Mind map generation method and device, electronic equipment and readable storage medium

Abstract

The invention provides a mind map generation method and device, electronic equipment and a readable storage medium, and relates to the technical field of document processing.The mind map generation method comprises the steps that target document content and a target mind map format are obtained; performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; and generating a mind map corresponding to the target document content based on the document key point data. According to the method, people can be helped to quickly understand document logics and key points, the reading efficiency and experience are improved, and the mind mapping accuracy can be improved.

Description

Technical Field

[0001] The present invention relates to the field of document processing technology, and in particular to a mind map generation method, device, electronic device and readable storage medium. Background Art

[0002] With the rapid development of computer technology, various document applications have emerged, and document processing technology has also risen rapidly. For example, in order to help users quickly understand the content of a document, document processing technology is used to process the document content to generate a document directory or summary.

[0003] However, since the document directory and summary are mainly a pure text representation method, the form is relatively simple and lacks divergent thinking in helping users understand the document content. Therefore, an effective solution is urgently needed to solve the above problems. Summary of the invention

[0004] In view of the problems existing in the prior art, embodiments of the present invention provide a mind map generating method, device, electronic device and readable storage medium.

[0005] The present invention provides a mind map generating method, comprising:

[0006] Get the target document content and target mind map format;

[0007] Performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format;

[0008] Based on the document key data, a mind map corresponding to the target document content is generated.

[0009] According to a mind map generation method provided by the present invention, generating a mind map corresponding to the content of the target document based on the document key data includes:

[0010] Determining the hierarchical relationship of each document key point in the document key point data according to the node structure attribute corresponding to the target mind map format;

[0011] A mind map corresponding to the content of the target document is generated according to the hierarchical relationship and the key points of each document.

[0012] According to a mind map generation method provided by the present invention, the step of extracting key points from the target document content to obtain key point data of the document corresponding to the target mind map format includes:

[0013] The target document content is input into a preset language processing model to perform document key point extraction processing to obtain document key point data corresponding to the target mind map format.

[0014] According to a mind map generation method provided by the present invention, the language processing model is a large-scale language model;

[0015] Before inputting the target document content into a preset language processing model to extract document key points and obtain document key point data corresponding to the target mind map format, the method further includes:

[0016] Get key extraction instructions;

[0017] The step of inputting the target document content into a preset language processing model to extract document key points and obtain document key point data corresponding to the target mind map format includes:

[0018] The key point extraction instruction and the target document content are input into the large-scale language model, and the large-scale language model is instructed to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format.

[0019] According to a mind map generation method provided by the present invention, the step of obtaining target document content and target mind map format includes:

[0020] receiving a content selection instruction, the content selection instruction carrying a target document identifier and a selection area; determining the target document content according to the target document identifier and the selection area;

[0021] The set mind map format example is input into the large-scale language model for format learning to determine the target mind map format.

[0022] According to a mind map generation method provided by the present invention, the step of inputting a set mind map format example into the large-scale language model for format learning and determining the target mind map format further includes:

[0023] receiving a format selection instruction, wherein the format selection instruction carries the mind map format example;

[0024] Generate a prompt instruction, wherein the prompt instruction includes the key point extraction instruction, the mind map format example and the target mind map format, and the prompt instruction is used to instruct the large-scale language model to perform a document key point extraction task.

[0025] According to a mind map generation method provided by the present invention, before generating a mind map corresponding to the target document content based on the document key point data, the method further includes:

[0026] Performing format detection on the key data of the document to obtain the degree of conformity of the format of the key data of the document;

[0027] The step of generating a mind map corresponding to the target document content based on the document key point data includes:

[0028] When the format compliance degree is format framework compliance, a mind map corresponding to the target document content is generated based on the document key point data.

[0029] According to a mind map generation method provided by the present invention, after performing format detection on the document key data and obtaining the format conformity degree of the document key data, the method further comprises:

[0030] When the format conformance degree is that the format framework does not conform, the prompt instruction is input into the large-scale language model to instruct the large-scale language model to re-extract document key points from the target document.

[0031] According to a mind map generation method provided by the present invention, when the format conformity degree is format framework conformity, generating a mind map corresponding to the content of the target document based on the document key point data includes:

[0032] When the format conformity degree is format framework conformity, each document key point in the document key point data is parsed according to the node structure attribute corresponding to the target mind map format to determine the hierarchical relationship between each document key point;

[0033] If there is an abnormal document key point in each of the document key points that does not match the node structure attribute, then the parsing of the abnormal document key point is skipped;

[0034] A mind map corresponding to the target document content is generated according to the hierarchical relationship and the document main points except the abnormal document main points.

[0035] According to a mind map generation method provided by the present invention, before generating a mind map corresponding to the content of the target document according to the hierarchical relationship and each of the document key points other than the abnormal document key points, the method further comprises:

[0036] Counting the number of key points of the abnormal document;

[0037] When the number is greater than or equal to a set threshold, the prompt instruction is input into the large-scale language model to instruct the large-scale language model to re-extract key points from the target document.

[0038] The present invention also provides a mind map generating device, comprising:

[0039] An acquisition module, configured to acquire target document content and target mind map format;

[0040] A document key point extraction module is configured to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format;

[0041] The generating module is configured to generate a mind map corresponding to the content of the target document based on the key points data of the document.

[0042] According to a mind map generation device provided by the present invention, the document key point extraction module includes a hierarchical relationship determination module and a mind map generation module;

[0043] The hierarchical relationship determination module is configured to determine the hierarchical relationship of each document key point in the document key point data according to the node structure attribute corresponding to the target mind map format;

[0044] The mind map generation module is configured to generate a mind map corresponding to the content of the target document according to the hierarchical relationship and each of the document key points.

[0045] According to a mind map generating device provided by the present invention, the document key point extraction module is further configured as follows:

[0046] The target document content is input into a preset language processing model to perform document key point extraction processing to obtain document key point data corresponding to the target mind map format.

[0047] According to a mind map generation device provided by the present invention, the language processing model is a large-scale language model;

[0048] The mind map generating device further includes a first receiving module;

[0049] The first receiving module is configured to obtain a key point extraction instruction;

[0050] The document key point extraction module is further configured to:

[0051] The key point extraction instruction and the target document content are input into the large-scale language model, and the large-scale language model is instructed to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format.

[0052] According to a mind map generation device provided by the present invention, the acquisition module includes a second receiving module, a target document content determination module and a target mind map format determination module;

[0053] The second receiving module is configured to receive a content selection instruction, wherein the content selection instruction carries a target document identifier and a selection area;

[0054] The target document content determination module is configured to determine the target document content according to the target document identifier and the selected area;

[0055] The target mind map format determination module is configured to input a set mind map format example into the large-scale language model for format learning to determine the target mind map format.

[0056] According to a mind map generation device provided by the present invention, the mind map generation device further includes a third receiving module and a prompt instruction generation module;

[0057] The third receiving module is configured to receive a format selection instruction, wherein the format selection instruction carries the mind map format example;

[0058] The prompt instruction generation module is configured to generate a prompt instruction, wherein the prompt instruction includes the key point extraction instruction, the mind map format example and the target mind map format, and the prompt instruction is used to instruct the large-scale language model to perform a document key point extraction task.

[0059] According to a mind map generation device provided by the present invention, the mind map generation device further includes a format detection module;

[0060] The format detection module is configured to perform format detection on the document key data to obtain the format conformity degree of the document key data;

[0061] The generation module is further configured to:

[0062] When the format compliance degree is format framework compliance, a mind map corresponding to the target document content is generated based on the document key point data.

[0063] According to a mind map generation device provided by the present invention, the mind map generation device further includes a first indication module;

[0064] The first instruction module is configured to input the prompt instruction to the large-scale language model when the format compliance level is that the format framework does not comply with the format framework, and instruct the large-scale language model to re-extract the key points of the target document.

[0065] According to a mind map generation device provided by the present invention, the generation module includes a hierarchical relationship determination module, a skip module and a mind map generation module;

[0066] The hierarchical relationship determination module is configured to parse each document key point in the document key point data according to the node structure attribute corresponding to the target mind map format when the format conformity degree is format framework conformity, and determine the hierarchical relationship of each document key point;

[0067] The skipping module is configured to skip parsing of an abnormal document key point if there is an abnormal document key point in each of the document key points that does not match the node structure attribute;

[0068] The mind map generation module is configured to generate a mind map corresponding to the target document content according to the hierarchical relationship and each of the document key points except the abnormal document key point.

[0069] According to a mind map generation device provided by the present invention, the generation module further includes a statistical module;

[0070] The statistical module is configured to count the number of key points of the abnormal document;

[0071] The mind map generating device further includes a second indication module;

[0072] The second instruction module is configured to input the prompt instruction to the large-scale language model when the number is greater than or equal to a set threshold, and instruct the large-scale language model to re-extract key points from the target document.

[0073] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a mind map generating method as described above is implemented.

[0074] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the mind map generating method as described in any one of the above is implemented.

[0075] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the mind map generating method as described above is implemented.

[0076] The present invention provides a mind map generation and device, including obtaining target document content and target mind map format; performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; and generating a mind map corresponding to the target document content based on the document key point data. By quickly understanding the target document content to extract document key points, obtaining document key point data corresponding to the target mind map format, and generating a mind map based on the document key point data, people can quickly understand document logic and key points, improve reading efficiency and experience, and also improve the accuracy of the mind map. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0078] Figure 1 It is one of the flowcharts of the mind map generation method provided by the present invention;

[0079] Figure 2 This is the second flow chart of the mind map generation method provided by the present invention;

[0080] Figure 3 It is a structural schematic diagram of a mind map generating device provided by the present invention;

[0081] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0082] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0083] In order to facilitate a clearer understanding of the embodiments of the present invention, some relevant background knowledge is first introduced as follows.

[0084] With the rapid development of computer technology, various document applications have emerged, such as text documents, presentation documents, and image documents. Users generally understand the content of documents through the document directory or summary. For some documents without a document directory and summary, users need to view the entire document content to understand it.

[0085] In order to help users quickly understand the content of a document without a document directory and a summary, the prior art uses document processing technology to process the document content to generate a document directory or a summary.

[0086] However, since the document directory and summary are mainly a pure text representation method, the form is relatively simple and lacks divergent thinking in helping users understand the document content, and there is no way to quickly generate a graphical representation method.

[0087] Mind mapping can express the relationship between the key points of each document by combining text and graphics in a hierarchical diagram of mutual subordination and correlation. It is an effective graphic thinking tool for expressing divergent thinking and is widely used in people's daily work, study and life.

[0088] Based on this, the present invention provides a mind map generation method, device, electronic device and readable storage medium, which obtains the target document content and the target mind map format; extracts the key points of the target document content to obtain the document key point data corresponding to the target mind map format; and generates a mind map corresponding to the target document content based on the document key point data. By quickly understanding the target document content and extracting the document key points, obtaining the document key point data corresponding to the target mind map format, and generating a mind map based on the document key point data, it can help people quickly understand the document logic and key points, improve reading efficiency and experience, and also improve the accuracy of the mind map.

[0089] Combine the following Figure 1-Figure 3 The present invention describes a mind map generation method, device, electronic device and readable storage medium.

[0090] Figure 1 This is one of the flowcharts of the mind map generation method provided by the present invention, see Figure 1 As shown, it includes steps 101 to 103, wherein:

[0091] Step 101: Obtain target document content and target mind map format.

[0092] First of all, it should be noted that the execution subject of the present invention can be any electronic device that generates a mind map, for example, it can be any one of a smart phone, a smart watch, a desktop computer, a laptop computer, etc.

[0093] Specifically, the target document content refers to the document content for which a mind map needs to be generated. The mind map format refers to the format of the data used in the generated mind map, such as the PlantUML mind map format, the FreeMind mind map format, and the Mermaid mind map format. The target mind map format refers to a pre-specified mind map format.

[0094] In practical applications, the target document content may be part of the document content of a document, may be the entire content of a document, or may be the document content formed by splicing part or all of the document contents of multiple documents, which is not limited in any way by the present invention.

[0095] There are many ways to obtain the target document content. For example, a user uploads the target document content through the upload page provided by the mind map generation platform, and accordingly, the execution subject obtains the target document content. For another example, the execution subject receives a target document content acquisition instruction or a mind map generation instruction, and accordingly, the execution subject obtains the target document content from the storage area pointed to by the target document content acquisition instruction or the mind map generation instruction. The present invention is not limited to this.

[0096] There are also many ways to obtain the target mind map format. For example, the mind map generation platform displays multiple initial mind map formats to the user through a display, and the user selects the target mind map format from the multiple initial mind map formats and confirms the selection. Accordingly, the execution subject obtains the target mind map format. For another example, if the target mind map format indicated or selected by the user is not received, the execution subject determines the default mind map format of the mind map generation platform as the target mind map format. For another example, the user uploads the target mind map format or the identifier of the target mind map format through the upload page provided by the mind map generation platform, and accordingly, the execution subject obtains the target mind map format. The present invention is not limited to this.

[0097] Step 102: extract key points from the target document content to obtain key point data of the document corresponding to the target mind map format.

[0098] Specifically, the document key points refer to the key points and / or highlights of the document content, mainly indicating the main content of the document. The document key point data refers to a collection of document key points arranged in a target mind map format.

[0099] In practical applications, the target document content can be parsed, the document key points can be extracted from the parsed results, and the extracted document key points can be organized according to the target mind map format to obtain document key point data corresponding to the target document content and in the target mind map format.

[0100] Step 103: Generate a mind map corresponding to the target document content based on the document key data.

[0101] In actual applications, on the basis of obtaining the document key points data corresponding to the target mind map format, further: a mind map structure can be constructed according to the arrangement of each document key point in the document key points data corresponding to the target mind map format, and each document key point can be added to the corresponding node in the mind map structure in turn to generate a mind map corresponding to the target document content; the document key points data corresponding to the target mind map format can also be input into the mind map generation model corresponding to the target mind map format to generate a mind map corresponding to the target document content.

[0102] The present invention provides a mind map generation method, including obtaining target document content and target mind map format; performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; and generating a mind map corresponding to the target document content based on the document key point data. By quickly understanding the target document content to extract document key points, obtaining document key point data corresponding to the target mind map format, and generating a mind map based on the document key point data, people can quickly understand document logic and key points, improve reading efficiency and experience, and also improve the accuracy of mind maps.

[0103] In one or more optional embodiments of the present invention, the hierarchical relationship of each document key point in the document key point data corresponding to the target mind map format can be determined, and a mind map corresponding to the target document content is generated according to the hierarchical relationship and the document key points. That is, the mind map corresponding to the target document content is generated based on the document key point data, and the specific implementation process can be as follows:

[0104] Determining the hierarchical relationship of each document key point in the document key point data according to the node structure attribute corresponding to the target mind map format;

[0105] A mind map corresponding to the content of the target document is generated according to the hierarchical relationship and the key points of each document.

[0106] Specifically, node structure attributes refer to the attributes that characterize nodes at different levels, such as setting different identifiers for nodes at different levels. For example, the node structure attributes corresponding to the PlantUML mind map format are: "*" (a single *) represents the root node, "**" (two connected *) represents the second-level node, and "***" (three connected *) represents the third-level node; another example is setting different indentation values ​​for nodes at different levels. Hierarchical relationship refers to the parent-child relationship between document points, that is, the subordinate relationship.

[0107] In practical applications, the node structure attributes corresponding to the target mind map format can be obtained, and then the node levels of each document key point in the document key point data can be identified based on the node structure attributes, and the hierarchical relationship between each document key point can be determined based on the arrangement of each document key point in the document key point data and the node level of each document key point. Furthermore, a mind map structure is generated based on the hierarchical relationship, and each document key point is added to the node corresponding to the mind map structure, thereby obtaining a mind map corresponding to the target document content. In this way, the accuracy and speed of the mind map can be improved.

[0108] For example, the document key data corresponding to the PlantUML mind map format is as follows:

[0109] @startmindmap

[0110] *Java

[0111] **Java Language

[0112] ***Write once, run anywhere

[0113] ***Java applications

[0114] **Development and Licensing

[0115] ***Java Virtual Machine

[0116] ***OpenJDK JVM

[0117] @endmindmap

[0118] Among them, @startmindmap is the start mark of the mind map; @endmindmap is the end mark of the mind map; * is a node, that is, the key points of the document.

[0119] The node structure attributes corresponding to the PlantUML mind map format are: 1 * is the root node, 2 * are the second-level nodes, 3 * are the third-level nodes, and so on.

[0120] By parsing the document key data according to the node structure attributes, it can be seen that: the document key Java is the root node or the first-level node, that is, the node level of the document key Java is first-level; the document key Java Language and the document key Development and Licensing are second-level nodes, that is, the node level of these two document key points is second-level; the document key Write once, run anywhere, the document key Java applications, the document key Java VirtualMachine and the document key OpenJDK JVM are third-level nodes, so the node level of these four document key points is third-level.

[0121] Since the third-level nodes Write once, run anywhere and Java applications are arranged after the second-level node Java Language and before the second-level node Development and Licensing, the third-level nodes Write once, run anywhere and Java applications are child nodes of the second-level node Java Language, that is, the second-level node Java Language is the parent node of the third-level nodes Write once, run anywhere and Java applications; since the third-level nodes Java Virtual Machine and OpenJDK JVM are arranged after the second-level node Development and Licensing, the third-level nodes Java Virtual Machine and OpenJDK JVM are child nodes of the second-level node Development and Licensing, that is, the second-level node Development and Licensing is the parent node of the third-level nodes Java Virtual Machine and OpenJDK JVM. The second-level nodes Java Language and Development and Licensing are child nodes of the root node Java, that is, the root node Java is the parent node of the second-level nodes JavaLanguage and Development and Licensing.

[0122] Based on the hierarchical relationship of the above-mentioned document key points and the key points of each document, a mind map corresponding to the content of the target document is generated.

[0123] In one or more optional embodiments of the present invention, the extracting key points of the target document content to obtain the key point data of the document corresponding to the target mind map format includes:

[0124] The target document content is input into a preset language processing model to perform document key point extraction processing to obtain document key point data corresponding to the target mind map format.

[0125] Specifically, the language processing model refers to a model for processing language and text, which may be a language processing model based on artificial intelligence, a model based on natural language processing technology, or other models, and the present invention does not impose any limitation on this.

[0126] In this way, by extracting key points of a document through a preset language processing model, the efficiency of extracting key points of a document and the accuracy of the key point data of the document can be improved.

[0127] In an achievable embodiment of the present invention, the language processing model is a model based on natural language processing, i.e., a natural language processing model. The target document content is input into a preset language processing model to perform document key point extraction processing to obtain document key point data corresponding to the target mind map format. The specific implementation process can be as follows:

[0128] Inputting the target document content into a content understanding unit of a preset natural language processing model for content understanding processing, obtaining semantic features of the target document content and semantic features of each content block in the target document content, wherein the content block is a paragraph and / or chapter; wherein content understanding, also known as semantic analysis, refers to the process of understanding the document content or content block and identifying its expression meaning (semantics);

[0129] Input each of the semantic features into the semantic association unit of the natural language processing model for semantic association processing, and obtain a first knowledge graph composed of each of the semantic features; wherein semantic association refers to the process of identifying the semantic relationship between semantic features; semantic relationship can be an alternative vertical relationship established based on the comparison of semantic features, including synonymy, antonym, similar meaning and heteronym, etc.; semantic relationship can also be a relationship formed by the mutual matching of semantic features in the language system and the speech flow, including giving and receiving relationship, possessive relationship, limiting relationship, parallel relationship, dominant relationship, judgment relationship and explanation supplement relationship, etc.; the first knowledge graph refers to a knowledge graph with semantic features as nodes and semantic relationships as edges;

[0130] The first knowledge graph is input into the text conversion unit of the natural language processing model, and each semantic feature in the first knowledge graph is converted into text by the text conversion unit (such as encoding the semantic features to obtain the text), so as to obtain a second knowledge graph composed of each of the texts; wherein the second knowledge graph refers to a knowledge graph with texts corresponding to the semantic features as nodes and semantic relationships as edges;

[0131] Inputting the second knowledge graph into the key point extraction unit of the natural language processing model, extracting the document key points of each text in the second knowledge graph through the text conversion unit, and obtaining a third knowledge graph composed of the document key points; wherein the third knowledge graph refers to a knowledge graph with document key points as nodes and semantic relationships as edges;

[0132] The third knowledge graph is input into the format conversion unit of the natural language processing model, and the format conversion unit performs format conversion processing on the third knowledge graph according to the target mind map format to obtain document key point data corresponding to the target mind map format. In this way, the accuracy of the document key point data can be further improved through natural language processing technology.

[0133] It should be noted that the document highlights of each text in the second knowledge graph are extracted by the text conversion unit. The specific implementation process may be: for each text, the text conversion unit determines the core content corresponding to the text according to the content of the text and the semantic relationship between the text and other texts, and uses the core content as the document highlights of the text, or uses the keywords corresponding to the core content as the document highlights of the text.

[0134] In addition, the format conversion unit performs format conversion processing on the third knowledge graph according to the target mind map format. The specific implementation process may be: according to each node (document key points) and each edge (semantic relationship) in the third knowledge graph, determine the node level of each node and its corresponding associated nodes, wherein the node levels include first level, second level, third level or above, etc.; associated nodes include child nodes and / or parent nodes; according to the node level of each node and its corresponding associated nodes, arrange each node according to the target mind map format to obtain the document key points data corresponding to the target mind map format.

[0135] In another achievable embodiment of the present invention, the language processing model is a large-scale language model; before inputting the target document content into a preset language processing model for document key point extraction processing to obtain document key point data corresponding to the target mind map format, the method further includes:

[0136] Get key extraction instructions;

[0137] The step of inputting the target document content into a preset language processing model to extract document key points and obtain document key point data corresponding to the target mind map format includes:

[0138] The key point extraction instruction and the target document content are input into the large-scale language model, and the large-scale language model is instructed to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format.

[0139] Specifically, the large-scale language model is a language processing model based on artificial intelligence, that is, an artificial intelligence (AI) large model. The key point extraction instruction is used to define the role of the large-scale language model as extracting key points, that is, to inform the large-scale language model that it needs to perform the task of document key point extraction.

[0140] In practical applications, since large-scale language models can perform multiple tasks, in order to avoid the situation where the tasks performed by large-scale language models are different from those required, resulting in unusable output results of large-scale language models, or large-scale language models perform multiple tasks at the same time, and there are many useless results in the output results, resulting in waste of resources, it is necessary to indicate the tasks that large-scale language models need to perform through key point extraction instructions. In this way, the efficiency of document key point data can be improved while avoiding waste of resources.

[0141] There are many ways to obtain key point extraction instructions. The user selects the task to be executed, that is, the key point extraction task, through the model task processing page provided by the mind map generation platform, and accordingly, the execution subject obtains the key point extraction instruction. For another example, the default execution task of the large-scale language model in the mind map generation platform is the key point extraction task. When the mind map generation instruction is started, the key point extraction instruction is automatically triggered, and accordingly, the execution subject obtains the key point extraction instruction. The present invention is not limited to this.

[0142] When the key point extraction instruction is obtained, the key point extraction instruction is input into the large-scale language model to inform the large-scale language model that the key point extraction task needs to be performed, that is, the large-scale language model is instructed to perform the key point extraction task, and the target document content is input into the large-scale language model to start the large-scale language model to perform the key point extraction task on the target document content, that is, to perform document key point extraction processing to obtain document key point data corresponding to the target mind map format.

[0143] In one or more optional embodiments of the present invention, the obtaining of the target document content and the target mind map format includes: obtaining the target document content, and obtaining the target mind map format.

[0144] The specific implementation process of obtaining the target document content may be as follows:

[0145] receiving a content selection instruction, wherein the content selection instruction carries a target document identifier and a selection area;

[0146] The target document content is determined according to the target document identifier and the selected area.

[0147] Specifically, the content selection instruction refers to an instruction for selecting the content of the target document. The target document identifier refers to an identifier corresponding to the selected document. The selection area refers to the location of the selected content in the document, which can be a coordinate, a page, or a paragraph number.

[0148] In actual applications, when viewing a target document, a user can choose to generate a mind map for part or all of the document content of the target document. When the user selects the document content for which a mind map needs to be generated by selecting an area in the target document and confirms it, the execution subject receives a content selection instruction carrying the target document identifier and the selection area.

[0149] Furthermore, the execution subject determines the target document through the target document identifier, and performs document content extraction processing at a position corresponding to the selected area in the target document, thereby obtaining the target document content. In this way, the corresponding target document content can be extracted based on user needs, thereby improving user satisfaction.

[0150] It should be noted that there can be only one or more target documents. When there are multiple target documents, there will also be multiple target document identifiers. For each target document, it can correspond to the same content selection instruction, that is, the content selection instruction carries the target document identifiers. At this time, in order to avoid confusing the selection area and causing errors in the extraction of the target document content, each selection area is associated with its corresponding target document identifier, that is, each selection area corresponds to a target document identifier, and each target document identifier corresponds to at least one selection area. It is also possible for one target document to correspond to one content selection instruction, and the content selection instruction carries the target document identifier and selection area of ​​the target document. After extracting the document content corresponding to the selection area in each target document, all the document contents constitute the target document content.

[0151] The specific implementation process of obtaining the target mind map format can be as follows:

[0152] The set mind map format example is input into the large-scale language model for format learning to determine the target mind map format.

[0153] Specifically, the mind map format examples refer to cases and examples of the target mind map format that are listed and displayed for reference.

[0154] In practical applications, the set mind map format example can be input into the large-scale language model for format learning, thereby determining the target mind map format. In this way, by inputting the mind map format example into the large-scale language model, the large-scale language model learns the mind map format example, which is conducive to the matching degree between the key data of the document output by the large-scale language model and the target mind map format.

[0155] For example, the PlantUML mind map format is as follows:

[0156] @startmindmap

[0157] *Debian

[0158] **Ubuntu

[0159] ***Linux Mint

[0160] ***Kubuntu

[0161] **LMDE

[0162] **SolydXK

[0163] SteamOS

[0164] **Raspbian with a very long name

[0165] ***Raspmbc

[0166] ***Raspyfi

[0167] @endmindmap

[0168] For example, the FreeMind mind map format is as follows:

[0169]

[0170]

[0171] For example, the Mermaid mind map format is as follows:

[0172]

[0173] In one or more optional embodiments of the present invention, the set mind map format example is a default mind map format example.

[0174] In one or more optional embodiments of the present invention, the set mind map format example is a mind map format example selected by a user, that is, the set mind map format example is input into the large-scale language model for format learning, and before determining the target mind map format, the method further includes:

[0175] A format selection instruction is received, wherein the format selection instruction carries the mind map format example.

[0176] Specifically, the format selection instruction refers to an instruction for selecting a mind map format example.

[0177] In actual applications, operators or customers can set the format of document key data, that is, by selecting one of the mind map format examples on a page displaying multiple different mind map format examples, and accordingly, the execution subject receives a format selection instruction carrying the mind map format example, so that the mind map format example can be input into the large-scale language model for format learning. In this way, the accuracy of determining the mind map format example and user stickiness can be improved.

[0178] In one or more optional embodiments of the present invention, after receiving the format selection instruction, the method further includes:

[0179] Generate a prompt instruction, wherein the prompt instruction includes the key point extraction instruction, the mind map format example and the target mind map format, and the prompt instruction is used to instruct the large-scale language model to perform a document key point extraction task.

[0180] Specifically, the prompt instruction is an instruction used to specify a large-scale language model to perform a document key point extraction task, such as the prompt instruction.

[0181] In practical applications, a prompt instruction can be set, where the prompt instruction includes key point extraction instructions, a mind map format example, and a target mind map format. The prompt instruction is input into the large-scale language model to instruct the large-scale language model to perform the document key point extraction task: that is, the key point extraction instruction in the prompt instruction is used to inform the large-scale language model that it is mainly a document reading expert and extracts the key points of the document, that is, it needs to perform the key point extraction task; the mind map format example in the prompt instruction is used to stipulate that the large-scale language model will output the document key point data according to the target mind map format requirements after learning the mind map format example; the target mind map format is used for the large-scale language model to perform the key point extraction task and output the document key point data corresponding to the target mind map format. In this way, the prompt instruction is used to control and guide the large-scale language model to the target document model.

[0182] In one or more optional embodiments of the present invention, before generating a mind map corresponding to the target document content based on the document key point data, the method further includes:

[0183] Performing format detection on the key data of the document to obtain the degree of conformity of the format of the key data of the document;

[0184] The step of generating a mind map corresponding to the target document content based on the document key point data includes:

[0185] When the format compliance degree is format framework compliance, a mind map corresponding to the target document content is generated based on the document key point data.

[0186] Specifically, format detection refers to detecting the format of the key data of the document. The degree of format conformity refers to the degree to which the key data of the document conforms to the format of the target mind map.

[0187] In actual applications, although the document key data is obtained in accordance with the target mind map format, some errors or mistakes are inevitable, resulting in the document key data not being output strictly in accordance with the target mind map format. In order to avoid errors in the generated mind map due to format errors, that is, to improve the accuracy of the mind map, after obtaining the document key data, it is necessary to perform format detection on the document key data, that is, to compare the actual format of the document key data with the target mind map format to obtain the format conformity of the document key data.

[0188] Among them, the degree of format compliance includes format framework compliance and format framework non-compliance. Format framework compliance means that the actual format of the document key data matches the target mind map format in the framework structure; format framework non-compliance means that the actual format of the document key data does not match the target mind map format in the framework structure, or the document key data is garbled.

[0189] Exemplarily, the framework structure of the target mind map format is "@startmindmap*@endmindmap". If the actual format of the document key point data starts with "@startmindmap", uses "*" to represent the document key point, and ends with "@endmindmap", then the format conformity of the document key point data is format framework conformity. If the actual format of the document key point data does not contain any of "@startmindmap", "*" and "@endmindmap", then the format conformity of the document key point data is format framework non-conformity.

[0190] When the format conformity is in accordance with the format framework, it means that the actual format of the document key data is the target mind map format. At this time, the mind map corresponding to the target document content can be generated based on the document key data. In this way, errors in the generated mind map caused by format errors can be avoided, that is, the accuracy of the mind map can be improved.

[0191] In one or more optional embodiments of the present invention, after performing format detection on the document key data to obtain the format conformity degree of the document key data, the method further includes:

[0192] When the format conformity level is that the format framework does not conform, the document key points are extracted from the target document content again to obtain document key point data corresponding to the target mind map format.

[0193] In actual applications, if the format conformance degree is that the format framework does not conform, it means that the document key data is basically not in the target mind map format, that is, the actual format of the document key data is not the target mind map format. Therefore, it is necessary to re-extract the document key points of the target document content to obtain the document key points data corresponding to the target mind map format, such as inputting the prompt instruction into the large-scale language model to instruct the large-scale language model to re-extract the document key points of the target document. In this way, errors in the generated mind map caused by format errors can be avoided, that is, the accuracy of the mind map can be improved.

[0194] In one or more optional embodiments of the present invention, when the format conformity degree is format framework conformity, a mind map corresponding to the content of the target document is generated based on the document key data, and the specific implementation process may be as follows:

[0195] When the format conformity degree is format framework conformity, each document key point in the document key point data is parsed according to the node structure attribute corresponding to the target mind map format to determine the hierarchical relationship between each document key point;

[0196] If there is an abnormal document key point in each of the document key points that does not match the node structure attribute, then the parsing of the abnormal document key point is skipped;

[0197] A mind map corresponding to the target document content is generated according to the hierarchical relationship and the document main points except the abnormal document main points.

[0198] In actual applications, if the format conformance degree is format framework conformance, the node structure attributes corresponding to the target mind map format can be obtained, and then the node level of each document point in the document point data can be identified based on the node structure attributes, and the hierarchical relationship between each document point can be determined based on the arrangement of each document point in the document point data and the node level of each document point. If there are abnormal document points in each document point, that is, document points that do not match the node structure attributes, there is no need to parse the abnormal document point, and the abnormal document point can be skipped.

[0199] Furthermore, a mind map structure is generated based on the hierarchical relationship, and the document points other than the abnormal document points are added to the nodes corresponding to the mind map structure, thereby obtaining a mind map corresponding to the target document content. In this way, the abnormal document points with problems are eliminated, and a mind map is generated based on the remaining normal document points, which can improve the accuracy of the mind map.

[0200] In one or more optional embodiments of the present invention, before generating a mind map corresponding to the target document content according to the hierarchical relationship and each of the document key points other than the abnormal document key points, the method further includes:

[0201] Counting the number of key points of the abnormal document;

[0202] When the number is greater than or equal to the set threshold, the target document content is re-processed for document key point extraction to obtain document key point data corresponding to the target mind map format.

[0203] Specifically, the threshold is set to a preset value, which is used to measure whether the number of abnormal document key points in the document key point data meets the requirements.

[0204] In actual applications, each time an abnormal document key point is skipped, the number of abnormal document key points is incremented by one. If after traversing all document key points in the document key point data, the number of abnormal document key points is not greater than or equal to the set threshold, it means that the completeness corresponding to the normal document key points in the document key point data meets the completeness requirements, and a mind map can be generated, that is, the step of generating a mind map corresponding to the target document content according to the hierarchical relationship and each of the document key points except the abnormal document key points is executed. If the number of abnormal document key points reaches the set threshold, it means that the completeness corresponding to the normal document key points in the document key point data does not meet the completeness requirements, and it is necessary to re-extract the document key points of the target document content to obtain the document key point data corresponding to the target mind map format, such as inputting the prompt instruction into the large-scale language model to instruct the large-scale language model to re-extract the document key points of the target document. In this way, the adaptation of the mind map to the target document content can be guaranteed, and the important content in the target document content can be avoided from being missed by the mind map, that is, the accuracy of the mind map can be improved.

[0205] It should be noted that the number of abnormal document key points can be counted while comparing the data with the set threshold. When the data reaches the set threshold, the parsing is stopped immediately and the document key points extraction process is performed on the target document content again. The number of abnormal document key points counted can also be compared with the set threshold after traversing each document key point in the document key point data.

[0206] In addition, the threshold value may be a directly set numerical value, such as 10 or 20. It may also be obtained by a preset percentage and the number of document key points in the document key point data. For example, if the preset percentage is 20% and the number of document key points in the document key point data is 80, the threshold value is set to 16. Other values ​​may also be used, and the present invention does not impose any limitation on this.

[0207] After the mind map, you can also draw a mind map and display it.

[0208] Specifically, the specific implementation process of drawing a mind map can be as follows:

[0209] Draw a node of the mind map, wherein the node includes node content and a text box storing the node content, and the node content is the key points of the document;

[0210] Draw lines between the nodes of the mind map.

[0211] Among them, the node can be directly drawn and displayed by using the method of providing a text box in the client system and displaying the node content in the text box, or other methods can be used. The connection relationship between nodes can be obtained by using a Bezier curve to select the plane position points between two text boxes to draw a curve to obtain a connection line, or by other methods. In this way, the tree node data of the generated mind map is finally drawn on the client to improve the user's document reading experience.

[0212] Combine the following Figure 2 The mind map generating method provided by the present invention is further explained.

[0213] First, the document content is parsed according to the document format of the target document, so as to obtain the target document content. There are different parsing methods for target documents of different document formats.

[0214] The user is currently reading a text document. He or she can select the full text or part of the document content, i.e. the target document content, based on the user's choice, and generate a mind map using an AI big model (large-scale language model).

[0215] Define a prompt instruction for the AI ​​big model, instructing the AI ​​big model to extract the key points of the document according to the target document content in the prompt instruction, and return the document key data based on the given mind map format example (such as the PlantUML mind map format), that is, determine the target mind map format according to the mind map format example, and output the document key data corresponding to the target mind map format.

[0216] Parse the mind map data returned by the AI ​​big model, that is, parse the key data of the document, generate a mind map and draw and display it on the client.

[0217] Specifically, based on the characteristics of the target mind map format, i.e., the node structure attribute, a node tree data structure is generated by parsing the text lines of the document key points data (each document key point) in sequence, and the mind map text nodes are drawn according to the node tree structure when drawing on the client. According to the parsed mind map node data, the document key points mind map is drawn and displayed on the client device through the client drawing method. The client can be a client device such as a user's mobile phone, computer, etc.

[0218] Since the data format returned by the AI ​​big model is not fixed, the prompt instruction is the key. It is necessary to strictly stipulate that the AI ​​big model returns the key data of the document according to the prompt instruction. Based on this, it can be agreed that the prompt has three clear responsibilities:

[0219] Role definition (key point extraction instructions), used to instruct the AI ​​model to be a document reading expert and extract the key points of the document;

[0220] Mind map format example: Give a specific format example in the prompt, and stipulate that after the AI ​​model learns the example, it will output the key data of the document according to the requirements (target mind map format);

[0221] Target document content: used for document key data output, and output based on the agreed target mind map format.

[0222] The role definition can be: "You are a document reading expert,according to the following document content quickly extract the key points in the document,thekey points strictly in PlantUML mindmap format output".

[0223] When the AI ​​big model returns the document key data corresponding to the target mind map format according to the prompt instruction, the AI ​​big model may also have a format mismatch problem. To address this problem, the following strategies can be adopted:

[0224] The document key data does not strictly meet the requirements. For this situation, you can try to adopt a compatible solution; if you find that the data results have @startmindmap and @endmindmap node outputs, and the middle part does not match the data, that is, the format compliance level is format framework compliance, you can choose to skip a document key and continue parsing the next document key; the example is as follows:

[0225] @startmindmap

[0226] *Java

[0227] **Java Language

[0228] ***Write once, run anywhere

[0229] ***Java applications

[0230] Development and Licensing

[0231] ***Java Virtual Machine

[0232] ***OpenJDK JVM

[0233] @endmindmap

[0234] In the above example, most of the formats meet the format requirements of the target mind map, but "Development and Licensing" does not match. You can try to skip this document point and continue to parse the next document point, and finally show the user a more accurate mind map; at the same time, statistics are made for the skipped document points (abnormal document points). If the number of abnormal document points is too large, that is, it reaches the set threshold, then the document point data is considered to be unqualified. In the client scenario, the user is prompted that the data format is wrong, and the user is instructed to make a new AI large model request, that is, to regenerate the mind map.

[0235] If the output result does not conform to the mind map format at all, that is, the format conformity level is not consistent with the format framework, in this case, the client can prompt the user that the data format is wrong and instruct the user to make a new AI large model request, that is, to regenerate the mind map.

[0236] Client mind map drawing: After parsing the key data of the document returned by the AI ​​large model in the mind map format and formatting it row by row, a tree structure data storage method will be generated. The mind map nodes will be drawn on the client through the client drawing method. The nodes can be drawn and displayed directly using the text box provided by the client system and the method of displaying the node content in the text box, or other methods. The connection relationship between the nodes can be obtained by drawing a curve by using a Bezier curve to select the plane position points between the two text boxes to obtain a connection line, or by other methods. In this way, the tree node data of the generated mind map will eventually be drawn on the client to improve the user's document reading experience.

[0237] User document reading is a daily behavior for people to acquire knowledge. This embodiment uses an AI big model to help understand and quickly generate document key data, and generates a mind map based on the document key data. The mind map is drawn on the client in a graphical visualization way, which can help people quickly understand the main content of the document and clarify the content logic, and help people quickly acquire document content knowledge. At the same time, it can also locate the document content where the key points are located based on the generated mind map, quickly find the document content, and improve people's document reading efficiency and experience.

[0238] The mind map generating device provided by the present invention is described below. The mind map generating device described below and the mind map generating method described above can be referred to each other.

[0239] Figure 3 : is a schematic diagram of the structure of the mind map generating device provided by the present invention, such as Figure 3 As shown, the mind map generation device 300 includes: an acquisition module 301, a document key point extraction model 302 and a generation module 303, wherein:

[0240] An acquisition module 301 is configured to acquire target document content and target mind map format;

[0241] The document key point extraction module 302 is configured to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format;

[0242] The generating module 303 is configured to generate a mind map corresponding to the content of the target document based on the document key data.

[0243] The present invention provides a mind map device, comprising obtaining target document content and target mind map format; performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; and generating a mind map corresponding to the target document content based on the document key point data. By quickly understanding the target document content to extract document key points, obtaining document key point data corresponding to the target mind map format, and generating a mind map based on the document key point data, people can quickly understand document logic and key points, improve reading efficiency and experience, and also improve the accuracy of mind maps.

[0244] In one or more optional embodiments of the present invention, the document key point extraction module 302 includes a hierarchical relationship determination module and a mind map generation module;

[0245] The hierarchical relationship determination module is configured to determine the hierarchical relationship of each document key point in the document key point data according to the node structure attribute corresponding to the target mind map format;

[0246] The mind map generation module is configured to generate a mind map corresponding to the content of the target document according to the hierarchical relationship and each of the document key points.

[0247] In one or more optional embodiments of the present invention, the document key point extraction module 302 is further configured to:

[0248] The target document content is input into a preset language processing model to perform document key point extraction processing to obtain document key point data corresponding to the target mind map format.

[0249] In one or more optional embodiments of the present invention, the language processing model is a large-scale language model;

[0250] The mind map generating device 300 further includes a first receiving module;

[0251] The first receiving module is configured to obtain a key point extraction instruction;

[0252] The document key point extraction module 302 is further configured to:

[0253] The key point extraction instruction and the target document content are input into the large-scale language model, and the large-scale language model is instructed to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format.

[0254] In one or more optional embodiments of the present invention, the acquisition module 301 includes a second receiving module, a target document content determination module and a target mind map format determination module;

[0255] The second receiving module is configured to receive a content selection instruction, wherein the content selection instruction carries a target document identifier and a selection area;

[0256] The target document content determination module is configured to determine the target document content according to the target document identifier and the selected area;

[0257] The target mind map format determination module is configured to input a set mind map format example into the large-scale language model for format learning to determine the target mind map format.

[0258] In one or more optional embodiments of the present invention, the mind map generating device 300 further includes a third receiving module and a prompt instruction generating module;

[0259] The third receiving module is configured to receive a format selection instruction, wherein the format selection instruction carries the mind map format example;

[0260] The prompt instruction generation module is configured to generate a prompt instruction, wherein the prompt instruction includes the key point extraction instruction, the mind map format example and the target mind map format, and the prompt instruction is used to instruct the large-scale language model to perform a document key point extraction task.

[0261] In one or more optional embodiments of the present invention, the mind map generating device 300 further includes a format detection module;

[0262] The format detection module is configured to perform format detection on the document key data to obtain the format conformity degree of the document key data;

[0263] The generating module 303 is further configured to:

[0264] When the format compliance degree is format framework compliance, a mind map corresponding to the target document content is generated based on the document key point data.

[0265] In one or more optional embodiments of the present invention, the mind map generating device 300 further includes a first indication module;

[0266] The first instruction module is configured to input the prompt instruction to the large-scale language model when the format compliance level is that the format framework does not comply with the format framework, and instruct the large-scale language model to re-extract the key points of the target document.

[0267] In one or more optional embodiments of the present invention, the generating module 303 includes a hierarchical relationship determining module, a skipping module and a mind map generating module;

[0268] The hierarchical relationship determination module is configured to parse each document key point in the document key point data according to the node structure attribute corresponding to the target mind map format when the format conformity degree is format framework conformity, and determine the hierarchical relationship of each document key point;

[0269] The skipping module is configured to skip parsing of an abnormal document key point if there is an abnormal document key point in each of the document key points that does not match the node structure attribute;

[0270] The mind map generation module is configured to generate a mind map corresponding to the target document content according to the hierarchical relationship and each of the document key points except the abnormal document key point.

[0271] In one or more optional embodiments of the present invention, the generating module 303 further includes a statistical module;

[0272] The statistical module is configured to count the number of key points of the abnormal document;

[0273] The mind map generating device 300 further includes a second indication module;

[0274] The second instruction module is configured to input the prompt instruction to the large-scale language model when the number is greater than or equal to a set threshold, and instruct the large-scale language model to re-extract key points from the target document.

[0275] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the mind map generation method, which includes: obtaining the target document content and the target mind map format; extracting the key points of the target document content to obtain the document key point data corresponding to the target mind map format; and generating the mind map corresponding to the target document content based on the document key point data.

[0276] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0277] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the mind map generation method provided by the above-mentioned methods, which method includes: obtaining the target document content and the target mind map format; performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; based on the document key point data, generating a mind map corresponding to the target document content.

[0278] On the other hand, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the mind map generation method provided by the above-mentioned methods, the method comprising: obtaining target document content and target mind map format; performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; and generating a mind map corresponding to the target document content based on the document key point data.

[0279] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0280] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0281] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these 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 invention.

Claims

1. A mind map generation method, characterized in that: include: Get the target document content and target mind map format; Performing document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; Based on the document key data, a mind map corresponding to the target document content is generated.

2. The mind map generation method according to claim 1, characterized in that: The step of generating a mind map corresponding to the target document content based on the document key point data includes: Determining the hierarchical relationship of each document key point in the document key point data according to the node structure attribute corresponding to the target mind map format; A mind map corresponding to the content of the target document is generated according to the hierarchical relationship and the key points of each document.

3. The mind map generation method according to claim 1 or 2, characterized in that: The extracting of key points from the target document content to obtain key point data of the document corresponding to the target mind map format includes: The target document content is input into a preset language processing model to perform document key point extraction processing to obtain document key point data corresponding to the target mind map format.

4. The mind map generation method according to claim 3, characterized in that: The language processing model is a large-scale language model; Before inputting the target document content into a preset language processing model to extract document key points and obtain document key point data corresponding to the target mind map format, the method further includes: Get key extraction instructions; The step of inputting the target document content into a preset language processing model to extract document key points and obtain document key point data corresponding to the target mind map format includes: The key point extraction instruction and the target document content are input into the large-scale language model, and the large-scale language model is instructed to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format.

5. The mind map generation method according to claim 4, characterized in that: The obtaining of the target document content and the target mind map format includes: receiving a content selection instruction, wherein the content selection instruction carries a target document identifier and a selection area; Determining the target document content according to the target document identifier and the selected area; The set mind map format example is input into the large-scale language model for format learning to determine the target mind map format.

6. The mind map generation method according to claim 5, characterized in that: The step of inputting the set mind map format example into the large-scale language model for format learning and determining the target mind map format further includes: receiving a format selection instruction, wherein the format selection instruction carries the mind map format example; Generate a prompt instruction, wherein the prompt instruction includes the key point extraction instruction, the mind map format example and the target mind map format, and the prompt instruction is used to instruct the large-scale language model to perform a document key point extraction task.

7. The mind map generation method according to claim 6, characterized in that: Before generating the mind map corresponding to the target document content based on the document key data, the method further includes: Performing format detection on the key data of the document to obtain the degree of conformity of the format of the key data of the document; The step of generating a mind map corresponding to the target document content based on the document key point data includes: When the format compliance degree is format framework compliance, a mind map corresponding to the target document content is generated based on the document key point data.

8. The mind map generation method according to claim 7, characterized in that: After the format detection is performed on the document key data to obtain the format conformity degree of the document key data, the method further includes: When the format conformance degree is that the format framework does not conform, the prompt instruction is input into the large-scale language model to instruct the large-scale language model to re-extract document key points from the target document.

9. The mind map generation method according to claim 7, characterized in that: When the format conformity degree is format framework conformity, generating a mind map corresponding to the target document content based on the document key point data includes: When the format conformity degree is format framework conformity, each document key point in the document key point data is parsed according to the node structure attribute corresponding to the target mind map format to determine the hierarchical relationship between each document key point; If there is an abnormal document key point in each of the document key points that does not match the node structure attribute, then the parsing of the abnormal document key point is skipped; A mind map corresponding to the target document content is generated according to the hierarchical relationship and the document main points except the abnormal document main points.

10. The mind map generation method according to claim 9, characterized in that: Before generating a mind map corresponding to the target document content according to the hierarchical relationship and each of the document key points other than the abnormal document key points, the method further includes: Counting the number of key points of the abnormal document; When the number is greater than or equal to a set threshold, the prompt instruction is input into the large-scale language model to instruct the large-scale language model to re-extract key points from the target document.

11. A mind map generating device, characterized in that: include: An acquisition module, configured to acquire target document content and target mind map format; A document key point extraction module is configured to perform document key point extraction processing on the target document content to obtain document key point data corresponding to the target mind map format; The generating module is configured to generate a mind map corresponding to the content of the target document based on the key points data of the document.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the mind map generating method according to any one of claims 1 to 10 is implemented.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the mind map generating method according to any one of claims 1 to 10 is implemented.

Images