Cooperative calling system and method for coal mine production informatization system

Abstract

The invention relates to a cooperative calling system and method for a coal mine production informatization system. The system comprises a plurality of external systems which are all coal mine production informatization systems with visual front ends and callable operation interfaces; the front-end plug-ins are deployed at the front ends of the external systems, upload callable operation interface information to the service registration and dispatching center, and receive execution instructions from the service registration and dispatching center; the service registration and dispatching center is used for receiving callable operation interface information, forming an operation interface list and distributing execution instructions to the corresponding front-end plug-ins; the large language model processing module is used for selecting a target operation interface from the operation interface list by utilizing a large language model based on the user instruction and generating an execution instruction; and the user interaction terminal receives an instruction input by a user. According to the method, the multiple different systems are controlled to execute the existing functions through the large language model, and the complexity and the personnel cost of multi-system operation are reduced.

Description

Technical Field

[0001] This invention relates to the field of intelligent coal mining technology; specifically, it relates to a collaborative invocation system and method for coal mine production information systems. Background Technology

[0002] Coal mining companies typically purchase and operate multiple systems from various vendors, requiring multiple personnel to learn and master the functions and operations of each system. This results in high labor costs, and the complex operations require specialized training to complete the tasks.

[0003] Existing technologies connect systems from multiple vendors to a unified platform, relying on keywords and hard-programmed bindings, which cannot effectively identify user intent and bind system commands through natural language. Summary of the Invention

[0004] In view of this, the present invention provides a collaborative invocation system and method for coal mine production information systems, thereby solving or at least alleviating one or more of the above-mentioned problems and other problems existing in the prior art.

[0005] To achieve the aforementioned objectives, a first aspect of the present invention provides a collaborative invocation system for a coal mine production information system, wherein the system comprises: Multiple external systems, which are coal mine production information systems, and have a visual front-end and callable operation interfaces; Multiple front-end plugins are deployed on the visual front-end of each of the external systems and establish long connections with the service registration and scheduling center. These plugins are used to obtain and send the callable operation interface information of the external system to the service registration and scheduling center, and to receive execution instructions containing the target operation interface identifier and input parameters from the service registration and scheduling center, and to call the corresponding operation interface according to the execution instructions. The service registration and scheduling center is used to receive and store the callable operation interface information sent by each of the front-end plugins, form an operation interface list, and communicate with the large language model processing module to receive and parse the execution instructions from the large language model processing module, and send the execution instructions to the corresponding front-end plugin through the long connection. The large language model processing module is used to obtain the list of operation interfaces, and use the large language model to select the target operation interface from the list of operation interfaces based on the user instruction, generate the execution instruction, and send the execution instruction to the service registration and scheduling center. The user interaction terminal is used to receive natural language information input by the user and send the natural language information as the user instruction to the large language model processing module.

[0006] In the system described above, optionally, the callable operation interface information includes interface identifier, interface name, interface function description, and interface input parameter information, wherein the interface input parameter information includes parameter name, parameter requirement, parameter description, and parameter data type. Both the callable operation interface information and the execution instructions are encapsulated in JSON format.

[0007] In the system described above, optionally, the visual front-end of the external system is a Web front-end, the front-end plugin is a JavaScript plugin, and the front-end plugin establishes a long connection with the service registration and scheduling center through the WebSocket communication protocol.

[0008] In the system described above, optionally, the service registration and scheduling center periodically performs heartbeat checks on the front-end plugin.

[0009] In the system described above, optionally, when the front-end plugin sends the callable operation interface information to the service registration and scheduling center, it also sends the plugin identifier of the front-end plugin to the service registration and scheduling center. The service registration and scheduling center is also used to establish a mapping relationship between each operation interface and the plugin identifier. When the execution instruction is received, the execution instruction is sent to the corresponding front-end plugin according to the mapping relationship.

[0010] Optionally, in the system described above, the user interaction terminal supports users to input the natural language information via text or voice.

[0011] Optionally, in the system described above, the external system includes a mining equipment operation and maintenance system and a coal mine personnel positioning system.

[0012] In the system described above, optionally, the large language model processing module can match one or more target operation interfaces for a single user instruction. When the number of target operation interfaces is greater than one, the calling order of the multiple target operation interfaces is set, and the execution instructions corresponding to each target operation interface are generated sequentially according to the calling order and sent to the service registration and scheduling center. The service registration and scheduling center sends each execution instruction to the corresponding front-end plugin in the order of invocation, thereby sequentially invoking each target operation interface and coordinating the execution of the user instructions.

[0013] To achieve the aforementioned objective, a second aspect of the present invention provides a collaborative invocation method for a coal mine production information system, wherein the method is based on a system as described in any of the preceding first aspects, and the method includes the following steps: Collect callable operation interface information from multiple external systems to form an operation interface list; Obtain user commands expressed in natural language; The target operation interface corresponding to the user instruction is determined from the operation interface list using a large language model; Based on the interface information of the target operation interface, an execution instruction for calling the target operation interface is generated using a large language model. The execution instruction is sent to the external system where the target operation interface is located, and the target operation interface is invoked.

[0014] In the method described above, optionally, the step of determining the target operation interface and generating the execution instructions includes: Fill the operation interface list and the user instructions into the first prompt word template to generate the first prompt word; The first prompt word is input into the large language model to parse the user's operation intention from the user command and identify the target operation interface that satisfies the user's intention from the operation interface list. The interface information of the target operation interface is confirmed by the operation interface list; Fill the interface information and the user command into the second prompt word template to generate the second prompt word; The second prompt word is input into the large language model, the interface information is parsed, and the interface call specification of the target operation interface is obtained. When the interface call specification contains input parameters, the values ​​of each input parameter are extracted from the user instruction. Generate execution instructions that conform to the interface call specification.

[0015] The coal mine production information system collaborative calling system of the present invention solves the problem of non-hard-coded intelligent conversion of user intent and system capabilities through natural language understanding and programming capabilities of large language models, effectively controls multiple different systems to execute existing functions, and reduces the complexity of multi-system operation and personnel costs.

[0016] The present invention further provides a collaborative invocation method for coal mine production information systems based on the above-mentioned system, and therefore this method also has the above-mentioned advantages. Attached Figure Description

[0017] The disclosure of this invention will become more apparent from the accompanying drawings. It should be understood that these drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings: Figure 1 This is a schematic block diagram of an embodiment of the collaborative calling system of the coal mine production information system of the present invention; Figure 2 for Figure 1 The module interaction diagram of the embodiment shown is as follows; Figure 3 This is a flowchart illustrating an embodiment of the collaborative invocation method for a coal mine production information system according to the present invention. Detailed Implementation

[0018] Referring to the accompanying drawings and specific embodiments, the structure, composition, features, and advantages of the coal mine production information system collaborative invocation system and method of the present invention will be described below by way of example. However, all descriptions should not be construed as limiting the present invention in any way.

[0019] Furthermore, for any single technical feature described or implied in the embodiments mentioned herein, or any single technical feature shown or implied in the various figures, the present invention still allows for any combination or deletion of these technical features (or their equivalents) without any technical obstacle, and thus these further embodiments according to the present invention should also be considered within the scope of this description.

[0020] It should also be noted that the terms "below", "left", "right", etc., indicate the orientation or positional relationship based on the orientation or positional relationship of the coal mine production information system collaborative calling system modules shown in the attached drawings. They are only for the convenience of describing this disclosure and simplifying the description, and do not indicate or imply that the modules referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this disclosure.

[0021] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0022] Figure 1 This is a schematic block diagram of an embodiment of the collaborative calling system of the coal mine production information system of the present invention.

[0023] The system includes multiple external systems, multiple front-end plugins, a service registration and scheduling center, a large language model processing module, and a user interaction terminal.

[0024] Figure 1The right side illustrates multiple external systems, schematically represented by systems A, B, and N. This does not represent the actual number of external systems; in practice, the number of external systems is not limited. In this embodiment, all the external systems shown have a web front-end and callable operation interfaces. In optional embodiments, the external system front-end can also be a non-web front-end (such as a desktop client, embedded human-machine interface, etc.), and front-ends from different platforms can collaboratively call each other through a unified service registration and scheduling center.

[0025] Figure 1 Below each external system web frontend, a JavaScript plugin adapted to that web frontend is shown; hereinafter referred to as the frontend plugin. Each frontend plugin is deployed on its respective external system web frontend and can call at least one function of its respective external system web frontend through at least one interface, thereby triggering predefined operations within that external system web frontend. Using frontend plugins enables non-intrusive integration without modifying the original web system code; interface registration and command reception can be completed simply by importing the script, resulting in low deployment costs.

[0026] Figure 1 The diagram also illustrates the Service Registration and Scheduling Center (represented in the diagram as "System Command Registration Center and Command Distribution Routing Service"), the Large Language Model Processing Module (represented in the diagram as "Large Model Intent Analysis and Command Generation Service"), the User Interaction Terminal (represented in the diagram as "Application Terminal"), and the user (illustratively represented by a figure on the left). The user can interact with the User Interaction Terminal, which communicates with the Large Language Model Processing Module. The Large Language Model Processing Module communicates with the Service Registration and Scheduling Center, which in turn communicates with each front-end plugin.

[0027] Figure 2 for Figure 1 The illustrated embodiment presents a schematic diagram of module interactions. Of the multiple external systems shown, only system N is illustrated as an example to demonstrate the interaction logic between the external system's web front-end and its front-end plugins, and between the front-end plugins and other modules.

[0028] like Figure 2 As shown, the interaction between the external system web front-end and the front-end plugin includes: the external system web front-end initializes and loads the front-end plugin, registers a list of available functional interfaces (i.e., callable operation interfaces) with the front-end plugin, and the front-end plugin returns the registration result.

[0029] For example, the information for each callable operation interface is encapsulated in JSON format. This JSON object can contain two parts: system information (info) and a list of available operation interfaces (functions). `info` describes the basic information of the current external system's web frontend and may include a `title` field describing the name of the external system. `functions` contains the definitions of multiple callable operation interfaces, each with a unique operation identifier as the key and an interface description object as its value. As an example, the interface description object may include fields such as `summary`, `description`, `operationId`, and `parameters`. `summary` provides a brief description of the interface's operation function, `description` provides a detailed description of the operation function, `operationId` is the unique identifier for the interface used for subsequent instruction matching, and `parameters` is an array listing the input parameters required to execute the interface. Each input parameter definition includes its name, parameter type (e.g., query parameter, path parameter, request header parameter, request body parameter, etc.), meaning description, whether it is required, data type, format, and default value. Using a unified JSON format enhances system interoperability; the structured description of the interfaces provides a standard for generating subsequent interface requests, supports parameter validation based on parameter definitions, and improves the success rate of interface calls.

[0030] For example, the external system (System N) is a coal mining equipment system, and its available functional interfaces include an interface for displaying equipment details. The JSON object encapsulating the available interface information of this coal mining equipment system includes "Coal Mining Equipment System Interfaces" in the system information, introducing the name of the external system. The interface information list includes interfaces for obtaining real-time data from coal mining machines and interfaces for displaying a list of coal mining machines, identified by "get_caimeiji_realtime_data" and "get_caimeiji_list" respectively. The description object for the `get_caimeiji_realtime_data` interface briefly describes it as "displaying real-time data of coal mining machines based on the equipment ID." Correspondingly, the field for required input parameters for this interface defines the parameter `equipId`, which stores the equipment ID and is a required string query parameter. The description object for the `get_caimeiji_list` interface describes it as "retrieving a list of coal mining machine data based on a keyword in the equipment name." Its input parameters include the optional query keyword parameter `equipName` and the number of records to read parameter `size`, where `size` defaults to 5 and ranges from 1 to 10.

[0031] like Figure 2As shown, the interaction between the front-end plugin and the service registration and scheduling center includes: after the front-end plugin is initialized, it registers itself with the service registration and scheduling center, establishes a WebSocket long connection, packages the aforementioned callable operation interface information and sends it to the service registration and scheduling center, and the service registration and scheduling center returns the registration result to the front-end plugin. The WebSocket long connection ensures bidirectional real-time communication, meeting the high real-time requirements of scenarios such as coal mine monitoring and alarm response. Through the front-end plugin and communication protocol, systems requiring expansion can be integrated with low coupling and minimal modification, effectively reducing implementation difficulty and cost.

[0032] For example, when packaging callable interface information, a unique identifier, socketId, for the communication channel between the front-end plugin and the service registration and scheduling center is added to the system information. Therefore, this identifier can also be the identifier of the front-end plugin. The service registration and scheduling center binds the communication channel identifier and the interface identifier through a mapping relationship, which is used to subsequently issue instructions to the corresponding front-end plugin based on the interface identifier. The service registration and scheduling center summarizes the available operation interface information submitted by multiple front-end plugins to generate an operation interface list.

[0033] like Figure 2 As shown, the front-end plugins and the service registration and scheduling center monitor the status of long-lived connections via heartbeat detection. The service registration and scheduling center periodically sends heartbeat check messages to each front-end plugin to confirm whether the connection with the front-end plugin is interrupted or abnormal. Upon receiving the heartbeat check message, the front-end plugin returns a health response message within a specified time. If the front-end plugin responds to the heartbeat promptly, it indicates that the communication channel is normal; otherwise, the heartbeat check is resent to the front-end plugin. If there are multiple unresponsive heartbeats, it is determined that the connection is lost, and the interface information bound to that communication channel can be removed from the operation interface list. Optionally, an alert is sent to operations and maintenance personnel after determining that a front-end plugin is lost. This heartbeat detection mechanism ensures the reliability of the system.

[0034] like Figure 2 As shown, the interaction between the user interaction terminal and the large language model processing module includes: after being activated, the user interaction terminal can receive user input; after the user inputs operation requirements in natural language form via text or voice, the terminal submits the user's instructions to the large language model processing module. For example, the user instruction might be "Show details of device number 1".

[0035] like Figure 2As shown, the interaction between the service registration and scheduling center and the large language model processing module includes: the large language model processing module requests a list of available interfaces from the service registration and scheduling center; the service registration and scheduling center sends the registered list of available interfaces to the large language model processing module; the large language model processing module, based on the user requirements and the list of available interfaces, determines the target interface, generates instructions, and then issues instructions to the service registration and scheduling center. For example, the large language model processing module constructs a prompt word template. Whenever it receives a user instruction, it fills the prompt word template with the user instruction and the list of available interface information. This instructs the large language model to understand the user requirements, select a suitable interface from the list of operation interfaces, and generate an execution instruction based on the interface's request specifications (such as parameter requirements). The execution instruction can be in JSON format, containing information about the target interface to be called and the values ​​of the input parameters. Optionally, after the large language model understands the user's instructions, if the user's request requires the collaborative completion of multiple interfaces, the large language model selects the multiple target interfaces for the user, determines whether there are any associations or constraints between the multiple interfaces, determines the interface call order, and sends execution instructions to the service registration and scheduling center in sequence, so that the service registration and scheduling center issues instructions to the corresponding front-end plugins in sequence, thereby calling each target interface in sequence.

[0036] like Figure 2 As shown, the interaction between the service registration and scheduling center and the front-end plugin also includes: after receiving an execution instruction and determining the instruction delivery channel, the service registration and scheduling center sends the instruction to the corresponding front-end plugin. For example, based on the unique information such as the identifier or name of the target interface in the execution instruction, the service registration and scheduling center retrieves the information of the interface from the operation interface list, determines the identifier of the front-end plugin corresponding to the interface, that is, determines the identifier of the communication channel with the front-end plugin, and then sends the execution instruction to the front-end plugin through this communication channel. This fine-grained instruction addressing mechanism ensures the accuracy of interface calls.

[0037] like Figure 2 As shown, the interaction between the front-end plugin and the external system's web front-end also includes: the front-end plugin calling the registered function according to the instruction. Specifically, after receiving the execution instruction issued by the service registration and scheduling center, the front-end plugin calls the registered interface in the execution instruction, thereby calling the function of the external system's web front-end corresponding to the registered interface to complete the user's instruction.

[0038] In an optional embodiment, the user interaction terminal, the large language model processing module, and the service registration and scheduling center can be integrated into an electronic device, which communicates with the plugins of the web front-end of each external system according to the above-described interaction logic.

[0039] Figure 3This is a flowchart illustrating an embodiment of the collaborative invocation method for a coal mine production information system according to the present invention.

[0040] like Figure 3 As shown, the method includes the following steps: S1, multiple system front-ends send available operation interface information and interface operation names; S2, obtain the user-input interaction command; S3, determine the target interface operation name corresponding to the interaction command information from the interface operation names using a large language model; S4, determine the interface request content corresponding to the interaction command information based on the target interface information corresponding to the target interface operation name using a large language model; S5, send the operation command to the system front-end based on the interface request content.

[0041] For example, multiple systems are included, such as a personnel precise positioning system and a coal mining machine remote operation and maintenance system. The personnel precise positioning system's visual front-end provides operation interfaces such as "displaying underground personnel statistics," "calling underground personnel," and "displaying personnel location details." The coal mining machine remote operation and maintenance system's visual front-end provides operation interfaces such as "displaying real-time data of the coal mining machine" and "displaying a list of coal mining machines."

[0042] S1. Multiple system front-ends submit available operation interface information and interface operation names. For example, these external system visualization front-ends first employ TCP-based Socket programming technology to establish a cross-domain, cross-language, bidirectional, and reliable data transmission channel with the service registration and scheduling center. These steps can be implemented through front-end plugins. Then, the available operation interface information is packaged in JSON format and sent to the service registration and scheduling center. For instance, a web front-end can use JavaScript programming and the WebSocket communication protocol to establish a long connection, while front-ends on other platforms can use corresponding network programming methods. Through front-end plugins and communication protocols, systems requiring expansion can be integrated with low coupling and minimal modification, effectively reducing implementation difficulty and cost.

[0043] As an example, the JSON object sent to the service registration and scheduling center can specifically include two parts: system information and available operation interface information. The system information section describes the basic information of the current external system frontend, which may include the name of the external system and the identifier of the communication channel between the service registration and scheduling center and the external system. The available operation interface information contains the definitions of multiple callable operation interfaces. Each operation interface has a unique operation identifier, and its interface definition also includes a functional description (natural language) and input parameter definitions. The definition of each input parameter may include its name, parameter type (e.g., query parameter, path parameter, request header parameter, request body parameter, etc.), meaning description, whether it is required, data type, format, default value, etc.

[0044] The service registration and scheduling center compiles the available operation interface information submitted by multiple system front-ends into an operation interface list.

[0045] S2, Obtain the user-inputted interaction instructions. For example, the user-inputted interaction instructions, i.e., user commands, can be text information in natural language. The user inputs natural language interaction instructions as commands to interact with software programs such as applications. Natural language typically refers to a language that naturally evolves with culture. For example, the interaction instructions could be "View the list of currently available coal mining machines" or "View the current operating status of coal mining machine ID 002," etc.

[0046] In optional embodiments, users can input interactive commands in various ways, such as through text input or voice input. Multiple input methods enhance the convenience and applicability of human-computer interaction. Inspectors can quickly issue commands via voice without manual input; furthermore, voice input is more practical in scenarios where typing is inconvenient, such as dust, noise, or when wearing gloves, adapting to the complex working environment of coal mines.

[0047] S3 uses a large language model to determine the target interface operation name corresponding to the interaction instruction information from the interface operation name. The large language model can be an artificial intelligence model used to understand and process natural language text. A prompt word generation template (hereinafter referred to as the first prompt word template) can be used, leveraging a structured and reusable prompt word framework, and through preset formats and placeholders, to quickly generate high-quality, target-specific prompt words.

[0048] For example, the first prompt word template may include a list of interfaces and placeholders for user interaction instructions. After filling the specific list of available interfaces and interaction instructions into the first prompt word template, a prompt word (hereinafter referred to as the first prompt word) is generated. After inputting the first prompt word into a large language model, the large language model will analyze and deduce the user's operational intent and output the target operation interface. For example, the first prompt word template can be: "You are an assistant for an internal tool, and you need to determine the user's target operation based on user input and context information;" You will extract the full names of all operations from the user's input in precise order and output them; Multiple extracted operation names are separated by commas; You must return only the full operation name, without any description; The possible operation list is as follows: enter: operate:" in," "This is a placeholder for the interface list, used to populate information about currently available operation interfaces, such as interface names;" "" serves as a placeholder for user commands, used to fill in interactive command information input by the user. In an optional embodiment, the interface list placeholder can also be filled with other interface information obtained from the operation interface list in S1. This interface information needs to include information that can locate a unique corresponding interface, as well as natural language information describing the interface operation function.

[0049] For example, if the user inputs the command "Check the current operating status of the coal mining machine with ID 002", then this command information will be filled into... If the available operation interfaces are "Display underground personnel statistics", "Call underground personnel", "Display personnel location details", "Display real-time data of coal mining machines", and "Display coal mining machine list", then fill in all of the above interface names in the field. After filling in the above information, the first prompt word is generated. After the first prompt word is input, the large language model will analyze it to determine the user's operation intention and output the target operation interface as "display the real-time data of the coal mining machine".

[0050] In optional embodiments, multiple interfaces can be collaboratively invoked to satisfy user operational intentions. These interfaces can originate from different external systems. For example, if the user command is "to compile statistics on current coal mining machines and personnel," then the interfaces for "displaying underground personnel statistics" and "displaying the list of coal mining machines" need to be invoked. Furthermore, the large language model can determine whether there are sequential relationships or constraints between the multiple interfaces to be invoked, thereby setting a reasonable interface invocation order. For example, when a user requests to add / delete / modify a piece of information about a coal mining machine and display the updated result, the interface for adding / deleting / modifying coal mining machine information needs to be invoked first, followed by the interface for displaying coal mining machine information. By supporting the collaborative invocation of multiple operation interfaces, multiple external systems can be automatically triggered to execute in conjunction based on a single user natural language command. The logical consistency of the business process is ensured by pre-setting the invocation order. This significantly improves the integration and intelligence level of the coal mine production information system.

[0051] S4, using a large language model, determines the interface request content corresponding to the interaction instruction information based on the target interface operation name and the target interface information corresponding to the target interface operation name. There are various specific implementation methods. For example, the target interface information corresponding to the target interface operation name and the interaction instruction information can be input into the large language model. The large language model identifies at least one request parameter name in the target interface information and the corresponding parameter value in the interaction instruction information. The interface request content corresponding to the interaction instruction information is then constructed based on the parameter name and parameter value corresponding to the request parameter. The target interface information corresponding to the target interface operation name can be obtained by querying the operation interface list in S1. Furthermore, the aforementioned request parameters are specific instantiated values ​​of the input parameters in the interface definition.

[0052] For example, a prompt word generation template (hereinafter referred to as the second prompt word template) is used to quickly generate high-quality, targeted prompt words. The second prompt word template may include target interface information and placeholders for user interaction instructions. The target interface information and user-input interaction instructions are filled into the second prompt word template to generate the second prompt word. After extraction through a large language model, the interface request content is obtained. For example, the second prompt word template can be: "You are an assistant for an internal tool, and you need to extract the relevant parameters based on user input and context information;" You must output all parameters in JSON format, where the keys are the parameter names and the values ​​are the extracted results. The operation and parameter information are as follows: enter: parameter:" in," "This can be the location for filling in the target interface information," "This can be the location to fill in interactive command information."

[0053] For example, if the target interface is named "Display real-time data of the coal mining machine", its complete interface definition is filled in " "At this point, the interface definition includes the required input parameter equipId, which represents the device ID;" The text inside the blank is "Check the current operating status of the coal mining machine with ID 002". The generated prompt is input into the large language model. After processing and extraction by the large language model, the interface request content is obtained as: {"equipId":"002"}, indicating that the request parameters include the device ID and the parameter value is "002".

[0054] S5: Based on the interface request content, an operation command is sent to the system front-end. According to the interface request content obtained in S4 and the interface request specification determined by the target interface information, a complete request message is formed as the operation command (execution command). Based on the communication channel identifier in the command, the command is sent to the corresponding external system front-end through the communication channel corresponding to that identifier. After receiving the message, the external system front-end checks whether it matches the communication channel identifier it sent, and then completes the execution of the command based on the operation name and parameters.

[0055] This embodiment leverages the natural language understanding and programming capabilities of a large model to achieve intelligent, non-hard-coded conversion between the operator's intent and the device's capabilities. This effectively controls multiple different systems to execute existing functions, reducing the complexity of multi-system operation and personnel costs.

[0056] Some embodiments of the present invention, through a large language model, accurately identify the corresponding interface requests based on the natural language input by the user and the operation interfaces of multiple visual interfaces of different systems that are dynamically recognized, and send them to the visual front end of the corresponding system for execution. This can effectively reduce the complexity of information interaction, effectively reuse the original visual operation interface, and improve the efficiency of information interaction.

[0057] The technical scope of this invention is not limited to the contents of the above specification. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the scope of this invention.

Claims

1. A coal mine production information system cooperative calling system, characterized in that, The system comprises: A plurality of external systems, which are coal mine production information systems and have visual front ends and callable operation interfaces; A plurality of front end plugins, each of which is deployed in the visual front end of each external system and establishes a long connection with a service registration and scheduling center, is used to obtain and send the callable operation interface information of the external system to the service registration and scheduling center, and is used to receive an execution instruction containing a target operation interface identifier and input parameters from the service registration and scheduling center and call the corresponding operation interface according to the execution instruction; The service registration and scheduling center is used to receive and store the callable operation interface information sent by each front end plugin, form an operation interface list, and communicate with a large language model processing module, receive and parse the execution instruction from the large language model processing module, and send the execution instruction to the corresponding front end plugin through the long connection; The large language model processing module is used to obtain the operation interface list, select a target operation interface from the operation interface list based on a user instruction using a large language model, generate the execution instruction, and send the execution instruction to the service registration and scheduling center; A user interaction terminal is used to receive natural language information input by a user and send the natural language information as the user instruction to the large language model processing module.

2. The system of claim 1, wherein, The callable operation interface information includes an interface identifier, an interface name, an interface function description, and interface input parameter information, and the interface input parameter information includes a parameter name, whether the parameter is necessary, a parameter description, and a parameter data type. The callable operation interface information and the execution instruction are both encapsulated in JSON format.

3. The system of claim 1, wherein, The visual front end of the external system is a Web front end, the front end plugin is a JavaScript plugin, and the front end plugin establishes a long connection with the service registration and scheduling center through the WebSocket communication protocol.

4. The system of claim 1, wherein, The service registration and scheduling center regularly performs heartbeat detection on the front end plugin.

5. The system of claim 1, wherein, When the front end plugin sends the callable operation interface information to the service registration and scheduling center, it also sends the plugin identifier of the front end plugin to the service registration and scheduling center, The service registration and scheduling center is also used to establish a mapping relationship between each operation interface and the plugin identifier, and when the execution instruction is received, the execution instruction is distributed to the corresponding front end plugin according to the mapping relationship.

6. The system of claim 1, wherein, The user interaction terminal supports the user to input the natural language information in the form of text or voice.

7. The system of claim 1, wherein, The external system includes a mine equipment operation and maintenance system and a coal mine personnel positioning system.

8. The system of claim 1, wherein, The large language model processing module can match one or more target operation interfaces for a single user instruction, when the number of target operation interfaces is greater than one, set the calling order of multiple target operation interfaces, and send the execution instruction corresponding to each target operation interface to the service registration and scheduling center in the calling order, The service registration and scheduling center sends each execution instruction to the corresponding front-end plug-in in turn according to the calling sequence, thereby calling each target operation interface in turn and cooperatively executing the user instruction.

9. A method for collaborative invocation of a coal mine production information system, characterized in that, The method is performed based on the system according to any one of claims 1-8, and the method comprises the following steps: collecting callable operation interface information of a plurality of external systems to form an operation interface list; obtaining a user instruction expressed in natural language; determining a target operation interface corresponding to the user instruction in the operation interface list through a large language model; generating an execution instruction for calling the target operation interface based on interface information of the target operation interface through a large language model; issuing the execution instruction to an external system where the target operation interface is located to call the target operation interface.

10. The method of claim 9, wherein, The steps of determining the target operation interface and generating the execution instruction comprise: filling the operation interface list and the user instruction into a first prompt word template to generate a first prompt word; inputting the first prompt word into a large language model to parse a user's operation intention from the user instruction and identify a target operation interface meeting the user's intention from the operation interface list; confirming interface information of the target operation interface through the operation interface list; filling the interface information and the user instruction into a second prompt word template to generate a second prompt word; inputting the second prompt word into a large language model to parse the interface information, obtain interface calling specifications of the target operation interface, and extract values of each input parameter from the user instruction when the interface calling specifications contain input parameters; generating an execution instruction conforming to the interface calling specifications.

Images