Query method, device and system, storage medium and computer program product

Abstract

The invention discloses a query method, device and system, a storage medium and a computer program product, and relates to the technical field of the Internet, and the method comprises the steps: obtaining the input content of a questioner in response to an input operation, and carrying out the intention recognition of the input content, and obtaining an intention recognition result; if the intention recognition result is to query the enterprise document of the target enterprise, matching the input content with the existing enterprise document to obtain a first query result; and if the intention identification result is to query the Internet of Things equipment information of the target enterprise, obtaining the Internet of Things equipment information in real time through the digital twin platform, and matching the input content with the Internet of Things equipment information to obtain a second query result. The information query efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to a query method, a query device, a query system, a storage medium, and a computer program product. Background Art

[0002] With the rapid development of information technology, enterprises are facing the challenge of massive data and information. Although the existing digital twin platform can monitor the operation of IoT devices, it requires users to manually query IoT device information on the digital twin platform, which makes information query inefficient. Summary of the invention

[0003] The main purpose of the present application is to provide a query method, a query device, a query system, a storage medium and a computer program product, aiming to solve the technical problem of low efficiency of information query.

[0004] To achieve the above purpose, the present application proposes a query method, which includes: In response to the input operation, the input content of the questioner is obtained and the intention recognition is performed on the input content to obtain an intention recognition result; If the intent recognition result is to query the enterprise documents of the target enterprise, the input content is matched with the existing enterprise documents to obtain the first query result; If the intention recognition result is to query the IoT device information of the target enterprise, the IoT device information is obtained in real time through the digital twin platform and the input content is matched with the IoT device information to obtain a second query result.

[0005] In one embodiment, the step of matching the input content with the existing enterprise documents to obtain the first query result includes: Perform multimodal analysis on existing enterprise documents to extract the text content of enterprise documents; Vectorizing the text content to obtain vectorized text content and storing it in a vector database; The input content is vectorized and matched with the vectorized text content stored in the vector database to obtain a first query result.

[0006] In one embodiment, the step of acquiring IoT device information in real time through the digital twin platform includes: By establishing an IoT device model and connecting the IoT device model to the digital twin platform, IoT device information can be monitored.

[0007] In one embodiment, the step of matching the input content with the IoT device information to obtain a second query result includes: Extract information from the questioner’s input content, and generate an input request that complies with the digital twin platform interface specification based on the extracted input content; Responding to the input request through the digital twin platform to obtain first IoT device information; Preprocessing the first IoT device information to obtain second IoT device information; With reference to the preset answer template, the second Internet of Things device information matching the input content is output as the second query result.

[0008] In one embodiment, the step of performing intent recognition on the input content to obtain an intent recognition result includes: The input content is classified into two categories to determine whether the questioner is querying the corporate documents of the target enterprise or the IoT device information of the target enterprise.

[0009] In one embodiment, after obtaining the first query result or obtaining the second query result, the method further comprises: Parsing the first query result or the second query result according to a large model of a preset prompt word template to generate initial answer content; The format of the initial answer content is converted to obtain a natural language text that complies with grammatical and semantic rules.

[0010] In addition, to achieve the above-mentioned purpose, the present application also proposes a query device, the query device comprising: an intention recognition module, for obtaining the input content of the questioner in response to the input operation and performing intention recognition on the input content to obtain an intention recognition result; A first matching module, for matching the input content with the existing enterprise documents to obtain a first query result if the intent recognition result is an enterprise document of the query target enterprise; The second matching module is used to obtain the IoT device information in real time through the digital twin platform and match the input content with the IoT device information to obtain the second query result if the intention recognition result is to query the IoT device information of the target enterprise.

[0011] In addition, to achieve the above-mentioned purpose, the present application also proposes a query system, which includes: a large language model for identifying intent based on input content and matching to obtain a first query result and a second query result, a digital twin platform for acquiring the IoT device information of the target enterprise in real time, and a built-in knowledge base for storing existing enterprise documents. The query system is configured to implement the steps of the query method described above.

[0012] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the query method described above are implemented.

[0013] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the query method described above are implemented.

[0014] Compared with the existing digital twin platform, which requires users to manually query IoT device information, users cannot interact with the digital twin platform through natural language, resulting in low information query efficiency. This application first determines the user's intention to query the target enterprise's IoT device information through intent recognition, and then uses the digital twin platform to obtain IoT device information in real time, and matches the user input content with the IoT device information obtained in real time. This allows users to use natural language to interact with the digital twin platform to query IoT device information, rather than being limited to manual queries using specific query formats or instructions, which greatly improves the efficiency of information query and improves the shortcomings of existing digital twin platforms in terms of interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A flowchart of the first embodiment of the query method of the present application is provided; Figure 2 A brief schematic diagram provided for the first embodiment of the query method of the present application; Figure 3 A brief flowchart diagram of the second embodiment of the query method of the present application is provided; Figure 4 This is a schematic diagram of the module structure of the query device for this application; Figure 5 Schematic diagram of the query system provided for this application.

[0018] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0020] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0021] It should be noted that the execution subject of this embodiment may be a query system, or a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device, a processor, etc. capable of implementing the above functions. The query system is taken as an example to illustrate this embodiment and the following embodiments.

[0022] Based on this, the present application embodiment provides a query method, referring to Figure 1 , Figure 1 This is a flowchart of the first embodiment of the query method of the present application.

[0023] In one embodiment, the query method of the present application is implemented based on a large language model (hereinafter referred to as a large model), such as LLM (Large Language Model).

[0024] In this embodiment, the query method includes steps S10 to S30: Step S10, in response to the input operation, obtaining the input content of the questioner and performing intent recognition on the input content to obtain an intent recognition result; It should be noted that when an input operation occurs (for example, the questioner inputs content in a certain interactive interface), the query system obtains the input content of the questioner, where the input content can be voice, text, pictures, etc. Exemplarily, the questioner can be a user. The large model analyzes the input content according to the prompt words, and identifies whether the input content is related to corporate documents or IoT device information through the prompt words in the input content. The analysis process includes lexical analysis, syntactic analysis, semantic understanding, etc.

[0025] In a feasible implementation manner, step S10 includes: The input content is classified into two categories to determine whether the questioner is querying the corporate documents of the target enterprise or the IoT device information of the target enterprise.

[0026] It should be noted that the consciousness recognizer of the large model divides the input content into two different categories through built-in prompt words, such as prompt words related to corporate documents such as "financial statements", "revenue", "cost", etc., and prompt words related to IoT devices such as the name, model, function, status, etc. of the device, to determine whether the input content is to query the corporate documents of the target enterprise or the IoT device information of the target enterprise.

[0027] Step S20, if the intention recognition result is to query the enterprise document of the target enterprise, then match the input content with the existing enterprise document to obtain a first query result; It should be noted that after the input content is recognized for intent based on the large model, it is identified that the questioner's query is related to corporate documents, where corporate documents include corporate policy documents, process manuals, etc. Matching the input content with the corporate documents will result in multiple matching results, which are sorted according to relevance. Exemplarily, the rerank model can be used for sorting, and the matching result with the highest relevance is selected as the first query result.

[0028] In a feasible implementation, step S20 includes steps A10 to A30: Step A10, performing multimodal analysis on the existing enterprise document to obtain the text content of the enterprise document; It should be noted that corporate documents include documents in various formats such as Word, PDF, EXCEL, PNG, etc., and there may be images in the documents. Perform multimodal analysis on existing corporate documents to obtain text information in corporate documents of different formats. Among them, multimodal analysis specifically refers to multimodal text analysis, that is, parsing text information from documents of different formats. For example, the text information parsed from a document in Word format may be the text part of the company's policies, process descriptions, project reports, etc.; perform text recognition on the image to extract the text information in the image.

[0029] Step A20, vectorizing the text content to obtain vectorized text content and storing it in a vector database; It should be noted that the parsed text information is converted into a vector through the Embedding model. The Embedding model refers to the process of mapping high-dimensional data (such as text, images, sounds, etc.) to a low-dimensional space to generate an N-dimensional real-valued vector so that computers can process and learn this data more efficiently.

[0030] Step A30, vectorize the input content and match it with the vectorized text content stored in the vector database to obtain a first query result.

[0031] It should be noted that in order to match the text content of the enterprise document, the questioner's input content also needs to be vectorized. The vectorized input content is searched in the vector database, and the most relevant content retrieved is combined with the questioner's input content and the built-in prompt words to generate an answer through a large model, which is converted into natural language form and fed back to the questioner as the first query result.

[0032] Exemplarily, one of the matching methods may be to calculate the similarity between vectors (eg, algorithms such as cosine similarity).

[0033] In this embodiment, by processing the text information in enterprise documents of different formats through multimodal analysis, the information content in the enterprise documents can be captured more comprehensively. When performing queries or analysis, a richer information basis can be provided to improve the accuracy and relevance of query results.

[0034] Step S30, if the intention recognition result is to query the IoT device information of the target enterprise, the IoT device information is obtained in real time through the digital twin platform and the input content is matched with the IoT device information to obtain a second query result.

[0035] It should be noted that after the input content is recognized based on the big model, it is recognized that the query is related to the information of IoT devices, where IoT devices include computers, servers, air conditioners, lights, etc. In the IoT, the digital twin platform can obtain information about IoT devices in real time, including device status data, historical data, operation logs, etc. Exemplarily, IoT devices are transmitted to the digital twin platform through sensors, network connections, etc.

[0036] The input content is matched with the IoT device information based on the big model. The matching process involves calling the interface of the digital twin platform according to the input content to obtain the IoT device information, performing data cleaning and analysis on the obtained IoT device information, and outputting the IoT device information that is most relevant to the input content. The second query result is obtained based on the matching result.

[0037] In one embodiment, a digital twin platform is first constructed, and then integrated with the large model through a unified interface format. When the intention is recognized that the questioner is querying the IoT device information, the IoT device information is obtained from the digital twin platform in real time through the interface, and then matched with the input content, and the device status, operating parameters or other relevant data are queried and fed back to obtain the second query result.

[0038] In a feasible implementation, step S30 includes: establishing an IoT device model and connecting the IoT device model to the digital twin platform to monitor and obtain IoT device information.

[0039] It should be noted that the IoT device model is an abstract representation of IoT devices. It contains a digital description of various properties, functions, behaviors and other information of IoT devices. The purpose of establishing an IoT device model is to better manage and control IoT devices. The digital twin platform is a software platform that can create digital models of IoT devices; the digital twin platform can interact and collaborate with actual IoT devices through digital IoT devices. The digital twin platform monitors the IoT device model and obtains information about the actual IoT device through the connection between the IoT device model and the actual IoT device (for example, connecting the IoT device model to the actual IoT device based on the communication protocol).

[0040] In another feasible implementation, after obtaining the first query result or the second query result, the method includes: parsing the first query result or the second query result according to a large model of a preset prompt word template to generate initial answer content; The format of the initial answer content is converted to obtain a natural language text that complies with grammatical and semantic rules.

[0041] It should be noted that the first query result or the second query result is parsed according to the preset prompt word template in the large model. Specifically, the large model identifies the key information in the query result and corresponds to the prompt word in the preset prompt word template, and then generates the initial answer content according to a certain logic and algorithm. For example, the initial answer content may be generated according to a specific technical logic or data structure, lacking the coherence and logic that natural language should have. Through format conversion, the initial answer content is adjusted, such as adding conjunctions, adjusting word order, supplementing missing grammatical components, etc., so that it becomes a natural language text that conforms to grammatical and semantic rules.

[0042] In this embodiment, the large model parses the query results according to the preset prompt word template to generate the initial answer content, and then performs format conversion to obtain natural language text. The large model with the preset prompt word template helps to accurately extract relevant information from the query results, avoids the interference of irrelevant information, and thus improves the accuracy of the initial answer content. During the format conversion process, the content is further sorted out to ensure accurate semantic expression, so that the final natural language answer can accurately reflect the key information in the query results.

[0043] For example, Figure 2 As shown, Figure 2A brief schematic diagram of the first embodiment of the present application. First, the questioner's input content is obtained, and then the questioner's intention is recognized based on the questioner's input content; the target enterprise's corporate documents or the target enterprise's IoT device information are queried based on the intention recognition result; and finally, the query result in natural text form (i.e., the first query result or the second query result) is obtained through the large model.

[0044] In this embodiment, compared with the traditional query system based on the traditional search engine or simple keyword matching technology, there are semantic understanding errors and the query results do not meet the user's needs, resulting in low efficiency when the user queries information. This application can identify the real intention behind the user's input content by performing intent recognition on the input content. When the intention recognition result is to query corporate documents, it is not just a simple keyword match, but a deep semantic analysis of the input content is performed, and then the query result is obtained by matching based on the feature vector in the vector database, which greatly improves the accuracy and efficiency of querying corporate documents and overcomes the defects of the traditional query system.

[0045] Compared with the existing digital twin platform, which requires users to manually query IoT device information, users cannot interact with the digital twin platform through natural language, resulting in low information query efficiency. This application first determines the user's intention to query the target enterprise's IoT device information through intent recognition, and then uses the digital twin platform to obtain IoT device information in real time, and matches the user input content with the IoT device information obtained in real time. This allows users to use natural language to interact with the digital twin platform to query IoT device information, rather than being limited to manual queries using specific query formats or instructions, which greatly improves the efficiency of information query and improves the shortcomings of existing digital twin platforms in terms of interaction.

[0046] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the first embodiment can refer to the above introduction, and will not be repeated in the following. Figure 3 , step S40, the query method further includes steps D10~D40: Step D10: extract information from the questioner's input content, and generate an input request that complies with the digital twin platform interface specification based on the extracted input content; It should be noted that the input content of the questioner may be in various forms, such as natural language expression, voice, etc. The input content does not directly meet the requirements of the digital twin platform interface specification, so information extraction is required so that the input content can be correctly understood and processed by the interface of the digital twin platform. For example, if the input content is text, the semantic information in the sentence is extracted and converted into the format information of the API interface specification (such as JSON).

[0047] Step D20, responding to the input request through the digital twin platform to obtain the first IoT device information; It should be noted that when the digital twin platform receives an input request that complies with its interface specifications, it parses the request content to determine the IoT device involved in the input request. For example, the input request is to query the current status of a specific IoT device in a workshop.

[0048] The digital twin platform obtains information related to the IoT device to be queried after processing the input request, including the status information of the IoT device (such as whether the device is online and whether it is working normally), the data stored in the IoT device, the historical operation data of the IoT device, or the configuration parameters of the IoT device.

[0049] For example, if the query is about the current status of an IoT device, the digital twin platform directly obtains the real-time status of the IoT device through the digital model of the IoT device; if the query is about the historical operating data of the IoT device, the digital twin platform retrieves relevant data records from its storage system.

[0050] Step D30, preprocessing the first IoT device information to obtain second IoT device information; It should be noted that preprocessing refers to operations such as data cleaning, data conversion, and data normalization of the first IoT device information to extract the second IoT device information related to the input content. The second IoT device information has higher data quality than the first IoT device information and is part of the first IoT device information.

[0051] Step D40, referring to the preset answer template, outputting the second IoT device information matching the input content as the second query result.

[0052] It should be noted that the preset answer template specifies the title format of the query results, the paragraph structure of the content, the order of data arrangement, etc., in order to form the final query results based on the IoT device information that matches the input content. The purpose of referring to the preset answer template is to ensure that the output second query result has a unified format and specification. This can improve the readability and comprehensibility of the query results, and facilitate users to interpret and use the results. Different application scenarios or user needs may require different result presentation methods, and the preset answer template can be customized according to these needs.

[0053] In this embodiment, by converting the format of the questioner's input content and generating an input request that complies with the digital twin platform interface specification, accurate IoT device information can be obtained from the digital twin platform to avoid erroneous responses due to semantic ambiguity; referring to the preset answer template, the second IoT device information matching the input content is output as the second query result, which can improve the readability of the second query result.

[0054] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the query method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0055] This application also provides a query device, please refer to Figure 4 , the query device comprises: The intention recognition module 10 is used to obtain the input content of the questioner in response to the input operation and perform intention recognition on the input content to obtain an intention recognition result; A first matching module 20 is used to match the input content with the existing enterprise documents to obtain a first query result if the intended recognition result is an enterprise document of the query target enterprise; The second matching module 30 is used to obtain the IoT device information in real time through the digital twin platform and match the input content with the IoT device information to obtain a second query result if the intention recognition result is to query the IoT device information of the target enterprise.

[0056] Optionally, the first matching module 20 is further used to perform multimodal analysis on existing enterprise documents to obtain text content of the enterprise documents; Vectorizing the text content to obtain vectorized text content and storing it in a vector database; The input content is vectorized and matched with the vectorized text content stored in the vector database to obtain a first query result.

[0057] Optionally, the second matching module 30 is further used to monitor and obtain IoT device information by establishing an IoT device model and connecting the IoT device model to the digital twin platform.

[0058] Optionally, the second matching module 30 is further used to extract information from the input content of the questioner, and generate an input request that complies with the interface specification of the digital twin platform based on the extracted input content; Responding to the input request through the digital twin platform to obtain first IoT device information; Preprocessing the first IoT device information to obtain second IoT device information; With reference to the preset answer template, the second Internet of Things device information matching the input content is output as the second query result.

[0059] Optionally, the intention recognition module 10 is further used to perform binary classification on the input content to determine whether the questioner is querying the enterprise documents of the target enterprise or the Internet of Things device information of the target enterprise.

[0060] Optionally, the first matching module 20 or the second matching module 30 is further configured to parse the first query result or the second query result according to a large model of a preset prompt word template to generate an initial answer content; The format of the initial answer content is converted to obtain a natural language text that complies with grammatical and semantic rules.

[0061] The query device provided by the present application adopts the query method in the above embodiment to solve the technical problem of low efficiency of information query. Compared with the prior art, the beneficial effects of the query device provided by the present application are the same as the beneficial effects of the query method provided by the above embodiment, and other technical features in the query device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0062] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device.

[0063] In addition, this application provides a query system, please refer to Figure 5 The query system includes: a large language model for identifying intent based on input content and matching to obtain a first query result and a second query result, a digital twin platform for obtaining IoT device information of a target enterprise in real time, and a built-in knowledge base for storing existing enterprise documents.

[0064] In one embodiment, the built-in knowledge base is constructed by the RAG (Retrieval-Augmented Generation) system. The RAG system indexes and hierarchically manages enterprise documents, and combined with semantic retrieval technology, can retrieve relevant document content based on input content, and generate natural language through a generative large model (LLM) to generate structured and easy-to-understand answers. The built-in knowledge base constructed by the RAG system ensures that sensitive data within the enterprise is not leaked, and takes into account data security and privacy protection.

[0065] The query system provided by the present application adopts the query method in the above embodiment to solve the technical problem of low efficiency of information query. Compared with the prior art, the beneficial effects of the query system provided by the present application are the same as the beneficial effects of the query method provided by the above embodiment, and other technical features of the query system are the same as those disclosed in the method of the previous embodiment, which will not be described in detail here.

[0066] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0067] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the query method in the above-mentioned embodiment.

[0068] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0069] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Python, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect through the Internet).

[0070] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0071] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0072] The readable storage medium provided in the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned query method, and can solve the technical problem of low efficiency of information query. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the present application are the same as the beneficial effects of the query method provided in the above-mentioned embodiment, and will not be described in detail here.

[0073] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned query method when executed by a processor.

[0074] The computer program product provided by the present application can solve the technical problem of low efficiency of information query. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the query method provided by the above embodiment, which will not be repeated here.

[0075] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A query method, characterized in that: The query method comprises: In response to the input operation, the input content of the questioner is obtained and the intention recognition is performed on the input content to obtain an intention recognition result; If the intent recognition result is to query the enterprise documents of the target enterprise, the input content is matched with the existing enterprise documents to obtain the first query result; If the intention recognition result is to query the IoT device information of the target enterprise, the IoT device information is obtained in real time through the digital twin platform and the input content is matched with the IoT device information to obtain a second query result.

2. The query method according to claim 1, characterized in that: The step of matching the input content with the existing enterprise documents to obtain the first query result includes: Perform multimodal analysis on existing enterprise documents to obtain the text content of enterprise documents; Vectorizing the text content to obtain vectorized text content and storing it in a vector database; The input content is vectorized and matched with the vectorized text content stored in the vector database to obtain a first query result.

3. The query method according to claim 1, characterized in that: The steps of obtaining IoT device information in real time through the digital twin platform include: By establishing an IoT device model and connecting the IoT device model to the digital twin platform, IoT device information can be monitored.

4. The query method according to claim 1, characterized in that: The step of matching the input content with the IoT device information to obtain a second query result includes: Extract information from the questioner’s input content, and generate an input request that complies with the digital twin platform interface specification based on the extracted input content; Responding to the input request through the digital twin platform to obtain first IoT device information; Preprocessing the first IoT device information to obtain second IoT device information; With reference to the preset answer template, the second Internet of Things device information matching the input content is output as the second query result.

5. The query method according to claim 1, characterized in that: The step of performing intent recognition on the input content to obtain an intent recognition result comprises: The input content is classified into two categories to determine whether the questioner is querying the corporate documents of the target enterprise or the IoT device information of the target enterprise.

6. The query method according to claim 1, characterized in that: After obtaining the first query result or the second query result, the method further comprises: Parsing the first query result or the second query result according to a large model of a preset prompt word template to generate initial answer content; The format of the initial answer content is converted to obtain a natural language text that complies with grammatical and semantic rules.

7. A query device, characterized in that: The query device comprises: An intention recognition module is used to obtain the input content of the questioner in response to the input operation and perform intention recognition on the input content to obtain an intention recognition result; A first matching module, for matching the input content with the existing enterprise documents to obtain a first query result if the intent recognition result is an enterprise document of the query target enterprise; The second matching module is used to obtain the IoT device information in real time through the digital twin platform and match the input content with the IoT device information to obtain the second query result if the intention recognition result is to query the IoT device information of the target enterprise.

8. A query system, characterized in that: The query system includes: a large language model for identifying intent based on input content and matching to obtain a first query result and a second query result, a digital twin platform for acquiring IoT device information of a target enterprise in real time, and a built-in knowledge base for storing existing enterprise documents. The query system is configured to implement the steps of the query method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the query method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the query method according to any one of claims 1 to 6 are implemented.

Images