Method and device for querying real estate service information and terminal equipment

Abstract

The invention provides a query method and device for real estate service information and terminal equipment, and is suitable for the technical field of data processing, and the method comprises the steps: receiving a question statement which is input by a user in a natural language and is related to the real estate service information; according to the role information of the user, matching an associated preset tool and determining an associated preset data range; according to tool information of a preset tool, the question statement and historical multi-round dialogue information of the user, performing model cue word generation to obtain a first target cue word; the first target cue word is analyzed based on a natural language processing model, so that a matched target tool and a target incoming parameter needing to be used are determined from preset tools; and executing the target tool based on the target incoming parameter to obtain a query result in a preset data range. According to the embodiment of the invention, the method can well adapt to the demands of a real estate digital marketing scene, and has the advantages of extremely low operation difficulty, higher query accuracy and safer data.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and in particular, relates to a method, device and terminal equipment for querying real estate business information. Background Art

[0002] With the development of science and technology, digitalization has become the first choice for many industries to improve their competitiveness. The real estate industry is also facing the need for digitalization. In the real estate industry, digital marketing has become an effective means to improve marketing efficiency.

[0003] Business information query technology based on SQL (Structured Query Language) (hereinafter referred to as SQL query technology) is a commonly used query technology for real estate digital marketing. In a typical real estate digital marketing scenario, technicians will first build a relevant business database, which stores a large amount of real estate business information (hereinafter referred to as business information), such as customer visit records, sales performance rankings, contract data, and regional housing data, and design corresponding query models based on SQL statements. On this basis, if business personnel want to obtain the latest customer visit records or sales performance rankings and other business information they need, they can enter the query information in the SQL query bar on the front end of the terminal device according to their needs, and then the query model will convert the query information into SQL query statements through table understanding, field understanding, table relationship understanding, condition filtering, sorting, etc., to obtain the final query results and display them on the front end of the terminal device.

[0004] Although SQL query technology can realize business information query to a certain extent, it is found in actual application that it is not adaptable enough to real estate digital marketing scenarios, and has defects such as high difficulty in operation, low query accuracy, and low data security. Therefore, there is an urgent need in the prior art for a query technology that can effectively meet the real estate digital marketing scenario, so as to reduce the difficulty of business information query and improve query accuracy and security. Summary of the invention

[0005] In view of this, the embodiments of the present application provide a method, device and terminal device for querying real estate business information, which can provide a business information query technology designed for real estate digital marketing scenarios with extremely low operating difficulty, higher query accuracy and more secure data, and is more suitable for the actual needs of real estate digital marketing scenarios.

[0006] A first aspect of an embodiment of the present application provides a method for querying real estate business information, including: Receive a user's question sentence about real estate business information input in natural language.

[0007] According to the role information of the user, a plurality of preset tools associated with the user are matched from the preset tool library, and a preset data range associated with the user is determined.

[0008] According to the tool information of each preset tool, the question statement and the historical multiple rounds of dialogue information with the user, model prompt words are generated for the natural language processing model to obtain the first target prompt word.

[0009] The first target prompt word is analyzed based on a natural language processing model to determine a target tool matching the first target prompt word from preset tools, and to determine target input parameters required for the target tool.

[0010] The target tool is executed based on the target input parameters to obtain the query results processed by the target tool within the preset data range based on the target input parameters.

[0011] In a possible implementation manner of the first aspect, the tool information includes: a tool name of the tool and attribute information of an input parameter.

[0012] Accordingly, according to the tool information of each preset tool, the question statement and the historical multiple rounds of dialogue information with the user, the model prompt word is generated for the natural language processing model to obtain the first target prompt word, including: Generate text description statements for all preset tools based on the tool name of each preset tool and the attribute information of the input parameters.

[0013] Based on the text description sentences, question sentences of all preset tools and multiple rounds of dialogue information with the user, model prompt words are generated for the natural language processing model to obtain the first target prompt word.

[0014] In a possible implementation manner of the first aspect, after obtaining the query result, the query method further includes: Determine the target method for displaying the query results based on the type of output data of the target tool.

[0015] The query results are processed based on the target method and the processed query results are displayed to the user. Different types of tools have different corresponding target methods.

[0016] In a possible implementation manner of the first aspect, the target mode includes: text display.

[0017] Accordingly, the target method for displaying the query results is determined according to the type of data output by the target tool. The query results are processed based on the target method, and the processed query results are displayed to the user, including: When the type of the output data of the target tool is indicator data, a prompt word generation of a natural language processing model is performed on the question statement and the query result to obtain a second target prompt word.

[0018] The second target prompt word is analyzed based on a natural language processing model, a natural language response text is generated for the second target prompt word, and the natural language response text is displayed to the user.

[0019] In a possible implementation manner of the first aspect, the query method further includes: Candidate questions similar to the questions in the question statement are screened out from a preset question library, and one or more candidate tools matching the candidate questions are screened out from preset tools.

[0020] Display candidate questions and links to detail pages of candidate tools on the same page as the processed query results.

[0021] In a possible implementation of the first aspect, after analyzing the first target prompt word based on the natural language processing model to determine a target tool matching the first target prompt word from preset tools and determining a target input parameter required to be used by the target tool, the method further includes: Determine the data or data type that the user needs to query based on the question statement, and determine whether the tool required to be used in the question statement is a preset tool.

[0022] If the data or data type to be queried does not belong to the preset data range, and / or the tool to be used does not belong to the preset tool, the user is prompted that the current authority is insufficient.

[0023] In a possible implementation manner of the first aspect, the query method further includes: If the data or data type to be queried intersects with the preset data range, and the tool to be used is a preset tool, the target tool is executed based on the target input parameters to obtain the query result obtained after the target tool processes the target input parameters within the preset data range.

[0024] In a possible implementation manner of the first aspect, the target manner includes: graphic or list display.

[0025] Accordingly, the target method for displaying the query results is determined according to the type of data output by the target tool. The query results are processed based on the target method, and the processed query results are displayed to the user, including: When the target tool outputs ranking data, the ranking display component is called and the query results are displayed to the user in the form of a preset ranking list in the ranking display component. and\or When the target tool outputs data of the type of trend data, the trend display component is called and the query results are displayed to the user in the form of a preset line chart in the trend display component. and\or When the type of the target tool output data is comparison data, the comparison display component is called, and the query result is displayed to the user in the comparison display component in the form of a preset bar chart.

[0026] A second aspect of an embodiment of the present application provides a real estate business information query device, including: A receiving module, used to receive a question sentence about real estate business information input by a user in natural language; A determination module, configured to match a plurality of preset tools associated with the user from a preset tool library according to the role information of the user, and determine a preset data range associated with the user; A prompt word generation module, configured to generate model prompt words for a natural language processing model according to tool information of each of the preset tools, the question statements, and information about multiple rounds of historical conversations with the user, to obtain a first target prompt word; An analysis module, configured to analyze the first target prompt word based on the natural language processing model, so as to determine a target tool matching the first target prompt word from the preset tools, and determine a target input parameter required to be used by the target tool; The query module is used to execute the target tool based on the target input parameters to obtain the query results of the target tool after processing the target input parameters within a preset data range.

[0027] A third aspect of an embodiment of the present application provides a terminal device, which includes a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, it implements the steps of the real estate business information query method as described in any one of the first aspects above.

[0028] The fourth aspect of an embodiment of the present application provides a computer-readable storage medium, comprising: storing a computer program, characterized in that when the computer program is executed by a processor, the steps of the real estate business information query method as described in any one of the first aspects above are implemented.

[0029] A fifth aspect of an embodiment of the present application provides a computer program product. When the computer program product runs on a terminal device, the terminal device executes the real estate business information query method described in any one of the first aspects above.

[0030] Compared with the prior art, the embodiments of the present application have the following beneficial effects: the query is based on the question statement input by the user in natural language, and the business information processing tool is adapted in combination with the user's role information, thereby realizing basic user authority adaptation. On this basis, on the one hand, the prompt word generation of the large model is generated based on the tool information of each processing tool, the question statement and the historical dialogue information with the user. On the other hand, the prompt word is reused based on the large model to perform secondary adaptation of the processing tool, so as to determine the target tool required for this query, and then the business information query is performed based on the prompt word and the target tool, and the required initial query result is obtained. The whole process has no operation threshold for the user, and the operation difficulty is extremely low; the mining and matching of the user's intention is more appropriate, so the query result page finally obtained is more accurate and more accurate; the embodiments of the present application can effectively adapt to the user role situation and prevent the business information beyond the user's authority from being queried. Therefore, the security of the data can be greatly improved. That is, the embodiments of the present application can adapt to the needs of real estate digital marketing scenarios very well, and have the advantages of extremely low operation difficulty, higher query accuracy and more secure data. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0032] Figure 1 It is a schematic diagram of the implementation process of the method for querying real estate business information provided in the embodiment of the present application; Figure 2 It is a flowchart of a user prompt in a method for querying real estate business information provided in an embodiment of the present application; Figure 3 It is a flow chart of the indicator data display in the real estate business information query method provided in the embodiment of the present application; Figure 4 This is another implementation flow diagram of the method for querying real estate business information provided by an embodiment of the present application; Figure 5 This is another implementation flow diagram of the method for querying real estate business information provided in the embodiment of the present application; Figure 6 It is a structural diagram of a device for querying real estate business information provided by an embodiment of the present application; Figure 7 It is a schematic diagram of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0033] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0034] For ease of understanding, some features that may be involved in the embodiments of the present application are first explained: Large models: usually refer to deep learning models with millions or even hundreds of billions of parameters in the field of artificial intelligence. They are able to capture and learn complex data features and patterns by training on large-scale data sets, thereby demonstrating excellent performance and generalization capabilities in tasks such as natural language processing, image recognition, and speech recognition.

[0035] Prompt words: refers to the text input to the large model. These texts serve as the context of the model and guide the model to generate the expected output results.

[0036] Historical conversation information: historical data used for conversations with the query system, in which a question and answer between the user and the system is recorded as one round of historical conversation.

[0037] With the development of science and technology, digitalization has become the first choice for many industries to improve their competitiveness. The real estate industry is also facing the need for digitalization. In the real estate industry, digital marketing has become an effective means to improve marketing efficiency. In order to achieve digital marketing, the query system has the following related technologies: Technology 1: Customized query technology In customized query technology, business personnel usually first clarify their query requirements, and then technical personnel perform a series of steps such as understanding the requirements, writing code, testing and deployment, providing a visual page to meet the specific query requirements of business personnel and provide real estate business information (hereinafter referred to as business information). However, as the business continues to evolve, these requirements may also change, forcing developers to repeat the above process to adapt to new business requirements, which is extremely costly. In addition, when business personnel perform data queries, they need to set the filtering conditions themselves, which is often inefficient.

[0038] Therefore, customized query technology has at least the following defects: 1. It requires multiple developers to complete a series of steps to meet the needs of business personnel, which is costly and has a long development cycle.

[0039] 2. As the business continues to evolve, the query requirements raised in the past may change accordingly, forcing developers to repeat the above process, resulting in high maintenance costs.

[0040] 3. When business personnel query data, they need to set the filtering conditions themselves, which is inefficient.

[0041] 4. When business personnel are performing another query requirement, they need to switch paths back and forth, which is inefficient. In addition, they need to be familiar with the functions of each page to match their own needs. The threshold for use is high and the operation is difficult.

[0042] Technology 2: SQL query technology In order to improve the defects of the above-mentioned customized query technology and meet more and more differentiated query needs, business information query technology based on SQL (Structured Query Language) (hereinafter referred to as SQL query technology) has begun to be promoted and applied and gradually become the mainstream practice. In a typical real estate digital marketing scenario, technicians will first build a relevant business database, which stores a large amount of business information, such as customer visit records, sales performance rankings, contract data, and regional housing data, and design corresponding query models based on SQL statements. On this basis, if business personnel want to obtain the latest customer visit records or sales performance rankings and other business information, they can enter the query information in the SQL query bar on the front end of the terminal device according to their needs, and then the query model will convert the query information into SQL query statements through table understanding, field understanding, table relationship understanding, condition filtering, sorting, etc., and obtain the final query results and display them on the front end of the terminal device.

[0043] SQL query technology has at least the following defects: 1. Extremely difficult to operate: The professionalism and learning curve of SQL query technology is a significant obstacle for non-professional users, and users need to allocate a lot of time to learn usage skills. Even users with a certain programming foundation may find it difficult to face complex SQL queries. Therefore, for real estate business personnel with high personnel turnover, on the one hand, the operation difficulty is extremely high for each business personnel, and the operability is extremely low. Technical personnel are often required to assist in the operation, which increases the cost of use and makes it difficult to improve efficiency. On the other hand, due to the high personnel turnover, learning and training for new business personnel is also the norm for real estate companies, which brings high time costs and direct economic costs.

[0044] Therefore, for real estate digital marketing, the difficult-to-operate SQL query technology will directly lead to a series of problems: low query efficiency, high direct manpower costs (including business personnel learning costs and technical personnel assistance costs) and maintenance manpower costs.

[0045] 2. Low query accuracy: When the query model converts SQL statements, the steps of data conversion and understanding are complex and logical, and are all based on the understanding and matching of the text content of the query information entered by the user, and cannot further mine the deep information in the query information. Due to the huge number of business tables, additional training models may be required to narrow down the selection of tables. In these complex operations, each process may lead to inaccurate final query results, which often makes the final query results less accurate.

[0046] 3. Low data security: There is no permission management mechanism, and the roles of business personnel cannot be adapted, resulting in some information outside the permission being queried. In addition, the query accuracy is not high, which makes it easier to leak customer information.

[0047] In summary, although SQL query technology can realize business information query to a certain extent, it is found in actual application that it is not adaptable enough to the real estate digital marketing scenario, and has defects such as high difficulty in operation, low query accuracy and low data security. Therefore, there is an urgent need in the prior art for a query technology that can effectively meet the real estate digital marketing scenario, so as to reduce the difficulty of business information query and improve query accuracy and security.

[0048] In order to provide a query technology that is more suitable for real estate digital marketing scenarios, in order to reduce the difficulty of business information query, while improving query accuracy and security. In an embodiment of the present application, the query statement input by the user in natural language is used as the query basis, and the business information processing tool is adapted in combination with the user's role information, thereby realizing basic user authority adaptation. On this basis, on the one hand, the prompt word generation of the large model is generated based on the tool attributes of each processing tool, the question statement, and the historical dialogue information with the user. On the other hand, based on the large model, the prompt word is reused to perform secondary adaptation of the processing tool, so as to determine the target tool required for this query, and then the business information query is performed based on the prompt word and the target tool, and the required initial query result is obtained. Finally, the initial query result is processed in a secondary data display mode according to the type of output result of the target tool, and the final query result after the final secondary processing is displayed to the user.

[0049] Compared with SQL query technology, the embodiment of the present application has at least the following advantages: 1. There is no operation threshold for users, and the operation difficulty is extremely low: In the embodiment of the present application, the original input of the query system is: the question sentence input by the user in natural language. That is, for the user, no relevant technical reserves are required, and only the content that he wants to query is needed as if he were talking to other humans in daily life. Therefore, for the user, the operation threshold is close to zero and the operation difficulty is extremely low.

[0050] At the same time, due to the extremely low operating difficulty, the business information query efficiency of users in the embodiments of the present application can be effectively improved, that is, users do not need to remember cumbersome operation pages, and can completely get rid of the complicated operation pages and obscure operation processes. And because the difficulty of getting started is extremely low, there is basically no technical reserve requirement for users, and users of any role (such as typical users in real estate digital marketing scenarios: business personnel) can directly start operating without prior training or technical assistance. Therefore, for the situation of extremely high personnel mobility in the real estate industry, the direct manpower (including business personnel learning costs and technical personnel assistance costs) and maintenance manpower costs required by the embodiments of the present application are extremely low.

[0051] 2. Higher query accuracy: On the one hand, the embodiment of the present application simplifies the steps of calling query information and the complexity of the query system, thereby reducing the possibility of errors in multiple complex operations, and thus effectively improving the query accuracy. For example, in SQL query technology, it is necessary to go through the processes of table understanding, field understanding, table relationship understanding, condition filtering, sorting, etc., each of which may lead to inaccurate final query results. In the embodiment of the present application, only tool understanding and tool parameter understanding are required to display the results.

[0052] On the other hand, in the embodiment of the present application, the user's question statements will be processed and mined for deep information to fully understand the real user intentions. On this basis, the prompt words required for the large model will be further generated according to the properties of the available query tools and the intentions in the user's question statements, so that the input and query of the large model are more accurate, thereby effectively improving the reliability of the final target tool screening and the accuracy of the parameters passed to the target tool. Finally, the target tool is executed based on these passed-in parameters, and the results obtained can better meet the real needs of the user. Compared with SQL query technology, the mining and matching of user intentions are more appropriate, so the final query result page is more accurate and more accurate.

[0053] In actual applications, the inventors collected 1,500 query samples, and the parsing accuracy of the embodiment of the present application was as high as 97.8%, while the probability that the query result of the first question was wrong and the second question was still wrong was only 0.33%, which is significantly higher than the existing query systems based on SQL query technology on the market.

[0054] 3. Higher data security: In the embodiment of the present application, the selection of tools will first be screened based on the user role information, and the data range that can be queried by different tools is relatively controllable, that is, the data range of each query will be limited to the available range of the current user role authority. Therefore, the embodiment of the present application can effectively adapt to the user role situation and prevent business information beyond the user's authority from being queried. Therefore, the security of the data can be greatly improved.

[0055] 4. The development workload of the query system is smaller, which greatly shortens the development cycle of the query system.

[0056] The inventor has verified in actual application that, through the embodiment of the present application, the development cycle of the query system can be reduced from the original average of 5 working days to 2 working days, reducing the development cycle by 60%. That is, the development cycle of the query system can be greatly shortened.

[0057] Based on the above analysis, it can be seen that the embodiments of the present application can adapt to the needs of real estate digital marketing scenarios very well: the operation difficulty is extremely low, the query accuracy is higher and the data is more secure.

[0058] In addition, the execution subject of the embodiment of the present application is explained. In the embodiment of the present application, the query system can be set in a single terminal device in the form of a software system for user use. At the same time, the query system can also be distributed in multiple terminal devices according to actual needs, some of which are used as human-computer interaction devices to perform information reception and output operations, and other devices are responsible for data processing operations other than human-computer interaction, such as the operation of query tools, the operation of large models, etc. The terminal device can be any electronic device with a certain data processing capability, including but not limited to: desktop computers, laptops, mobile phones, tablet computers, servers, etc.

[0059] In order to illustrate the technical solution described in this application, a specific embodiment is provided below for illustration.

[0060] Figure 1 The following is a flowchart of the method for querying real estate business information provided in the first embodiment of the present application, which is described in detail as follows: S101, receiving a question sentence about real estate business information input by a user in natural language.

[0061] The user refers to a person who operates a terminal device to query business information, and is mainly a real estate industry business person. The following description will be based on the business person as an example. The input method of the question statement can be text input or voice input, etc., and there is no excessive limitation here.

[0062] In an embodiment of the present application, users can ask questions to the terminal device according to actual needs. When asking questions, they only need to say what they want to query in the same way as in daily conversations with other humans. For example, suppose business person A wants to query the contract data in area A, and business person B wants to query the real estate sales data of area A + area B. For business person A, you can enter "Help me check the contract data of area A", and business person B can enter "I want to see the real estate sales data of area A + area B". At this point, different business personnel can freely enter question statements according to their own language habits and the needs of this query, and the entire input process does not require too many complex and professional operations in the operation page.

[0063] S102: According to the role information of the user, a plurality of preset tools associated with the user are matched from a preset tool library, and a preset data range associated with the user is determined.

[0064] In the embodiment of the present application, in order to achieve accurate control of user rights, different usable tools and the data range that each role can query will be allocated in advance according to the actual situation of different roles. On this basis, when each user queries business information, the embodiment of the present application will match all the tools (i.e., preset tools) that can be used within the scope of their authority based on the role information of the user, and at the same time determine the data range (i.e., preset data range) that the user role has the authority to query, as the basis for subsequent business information queries.

[0065] Take an example to illustrate. Assume that the preset tool library contains: tool a, tool b, tool c, tool d and tool e. At the same time, all business information is divided according to the region to which the data belongs, such as all real estate business information in region A, all real estate business information in region B, all real estate business information in region C, all real estate business information in region D, and all real estate business information in region E. At the same time, it is assumed that there are three types of roles: ordinary personnel who can query a single region, supervisors who can manage multiple regions, and managers who can manage all regions. At this time, tools and data ranges can be allocated according to the specific regions associated with each type of role and the specific tasks undertaken in real estate digital marketing. For example, it is assumed that for ordinary personnel in region A, it can be set to use tools a, tool b and tool c, and the data range that can be queried includes all real estate business information in region A; for supervisors who manage regions A and B at the same time, it can be set to use tools a, tool b, tool c and tool d, and the data range that can be queried includes all real estate business information in regions A and B; for managers who can manage all regions, it can be set to use all tools and query real estate business information in all regions.

[0066] Based on the above settings, assuming that the role of business personnel A belongs to an ordinary staff in region A, then at this time, it can be determined through matching that the associated preset tools include: tool a, tool b and tool c, and the associated preset data range is: all real estate business information in region A.

[0067] Alternatively, a more detailed division of permissions can be performed, and the matching tools and scopes can be further subdivided for each role. For example, ordinary personnel can also be subdivided into multiple roles with finer granularity, so as to achieve further control of permissions with finer granularity. For example, for ordinary personnel a, it can be set to use basic query tools and data comparison tools, and other more advanced tools such as ranking tools and trend analysis tools are prohibited from use, and the data range is all contract data in region A, and other data are not allowed to be queried. Ordinary personnel b, on the other hand, can be set to use all basic query tools, data comparison tools, ranking tools, and trend analysis tools, and the data range is all real estate business information in region A. At this time, assuming that business personnel B belongs to ordinary personnel a in region A, the corresponding preset tools include basic query tools and data comparison tools, and the preset data range is all contract data in region A.

[0068] Based on the available tools and data range allocation of the above roles, it can be ensured that the data and tools in all subsequent operations are within the scope of the current user's authority and will not involve content that exceeds the user's authority, thereby ensuring data security.

[0069] S103, generating text description statements for all preset tools according to the tool names of the preset tools and the attribute information of the input parameters.

[0070] After determining all preset tools, the embodiment of the present application further obtains the tool name of each preset tool and the attribute information of the input parameters. The attribute information of the input parameters includes: the name of the tool parameter, the description of the tool parameter, and the type of the tool parameter, etc. When a single tool contains multiple input parameters, the attribute information of all the input parameters of the tool is obtained.

[0071] Let's take an example to illustrate this. For the contract data query tool: Tool name: Contract data query; Incoming parameter 1: Parameter name: start_time, Parameter description: the start time of data query, Parameter type: string; Pass in parameter 2: Parameter name: end_time, Parameter description: the end time of data query, Parameter type: string; Pass in parameter 3: Parameter name: location, Parameter description: The area for data query. The area type is limited to province, city, district, and project name. Each area is an inclusion relationship. For example, if you filter Shenzhen, all real estate projects in all areas of Shenzhen will be selected. Parameter type: String.

[0072] After obtaining the tool name of each preset tool and the attribute information of the input parameters, a text description statement for all preset tools is generated based on these data, so that more accurate prompt word information for improving the large model can be generated later. Among them, the method for generating the text description statement is not limited here, including but not limited to: for a single preset tool, the tool name and the attribute information of the input parameters are listed in sequence to obtain a text description of the preset tool. For all preset tools, the text descriptions of all preset tools are listed in sequence in the form of tool name: text description of the tool. For example: Tool 1 Name: Text description of Tool 1; Tool 2 Name: Text description of Tool 2; Tool 3 Name: Text description of Tool 3.

[0073] In some embodiments, if the preset tool has more tool information besides the tool name and input parameters, you can also choose to synchronously obtain this information to generate a text description statement, or you can choose to still use the attribute information of the tool name and input parameters to generate a text description statement.

[0074] S104, based on the preset model, generate model prompt words for the natural language processing model for the text description sentences of all preset tools, the user's question sentences, and the historical multiple rounds of dialogue information with the user, and obtain the first target prompt word.

[0075] In another optional embodiment of the present application, the operations of S103 and S104 can also be summarized as follows: According to the tool information of each preset tool, the question statement and the historical multiple rounds of dialogue information with the user, model prompt words are generated for the natural language processing model to obtain the first target prompt word.

[0076] The tool information at least includes: the tool name and the attribute information of the input parameters. Please refer to the above related instructions for details, which will not be repeated here.

[0077] S105, analyzing the first target prompt word based on a natural language processing model, determining a target tool matching the first target prompt word from a plurality of preset tools, and determining target input parameters required for the target tool.

[0078] In the embodiment of the present application, the natural language processing model is a large model pre-trained for natural language learning. In practical applications, the natural language processing model has extremely strong deep learning and reasoning capabilities, and can better learn the deep semantics of question sentences and understand the user's intentions.

[0079] Limited by the common drawbacks of current large models, existing large models have high requirements for input prompt words. Two paragraphs of text with the same substantive content have different description methods for their prompt words, and different guiding capabilities for the large model, which may lead to large differences in the final large model output results. Based on this, in order to improve the ability of natural language processing model analysis, after obtaining the text description sentences for all preset tools, the embodiment of the present application does not directly use this information to input into the natural language processing model to analyze the tools required by the user and the parameter information required by the tools. Instead, the text description sentences of all preset tools, the user's question sentences, and the recent multiple rounds of dialogue information with the user are first sorted out in a text description method to generate a first target prompt word suitable for the natural language processing model. Among them, the embodiment of the present application does not make too many restrictions on the specific text description method in the prompt word, and can be set according to the actual natural language processing model situation.

[0080] As an optional embodiment of the present application, a text description template suitable for a natural language processing model can be pre-set in the preset model, and when necessary, all the text description sentences of the preset tools, the user's question sentences, and the recent multiple rounds of dialogue information with the user are inserted into the corresponding positions in the text description template to obtain the required target prompt words. For example, in some specific embodiments, all information is logically connected in series using natural language, and a guiding logic is set to guide the model to output the required data. For example, the text description template can be: "You will receive a string containing a list of callable tools. Please parse this string and return an object containing the tool name and tool parameters based on the context information of the user's question. Here is the list of tools you can use: [text description of all preset tools], the user's question is: [the user's question statement], the user's historical conversation information: [multiple rounds of historical conversation information with the user], the return format is as follows: tool name, the value of each tool's passed parameter." Based on the text description template, the preset model can be further filled with data to obtain the required first target prompt word.

[0081] After obtaining the first target prompt word, the embodiment of the present application will input it into the natural language processing model, and the natural language processing model will perform a round of natural language analysis on the user's question statement and the recent multiple rounds of dialogue information with the user, so as to parse out the deep semantic information therein and achieve a deep understanding of the user's intention. On this basis, the target tool required for the user's intention is matched from the preset tool, and the target input parameters required for the determined target tool are used. Among them, the dialogue information of multiple rounds with the user's history can provide richer semantic information, which is conducive to in-depth analysis of the user's deep-level true intention. On this basis, combined with the user's current question statement, the tool that the user really needs can be obtained. Relative to the analysis of SQL, this can achieve a deep analysis of user needs, thereby effectively improving the accuracy of the analysis, and then improving the accuracy of the final query result. The embodiment of the present application does not make too many restrictions on the specific number of historical dialogue rounds used, and can be set according to actual needs. For example, in some optional embodiments, it can be set to any value among 4, 5, 6, 7, and 8.

[0082] It should be noted that the preset model and the natural language processing model can be the same large model or different models, which can be set according to actual needs, and no excessive restrictions are made here. For example, in some embodiments, the two are the same large model, and the large model can implement different functions according to different input data. When the text description sentences of all preset tools, the user's question sentences and the recent rounds of dialogue information with the user are directly input, the corresponding prompt words are generated. When a specific prompt word is entered, a specific result is output according to the prompt word, such as determining the appropriate target tool and the input parameters required for the tool based on the prompt word.

[0083] As an optional embodiment of the present application, on the basis of determining the preset tool and the preset data range, the embodiment of the present application can also determine the required tool and the data type involved for the question statement, and match the preset tool and the preset data range. If the required tool is within the preset tool range, and the data of the data type involved is included in the preset data range, the data matching and query work will be carried out normally. If the required tool does not belong to the preset tool range, and / or the data of the data type involved is not included in the preset data range, it will be prompted that the current user has insufficient authority and cannot be queried. Reference Figure 2 Specifically, the embodiments of the present application include: S201, determining the data or data type that the user needs to query according to the question statement, and determining whether the tool required to be used in the question statement is a preset tool.

[0084] S202: If the data or data type to be queried does not belong to the preset data range, and / or the tool to be used does not belong to the preset tool, the user is prompted that the current authority is insufficient.

[0085] S203: If the data or data type to be queried overlaps with the preset data range, and the tool to be used belongs to the preset tool, then continue to perform subsequent operations.

[0086] If there is no mismatch between the tool and the data range, or if the tool matches but the data range does not completely match, that is, there is a partial intersection or full overlap with the preset data range, the next step S106 of business information query can be carried out normally.

[0087] S106, executing the target tool based on the target input parameters, and obtaining the query results output by the target tool after processing the target input parameters within a preset data range.

[0088] After obtaining the target input parameters, the parameters are input into the target tool and the target tool is executed. At this time, the target tool performs a specific analysis on the preset data range within the user's authority based on the target input parameters, thereby obtaining the corresponding query results. Among them, the execution logic of the target tool is not described in detail here. The focus is that when the target tool is running, the range of data samples that can be filtered is the characteristic range that matches the user role to improve the security of the data. As an embodiment of the present application, each tool may need to be limited to a certain type of output format, such as indicator data, ranking data, trend data, comparison data, etc.

[0089] For example, assuming that the target tool is contract data query, the scope that user A can query is all contract data in region A, and assuming that user A needs to query contract data within the past year. At this time, based on the actual authority of user A, the embodiment of the present application will use all contract data within the past year in region A as the query result. Among them, areas outside region A are outside the preset data range of user A and will not respond at this time. That is, when the tool matches, the embodiment of the present application will only analyze and query the data within the preset data range.

[0090] After obtaining the query results, you can choose to display the query results directly on the front end, or you can choose to further process the query results before displaying them on the front end, which is not limited here. For example, when the query result is report data, you can choose to display the report data directly, or perform secondary processing (such as visualization) on the report data before displaying it. At the same time, in some optional embodiments, you can also choose to display the detailed links corresponding to the query results together, so that users can view them when needed.

[0091] Compared with SQL query technology, the embodiment of the present application has at least the following advantages: 1. There is no operation threshold for users, and the operation difficulty is extremely low: In the embodiment of the present application, the original input of the query system is: the question sentence input by the user in natural language. That is, for the user, no relevant technical reserves are required, and only the content that he wants to query is needed as if he were talking to other humans in daily life. Therefore, for the user, the operation threshold is close to zero and the operation difficulty is extremely low.

[0092] At the same time, due to the extremely low operating difficulty, the business information query efficiency of users in the embodiments of the present application can be effectively improved, that is, users do not need to remember cumbersome operation pages, and can completely get rid of the complicated operation pages and obscure operation processes. And because the difficulty of getting started is extremely low, there is basically no technical reserve requirement for users, and users of any role (such as typical users in real estate digital marketing scenarios: business personnel) can directly start operating without prior training or technical assistance. Therefore, for the situation of extremely high personnel mobility in the real estate industry, the direct manpower (including business personnel learning costs and technical personnel assistance costs) and maintenance manpower costs required by the embodiments of the present application are extremely low.

[0093] 2. Higher query accuracy: On the one hand, the embodiment of the present application simplifies the steps of calling query information and the complexity of the query system, thereby reducing the possibility of errors in multiple complex operations, and thus effectively improving the query accuracy. For example, in SQL query technology, it is necessary to go through the processes of table understanding, field understanding, table relationship understanding, condition filtering, sorting, etc., each of which may lead to inaccurate final query results. In the embodiment of the present application, only tool understanding and tool parameter understanding are required to display the results.

[0094] On the other hand, in the embodiment of the present application, the user's question statements will be processed and mined for deep information to fully understand the real user intentions. On this basis, the prompt words required for the large model will be further generated according to the properties of the available query tools and the intentions in the user's question statements, so that the input and query of the large model are more accurate, thereby effectively improving the reliability of the final target tool screening and the accuracy of the parameters passed to the target tool. Finally, the target tool is executed based on these passed-in parameters, and the results obtained can better meet the real needs of the user. Compared with SQL query technology, the mining and matching of user intentions are more appropriate, so the final query result page is more accurate and more accurate.

[0095] In actual applications, the inventors collected 1,500 query samples, and the parsing accuracy of the embodiment of the present application was as high as 97.8%, while the probability that the query result of the first question was wrong and the second question was still wrong was only 0.33%, which is significantly higher than the existing query systems based on SQL query technology on the market.

[0096] 3. Higher data security: In the embodiment of the present application, the selection of tools will first be screened based on the user role information, and the data range that can be queried by different tools is relatively controllable, that is, the data range of each query will be limited to the available range of the current user role authority. Therefore, the embodiment of the present application can effectively adapt to the user role situation and prevent business information beyond the user's authority from being queried. Therefore, the security of the data can be greatly improved.

[0097] Based on the above analysis, it can be seen that the embodiments of the present application can adapt to the needs of real estate digital marketing scenarios very well: the operation difficulty is extremely low, the query accuracy is higher and the data is more secure.

[0098] 4. The development workload of the query system is smaller, which greatly shortens the development cycle of the query system.

[0099] On the one hand, the embodiments of the present application rely on the definition, arrangement and calling of tools, which can greatly simplify the overall query process, thereby reducing the workload of development. Specifically, business personnel need to develop a report for visit trend query. The original process is that business personnel put forward query requirements, product managers design display plans and operation instructions, and after the back-end developers understand the requirements, they develop the code. After verification by testers, it is deployed and launched as a query interface. At the same time, front-end developers need to develop corresponding visualization pages, and only after the joint debugging with the back-end is passed can the project be launched. When using the embodiments of the present application, only the back-end personnel need to develop the query interface. At this time, the query interface can be defined as a tool for use, omitting the work of product managers and front-end developers.

[0100] On the other hand, the embodiment of the present application limits the output type of each tool, so that all tools (N) need to be displayed according to a certain type of output format (M), so that the front-end development workload is reduced from N to M. That is, all data queries (set to N types) will eventually become one of the M outputs, so the amount of front-end development can be greatly reduced. Wherein N and M are both positive integers greater than 1. For example, in the traditional development process, the development of a requirement requires the front-end to develop a separate display chart for the requirement. In this solution, the development of N tools requires the front-end to develop N display charts for the tool, but through the output format restriction, they are ultimately displayed in one of the M formats, so at this time, only M development times are required.

[0101] The inventor has verified in actual application that, through the embodiment of the present application, the initial development cycle of the query system can be reduced from the original average of 5 working days to 2 working days, reducing the development cycle by 60%. That is, the development cycle of the query system can be greatly shortened.

[0102] 5. It has strong adaptability to new business scenarios, and the cycle for updating and developing the query system is extremely short and the cost is low.

[0103] The embodiments of the present application can better adapt to the changes in the demand for new scenarios of digital marketing in the real estate industry. When there are new business scenario requirements (business scenarios have changed), it is only necessary to modify certain configuration options based on the existing query system to meet the needs. Or when there are demand adjustments to the current business scenario (including demand increases, decreases, or partial modifications), subsequent development needs can be achieved only through configuration.

[0104] The inventors have verified in actual applications that, through the embodiments of the present application, the development cost of the system update development cycle of the query system can be greatly reduced and the development cycle can be shortened by more than 90% when adapting to the needs of new business scenarios (including adjustments to the needs of the current business scenarios or new business scenarios).

[0105] As an optional embodiment of the present application, considering the large differences in data analysis levels among real estate industry business personnel, in order to further improve the effectiveness of digital marketing, the present embodiment of the present application will further optimize the display method of the query results. Specifically, the present embodiment of the present application will adaptively select and analyze the display method based on the characteristics of the data output by each tool, and analyze and display based on the selected display method, thereby improving the data visibility of the query results and reducing the difficulty of data understanding. Figure 1 , the embodiment of the present application also includes: S107, determining a target method for displaying the results according to the type of data output by the target tool, processing the query results based on the target method, and displaying the processed query results to the user; wherein different types of tools correspond to different target methods.

[0106] In the embodiments of the present application, in order to fully adapt to the characteristics of different data, the most appropriate display method is designed for each type of data. The display method includes but is not limited to: text, different visual charts, etc., so as to reduce the difficulty of data understanding as much as possible, improve the visibility of data, and help various business personnel better understand and use the queried business data.

[0107] As an optional embodiment of the present application, in the embodiment of the present application, for indicator data, considering that the tool content output is mainly text content, the readability of the text can be improved as much as possible, that is, it can be displayed to the user in a natural language as much as possible, so that the user can read and understand the query results. Figure 3 , the embodiments of the present application include: S10711, when the type of the output data of the target tool is indicator data, a prompt word generation of a natural language processing model is performed on the user's question statement and the query result of the target tool to obtain a second target prompt word.

[0108] S10712, analyzing the second target prompt word based on the natural language processing model, generating a natural language response text for the second target prompt word, and displaying the natural language response text to the user.

[0109] In the embodiment of the present application, in order to further improve the readability of the query results of indicator data, after obtaining this type of query results, they are not directly displayed, but a second target prompt word for the natural language processing model is generated based on the user's question statement and query results to guide the natural language processing model to process these contents in natural language. After receiving this second target promotion word, the natural language processing model will further process the question statement and the query results of the target tool in natural language, and generate natural language answer text that is closer to human language habits, so that the final query results will not be too stiff or obscure, which is conducive to improving the readability and friendliness of the query result content and reducing the difficulty of user reading.

[0110] Take an example to illustrate, for example, a text description template suitable for a natural language processing model can be pre-set, and when necessary, the user's question sentence and query result are inserted into the corresponding position of the text description template to obtain the required second target prompt word. For example, in some specific embodiments, natural language is used to logically connect all information in series, and a guiding logic is set to guide the model to output the required data. For example, the text description template can be: “User consultation: [user’s question], use [target tool name] tool, tool call parameters are: [input parameters of target tool], tool call result is [query result], please give a reply based on the user’s question and tool call result, just answer directly.” Based on the text description template, data can be filled in to obtain the required second target prompt word.

[0111] For example, in some specific embodiments, the filled second target prompt word may be: "User inquiry: [Query the contract data of Shenzhen City from January 1 to January 31, 2024], use the [Contract Data Query] tool, the tool call parameters are: [Start time: January 1, 2024, End time: January 31, 2024, Region: Shenzhen City], the tool call result is

[100] . Please give a reply based on the user's question and the tool call result. Just answer directly." Finally, the second target prompt word is used to call the natural language processing model, which describes and outputs a natural language reply text in sequence and displays it in the form of text on the front end.

[0112] As an optional embodiment of the present application, for ranking data, the present application embodiment includes: S1072: When the type of the output data of the target tool is ranking data, call the ranking display component, and display the query result of the target tool to the user in the ranking display component in the form of a preset ranking list.

[0113] For ranking data, the embodiment of the present application uses a list-style ranking list display component for display, which allows business personnel to more clearly and directly understand the ranking position of each data in the query results, thereby improving the readability of the data.

[0114] As an embodiment of the present application, for a ranking target tool, the configured tool return should include: ranking, field name used for sorting, and other display fields, such as ranking, number of visits, and project name. For example, a certain return: ranking, number of visits, project name\n1, 100, B building\n2, 80, C building\n3, 50, A building. Based on this, the front end will call the pre-defined ranking display component (list form) for display according to the return format.

[0115] As an optional embodiment of the present application, for trend data, the present application embodiment includes: S1073, when the type of the output data of the target tool is trend data, the trend display component is called, and the query result of the target tool is displayed to the user in the trend display component in the form of a preset line chart.

[0116] For trend data, a line graph can better display its trend direction, so the embodiment of the present application will display the query results of the target tool to the user in the form of a line graph in the trend display component.

[0117] As an embodiment of the present application, if the tool is trend data, the configured tool return should include: time point, indicator name, indicator value, indicator unit. At this time, the front end will call the pre-defined trend display component (line chart form) for display according to the return format.

[0118] As an optional embodiment of the present application, for comparison data, the present application embodiment includes: S1074: When the type of the output data of the target tool is comparison data, the comparison display component is called, and the query result of the target tool is displayed to the user in the comparison display component in the form of a preset bar chart.

[0119] For comparative data, the comparison effect is more obvious when displayed and compared through bar graphs. Therefore, the embodiment of the present application uses a preset bar graph to display comparative data.

[0120] As an embodiment of the present application, if the tool is for comparison data, the configured tool return should include: comparison item name, comparison item value, unit. Take a certain return as an example, comparison item name, comparison item value, unit\nA building, 100, sets\nB building, 80, sets\nC building, 50, sets. At this time, the front end will call the pre-defined comparison display component (bar chart form) for display according to the return format.

[0121] As an optional embodiment of the present application, in order to further improve the efficiency of business personnel's query of business information, based on the above embodiments, while displaying the query results, the present application embodiment will further match questions similar to those asked by the user and related tools and display them to the user. Since the questions and tools are highly relevant to the user's current needs, the user is relatively likely to be interested in these tools and questions, which can provide an effective reference for the user. Continue to refer to Figure 1 , the embodiments of the present application include: S108, screening out candidate questions similar to the questions in the question statement from a preset question library, and screening out one or more candidate tools matching the candidate questions from a plurality of preset tools.

[0122] In the embodiment of the present application, the technicians will collect and store a large number of user questions in advance. On this basis, the embodiment of the present application will match these user questions based on the user's question sentences, so as to select questions similar to the user's question sentences (i.e., candidate questions). Among them, the embodiment of the present application does not make too many restrictions on the specific question matching method.

[0123] As an optional embodiment of the present application, a method of multi-way recall + large model precise matching can be used to achieve precise matching of similar questions. Multi-way recall refers to the simultaneous use of multiple methods to match similar questions, and obtain the matching results of each matching method. For example, in some embodiments, on the one hand, the question statement can be processed using an algorithm such as TFIDF to obtain the keywords of the user's question, and then compared with the keywords of each preset question to screen out the five similar questions with the largest number of keywords; on the other hand, through a vector model such as BGE-M3, the similarity between the user's question statement and each preset question is calculated to find the five most similar questions. On the basis of obtaining the similar question results of multi-way matching, the natural language processing model is used to perform a secondary screening of all similar questions, and the similar questions with the highest similarity are selected as candidate questions.

[0124] After the candidate questions are determined, several tools that best match the candidate questions are selected from the preset tools as candidate tools. The embodiments of the present application do not place too many restrictions on the specific problem matching scheme and tool matching scheme used. For example, one or more tools can be associated with each question in advance, and on this basis, one or more candidate tools that match the candidate questions can be directly determined through the mapping relationship.

[0125] S109, displaying the candidate questions and detail page links of the candidate tools on the same page as the processed query results.

[0126] After determining the candidate questions and candidate tools, the embodiment of the present application will further display the detail page links of the candidate questions and candidate tools together with the query results on the front page of the terminal device for the user to view together, and to view or use the candidate tools through the detail page links when necessary. That is, data query can be performed on the same operation page, plus guessing what you want to ask, the display of specific detail pages is conducive to divergent thinking and is more flexible than traditional methods.

[0127] The embodiment of the present application can intelligently recommend possible subsequent business query directions based on the query requirements proposed by the user, thereby improving the efficiency and accuracy of business decision-making.

[0128] It should be specially explained that, for a single terminal device as the execution subject, the above S101 to S109 are all executed by the terminal device, and the human-computer interaction and background data processing are all performed by the same terminal device. For a system where the execution subject is composed of multiple terminal devices, S101 to S109 can still be executed by the same terminal device, but the difference is that the receiving question statement of S101 is not the question statement directly input by the user, but the question statement of the user sent by other terminal devices. Similarly, the operation of displaying the results in S107 to S109 is to send the results to other terminal devices and the other terminal devices will perform the final display. The form of the overall multi-device system can refer to the common system composed of local computers + cloud servers, with the local computer as the human-computer interaction device and the cloud server as the data processing device, so the specific details are not repeated here.

[0129] As an optional specific embodiment of the present application, refer to Figure 4 Taking a single query as an example, in the embodiment of the present application, when a business person asks a question about a certain requirement, the solution will determine which tool's query corresponds to the question, and automatically parse the parameters contained in the question, call the corresponding data query tool, and find the result report. At this time, the detailed link corresponding to this report and similar questions to this query will also be displayed. The specific display method of the query report is combined with the above embodiment and is set according to the data type output by the target tool, which will not be repeated here.

[0130] Among them, the specific implementation details and beneficial effects of the embodiments of the present application can be referred to the other embodiments mentioned above and will not be repeated here.

[0131] As another optional specific embodiment of the present application, refer to Figure 5 , in the embodiments of the present application: First, the user will ask the system questions based on their needs. The system will obtain the tools available under the user role. For example, role A can only query the contract status in a certain area and cannot call the ranking tool for query. When A wants to query the contract data in a non-role area, the system will prompt that the authority is insufficient. Role B can have the contract query authority for all areas and the authority to use the ranking tool.

[0132] The prompt word management module will splice multiple tool descriptions based on the description of each tool and the parameters that need to be passed in. For example, the text description of a tool is as follows: Tool name: Contract data query; Incoming parameter 1: Parameter name: start_time, Parameter description: the start time of data query, Parameter type: string; Pass in parameter 2: Parameter name: end_time, Parameter description: the end time of data query, Parameter type: string; Pass in parameter 3: Parameter name: location, Parameter description: The area for data query. The area type is limited to province, city, district, and project name. Each area is an inclusion relationship. For example, if you filter Shenzhen, all real estate projects in all areas of Shenzhen will be selected. Parameter type: String.

[0133] Finally, the text descriptions of multiple sets of tools are as follows: Tool 1: Text description of Tool 1; Tool 2: Text description of Tool 2. At the same time, each tool needs to be limited to a certain type of output format, such as indicator data, ranking data, trend data, comparative data, etc. At the same time, it is also necessary to configure a list of tools similar to this tool and the corresponding detail page link of the tool (if any).

[0134] At this point, the big model will form a prompt word for the big model to call based on the text descriptions of multiple sets of tools, user questions, and the historical five-round dialogue information. An example is as follows: You will receive a string containing a list of callable tools. Please parse this string and return an object containing the tool name and tool parameters based on the context information of the user's question. Here is the list of tools you can use: [combination of text descriptions of multiple tools], the return format is as follows: tool name, the value of each passed parameter, For example, if the tool name is: Contract Data Query, parameter 1 is: January 1, 2024, parameter 2 is: January 31, 2024, parameter 3 is: Shenzhen City, the user's question is: {the content of the user's text input}, and the user's historical conversation information is: [the historical 5 rounds of conversation information]. Return:.

[0135] Historical dialogue information: The question and answer text between the user and the system is recorded as one round of historical dialogue.

[0136] At this time, the big model will judge which tool is most appropriate to call based on this prompt word and its own understanding ability, and give the parameter values ​​required by the tool. At this time, the system will execute the tool to obtain the result. For example, when calling the contract data, the query result is as follows: 100.

[0137] At this time, if the tool is indicator data, the system will combine the user's questions and data query results into a text description. For example, the user asks: [content entered by the user], uses the [Contract Data Query] tool, and the tool call parameters are: [Start time: January 1, 2024, End time: January 31, 2024, Region: Shenzhen City], and the tool call result is

[100] . Please give a reply based on the user's question and the tool call result. Just answer directly. Use this text description as a prompt to call the big model. The big model outputs a reply text based on this description and displays it in text form on the front end.

[0138] If the tool is for ranking data, the configured tool return should include: ranking, field name used for sorting, and other display fields, such as ranking, number of visits, and project name. Take a certain return as an example: ranking, number of visits, project name\n1, 100, Building B\n2, 80, Building C\n3, 50, Building A. At this time, the front end will call the pre-defined ranking display component (list form) for display based on the return format.

[0139] If the tool is trend data, the configured tool return should include: time point, indicator name, indicator value, indicator unit. At this time, the front end will call the pre-defined trend display component (line chart form) for display according to the return format.

[0140] If the tool is for comparison data, the configured tool return should include: comparison item name, comparison item value, unit. Take a certain return as an example, comparison item name, comparison item value, unit\nA building, 100, sets\nB building, 80, sets\nC building, 50, sets. At this time, the front end will call the pre-defined comparison display component (bar chart format) for display according to the return format.

[0141] In addition, when the big model selects which tool to call, the corresponding similar questions and tool details page links will be selected based on the pre-configured items and displayed on the front end. The tool details page link and similar questions will appear below the answer results.

[0142] Among them, the specific implementation details and beneficial effects of the embodiments of the present application can be referred to the other embodiments mentioned above and will not be repeated here.

[0143] Corresponding to the method of the above embodiment, Figure 6 A structural block diagram of a real estate business information query device provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown. Figure 6 The exemplary real estate business information query device may be an execution subject of the real estate business information query method provided in the aforementioned first embodiment.

[0144] Reference Figure 6 , the real estate business information query device includes: The receiving module 61 is used to receive a question sentence about real estate business information input by a user in natural language; A determination module 62 is used to match a plurality of preset tools associated with the user from a preset tool library according to the role information of the user, and determine a preset data range associated with the user; A prompt word generation module 63 is used to generate model prompt words for a natural language processing model according to the tool information of each of the preset tools, the question statement, and the historical multiple rounds of dialogue information with the user, so as to obtain a first target prompt word; An analysis module 64 is used to analyze the first target prompt word based on the natural language processing model to determine a target tool matching the first target prompt word from the preset tools, and determine target input parameters required to be used by the target tool; The query module 65 is used to execute the target tool based on the target input parameters to obtain the query result of the target tool after processing the target input parameters within a preset data range.

[0145] In some embodiments of the present application, the tool information includes: the tool name of the tool and attribute information of the input parameters.

[0146] Accordingly, according to the tool information of each preset tool, the question statement and the historical multiple rounds of dialogue information with the user, the model prompt word is generated for the natural language processing model to obtain the first target prompt word, including: Generate text description statements for all preset tools based on the tool name of each preset tool and the attribute information of the input parameters.

[0147] Based on the text description sentences, question sentences of all preset tools and multiple rounds of dialogue information with the user, model prompt words are generated for the natural language processing model to obtain the first target prompt word.

[0148] In some embodiments of the present application, after obtaining the query result, the query method further includes: Determine the target method for displaying the query results based on the type of output data of the target tool.

[0149] The query results are processed based on the target method and the processed query results are displayed to the user. Different types of tools have different corresponding target methods.

[0150] In some embodiments of the present application, the target mode includes: text display.

[0151] Accordingly, the target method for displaying the query results is determined according to the type of data output by the target tool. The query results are processed based on the target method, and the processed query results are displayed to the user, including: When the type of the output data of the target tool is indicator data, a prompt word generation of a natural language processing model is performed on the question statement and the query result to obtain a second target prompt word.

[0152] The second target prompt word is analyzed based on a natural language processing model, a natural language response text is generated for the second target prompt word, and the natural language response text is displayed to the user.

[0153] In some embodiments of the present application, the query method further includes: Candidate questions similar to the questions in the question statement are screened out from a preset question library, and one or more candidate tools matching the candidate questions are screened out from preset tools.

[0154] Display candidate questions and links to detail pages of candidate tools on the same page as the processed query results.

[0155] In some embodiments of the present application, after analyzing the first target prompt word based on the natural language processing model to determine a target tool matching the first target prompt word from preset tools and determining the target input parameters required to be used by the target tool, the method further includes: Determine the data or data type that the user needs to query based on the question statement, and determine whether the tool required to be used in the question statement is a preset tool.

[0156] If the data or data type to be queried does not belong to the preset data range, and / or the tool to be used does not belong to the preset tool, the user is prompted that the current authority is insufficient.

[0157] In some embodiments of the present application, the query method further includes: If the data or data type to be queried intersects with the preset data range, and the tool to be used is a preset tool, the target tool is executed based on the target input parameters to obtain the query result obtained after the target tool processes the target input parameters within the preset data range.

[0158] In some embodiments of the present application, the target mode includes: graphic or list display.

[0159] Accordingly, the target method for displaying the query results is determined according to the type of data output by the target tool. The query results are processed based on the target method, and the processed query results are displayed to the user, including: When the target tool outputs ranking data, the ranking display component is called and the query results are displayed to the user in the form of a preset ranking list in the ranking display component. and\or When the target tool outputs data of the type of trend data, the trend display component is called and the query results are displayed to the user in the form of a preset line chart in the trend display component. and\or When the type of the target tool output data is comparison data, the comparison display component is called, and the query result is displayed to the user in the comparison display component in the form of a preset bar chart.

[0160] The process of each module implementing its own function in the real estate business information query device provided in the embodiment of the present application can be specifically referred to in the aforementioned Figures 1 to 5 The description of the first embodiment is not repeated here.

[0161] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0162] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0163] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0164] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0165] In addition, in the description of the present specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish descriptions, and cannot be understood as indicating or suggesting relative importance. It should also be understood that although the terms "first", "second", etc. are used to describe various elements in some embodiments of the present application in the text, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, the first table can be named as the second table, and similarly, the second table can be named as the first table without departing from the scope of the various described embodiments. The first table and the second table are both tables, but they are not the same table.

[0166] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0167] The real estate business information query method provided in the embodiment of the present application can be applied to terminal devices such as mobile phones, tablet computers, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptop computers, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA), etc. The embodiment of the present application does not impose any restrictions on the specific type of the terminal device.

[0168] For example, the terminal device can be a station (STAION, ST) in a WLAN, a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA) device, a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a vehicle networking terminal, a computer, a laptop computer, a handheld communication device, a handheld computing device, a satellite wireless device, a wireless modem card, a TV set top box (STB), a customer premises equipment (CPE) and / or other devices for communicating on a wireless system and a next-generation communication system, such as a mobile terminal in a 5G network or a mobile terminal in a future evolved Public Land Mobile Network (PLMN) network, etc.

[0169] As an example but not limitation, when the terminal device is a wearable device, the wearable device can also be a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not just hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are fully functional, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0170] Figure 7Schematic diagram of the structure of a terminal device provided by an embodiment of the present application. Figure 7 As shown, the terminal device 7 of this embodiment includes: at least one processor 70 ( Figure 7 Only one is shown in the figure), a memory 71, wherein the memory 71 stores a computer program 72 that can be run on the processor 70. When the processor 70 executes the computer program 72, the steps in the above-mentioned query method for real estate business information are implemented, such as Figure 1 Alternatively, when the processor 70 executes the computer program 72, the functions of each module / unit in the above-mentioned device embodiments are realized, for example Figure 6 The functions of the modules 61 to 65 are shown.

[0171] The terminal device 7 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal device may include, but is not limited to, a processor 70 and a memory 71. Those skilled in the art will appreciate that Figure 7 It is only an example of the terminal device 7 and does not constitute a limitation on the terminal device 7. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include an input sending device, a network access device, a bus, etc.

[0172] The processor 70 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0173] In some embodiments, the memory 71 may be an internal storage unit of the terminal device 7, such as a hard disk or memory of the terminal device 7. The memory 71 may also be an external storage device of the terminal device 7, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 7. Further, the memory 71 may include both an internal storage unit and an external storage device of the terminal device 7. The memory 71 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program. The memory 71 may also be used to temporarily store data that has been sent or is to be sent.

[0174] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0175] An embodiment of the present application also provides a terminal device, which includes at least one memory, at least one processor, and a computer program stored in the at least one memory and executable on the at least one processor, wherein when the processor executes the computer program, the terminal device implements the steps in any of the above-mentioned method embodiments.

[0176] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0177] An embodiment of the present application provides a computer program product. When the computer program product runs on a terminal device, the terminal device can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0178] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0179] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0180] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0181] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0182] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for querying real estate business information, characterized in that: include: Receiving a question sentence about real estate business information input by a user in natural language; According to the role information of the user, a plurality of preset tools associated with the user are matched from a preset tool library, and a preset data range associated with the user is determined; Generate a model prompt word for a natural language processing model according to the tool information of each of the preset tools, the question statement, and the historical conversation information of multiple rounds with the user, to obtain a first target prompt word; Analyzing the first target prompt word based on the natural language processing model to determine a target tool matching the first target prompt word from the preset tools, and determining target input parameters required to be used by the target tool; The target tool is executed based on the target input parameters to obtain a query result after the target tool processes the target input parameters within a preset data range.

2. The method for querying real estate business information according to claim 1, characterized in that: The tool information includes: the tool name and attribute information of the input parameters; Accordingly, the generating of model prompt words for the natural language processing model according to the tool information of each of the preset tools, the question statement, and the historical multiple rounds of dialogue information with the user to obtain the first target prompt word includes: Generate text description statements for all the preset tools according to the tool names of the preset tools and the attribute information of the input parameters; Based on the text description sentences of all the preset tools, the question sentences and multiple rounds of dialogue information with the user, model prompt words are generated for the natural language processing model to obtain the first target prompt words.

3. The method for querying real estate business information according to claim 1, characterized in that: After obtaining the query result, the query method further includes: Determining a target method for displaying the query results according to the type of output data of the target tool; The query result is processed based on the target method, and the processed query result is displayed to the user; wherein different types of tools correspond to different target methods.

4. The method for querying real estate business information according to claim 3, characterized in that: The target methods include: text display; Accordingly, determining a target method for displaying the query result according to the type of data output by the target tool; processing the query result based on the target method, and displaying the processed query result to the user, includes: When the type of the output data of the target tool is indicator data, generating prompt words of the natural language processing model on the question statement and the query result to obtain a second target prompt word; The second target prompt word is analyzed based on the natural language processing model, the natural language response text for the second target prompt word is generated, and the natural language response text is displayed to the user.

5. The method for querying real estate business information according to any one of claims 1 to 4, characterized in that: The query method further includes: Selecting candidate questions similar to the questions in the question statement from a preset question library, and selecting one or more candidate tools matching the candidate questions from the preset tools; The candidate questions and detail page links to the candidate tools are displayed on the same page as the processed query results.

6. The method for querying real estate business information according to any one of claims 1 to 4, characterized in that: After the operation of analyzing the first target prompt word based on the natural language processing model to determine a target tool matching the first target prompt word from the preset tools and determining a target input parameter required to be used by the target tool, the method further includes: Determine the data or data type that the user needs to query according to the question statement, and determine whether the tool required to be used in the question statement belongs to the preset tool; If the data or data type to be queried does not belong to the preset data range, and / or the tool to be used does not belong to the preset tool, the user is prompted that the current authority is insufficient.

7. The method for querying real estate business information according to claim 6, characterized in that: The query method further includes: If the data or data type to be queried intersects with the preset data range, and the tool to be used belongs to the preset tool, then the operation of executing the target tool based on the target input parameters to obtain the query result obtained by the target tool after processing the target input parameters within the preset data range.

8. The method for querying real estate business information according to claim 3 or 4, characterized in that: The target modes include: graphical or list display; Accordingly, determining a target method for displaying the query result according to the type of data output by the target tool; processing the query result based on the target method, and displaying the processed query result to the user, includes: When the type of data output by the target tool is ranking data, calling a ranking display component, and displaying the query result to the user in the ranking display component in the form of a preset ranking list; and\or When the type of the output data of the target tool is trend data, calling a trend display component, and displaying the query result to the user in the trend display component in a preset line chart form; and\or When the type of the output data of the target tool is comparison data, a comparison display component is called, and the query result is displayed to the user in the comparison display component in a preset bar chart format.

9. A device for querying real estate business information, characterized in that: include: A receiving module, used to receive a question sentence about real estate business information input by a user in natural language; A determination module, configured to match a plurality of preset tools associated with the user from a preset tool library according to the role information of the user, and determine a preset data range associated with the user; A prompt word generation module, configured to generate model prompt words for a natural language processing model according to tool information of each of the preset tools, the question statements, and information about multiple rounds of historical conversations with the user, to obtain a first target prompt word; An analysis module, configured to analyze the first target prompt word based on the natural language processing model, so as to determine a target tool matching the first target prompt word from the preset tools, and determine a target input parameter required to be used by the target tool; The query module is used to execute the target tool based on the target input parameters to obtain the query results of the target tool after processing the target input parameters within a preset data range.

10. A terminal device, characterized in that: The terminal device includes a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and the processor implements the steps of the method according to any one of claims 1 to 8 when executing the computer program.

Images