position is this candidate suitable for?" along with the candidate's information, the intelligent recruitment assistant analyzes the candidate's information and outputs recommended positions to the user. If the user reports that "the intelligent recruitment assistant recommended a completely irrelevant position," the following steps can be taken: Locate the complete conversation history using the application management component; monitor the process to find the abnormally high time consumption of the LLM call; debug and replay the process to discover that the RAG search mistakenly treated "interests and hobbies" as "work experience"; conduct evaluation and analysis; use the evaluation engine for batch testing to find similar issues in 5% of requests; and finally, fix and test the solution by optimizing the search logic and then using a test set to verify the improved performance.

[0082] In summary, by utilizing application management components, the stability and maintainability of target service applications can be improved through debugging, testing, logging, monitoring, and evaluation.

[0083] In practical applications, the step of extracting service information corresponding to the target service task from the service construction request in response to the service construction request for the target service task includes: In response to a service construction request sent by a client for the target service task, service information corresponding to the target service task is extracted from the service construction request; After constructing the target service application based on the target service component and the service link, the data processing method further includes: The service interface corresponding to the target service application is sent to the client, so that the client can call the target service application to execute the target service task based on the service interface.

[0084] In practical applications, service development platforms can provide API interfaces, which can be the service interfaces corresponding to the target service application.

[0085] Specifically, after receiving the service construction request sent by the client, the target service application can be constructed based on the above process, and the service interface corresponding to the target service application can be sent to the client so that the client can directly call the target service application to execute the target service task based on the service interface.

[0086] Taking the construction of an intelligent customer service robot AI application as an example, the service information (i.e., the application goal) corresponding to this target service application can be "users enter questions on web pages, and AI automatically finds answers from the knowledge base and replies, supporting multi-turn conversations." Therefore, to build this target service application, the target service components selected from multiple candidate service components in the service development platform can include input components, natural language understanding components, knowledge retrieval components, large model generation components, dialogue management components, and output components. Among these, the input component, as the starting point for receiving user input, can monitor user questions and trigger subsequent AI processing flows when a user sends a message. The Natural Language Understanding component can be used to identify user intent and extract key information; the Knowledge Retrieval component can be used to convert user questions into vectors and search for relevant documents in the knowledge base; the Large Model Generation component can be used to call a large model through a model call interface to generate natural language responses. It can input retrieved knowledge, key information extracted from user input, and prompts into the large model to obtain the model's output answer; the Dialogue Management component can be used to manage multi-turn dialogue contexts. It can retrieve dialogue history, determine whether follow-up questions are needed, and support state machines or rule transitions. It can set dialogue states, configure conditional branches, and set timeout, retry, and other strategies to achieve multi-turn dialogue management. The Output component can be used to display the generated answers to the user.

[0087] Furthermore, additional components can be selected, such as tool invocation components, sentiment analysis components, log monitoring components, and risk identification components. Tool invocation components allow AI to call external APIs, sentiment analysis components can be used to determine user emotions, log monitoring components can be used to record conversation logs, and risk identification components can be used to filter sensitive words and prevent AI from generating abnormal content. On the service development platform, this entire process is visualized and constructed through node connections, resulting in a service chain, with each node being a configurable target service component.

[0088] In summary, the above method, in response to a service construction request for a target service task, determines the service information corresponding to the target service task, and selects the target service component from multiple candidate service components based on the service information. Since a task execution tool is pre-written as a candidate service component, it does not need to be written when selecting the target service component. The task execution tool corresponding to the service information can be directly selected as the target service component, and the service link corresponding to the target service component is determined based on the service information. This allows the target service application to be built directly based on the target service component and service link, further reducing development costs while meeting the application requirements corresponding to the target service task.

[0089] The following is in conjunction with the appendix Figure 4 Taking the data processing method provided in this specification as an example in the construction of a service application, the data processing method will be further explained. Figure 4 A flowchart illustrating the processing procedure of a data processing method according to an embodiment of this specification is shown, specifically including the following steps.

[0090] Step 402: In response to the service construction request sent by the client for the target service task, determine the service information corresponding to the target service task.

[0091] Step 404: Based on the service information, determine at least one sub-service task corresponding to the target service task.

[0092] Step 406: Based on at least one sub-service task, determine the target service component corresponding to at least one sub-service task from multiple candidate service components.

[0093] Step 408: Determine the execution order among multiple target service components based on the service information.

[0094] Step 410: Construct service links corresponding to multiple target service components according to the execution order.

[0095] Step 412: Build the target service application based on the target service components and service links.

[0096] Step 414: Monitor the target service application based on the application management component.

[0097] Step 416: Send the service interface corresponding to the target service application to the client, so that the client can call the target service application to execute the target service task based on the service interface.

[0098] In the above method, in response to a service construction request for a target service task, the service information corresponding to the target service task is determined, and the target service component is determined from multiple candidate service components based on the service information. Since a task execution tool is pre-written as a candidate service component, it is not necessary to write one when selecting the target service component. The task execution tool corresponding to the service information can be directly selected as the target service component, and the service link corresponding to the target service component is determined based on the service information. This allows the target service application to be built directly based on the target service component and service link, further reducing development costs while meeting the application requirements corresponding to the target service task. Corresponding to the above method embodiments, this specification also provides data processing apparatus embodiments. Figure 5 A schematic diagram of the structure of a data processing apparatus according to one embodiment of this specification is shown. Figure 5 As shown, the device includes: The first determining module 502 is configured to extract service information corresponding to the target service task from the service construction request in response to a service construction request for the target service task. The second determining module 504 is configured to determine a target service component from multiple candidate service components based on the service information, wherein the candidate service components are pre-written task execution tools, and the multiple candidate service components are used to provide different service capabilities; The third determining module 506 is configured to determine the service link corresponding to the target service component based on the service information. The construction module 508 is configured to construct a target service application based on the target service component and the service link, so as to execute the target service task based on the target service application.

[0099] In an optional embodiment, the second determining module 504 is further configured to: Based on the service information, at least one sub-service task corresponding to the target service task is determined; Based on the at least one sub-service task, determine the target service component corresponding to the at least one sub-service task from the plurality of candidate service components.

[0100] In one optional embodiment, the target service components are multiple; The third determining module 506 is further configured as follows: Based on the service information, determine the execution order among multiple target service components; Based on the execution order, construct the service links corresponding to the multiple target service components.

[0101] In an optional embodiment, the second determining module 504 is further configured to: If the task processing model corresponding to the target service task is determined based on the service information, the model call interface corresponding to the task processing model is determined, wherein the model call interface is used by the target service component to call the task processing model.

[0102] In an optional embodiment, the second determining module 504 is further configured to: The multiple candidate service components and / or the service information are displayed through a display interface; In response to a component selection instruction for the plurality of candidate service components, the candidate service component corresponding to the component selection instruction is determined as the target service component.

[0103] In an optional embodiment, the device further includes a monitoring module configured to: The target service application is monitored based on the application management component.

[0104] In an optional embodiment, the first determining module 502 is further configured to: In response to a service construction request sent by a client for the target service task, service information corresponding to the target service task is extracted from the service construction request; The construction module 508 is further configured as follows: The service interface corresponding to the target service application is sent to the client, so that the client can call the target service application to execute the target service task based on the service interface.

[0105] In the aforementioned device, in response to a service construction request for a target service task, the service information corresponding to the target service task is determined, and based on the service information, a target service component is determined from multiple candidate service components. Since a task execution tool is pre-written as a candidate service component, it is not necessary to write one when selecting the target service component. The task execution tool corresponding to the service information can be directly selected as the target service component, and the service link corresponding to the target service component is determined based on the service information. This enables the direct construction of the target service application based on the target service component and service link, further reducing development costs while meeting the application requirements corresponding to the target service task.

[0106] The above is an illustrative scheme of a data processing apparatus according to this embodiment. It should be noted that the technical solution of this data processing apparatus and the technical solution of the data processing method described above belong to the same concept. For details not described in detail in the technical solution of the data processing apparatus, please refer to the description of the technical solution of the data processing method described above.

[0107] Corresponding to the above method embodiments, this specification also provides embodiments of task processing methods. Figure 6 A flowchart of a task processing method provided in one embodiment of this specification is shown, which specifically includes the following steps.

[0108] Step 602: In response to the task processing request, determine the task to be processed; Step 604: Execute the task to be processed according to the target service application, and obtain the task processing result corresponding to the task to be processed; The target service application is constructed according to the data processing method provided in the embodiments of this specification.

[0109] In this context, the target service application and the task to be processed can be corresponding. For example, if the target service application is a text generation application, then this text generation application can be used to provide services such as report generation and article generation. The task to be processed can be a task to generate a report or a task to generate an article. When executing the task to be processed, the task information of the task to be processed can be sent to the target service application. The task information can include the task content and task prompts of the task to be processed. For example, the task information can be "Generate an article describing autumn, no less than 400 words".

[0110] Specifically, after the target service application is built, it can respond to a user's task processing request, determine the task to be processed corresponding to the request, execute the task according to the target service application, obtain the task processing result, and send the result to the user. The target service application can be built according to the above method, and this embodiment will not be repeated here.

[0111] Corresponding to the above method embodiments, this specification also provides a task processing apparatus, including: The determination module is configured to determine the tasks to be processed in response to task processing requests; The execution module is configured to execute the task to be processed according to the target service application and obtain the task processing result corresponding to the task to be processed. The target service application is constructed according to the data processing method provided in the embodiments of this specification.

[0112] The above is an illustrative scheme of a task processing device according to this embodiment. It should be noted that the technical solution of this task processing device and the technical solution of the task processing method described above belong to the same concept. For details not described in detail in the technical solution of the task processing device, please refer to the description of the technical solution of the task processing method described above.

[0113] Corresponding to the above method embodiments, this specification also provides a service development platform for executing the data processing method provided in this specification embodiment, and for managing multiple candidate service components, wherein the candidate service components are pre-written task execution tools, and the multiple candidate service components are used to provide different service capabilities.

[0114] The above is an illustrative scheme of a service development platform according to this embodiment. It should be noted that the technical solution of this service development platform and the technical solution of the data processing method described above belong to the same concept. For details not described in detail in the technical solution of the service development platform, please refer to the description of the technical solution of the data processing method described above.

[0115] Figure 7 A structural block diagram of a computing device 700 according to one embodiment of this specification is shown. The components of the computing device 700 include, but are not limited to, a memory 710 and a processor 720. The processor 720 is connected to the memory 710 via a bus 730, and a database 750 is used to store data.

[0116] The computing device 700 also includes an access device 740, which enables the computing device 700 to communicate via one or more networks 760. Examples of these networks include Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or combinations of communication networks such as the Internet. The access device 740 may include one or more of any type of wired or wireless network interface (e.g., a network interface card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) wireless interface, a Wi-MAX (Worldwide Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, or a Near Field Communication (NFC) interface.

[0117] In one embodiment of this specification, the above-described components of the computing device 700 and Figure 7 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 7 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0118] The computing device 700 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or personal computers (PCs). The computing device 700 can also be a mobile or stationary server.

[0119] The processor 720 is used to execute the following computer program / instructions, which, when executed by the processor, implement the steps of the above method.

[0120] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the computing device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0121] An embodiment of this specification also provides a computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the above-described method.

[0122] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the computer-readable storage medium embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0123] An embodiment of this specification also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described method.

[0124] The above is an illustrative scheme of a computer program product according to this embodiment. It should be noted that the technical solution of this computer program product and the technical solution of the above method belong to the same concept, and all details not described in detail in the technical solution of the computer program product can be referred to the description of the technical solution of the above method.

[0125] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0126] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added or removed according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0127] 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 the embodiments in this specification are not limited to the described order of actions, because according to the embodiments in this specification, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments in this specification.

[0128] In the above embodiments, 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.

[0129] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments described herein. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. A data processing method, comprising: In response to a service construction request for a target service task, service information corresponding to the target service task is extracted from the service construction request. Based on the service information, a target service component is determined from multiple candidate service components, wherein the candidate service components are pre-written task execution tools, and the multiple candidate service components are used to provide different service capabilities; Based on the service information, determine the service link corresponding to the target service component; Based on the target service component and the service link, a target service application is constructed to execute the target service task.

2. The method according to claim 1, wherein determining the target service component from a plurality of candidate service components based on the service information includes: Based on the service information, at least one sub-service task corresponding to the target service task is determined; Based on the at least one sub-service task, determine the target service component corresponding to the at least one sub-service task from the plurality of candidate service components.

3. The method according to claim 1, wherein the target service component is multiple; The step of determining the service link corresponding to the target service component based on the service information includes: Based on the service information, determine the execution order among multiple target service components; Based on the execution order, construct the service links corresponding to the multiple target service components.

4. The method according to claim 1, further comprising, after determining the target service component from multiple candidate service components based on the service information: If the task processing model corresponding to the target service task is determined based on the service information, the model call interface corresponding to the task processing model is determined, wherein the model call interface is used by the target service component to call the task processing model.

5. The method according to claim 1, wherein determining the target service component from a plurality of candidate service components based on the service information comprises: The multiple candidate service components and / or the service information are displayed through a display interface; In response to a component selection instruction for the plurality of candidate service components, the candidate service component corresponding to the component selection instruction is determined as the target service component.

6. The method according to any one of claims 1-5, after constructing the target service application based on the target service component and the service link, further comprising: The target service application is monitored based on the application management component.

7. The method according to any one of claims 1-5, wherein extracting service information corresponding to the target service task from the service construction request in response to the service construction request for the target service task includes: In response to a service construction request sent by a client for the target service task, service information corresponding to the target service task is extracted from the service construction request; After constructing the target service application based on the target service component and the service link, the process further includes: The service interface corresponding to the target service application is sent to the client, so that the client can call the target service application to execute the target service task based on the service interface.

8. A task processing method, comprising: In response to a task processing request, identify the tasks to be processed; Based on the target service application, execute the task to be processed and obtain the task processing result corresponding to the task to be processed; The target service application is constructed according to any one of claims 1-7.

9. A service development platform for performing the method of any one of claims 1-7, and for managing a plurality of candidate service components, wherein the candidate service components are pre-written task execution tools, and the plurality of candidate service components are used to provide different service capabilities.

10. A computing device, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, which, when executed by the processor, implement the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 8.

12. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1 to 8.

Images