A method and apparatus for visualizing parameter configuration of a workflow

Abstract

This disclosure provides a method and apparatus for configuring visual parameters of a workflow, relating to the field of artificial intelligence technology, and particularly to the field of intelligent agent technology. The specific implementation scheme is as follows: A workflow file is obtained, which represents the logical flow executed by a target intelligent agent when responding to a request. The logical flow includes multiple logical nodes. The workflow file is parsed to obtain configurable parameters of the multiple logical nodes and the corresponding functional semantics. Visual components corresponding to the configurable parameters are determined. User interface (UI) data is obtained based on the visual components corresponding to the configurable parameters. The UI data is used to render a parameter configuration interface, which includes the visual components and functional semantics corresponding to the configurable parameters. The visual components are used to configure the parameter values ​​of the configurable parameters in the workflow file. This automates the generation of UI data for rendering the parameter configuration interface, improving the efficiency of UI data generation.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and more particularly to the field of intelligent agent technology. Specifically, this disclosure relates to a method and apparatus for configuring visual parameters for workflow. Background Technology

[0002] With the rapid development of intelligent agent technology, workflow-based intelligent agents have become the core tool for realizing intelligent interaction. The execution logic of such intelligent agents is defined by the underlying workflow, and their execution effect depends on the fine-grained configuration of the parameters of each logical node in the workflow.

[0003] Current methods for configuring parameters in workflows typically require users to understand the meaning of the lowest-level parameters, which cannot meet the needs of users without a technical background to configure parameters independently. Summary of the Invention

[0004] This disclosure provides a method and apparatus for configuring visual parameters for a workflow, thereby automating the process of generating UI data for rendering the parameter configuration interface and improving the efficiency of UI data generation.

[0005] According to one aspect of this disclosure, a method for configuring visual parameters for a workflow is provided, comprising: Obtain a workflow file, which is used to characterize the logical flow executed by the target intelligent agent when responding to a request, and the logical flow includes multiple logical nodes; The workflow file is parsed to obtain the configurable parameters of the multiple logical nodes and the functional semantics corresponding to the configurable parameters; Determine the visualization component corresponding to the configurable parameters; User interface (UI) data is obtained based on the visualization component corresponding to the configurable parameter. The UI data is used to render the parameter configuration interface. The parameter configuration interface includes the visualization component and functional semantics corresponding to the configurable parameter. The visualization component is used to configure the parameter value of the configurable parameter in the workflow file.

[0006] According to another aspect of this disclosure, a method for configuring visual parameters for a workflow is provided, comprising: In response to triggering a parsing operation on a workflow file, the workflow file being used to characterize the logical flow executed by the target agent when responding to a request, the logical flow including multiple logical nodes, the configurable parameters of the multiple logical nodes parsed from the workflow file and the functional semantics corresponding to the configurable parameters are obtained, and the visualization components corresponding to the configurable parameters are obtained. The component configuration interface displays the configurable parameters of the multiple logical nodes, the functional semantics corresponding to the configurable parameters, and the visual components. In response to a confirmation operation triggered in the component configuration interface, a confirmed configurable parameter, the corresponding functional semantics of the configurable parameter, and a visualization component are sent to the server to generate UI data. The UI data is used to render the parameter configuration interface, which includes the visualization component and functional semantics corresponding to the configurable parameter. The visualization component is used to configure the parameter value of the configurable parameter in the workflow file.

[0007] According to another aspect of this disclosure, a visual parameter configuration apparatus for a workflow is provided, comprising: The first acquisition unit is configured to acquire a workflow file, which is used to characterize the logical flow executed by the target intelligent agent when responding to a request, and the logical flow includes multiple logical nodes. The parsing unit is configured to parse the workflow file to obtain the configurable parameters of the plurality of logical nodes and the functional semantics corresponding to the configurable parameters; The determining unit is configured to determine the visualization component corresponding to the configurable parameter; The obtaining unit is configured to obtain user interface (UI) data based on the visualization component corresponding to the configurable parameter. The UI data is used to render the parameter configuration interface, which includes the visualization component and functional semantics corresponding to the configurable parameter. The visualization component is used to configure the parameter value of the configurable parameter in the workflow file.

[0008] According to another aspect of this disclosure, a visual parameter configuration apparatus for a workflow is provided, comprising: The second acquisition unit is configured to respond to triggering a parsing operation on a workflow file, the workflow file being used to characterize the logical flow executed by the target agent when responding to a request, the logical flow including multiple logical nodes, acquiring configurable parameters of the multiple logical nodes parsed from the workflow file and the functional semantics corresponding to the configurable parameters, and acquiring the visualization components corresponding to the configurable parameters. The display unit is configured to display the configurable parameters of the multiple logical nodes, the functional semantics corresponding to the configurable parameters, and the visual components in the component configuration interface; The sending unit is configured to send configurable parameters, functional semantics corresponding to the configurable parameters, and a visualization component in response to a confirmation operation triggered on the component configuration interface, so that the server can generate UI data. The UI data is used to render the parameter configuration interface, which includes the visualization component and functional semantics corresponding to the configurable parameters. The visualization component is used to configure the parameter value of the configurable parameters in the workflow file.

[0009] According to another aspect of this disclosure, an electronic device is provided, comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the methods described in the embodiments of this disclosure.

[0010] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are configured to cause the computer to perform the methods described in embodiments of this disclosure.

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

[0012] This disclosure can parse the underlying workflow file to obtain configurable parameters and determine the corresponding visualization components for each configurable parameter. Based on these visualization components, UI data for rendering the parameter configuration interface can be obtained. This disclosure can automatically generate a parameter configuration interface based on the functional semantics of the configurable parameters and the visualization components. On the one hand, automated UI data generation improves the efficiency of the first user and saves significant time. On the other hand, the generated parameter configuration interface allows the second user to directly configure parameter values ​​based on the functional semantics through the visualization components, without needing to understand the underlying technical logic. This effectively lowers the operational threshold for the second user and reduces configuration errors caused by directly manipulating parameters in the workflow file.

[0013] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0014] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure.

[0015] Figure 1 This is the system architecture diagram to which this disclosure applies.

[0016] Figure 2 This is a flowchart of a visual parameter configuration method for workflows provided in this disclosure.

[0017] Figure 3 This is a flowchart of another method for configuring visual parameters for workflows provided in this disclosure.

[0018] Figure 4 This is a schematic diagram of the parsing interface provided in this publication.

[0019] Figure 5 This is a schematic diagram of the parameter selection interface provided in this publication.

[0020] Figure 6 This is a schematic diagram of the component configuration interface provided in this publication.

[0021] Figure 7 This is a schematic block diagram of a visualization parameter configuration device for workflow provided in this disclosure.

[0022] Figure 8 This is a schematic block diagram of another visualization parameter configuration device for workflow provided in this disclosure.

[0023] Figure 9 This is a block diagram of an electronic device used to implement the visualization parameter configuration method for workflow according to the embodiments of this disclosure. Detailed Implementation

[0024] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0025] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0027] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0028] In the field of intelligent agent technology, to enable users to personalize the behavior of intelligent agents, a common practice is to provide a graphical workflow orchestration platform. Specifically, users can define the logical flow executed by the intelligent agent by dragging and connecting different logical nodes, and configure the parameters of each logical node by clicking on it.

[0029] However, because this solution directly configures parameters in the underlying logical nodes, it requires users to have a certain level of technical expertise. For example, in the scenario of configuring a medical agent, when clinicians need to adjust the consultation process or response style of the medical agent, they often find themselves at a loss due to a lack of understanding of the technical parameters in the logical nodes, such as those related to the creativity of large language models, knowledge base retrieval thresholds, and branch condition expressions. This can easily lead to malfunctions in the medical agent's behavior due to incorrect configuration. Therefore, this solution cannot meet the needs of users without a technical background for autonomous and secure configuration.

[0030] In view of this, this disclosure provides a new approach. To facilitate understanding of this disclosure, the system architecture on which this disclosure is based will first be described. Figure 1 Exemplary system architectures that can be applied to embodiments of this disclosure are shown, such as Figure 1 As shown, the system architecture may include: a client for the first user, a client for the second user, and a server.

[0031] The server side and the client side are the two main components of an application service. The server side uses a server as its primary hardware infrastructure and may include one or more software service modules. The server side and the client side form a collaborative front-end and back-end.

[0032] The client can be set on the terminal device. In this embodiment of the disclosure, the client can be a local application, a mini-program, or a web application running through a browser on the terminal device.

[0033] Terminal devices can include, but are not limited to, smart mobile terminals, wearable devices, PCs (Personal Computers), and smart home devices. Smart mobile devices can include devices such as mobile phones, tablets, laptops, PDAs (Personal Digital Assistants), and connected car terminals. Wearable devices can include devices such as smartwatches, smart glasses, smart bracelets, VR (Virtual Reality) devices, AR (Augmented Reality) devices, and mixed reality devices (devices that support both virtual and augmented reality). Smart home devices can include devices such as smart TVs and smart refrigerators with displays.

[0034] A server can be a single server, a server cluster consisting of multiple servers, or a cloud server. A cloud server, also known as a cloud computing server or cloud host, is a hosting product in the cloud computing service system, designed to address the shortcomings of traditional physical hosts and Virtual Private Servers (VPS) services, such as high management difficulty and weak service scalability.

[0035] It should be understood that Figure 1 The number of client terminals for the first user, the client terminals for the second user, and the server terminals shown in the diagram are merely illustrative. Depending on implementation needs, there can be any number of client terminals for the first user, the client terminals for the second user, and the server terminals.

[0036] As one embodiment, the first user refers to a user with a certain technical background (such as a technician), while the second user refers to a user without a technical background (such as a business person). In this disclosure, the first user uses its corresponding client and server to generate user interface (UI) data for rendering the parameter configuration interface, and then sends the UI data to the client of the second user. Based on this, the second user can use the parameter configuration interface rendered based on the UI data to adjust the parameters in the underlying workflow.

[0037] Specifically, in response to triggering the parsing operation for the workflow file, the client of the first user sends a parsing request to the server.

[0038] After receiving the parsing request, the server parses the workflow file to obtain the configurable parameters of multiple logical nodes and the corresponding functional semantics, and determines the visualization components corresponding to the configurable parameters. Then, the server sends the configurable parameters, their corresponding functional semantics, and the visualization components to the client of the first user.

[0039] The first user's client obtains the configurable parameters of multiple logical nodes in the workflow file and the corresponding functional semantics from the server. It also obtains the corresponding visual components and displays the configurable parameters, functional semantics, and visual components of the multiple logical nodes in the component configuration interface. In response to the confirmation operation triggered in the component configuration interface, it sends the confirmed configurable parameters, corresponding functional semantics, and visual components to the server.

[0040] After receiving the configurable parameters, the corresponding functional semantics, and the visual components, the server generates UI data.

[0041] The first user can send a publish request for UI data through their corresponding client. After receiving the publish request, the server will send the UI data to the second user's client. At this time, the second user's client can display the parameter configuration interface based on the UI data. The second user can use the parameter configuration interface on their corresponding client to adjust the parameter values ​​of configurable parameters in the workflow file.

[0042] Figure 2 This is a flowchart illustrating a method for configuring visual parameters for a workflow, provided in an embodiment of this disclosure. This method can be implemented by... Figure 1 The server-side execution in the system shown. For example... Figure 2 As shown, the method may include the following steps: Step 201: Obtain the workflow file. The workflow file is used to characterize the logical flow executed by the target intelligent agent when responding to a request. The logical flow includes multiple logical nodes.

[0043] Step 202: Parse the workflow file to obtain the configurable parameters of multiple logical nodes and the corresponding functional semantics of the configurable parameters.

[0044] Step 203: Determine the visualization component corresponding to the configurable parameters.

[0045] Step 204: Obtain user interface (UI) data based on the visualization components corresponding to the configurable parameters. The UI data is used to render the parameter configuration interface. The parameter configuration interface includes the visualization components and functional semantics corresponding to the configurable parameters. The visualization components are used to configure the parameter values ​​of the configurable parameters in the workflow file.

[0046] As can be seen from the above process, this disclosure can parse the underlying workflow file to obtain the configurable parameters and determine the corresponding visualization components. Based on the visualization components, UI data for rendering the parameter configuration interface is obtained. This disclosure can automatically generate a parameter configuration interface based on the functional semantics of the configurable parameters and the visualization components. On the one hand, automatically generating UI data can improve the work efficiency of the first user and save a significant amount of time. On the other hand, the generated parameter configuration interface allows the second user to directly configure the parameter values ​​of the configurable parameters through the visualization components based on the functional semantics, without needing to understand the underlying technical logic. This effectively lowers the operational threshold for the second user and reduces configuration errors caused by directly manipulating parameters in the workflow file.

[0047] The following describes in detail each step of the above process and the effects that can be produced, with reference to the embodiments.

[0048] First, the above step 201, "obtaining workflow files", will be described in detail with reference to the embodiments.

[0049] In the embodiments of this disclosure, the workflow file to be parsed can first be obtained. This workflow file can be uploaded directly by the first user through their corresponding user terminal, or it can be obtained online by the server by calling the open interface provided by the intelligent agent platform, based on the identifier of the target intelligent agent selected by the first user. The format of the workflow file can be JSON, YAML, XML, or any other structured data serialization format.

[0050] The workflow file represents the logical flow executed by the target agent when responding to a request. This logical flow includes multiple logical nodes, each containing corresponding parameters. A logical node can be understood as the basic execution unit constituting the workflow file, such as a condition checker, API caller, database queryer, or large language model caller. For example, a workflow file can describe the complete process starting from the query input node, passing through the intent recognition node, knowledge base retrieval node, large language model generation node, and finally reaching the response output node, where each logical node contains its own parameters.

[0051] The following describes step 202, namely "parse the workflow file to obtain the configurable parameters of multiple logical nodes and the functional semantics corresponding to the configurable parameters", in detail with reference to the embodiments.

[0052] In this embodiment of the disclosure, the server parses the workflow file and extracts the attribute information of each logical node. The attribute information includes node type, node ID, node name, and parameter information. The attribute information of the logical nodes is identified by a preset rule base and keywords to obtain the configurable parameters of multiple logical nodes and the functional semantics corresponding to the configurable parameters.

[0053] It should be noted that configurable parameters of a logical node can refer to all parameters included in the logical node's parameter information. In other words, the server can treat all parameters parsed from the logical nodes included in the workflow file as configurable parameters. Configurable parameters of a logical node can also refer to important parameters included in the logical node. That is, the server can identify important parameters from all parsed parameters and treat them as configurable parameters. For example, the server can determine whether a parameter is important based on the node type of the logical node to which it belongs. Configurable parameters of a logical node can also refer to parameters selected by the first user from all parameters through their corresponding client. In other words, the server will display all parameters parsed from the workflow file to the first user through their client. The first user's client, in response to the first user's selection operation on all displayed parameters, will return the selected parameter to the server, and the server will treat the selected parameter as a configurable parameter.

[0054] As a specific implementation, the configurable parameters of logical nodes can be obtained as follows: The server parses the workflow file, extracts the attribute information of each logical node, and, based on the node type and node name included in the attribute information of each logical node, and based on the service information of the target service used by the target agent (such as service name, service purpose, service scenario, etc.), determines whether each logical node is an important node of the target agent. Then, all parameters of the important nodes are extracted as configurable parameters.

[0055] This approach can identify the more important logical nodes for the target agent based on the node type and node name of each logical node, and treat the parameters of the important nodes as configurable parameters. In other words, this disclosure does not blindly identify all parameters, but selects the more important parameters for visualization, thereby improving the efficiency of subsequent configuration.

[0056] As another specific embodiment, the configurable parameters of logical nodes can be obtained in the following way: The server parses the workflow file, extracts the attribute information of each logical node, identifies the parameter information included in the attribute information of the logical node, obtains all parameters of multiple logical nodes and the correlation between all parameters, and based on the correlation between all parameters, groups all parameters (i.e., divides the parameters with correlation into a group) to obtain multiple candidate sets, selects the target set from the multiple candidate sets whose correlation with the service information of the target service used by the target agent exceeds the correlation threshold, and uses the parameters included in the target set as configurable parameters.

[0057] The configurable parameters selected in this way are related. When the related configurable parameters are converted into a parameter configuration interface, a second user can jointly adjust the related configurable parameters, which ensures the accuracy and consistency of the configurable parameters and improves the robustness of the target intelligent agent.

[0058] It should also be noted that the functional semantics corresponding to configurable parameters can be text information describing the role of the configurable parameter in the workflow file (i.e., what the configurable parameter does). Specifically, the server can parse the workflow file to obtain the parameter information (e.g., parameter name, parameter value, parameter type, etc.) included in each logical node. The parameter name of the configurable parameter can be directly used as its corresponding functional semantics, or the parameter name of the configurable parameter can be translated to obtain the functional semantics of the configurable parameter. Alternatively, the first user can determine the business alias of the configurable parameter based on the parameter information (e.g., parameter name) displayed on their corresponding client, and the server can use the business alias of the configurable parameter as its corresponding functional semantics.

[0059] In this context, the business alias for configurable parameters can be understood as a more business-readable name set by the first user for the configurable parameter, replacing the original parameter name. Using the business alias set by the first user as functional semantics can eliminate cognitive barriers caused by technical jargon and realize the business semanticization of the parameter configuration interface. For example, the first user can set the business alias of the configurable parameter "query_follow_up" to "whether to enable follow-up question function".

[0060] The following describes step 203, namely "determining the visualization component corresponding to the configurable parameters", in detail with reference to the embodiments.

[0061] Next, the server can determine the visual components corresponding to the configurable parameters. These visual components can refer to any UI element that can be presented on the user interface and receive user input or selection, such as input boxes, sliders, switches, dropdown selection boxes, date pickers, etc.

[0062] Specifically, the server can determine the component type of the visualization component corresponding to the configurable parameter based on at least one of the following: parameter name, business alias, functional semantics, historically set business alias, and parameter type.

[0063] The business alias for configurable parameters can be set by the first user. One possible approach is for the first user to convert the parameter name of the configurable parameter into a simple and easy-to-understand business alias. For example, the first user could set the business alias for the configurable parameter "temperature" to "Answer Style (Conservative / Open)". Another possible approach is for the first user to first obtain the translated parameter name of the configurable parameter, and then further refine the translated parameter name to obtain the business alias. For example, the translated configurable parameter "temperature" could be simply "temperature", and the first user could set the business alias corresponding to "temperature" to "Answer Style (Conservative / Open)" based on common knowledge in the field.

[0064] Then, based on the component type of the visualization component corresponding to the configurable parameter, the visualization component corresponding to the configurable parameter is determined. That is, based on the component type of the visualization component corresponding to the configurable parameter, the definition of the corresponding visualization component is instantiated or referenced, thereby determining the final visualization component.

[0065] For example, the component type of the visualization component corresponding to the configurable parameter can be determined based on the parameter type of the configurable parameter (such as Boolean type, numeric type, string type, enumeration type). For example, the component type of the visualization component corresponding to the Boolean type configurable parameter is a switch, the component type of the visualization component corresponding to the numeric type configurable parameter is a number input box, the component type of the visualization component corresponding to the string type configurable parameter is a text input box, and the component type of the visualization component corresponding to the enumeration type configurable parameter is a checkbox.

[0066] For example, based on the business alias of the configurable parameter, the component type of the corresponding visualization component can be determined. For instance, if the business alias of the configurable parameter is "Enable Smart Follow-up Questions", the component type of the corresponding visualization component is a switch. If the business alias of the configurable parameter is "Number of Follow-up Questions", the component type of the corresponding visualization component is a numeric input box or a slider.

[0067] As can be seen, the component type of the corresponding visual component can be determined based on at least one of the following: parameter name, business alias, functional semantics, historically set business alias, and parameter type. This method automates the mapping of configurable parameters to visual components, improving the efficiency of the first user obtaining the parameter configuration interface and the accuracy of the determined visual component, thereby increasing the reliability of the entire mapping process.

[0068] In addition, this disclosure can also obtain visualization component information of historical mappings of configurable parameters, and determine the component type of the visualization component corresponding to the configurable parameter based on the historical mapping visualization component information. For example, if a certain configurable parameter has been mapped to a visualization component of component type "switch" multiple times in history, then when it is necessary to determine the component type of the visualization component corresponding to the configurable parameter, the server can directly use "switch" as the component type of the visualization component corresponding to the configurable parameter.

[0069] As one feasible implementation method, based on at least one of the following: parameter name, business alias, functional semantics, business alias of historical settings, parameter type, and visualization component information of historical mapping of configurable parameters, multiple recommended component types are generated for configurable parameters. The multiple recommended component types are displayed to the first user through the user terminal of the first user, and one of the multiple recommended component types selected by the first user is obtained as the component type of the visualization component corresponding to the configurable parameter.

[0070] For example, the first user can set the configurable parameter "max_tokens" as the business alias to "maximum number of characters in the answer". The server recognizes that the word "maximum" in the business alias is usually related to a numerical range, and the parameter type of this configurable parameter is numeric. Based on this, multiple recommended component types are generated. These recommended component types can be numeric input boxes and sliders. In response to the first user triggering the display of recommended component types, a list of recommended component types is displayed, including two options: "numeric input box" and "slider". The first user can choose one of the two options as the component type of the visualization corresponding to the configurable parameter according to their actual needs (such as precise control or approximate range control).

[0071] This approach can generate multiple recommended component types simultaneously, allowing the first user to choose which type to use. It provides automated intelligent suggestions to improve efficiency while respecting the first user's final decision-making power, ensuring the flexibility of the visual component configuration method and meeting the preferences of different users and the needs of specific scenarios.

[0072] As another feasible implementation, the server can generate multiple recommended component types for the configurable parameters based on at least one of the following: parameter name, business alias, functional semantics, business alias of historical settings, parameter type, and visualization component information of historical mapping. The multiple recommended component types are then input into a large language model so that the large language model can predict the recommended component type that the first user is most likely to choose, which will then be used as the component type of the visualization component corresponding to the configurable parameters.

[0073] The following describes step 204, namely "obtaining UI data based on the visualization component corresponding to the configurable parameters", in detail with reference to the embodiments.

[0074] In this embodiment, the server can obtain UI data based on the visual components corresponding to the configurable parameters. Specifically, the server can acquire various preset interface templates (such as two-column, wizard-style, and tiled templates), and then select a suitable interface template according to the number and importance of the configurable parameters. The visual components are then populated into the selected interface template to generate the final UI data. UI data is structured data describing the interface structure and behavior, such as data conforming to the JSON Schema specification or configuration descriptors that satisfy a specific front-end framework. The UI data in this disclosure is used to render the parameter configuration interface. The parameter configuration interface includes visual components and functional semantics corresponding to the configurable parameters. The visual components are used to configure the parameter values ​​of the configurable parameters in the workflow file, and the functional semantics help users of the parameter configuration interface understand the function of each configurable parameter controlled by each visual component.

[0075] It's important to note that UI data is an intermediate data format used to describe how the parameter configuration interface is constructed; it is not the final rendered pixel-based interface itself. The parameter configuration interface, on the other hand, refers to the graphical user interface ultimately presented to the second user for completing parameter configuration tasks. In other words, based on UI data, an interactive parameter configuration interface can be rendered on different terminals (i.e., the second user's client) such as browsers and mini-programs. This interface directly displays the functional semantics of each configurable parameter and its corresponding visual components. The second user can configure the parameter values ​​of the configurable parameters in the workflow file by manipulating these visual components.

[0076] To lower the configuration barrier for second users while ensuring the rigor and security of the configuration, effective constraints and validations of the second user's input are necessary. Specifically, parsing the workflow file can yield constraint rules for configurable parameters. When obtaining UI data based on the visualization component corresponding to the configurable parameters, these constraint rules can be written into the configuration properties of the visualization component and registered with the validation rules of that visualization component to obtain the UI data.

[0077] The constraint rules include at least one of the following: default values ​​for configurable parameters, value ranges, validation rules, and inter-parameter linkage rules. It should be noted that default values ​​refer to the default parameter values ​​corresponding to the configurable parameters; value ranges refer to the range of parameter values ​​that can be configured for the configurable parameters; validation rules refer to the format validation of the parameter values ​​(e.g., whether a configurable parameter is an email address); and inter-parameter linkage rules refer to the association rules between the parameter values ​​of two configurable parameters (e.g., when the parameter value of configurable parameter A is X, the visualization component corresponding to configurable parameter B can be displayed and set as a required field).

[0078] By injecting constraint rules into the UI data as described above, basic restrictions (such as the maximum number of calls allowed by the interface and the valid range of parameter values) or business restrictions (such as the parameter values ​​of two configurable parameters cannot be true at the same time) of configurable parameters in the underlying workflow file are applied to the parameter configuration interface. This prevents second users from entering incorrect parameter values ​​in the visual components of the parameter configuration interface, thereby improving the stability and security of parameter configuration.

[0079] As one feasible approach, constraint rules can be parsed from the workflow file. For example, during the parsing of the workflow file, when the "max_tokens" parameter is encountered, the server can obtain that the value range of "max_tokens" is [1, 300], and the default value is 200. As another example, during the parsing of the workflow file, the server can parse "When the knowledge base retrieval switch is enabled, a similarity threshold must be set" (i.e., a parameter linkage rule).

[0080] As another possible approach, constraint rules can also be obtained from the first user's client. Specifically, a component configuration interface can be displayed on the first user's client. The component configuration interface includes rule configuration components corresponding to configurable parameters. The first user can enter the rule configuration interface by triggering the rule configuration component, configure constraint rules in the rule configuration interface, and after the configuration is completed, the first user's client sends the constraint rules to the server.

[0081] As another possible approach, the constraint rules can also be obtained by the first user adjusting the constraint rules parsed from the workflow file. Specifically, a component configuration interface is displayed on the first user's client. The component configuration interface includes rule configuration components corresponding to configurable parameters. In response to the operation of triggering the rule configuration component, the constraint rules of configurable parameters are parsed from the workflow file. Based on the obtained constraint rules, the rule configuration interface is displayed. In response to the configuration operation of the constraint rules in the rule configuration interface, the configured constraint rules are sent to the server.

[0082] In addition, during the configuration of configurable parameters using visual components, the location of configurable parameters in the workflow file can be located through structural identifiers. Generally, structural identifiers are determined based on the structure of the workflow, for example, "nodes[3].parameters.temperature", which is the "temperature" parameter of the third logical node. To further ensure the accuracy and traceability of the configuration, especially to accurately locate configurable parameters when the structure of the workflow changes, this disclosure provides a method for generating semantic identifiers for configurable parameters.

[0083] Specifically, during the parsing of the workflow file, the functional semantics of multiple logical nodes can be obtained. Based on the functional semantics of these logical nodes, semantic identifiers for configurable parameters are determined. These semantic identifiers are used to locate the position of the configurable parameters in the workflow file during the configuration of the configurable parameters using the visual components.

[0084] For example, the semantic identifier for the "temperature" parameter in a logical node with the functional semantics of "large language model call" could be "llm.parameters.temperature". This way, even if the structural position of the logical node to which the configurable parameter belongs changes in the workflow file, as long as the functional semantics of the logical node remain unchanged, the server can still locate the configurable parameter's position in the workflow file using the semantic identifier.

[0085] Of course, in addition to determining the semantic identifier of configurable parameters based on the functional semantics of multiple logical nodes, a globally unique identifier can be set for each configurable parameter, or a dynamic mapping table from semantic identifier to structural identifier can be established and updated when the workflow structure changes. All of the above methods can accurately locate the position of configurable parameters in the workflow file. This disclosure will not provide examples of each method here.

[0086] As can be seen in the above embodiments, when logical nodes in the workflow file are added, deleted, or their order is adjusted, semantic identifiers can be used to locate the position of configurable parameters in the workflow file. This ensures the stability of the mapping relationship between configurable parameters and visualization components during workflow iteration, and avoids the failure of visualization component configuration due to structural adjustments of underlying logical nodes.

[0087] To fully verify the correctness of UI data before configuration release and minimize online risks, the first user can perform simulated tests on the parameter configuration interface. Specifically, the parameter configuration interface can be rendered on the first user's client based on UI data. The first user configures simulated values ​​for the corresponding configurable parameters in the visual components included in the parameter configuration interface and sends these simulated values ​​to the server. The server receives the simulated values, loads a copy of the workflow file in the isolated environment, and replaces the parameter values ​​of the configurable parameters in the copy of the workflow file with the simulated values. Next, based on the acquired test data, the server executes test tasks using the copy of the workflow file to obtain test results. Finally, based on the test results, the success of the configuration is verified.

[0088] Executing a test task is essentially a complete simulation of the workflow file. Test data can be manually entered by the user or automatically generated by the server. Verifying successful configuration includes not only whether the parameter configuration interface can be displayed correctly and used to configure configurable parameters, but also whether a copy of the workflow file can be executed normally without runtime errors after configuring the configurable parameters.

[0089] This allows for testing the effectiveness of the parameter configuration interface based on test data before its official release, further improving the stability of the entire method and greatly reducing release risks.

[0090] Furthermore, the server can also determine the simulation report corresponding to the copy of the workflow file during the execution of the test task. The simulation report includes the response time of each logical node and the cost of calling the large model (e.g., the cost obtained based on the pricing strategy of the interface of the large model being called). Then, based on the response time and cost, a visual heatmap is determined. The visual heatmap is used to prompt the first user to adjust the UI data based on the visual heatmap, such as adjusting the value range of the configurable parameter for the context length, so that the configurable parameter for the context length is kept at a low value, thereby reducing subsequent costs.

[0091] This method of identifying visual heatmaps can intuitively display the logical nodes with the longest response times or the highest costs, guiding first-time users to make targeted optimizations, assisting them in making more informed configuration decisions, and improving their user experience.

[0092] In addition, if the response time of a logical node is abnormal or the cost exceeds the warning threshold, the server can highlight the response time and cost of that logical node in a visual heatmap to prompt users to pay close attention to that logical node.

[0093] Furthermore, in response to the first user triggering the display of the logical node (hereinafter referred to as the target node) in the visual heatmap, a component configuration interface is displayed. At this time, the component configuration interface displays the rule configuration components corresponding to the configurable parameters included in the target node. The first user can enter the rule configuration interface of the corresponding configurable parameter by triggering the rule configuration component, and configure the constraint rules of the configurable parameter in the rule configuration interface. In response to the configuration, the first user's client sends the constraint rules corresponding to the configurable parameters included in the target node to the server.

[0094] As one feasible approach, the server adjusts the UI data based on the constraints of the configurable parameters included in the received target node. On the first user's client, the parameter configuration interface is rendered based on the adjusted UI data. This interface can then display only the visual components corresponding to the configurable parameters of the target node. Based on this, the server can obtain simulated values ​​of the configurable parameters of the target node configured by the first user through this interface. Then, a copy of the target node from the workflow file is loaded into the isolated environment, and the simulated values ​​of the target node's configurable parameters are used to replace the parameter values ​​in the copy. Based on the obtained test data for the target node, the target node's test tasks are executed using the copy to obtain test results. Based on these test results, the success of the target node's configuration is verified.

[0095] In other words, if UI data is adjusted based on a visual heatmap, and the adjusted UI data is then simulated and verified again, testing can be performed only on the problematic logical node. This approach allows the first user to quickly adjust the problematic logical node without having to pay too much attention to other logical nodes, improving adjustment efficiency. Furthermore, subsequent tests only need to test the logical node that needs adjustment, significantly saving testing resources.

[0096] As another possible approach, the server adjusts the constraint rules of the configurable parameters in other logical nodes (excluding the target node) based on the constraint rules of the received target node's configurable parameters. This is because the constraint rules between configurable parameters in different logical nodes may be related; for example, if the default value of configurable parameter 1 is changed to 0.1, then the default value of configurable parameter 2 should be changed to 0.5. Based on the constraint rules of the received target node's configurable parameters and the adjusted constraint rules of the configurable parameters in other logical nodes, the UI data is adjusted to obtain the adjusted UI data.

[0097] Then, on the first user's client, the parameter configuration interface is rendered based on the adjusted UI data. The simulated values ​​of the configurable parameters configured by the first user through the visual components included in the parameter configuration interface are obtained. A copy of the workflow file is loaded in the isolated environment, and the parameter values ​​of the configurable parameters in the copy of the workflow file are replaced with the simulated values ​​of the configurable parameters. Based on the obtained test data, the test task is executed using the copy of the workflow file to obtain the test results. Based on the test results, the success of this configuration is verified.

[0098] In other words, this disclosure can adjust the constraint rules of configurable parameters of other logical nodes based on the constraint rules of configurable parameters included in the target node, ensuring the consistency of the adjustment process.

[0099] For example, firstly, the parameter configuration interface is rendered on the user's client based on UI data. The first user configures the simulated values ​​of the configurable parameters in the copy of the workflow file through the parameter configuration interface (such as configuring the answer creativity to 0.2 and the maximum answer length to 500). The server obtains the simulated values ​​of the configurable parameters, loads the copy of the workflow file in the isolated environment, and replaces the parameter values ​​of the configurable parameters in the copy of the workflow file with the simulated values ​​of the configurable parameters.

[0100] Then, the first user can input test data (e.g., a child has had a fever for three days) through their corresponding client. Based on the obtained test data, the server uses a copy of the workflow file to execute the test task to obtain test results (e.g., a recommendation to go to a fever clinic first) and a simulation report, and then returns the test results and simulation report to the first user's client.

[0101] The first user's client receives and displays the test results and simulation report. Next, the first user can adjust the UI data based on the test results and simulation report through the component configuration interface displayed on the first user's client. For example, the default value or range of the maximum answer length can be reduced to reduce the cost and response time of the relevant logical node. Then, the test is performed again until the parameter configuration interface passes the test task.

[0102] Next, in response to a successful test result, the server can encapsulate the UI data into a configuration snapshot with a unique version identifier. Then, the configuration snapshot is managed based on its release status, which includes at least draft, pending approval, pending release, live, and offline.

[0103] Specifically, in response to a successful test result, a save component can be displayed on the first user's client. The first user's client, responding to the user's triggering of the save component, packages the UI data and its corresponding workflow file identifier into a configuration snapshot with a unique version identifier. The first user can then manage different versions of the configuration snapshot through the version management list displayed in the publishing interface. For example, only configuration snapshots with a publishing status of "online" will provide the second user's client with the service of configuring parameters through the parameter configuration interface via the interface. Furthermore, if the first user wants to publish a new version of the configuration snapshot, they need to set the publishing status of the old version's configuration snapshot to "offline" and then set the publishing status of the new version's configuration snapshot to "online." Additionally, the server can set version rules, such as allowing only one version of the configuration snapshot to have a publishing status of "online" at a time, to avoid version conflicts.

[0104] The server can also record changes to the release status of configuration snapshots for different versions in a preset log, which facilitates subsequent backtracking and improves the reliability of the configuration process.

[0105] This approach strictly manages configuration snapshots for different versions through release status, ensuring the stability of the online environment and the traceability of changes, and greatly guaranteeing the stability and maintainability of the parameter configuration interface.

[0106] Based on this, the first user manages different versions of configuration snapshots through their corresponding client. If the second user requests to display the parameter configuration interface through their corresponding client, the server will send a configuration snapshot with a "released" status to the second user's client. Different types of terminals (such as web pages, mini-programs, applications, etc.) will send their corresponding configuration snapshots. After receiving the configuration snapshot, the second user's client can render the parameter configuration interface based on the UI data included in the configuration snapshot. The second user can adjust the parameter values ​​of configurable parameters in the workflow file through the parameter configuration interface. If the first user releases a new version of the configuration snapshot on their corresponding client, the second user's client can use hot update, meaning the second user can see the parameter configuration interface corresponding to the new version of the configuration snapshot without closing the currently displayed parameter configuration interface.

[0107] The above describes the methods executed on the server side. Next, we will further explain the methods executed on the client side for the first user. Figure 3 This is a flowchart illustrating a method for configuring visual parameters for a workflow, provided in an embodiment of this disclosure. This method can be implemented by... Figure 1 The first user's client in the system shown executes the command. For example... Figure 3 As shown, the method may include the following steps: Step 301: In response to triggering a parsing operation on the workflow file, the workflow file is used to characterize the logical flow executed by the target agent when responding to a request. The logical flow includes multiple logical nodes. The configurable parameters of the multiple logical nodes parsed from the workflow file and the functional semantics corresponding to the configurable parameters are obtained, as well as the visualization components corresponding to the configurable parameters are obtained.

[0108] Step 302: Display the configurable parameters of multiple logical nodes, the functional semantics corresponding to the configurable parameters, and the visual components in the component configuration interface.

[0109] Step 303: In response to the confirmation operation triggered in the component configuration interface, send the confirmed configurable parameters, the corresponding functional semantics of the configurable parameters, and the visual components for the server to generate UI data. The UI data is used to render the parameter configuration interface. The parameter configuration interface includes the visual components and functional semantics corresponding to the configurable parameters. The visual components are used to configure the parameter values ​​of the configurable parameters in the workflow file.

[0110] As can be seen from the above process, this disclosure can respond to the parsing operation of a workflow file, obtain configurable parameters of multiple logical nodes, the functional semantics corresponding to the configurable parameters, and the visual components corresponding to the configurable parameters. The obtained information is then displayed in the component configuration interface, and subsequently sent to the server to generate UI data. On the one hand, the component configuration interface can display the visual components and functional semantics of the configurable parameters to the first user, facilitating confirmation and ensuring the accuracy of the visual components and functional semantics, thus improving the first user's experience. On the other hand, when the generated parameter configuration interface is released to a second user without a technical background, it effectively lowers the operational threshold for the second user and reduces configuration errors caused by directly manipulating parameters in the workflow file.

[0111] In the embodiments of this disclosure, the first user can first upload a workflow file on the parsing interface, that is, the first user's client responds to the operation of the first user uploading the workflow file and obtains the workflow file. Alternatively, the first user can select a candidate agent as the target agent on the parsing interface, that is, the first user's client responds to the operation of selecting one of multiple candidate agents and uses the selected candidate agent as the target agent so that the server can obtain the workflow file of the target agent.

[0112] The parsing interface can be like Figure 4As shown, the first user can click the file upload component displayed in the parsing interface to select the workflow file to upload. The user's client can then send the uploaded workflow file to the server. Alternatively, the first user can click the agent selection component to select the target agent. The user's client can then send the identifier of the selected target agent to the server. Based on the target agent's identifier, the server can call the open interface provided by the agent platform to obtain the target agent's workflow file.

[0113] In this way, the first user can obtain workflow files in multiple ways, enriching the sources of workflow files. Furthermore, the first user can make the server obtain the workflow files by using the parsing interface, reducing the operational complexity for the first user and improving the work efficiency of the first user.

[0114] As one possible approach, the first user's client responds by triggering a parsing operation on the workflow file (such as clicking). Figure 4 The "Start Parsing" button in the document retrieves the configurable parameters of multiple logical nodes parsed from the workflow file, the corresponding functional semantics of the configurable parameters, and the corresponding visual components.

[0115] Then, the user's client can jump directly from the parsing interface to the component configuration interface, and the component configuration interface displays the configurable parameters of multiple logical nodes, as well as the functional semantics and visual components corresponding to the configurable parameters.

[0116] As another possible approach, the first user's client responds to triggering a parsing operation on the workflow file (such as clicking). Figure 4 The "Start Parsing" button in the document retrieves all parameters of multiple logical nodes parsed from the workflow file and jumps to the parameter selection interface to display all the parsed parameters. For example, the parameter selection interface can display the parameter name, parameter type, node ID of the logical node to which it belongs, node name of the logical node, node type of the logical node to which it belongs, etc.

[0117] like Figure 5 As shown, Figure 5 The parameter selection interface displays a parameter table. Each row of the parameter table corresponds to a parameter, and each column corresponds to an attribute of the parameter (such as parameter name, parameter type, node ID of the logical node to which it belongs, node name of the logical node, node type of the logical node to which it belongs, etc.).

[0118] In response to the first user's selection action on all parameters (such as clicking) Figure 5The leftmost column of the parameter table contains the selected components, and the first user triggers the component configuration function (such as clicking...). Figure 5 The "Start Component Configuration" button will select the parameter as a configurable parameter and redirect to the component configuration interface, which displays the configurable parameter, its corresponding functional semantics, and the visual component.

[0119] in addition, Figure 5 The parameter selection interface also displays a first filter component. In response to the operation of triggering the first filter component, multiple node types are displayed. In response to selecting one of the node types, the attribute information of the parameters corresponding to the selected node type can be displayed in the parameter table.

[0120] In some cases, the number of parameters obtained from the workflow file may be large. To make it easier for the first user to locate the required parameters among all the parameters, Figure 5 The parameter selection interface also provides a second search component. In response to entering a node name or parameter name in the second search component, the attribute information of the parameter corresponding to the entered node name or parameter name is displayed.

[0121] This disclosure provides a component configuration interface, which displays configurable parameters, corresponding functional semantics, and visual components. The visual components displayed here may refer to the component diagram, component ID, component name, etc. of the visual components corresponding to the configurable parameters.

[0122] As a feasible approach, such as Figure 6 As shown, the actual display in the component configuration interface can be a configurable parameter table. Each row of the configurable parameter table corresponds to a parameter, and each column corresponds to an attribute information of the configurable parameter (such as parameter name, parameter type, node ID of the logical node to which it belongs, node name of the logical node, node type of the logical node to which it belongs, etc.) and configuration information (such as functional semantics, business alias and component type of the visualization component, etc.).

[0123] The component type of the visualization component is determined based on at least one of the following: the business alias of the configurable parameters, the functional semantics, the historically set business alias, and the parameter type. Additionally, if the first user is not satisfied with the determined component type of the visualization component, a drop-down menu can be triggered. At this time, a list of recommended component types corresponding to the visualization component is displayed on the first user's client. The recommended component list includes multiple recommended component types, and the first user can select one from these as the component type for the visualization component corresponding to the configurable parameters. The business alias of the configurable parameters is set by the first user on the component configuration interface. The specific methods for determining the business alias of the configurable parameters and the component type of the visualization component have been explained in steps 202 and 203, and will not be repeated here.

[0124] in addition, Figure 6 The component configuration interface also displays a component deletion option. In response to the triggering of the component deletion operation, the parameters corresponding to the component to be deleted are removed from the configurable parameters. Figure 6 The component configuration interface also displays a second filter component. In response to the operation of triggering the second filter component, multiple parameter types are displayed. In response to selecting one of the parameter types, only the attribute information and configuration information of the parameter corresponding to the selected parameter type can be displayed in the configurable parameter table.

[0125] In some cases, the number of configurable parameters obtained from the workflow file may be large. To make it easier for the first user to locate the required parameters among the configurable parameters, Figure 6 The component configuration interface also provides a second search component. In response to the input of a function semantic or parameter name in the second search component, the attribute information and configuration information of the parameter corresponding to the input function semantic or parameter name are displayed.

[0126] As can be seen, the component type of the corresponding visual component can be determined based on at least one of the following: parameter name, business alias, functional semantics, historically set business alias, and parameter type. This method automates the mapping of configurable parameters to visual components, improving the efficiency of the first user obtaining the parameter configuration interface and the accuracy of the determined visual component, thereby increasing the reliability of the entire mapping process.

[0127] In addition, the component configuration interface can also display rule configuration components with configurable parameters. As one possible approach, in response to the operation of triggering the rule configuration component, the rule configuration interface is displayed. In response to the configuration operation performed by the first user in the rule configuration interface, the constraint rules are obtained and sent so that the server can write the constraint rules into the configuration properties of the corresponding visualization component and register the validation rules of the visualization component when generating UI data.

[0128] As another possible approach, in response to the operation of triggering the rule configuration component, the constraint rules of the configurable parameters parsed from the workflow file are obtained, and based on the constraint rules of the configurable parameters, the rule configuration interface corresponding to the configurable parameters is displayed. The constraint rules include at least one of the default values, value ranges, verification rules, and inter-parameter linkage rules of the configurable parameters.

[0129] In response to the configuration operation of constraint rules in the rule configuration interface, the configured constraint rules are sent so that the server can write the configured constraint rules into the configuration properties of the corresponding visualization component and register the validation rules of the visualization component when generating UI data.

[0130] The rule configuration interface is used to configure the parsed constraint rules, making the constraint rules more in line with the needs of the first user, so as to obtain more accurate and complete constraint rules. This improves the stability and security of the parameter configuration process for the second user, and the first user does not need to configure the constraint rules from scratch, thus improving the configuration efficiency.

[0131] Furthermore, the first user's client responds to a confirmation action triggered in the component configuration interface (such as clicking). Figure 6 The confirmation component sends configurable parameters, corresponding functional semantics, and visualization components to the server for the server to generate UI data.

[0132] For example, when the first user's client displays a component configuration interface, the component configuration interface displays a table of configurable parameters. Each column of the configurable parameter table is used to represent the parameter name, functional semantics (such as the parameter name after translation processing), parameter type, business alias, and component type of the visualization component of the configurable parameter. The first user can set the business alias for each configurable parameter in the column where the business alias is located. In response to the confirmation operation triggered in the component configuration interface, the first user sends the confirmed configurable parameters, the corresponding functional semantics of the configurable parameters, and the component type of the visualization component to the server so that the server can generate UI data for rendering the parameter configuration interface.

[0133] It should be noted that if the first user has not set a corresponding business alias for a configurable parameter, the functional semantics corresponding to the configurable parameter sent to the server in response to the confirmation operation triggered in the component configuration interface are the same as the functional semantics displayed in the component configuration interface. If the first user has set a corresponding business alias for a configurable parameter, the functional semantics corresponding to the configurable parameter will be updated based on the business alias in response to the confirmation operation triggered in the component configuration interface, and the updated functional semantics will be sent to the server.

[0134] The UI data used to render the parameter configuration interface can be obtained through the above method. In order to fully verify the correctness of the UI data before the configuration is released, the first user's client responds to the confirmation operation triggered in the component configuration interface (such as clicking). Figure 6 The first user's client receives the UI data and renders the parameter configuration interface based on the UI data. The first user configures the simulated value of the corresponding configurable parameter in the visualization component included in the parameter configuration interface and inputs test data in the parameter configuration interface, and then sends the simulated value of the configurable parameter and the test data to the server.

[0135] After obtaining the simulated values ​​of the configurable parameters, the server loads a copy of the workflow file in the isolated environment and replaces the parameter values ​​in the copy with the simulated values. Next, based on the acquired test data, the server executes a test task using the copy of the workflow file to obtain the test results, which are then returned to the first user's client. The first user can then determine whether the configuration was successful based on the test results.

[0136] Additionally, during the execution of test tasks, the server can determine the simulation report corresponding to the copy of the workflow file. The simulation report includes the response time of each logical node and the cost of calling the large model. Then, based on the response time and cost, a visual heatmap is generated and sent to the client of the first user.

[0137] The first user's client receives and displays a visual heatmap, allowing the first user to return to the component configuration interface to adjust UI data based on the visual heatmap.

[0138] In response to the test result being passed, the save component can be displayed on the first user's client. In response to the first user triggering the save component operation, the first user's client packages the UI data and its corresponding workflow file identifier into a configuration snapshot with a unique version identifier.

[0139] This disclosure also provides a management interface for different versions of UI data. The management interface includes multiple configuration snapshots with unique version identifiers. These configuration snapshots are managed by a first user through the management interface, allowing them to control the release status of different versions of the snapshots. Release statuses include at least draft, pending approval, pending release, live, and offline. Subsequently, the server can distribute the live configuration snapshot to a second user's client, enabling the second user's client to display the parameter configuration interface corresponding to the live snapshot. The parameter configuration interface includes visual components and functional semantics corresponding to the configurable parameters. The functional semantics prompt the second user to use the visual components to configure the parameter values ​​of the configurable parameters in the workflow file.

[0140] To illustrate the method of the present invention more clearly, the following description is based on a configuration scenario of a triage agent in the medical field.

[0141] The first user selects a triage agent in the parsing interface displayed on their corresponding user terminal and triggers a parsing operation on the workflow file. The first user's user terminal sends the identifier of the triage agent to the server. Based on the identifier of the triage agent, the server obtains the workflow file of the triage agent and completes the parsing, obtaining the configurable parameters of multiple logical nodes and the functional semantics corresponding to the configurable parameters, as well as the visualization components corresponding to the configurable parameters. The server returns the configurable parameters of multiple logical nodes, the functional semantics corresponding to the configurable parameters, and the visualization components to the first user's user terminal.

[0142] The first user's client displays configurable parameters, corresponding functional semantics, and visual components for multiple logical nodes in the component configuration interface. The first user can set business aliases for configurable parameters in the component configuration interface. In response to the confirmation operation triggered in the component configuration interface, the functional semantics of the configurable parameters are updated based on the business aliases of the configurable parameters. The confirmed configurable parameters, the updated functional semantics, and visual components are sent to the server for generating UI data.

[0143] Subsequently, the first user can conduct a simulation test on the parameter configuration interface rendered based on the UI data on their corresponding user client. For the parameter configuration interface that passes the simulation test, the server will encapsulate the UI data into a configuration snapshot with a unique version identifier.

[0144] When the second user's client requests to display the parameter configuration interface, the server can send a configuration snapshot with a published status of "online" to the second user's client. After receiving the configuration snapshot, the second user's client can render the parameter configuration interface based on the UI data included in the configuration snapshot. The second user can adjust the parameter values ​​of configurable parameters in the workflow file through the parameter configuration interface.

[0145] In this context, the first user can refer to technical personnel with a technical background, while the second user can refer to business personnel such as doctors and nurses who use the triage agent. This way, even if business personnel don't understand the parameter names of the configurable parameters, they can still optimize the triage agent's performance based on the functional semantics and visual components corresponding to the configurable parameters. It should be noted that the method disclosed herein can be applied not only to triage agents in the medical field, but also to customer service agents in the shopping field, image analysis agents in the industrial field, question-answering agents in the educational field, and various other intelligent agents in various other fields, which will not be listed here.

[0146] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0147] The foregoing has described specific embodiments of this disclosure. 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 also possible or may be advantageous.

[0148] According to another embodiment, a visual parameter configuration device for workflow is provided. Figure 7 A schematic block diagram of a workflow visualization parameter configuration apparatus according to one embodiment is shown, the apparatus being disposed in... Figure 1 The server side in the illustrated architecture. For example... Figure 7 As shown, the workflow visualization parameter configuration device 700 includes a first acquisition unit 701, a parsing unit 702, a component determination unit 703, and a UI data determination unit 704, and further includes an identifier determination unit 705, a simulation unit 706, and a management unit 707. The main functions of each component are as follows: The first acquisition unit 701 is configured to acquire a workflow file, which is used to characterize the logical flow executed by the target intelligent agent when responding to a request. The logical flow includes multiple logical nodes.

[0149] The parsing unit 702 is configured to parse the workflow file to obtain the configurable parameters of multiple logical nodes and the functional semantics corresponding to the configurable parameters.

[0150] The component determination unit 703 is configured to determine the visual component corresponding to the configurable parameters.

[0151] The UI data determination unit 704 is configured to obtain user interface (UI) data based on the visualization component corresponding to the configurable parameter. The UI data is used to render the parameter configuration interface, which includes the visualization component and functional semantics corresponding to the configurable parameter. The visualization component is used to configure the parameter value of the configurable parameter in the workflow file.

[0152] As one possible implementation, the parsing unit 702 can also be configured to parse the workflow file and obtain the functional semantics of multiple logical nodes.

[0153] Furthermore, the identifier determination unit 705 can be specifically configured to: determine the semantic identifier of the configurable parameter based on the functional semantics of multiple logical nodes. The semantic identifier is used to locate the position of the configurable parameter in the workflow file during the process of configuring the configurable parameter using the visualization component.

[0154] As one possible implementation method, the component determination unit 703, when determining the visualization component corresponding to the configurable parameter, can be specifically configured to: obtain and store the business alias of the configurable parameter set by the first user; determine the component type of the visualization component corresponding to the configurable parameter based on at least one of the parameter name, business alias, functional semantics, historically set business alias, and parameter type of the configurable parameter; and determine the visualization component corresponding to the configurable parameter based on the component type of the visualization component corresponding to the configurable parameter.

[0155] As one possible implementation method, when determining the component type of the visualization component corresponding to the configurable parameter based on at least one of the configurable parameter's parameter name, business alias, functional semantics, historically set business alias, and parameter type, the component determination unit 703 can be specifically configured as follows: generating multiple recommended component types for the configurable parameter based on at least one of the configurable parameter's parameter name, business alias, functional semantics, historically set business alias, and parameter type; and obtaining the first user's selection from the multiple recommended component types as the component type of the visualization component corresponding to the configurable parameter.

[0156] As one possible implementation, the parsing unit 702 can also be configured to parse the workflow file and obtain constraint rules with configurable parameters.

[0157] The UI data determination unit 704, when obtaining UI data based on the visualization component corresponding to the configurable parameters, can be specifically configured to: write constraint rules into the configuration properties of the visualization component corresponding to the configurable parameters and register them with the validation rules of the visualization component to obtain UI data. The constraint rules include at least one of the default values, value ranges, validation rules, and parameter linkage rules of the configurable parameters.

[0158] Furthermore, the simulation unit 706 can be specifically configured to: obtain simulated values ​​of configurable parameters configured by the first user through the visualization components included in the parameter configuration interface; load a copy of the workflow file in the isolated environment and replace the parameter values ​​of the configurable parameters in the copy of the workflow file with the simulated values ​​of the configurable parameters; execute test tasks using the copy of the workflow file based on the obtained test data to obtain test results; and verify whether the configuration is successful based on the test results.

[0159] Simulation unit 706 can also be configured to: determine the simulation report corresponding to the copy of the workflow file during the execution of the test task, the simulation report including the response time of each logical node and the cost of calling the large model; and determine a visual heatmap based on the response time and cost, the visual heatmap being used to prompt the first user to adjust the UI data based on the visual heatmap.

[0160] Furthermore, management unit 707 can be specifically configured to: encapsulate UI data into a configuration snapshot with a unique version identifier in response to a test result of "test passed"; and manage the configuration snapshot based on its release status, which includes draft, pending approval, pending release, live, and offline.

[0161] According to another embodiment, a visual parameter configuration device for workflow is provided. Figure 8 A schematic block diagram of a workflow visualization parameter configuration apparatus according to one embodiment is shown, the apparatus being disposed in... Figure 1 The client-side interface of the first user in the illustrated architecture. For example... Figure 8 As shown, the workflow visualization parameter configuration device 800 includes a second acquisition unit 801, a display unit 802, and a sending unit 803, and further includes a selection unit 804. The main functions of each component are as follows: The second acquisition unit 801 is configured to respond to a triggering operation on a workflow file. The workflow file is used to characterize the logical flow executed by the target agent when responding to a request. The logical flow includes multiple logical nodes. The unit acquires the configurable parameters of the multiple logical nodes parsed from the workflow file and the functional semantics corresponding to the configurable parameters, as well as the visualization components corresponding to the configurable parameters.

[0162] Display unit 802 is configured to display configurable parameters of multiple logical nodes, functional semantics corresponding to the configurable parameters, and visual components in the component configuration interface.

[0163] Sending unit 803 is configured to respond to a confirmation operation triggered in the component configuration interface by sending the confirmed configurable parameters, the corresponding functional semantics of the configurable parameters, and the visual components for the server to generate UI data. The UI data is used to render the parameter configuration interface, which includes the visual components and functional semantics corresponding to the configurable parameters. The visual components are used to configure the parameter values ​​of the configurable parameters in the workflow file.

[0164] As one possible implementation, the selection unit 804, before responding to triggering the parsing operation for the workflow file, can be specifically configured to: obtain the workflow file in response to the operation of the first user uploading the workflow file; or, in response to the operation of selecting one of multiple candidate agents, use the selected candidate agent as the target agent and obtain the workflow file of the target agent.

[0165] As one possible implementation method, when displaying the visual component corresponding to the configurable parameter in the component configuration interface, the display unit 802 can be specifically configured as follows: the component type of the visual component corresponding to the configurable parameter is displayed in the component configuration interface. The component type of the visual component is determined based on at least one of the business alias, functional semantics, historically set business alias and parameter type of the configurable parameter. The business alias of the configurable parameter is set by the user based on the functional semantics of the configurable parameter.

[0166] As one possible implementation, the display unit 802 can also be configured to: display the rule configuration component corresponding to the configurable parameter on the component configuration interface; in response to the operation of triggering the rule configuration component, obtain the constraint rules of the configurable parameter parsed from the workflow file; based on the constraint rules of the configurable parameter, display the rule configuration interface, wherein the constraint rules include at least one of the default value, value range, validation rules, and parameter linkage rules of the configurable parameter; in response to the configuration operation of the constraint rules on the rule configuration interface, send the configured constraint rules so that the server can write the configured constraint rules into the configuration attributes of the corresponding visualization component and register the validation rules of the visualization component when generating UI data.

[0167] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0168] Figure 9A schematic block diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0169] like Figure 9 As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded from storage unit 908 into random access memory (RAM) 903. RAM 903 may also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0170] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0171] The computing unit 901 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as a workflow visualization parameter configuration method. For example, in some embodiments, the workflow visualization parameter configuration method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by the computing unit 901, one or more steps of the workflow visualization parameter configuration method described above can be performed. Alternatively, in other embodiments, the computing unit 901 may be configured by any other suitable means (e.g., by means of firmware) to perform a visual parameter configuration method for the workflow.

[0172] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0173] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0174] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0175] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0176] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0177] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0178] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0179] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for configuring visual parameters for workflows, comprising: Obtain a workflow file, which is used to characterize the logical flow executed by the target intelligent agent when responding to a request, and the logical flow includes multiple logical nodes; The workflow file is parsed to obtain the configurable parameters of the multiple logical nodes and the functional semantics corresponding to the configurable parameters; Determine the visualization component corresponding to the configurable parameters; User interface (UI) data is obtained based on the visualization component corresponding to the configurable parameter. The UI data is used to render the parameter configuration interface. The parameter configuration interface includes the visualization component and functional semantics corresponding to the configurable parameter. The visualization component is used to configure the parameter value of the configurable parameter in the workflow file.

2. The method according to claim 1, wherein, The process of parsing the workflow file also yields the functional semantics of the multiple logical nodes; The method further includes: Based on the functional semantics of the multiple logical nodes, a semantic identifier for the configurable parameter is determined. The semantic identifier is used to locate the position of the configurable parameter in the workflow file during the process of configuring the configurable parameter using the visualization component.

3. The method according to claim 1, wherein, The step of determining the visualization component corresponding to the configurable parameter includes: Obtain and store the service alias of the configurable parameters set by the first user; Based on at least one of the parameter name, business alias, functional semantics, historically set business alias, and parameter type of the configurable parameter, determine the component type of the visualization component corresponding to the configurable parameter; Based on the component type of the visualization component corresponding to the configurable parameter, determine the visualization component corresponding to the configurable parameter.

4. The method according to claim 3, wherein, The step of determining the component type of the visualization component corresponding to the configurable parameter based on at least one of the parameter name, business alias, functional semantics, historically set business alias, and parameter type of the configurable parameter includes: Based on at least one of the parameter name, business alias, functional semantics, historically set business alias, and parameter type of the configurable parameters, multiple recommended component types are generated for the configurable parameters. Obtain the component type from which the first user selects one of the multiple recommended component types as the visualization component corresponding to the configurable parameter.

5. The method according to claim 1, wherein, The process of parsing the workflow file also yields the constraint rules for the configurable parameters; The process of obtaining UI data based on the visualization component corresponding to the configurable parameters includes: The constraint rules are written into the configuration attributes of the visualization component corresponding to the configurable parameter and registered to the validation rules of the visualization component to obtain UI data. The constraint rules include at least one of the default value, value range, validation rules, and parameter linkage rules of the configurable parameter.

6. The method according to claim 1, further comprising: Obtain simulated values ​​of the configurable parameters configured by the first user through the visualization components included in the parameter configuration interface; Load a copy of the workflow file in an isolated environment and replace the parameter values ​​of the configurable parameters in the copy of the workflow file with simulated values ​​of the configurable parameters; Based on the acquired test data, a copy of the workflow file is used to execute test tasks to obtain test results; Based on the test results, verify whether the configuration is successful.

7. The method according to claim 6, further comprising: Determine the simulation report corresponding to the copy of the workflow file during the execution of the test task, the simulation report including the response time of each logical node and the cost of calling the large model; Based on the response time and the cost, a visual heatmap is determined, which is used to prompt the first user to adjust the UI data based on the visual heatmap.

8. The method according to claim 6, further comprising: In response to the test result being a pass, the UI data is encapsulated into a configuration snapshot with a unique version identifier; The configuration snapshot is managed based on its release status, which includes draft, pending approval, pending release, online, and offline.

9. A method for configuring visual parameters for workflows, comprising: In response to triggering a parsing operation on a workflow file, the workflow file being used to characterize the logical flow executed by the target agent when responding to a request, the logical flow including multiple logical nodes, the configurable parameters of the multiple logical nodes parsed from the workflow file and the functional semantics corresponding to the configurable parameters are obtained, and the visualization components corresponding to the configurable parameters are obtained. The component configuration interface displays the configurable parameters of the multiple logical nodes, the functional semantics corresponding to the configurable parameters, and the visual components. In response to a confirmation operation triggered in the component configuration interface, a confirmed configurable parameter, the corresponding functional semantics of the configurable parameter, and a visualization component are sent to the server to generate UI data. The UI data is used to render the parameter configuration interface, which includes the visualization component and functional semantics corresponding to the configurable parameter. The visualization component is used to configure the parameter value of the configurable parameter in the workflow file.

10. The method of claim 9, further comprising, before triggering the parsing operation for the workflow file: In response to the first user's operation of uploading a workflow file, the workflow file is obtained; or, In response to the operation of selecting one of multiple candidate agents, the selected candidate agent is taken as the target agent, and the workflow file of the target agent is obtained.

11. The method according to claim 9, wherein, The component configuration interface displays the visual components corresponding to the configurable parameters, including: The component configuration interface displays the component type of the visual component corresponding to the configurable parameter. The component type of the visual component is determined based on at least one of the business alias, functional semantics, historically set business alias, and parameter type of the configurable parameter. The business alias of the configurable parameter is set by the user based on the functional semantics of the configurable parameter.

12. The method according to claim 9, wherein, The method further includes: The rule configuration component corresponding to the configurable parameter is displayed on the component configuration interface; In response to the operation of triggering the rule configuration component, the constraint rules of the configurable parameters parsed from the workflow file are obtained; Based on the constraint rules of the configurable parameters, a rule configuration interface is displayed. The constraint rules include at least one of the default values, value ranges, verification rules, and inter-parameter linkage rules of the configurable parameters. In response to the configuration operation of the constraint rules in the rule configuration interface, the configured constraint rules are sent so that the server can write the configured constraint rules into the configuration properties of the corresponding visualization component and register the validation rules of the visualization component when generating UI data.

13. A device for configuring visual parameters for workflows, comprising: The first acquisition unit is configured to acquire a workflow file, which is used to characterize the logical flow executed by the target intelligent agent when responding to a request, and the logical flow includes multiple logical nodes. The parsing unit is configured to parse the workflow file to obtain the configurable parameters of the plurality of logical nodes and the functional semantics corresponding to the configurable parameters; The determining unit is configured to determine the visualization component corresponding to the configurable parameter; The obtaining unit is configured to obtain user interface (UI) data based on the visualization component corresponding to the configurable parameter. The UI data is used to render the parameter configuration interface, which includes the visualization component and functional semantics corresponding to the configurable parameter. The visualization component is used to configure the parameter value of the configurable parameter in the workflow file.

14. A device for configuring visual parameters for workflows, comprising: The second acquisition unit is configured to respond to triggering a parsing operation on a workflow file, the workflow file being used to characterize the logical flow executed by the target agent when responding to a request, the logical flow including multiple logical nodes, acquiring configurable parameters of the multiple logical nodes parsed from the workflow file and the functional semantics corresponding to the configurable parameters, and acquiring the visualization components corresponding to the configurable parameters. The display unit is configured to display the configurable parameters of the multiple logical nodes, the functional semantics corresponding to the configurable parameters, and the visual components in the component configuration interface; The sending unit is configured to send configurable parameters, functional semantics corresponding to the configurable parameters, and a visualization component in response to a confirmation operation triggered on the component configuration interface, so that the server can generate UI data. The UI data is used to render the parameter configuration interface, which includes the visualization component and functional semantics corresponding to the configurable parameters. The visualization component is used to configure the parameter value of the configurable parameters in the workflow file.

15. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-12.

16. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-12.

17. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-12.

Images