CN118377912A - Electronic manual processing method, interactive system, electronic device and readable storage medium - Google Patents

Abstract

The invention discloses an electronic manual processing method, an interaction system, electronic equipment and a readable storage medium in the technical field of computer application, wherein the method comprises the following steps: acquiring a technical document, and performing chapter decomposition on the technical document to acquire chapter text; extracting text feature vectors of the chapter text, and classifying article contents of the chapter text based on the text feature vectors to obtain article classification; extracting the topics of the chapter texts of different classes to obtain text topics; extracting keywords from the chapter text to obtain article keywords; storing article classification, text subject and article keywords as text associated information of corresponding chapter text; in response to the operational behavior of the interactive electronic manual client, outputting a target chapter text associated with the operational behavior based on the text-related information. The technical effects are as follows: the interaction efficiency with respect to the technical document can be improved.

Description

Electronic manual processing method, interactive system, electronic device and readable storage medium

Technical Field

The present invention relates to the field of computer application technology, and in particular to an electronic manual processing method, an interactive system, an electronic device and a readable storage medium.

Background technique

Interactive Electronic Technical Manual (IETM) is a digital interactive system of technical manuals designed for equipment, devices or large-scale engineering systems. It meets the urgent need for digitization of massive technical documents of equipment. It can effectively improve the efficiency of system use and maintenance support. It is an important component and key technical means of equipment support.

As the amount of data increases, it is often difficult to quickly and accurately query electronic manuals. Although keywords can be extracted from articles by manually creating rules, so that keywords can be queried, querying based on keywords alone cannot effectively handle the processing of complex article language logic, thereby achieving efficient interaction.

In summary, how to effectively solve problems such as document processing in interactive electronic manuals is a technical problem that those skilled in the art urgently need to solve.

Summary of the invention

The purpose of the present invention is to provide an electronic manual processing method, an interactive system, an electronic device and a readable storage medium, which can effectively respond to the operation behavior of an interactive electronic manual client based on the text association information by extracting text association information from technical documents.

In order to solve the above technical problems, the present invention provides the following technical solutions:

An electronic manual processing method, comprising:

Acquire a technical document, and decompose the technical document into chapters to obtain chapter texts;

Extracting a text feature vector of the chapter text, and classifying the chapter text by article content based on the text feature vector to obtain an article classification;

Performing topic extraction on the chapter texts of different categories to obtain text topics;

Extract keywords from the chapter text to obtain article keywords;

The article classification, the text theme and the article keywords are stored as text association information of the corresponding chapter text;

In response to an operation behavior of the interactive electronic manual client, a target chapter text associated with the operation behavior is output based on the text association information.

Preferably, decomposing the technical document into chapters to obtain chapter texts includes:

Obtaining technical documents, and parsing the technical documents to obtain article titles and title levels;

The technical document is decomposed into chapters based on the article titles and the title levels to obtain chapter texts.

Preferably, keyword extraction is performed on the chapter text to obtain article keywords, including:

Performing word segmentation processing on the chapter text to obtain a number of words constituting the chapter text;

Calculating the word frequency and the reverse document frequency of the word, and taking the product of the word frequency and the reverse document frequency as the word score;

The article keywords are selected from a plurality of the words using the word scores.

Preferably, subject extraction is performed on the chapter texts of different categories to obtain text subjects, including:

The chapter texts of different categories are regarded as text groups;

Using a topic model algorithm, training the topic of the text group;

After convergence, the machine learning text similarity algorithm is used to determine the topic similarity of the articles;

Based on the article topic similarity, the text topic is determined from the trained topics.

Preferably, classifying the chapter text by article content based on the text feature vector to obtain article classification includes:

The naive Bayes algorithm is used and based on the Bayesian theorem, the probability of the features in the text feature vector appearing under the condition of a given category is calculated to obtain the text classification.

Preferably, obtaining technical documents includes:

Receiving the technical document from a specified interface;

The designated interface is used to obtain title format information of the technical document; the title format information includes a title level.

Preferably, using the designated interface to obtain the title format information of the technical document includes:

In the process of receiving the technical document by using the designated interface, creating a chapter object corresponding to the technical document; wherein the chapter object stores chapter title content, chapter text content and a sub-chapter list;

Loop through each text in the technical document, and use the title format information to determine whether the current text is a title;

If not, it is determined that the current text is text content, and the current text is written into the chapter text content of the last generated title object;

If so, if the current text is a title of the same level as the title object generated last time or is a higher level title of the title object generated last time, it is determined that the text content of the title object generated last time has been completed, a new title object is generated, and the current text is written into the chapter title content; if the current text is a lower level title of the title object generated last time, a new title object is generated and the current text is written into the chapter title content, and the new title object is added to the sub-chapter list of the title object generated last time;

Using the chapter object, the technical document is converted into a JSON format object array divided according to the document structure; wherein each value in the array corresponds to each chapter object of the technical document, and if the chapter object has sub-chapter, the chapter object has a sub-chapter object list as an attribute.

An interactive system, comprising:

Interactive electronic manual client, file management server, file parsing server;

The file parsing server is used to obtain technical documents, decompose the technical documents into chapters, and obtain chapter texts; extract text feature vectors of the chapter texts, classify the chapter texts into article content based on the text feature vectors, and obtain article classifications; extract topics from the chapter texts of different categories to obtain text topics;

Extract keywords from the chapter text to obtain article keywords; use the article classification, the text theme and the article keywords as text association information of the corresponding chapter text; send the text association information and the technical document to the file management server;

The file management server is used to receive and store the text association information and the technical document; in response to the operation behavior of the interactive electronic manual client, based on the text association information, output the target chapter text associated with the operation behavior;

The interactive electronic manual client is used to provide an operation interface, interact with the file management server, and output the target chapter text fed back by the file management server.

An electronic device, comprising:

Memory for storing computer programs;

A processor is used to implement the steps of the above-mentioned electronic manual processing method when executing the computer program.

A readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the electronic manual processing method are implemented.

The method provided by the embodiment of the present invention includes: acquiring technical documents, and decomposing the technical documents into chapters to obtain chapter texts; extracting text feature vectors of the chapter texts, and classifying the chapter texts into article categories based on the text feature vectors to obtain article categories; extracting topics from chapter texts of different categories to obtain text topics; extracting keywords from chapter texts to obtain article keywords; storing article categories, text topics, and article keywords as text association information of corresponding chapter texts; and outputting target chapter texts associated with the operation behavior based on the text association information in response to the operation behavior of the interactive electronic manual client.

In the present invention, in order to achieve effective processing of each part in the technical document, the complete technical document is first decomposed into chapters to obtain chapter texts. By classifying the chapter texts, article classifications can be obtained; by extracting keywords from the chapter texts, keywords of the chapter texts can be obtained; by extracting topics from chapter texts of different categories, text topics can be obtained. The article classifications, text topics and article keywords corresponding to the chapter texts are stored as text association information of the chapter texts. When the user operates on the interactive electronic manual client, in response to the operation behavior of the interactive electronic manual client, the target chapter text associated with the operation behavior can be output based on the text association information. Since the text association information is stored based on the chapter text dimension, and the text association information includes article classifications, text topics and article keywords, compared with realizing interactive responses based only on keywords in the complete document dimension, the present invention can provide more accurate output content during the interactive process.

Technical effect: The technical document is broken down chapter by chapter, and key information of each chapter is extracted and classified by combining different algorithms, so as to obtain text association information including article classification, text theme and article keywords. Based on this text association information, the query and interaction efficiency of the electronic manual can be improved.

Correspondingly, the embodiments of the present invention also provide a system, a device and a readable storage medium corresponding to the above-mentioned artificial intelligence-based electronic manual processing method, which has the above-mentioned technical effects and will not be repeated here.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only 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.

FIG1 is a flowchart of an implementation method of an electronic manual processing method in an embodiment of the present invention;

FIG2 is a schematic diagram of the structure of a system in an embodiment of the present invention;

FIG3 is a specific implementation flow chart of an electronic manual processing method according to an embodiment of the present invention;

FIG4 is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of a specific structure of an electronic device in an embodiment of the present invention.

Detailed ways

In order to enable those skilled in the art to better understand the scheme of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

Please refer to FIG. 1 , which is a flow chart of an electronic manual processing method in an embodiment of the present invention. The method can be applied to the interactive system (technical manual digitization interactive system) shown in FIG. 2 . The method includes the following steps:

S101. Obtain a technical document, and decompose the technical document into chapters to obtain chapter texts.

In the technical manual digital interaction system, there are a large number of technical documents. In order to achieve efficient interaction based on these technical documents, in this embodiment, it is proposed to split the technical documents into chapter text units for processing.

Specifically, the technical document may be obtained by receiving the technical document, downloading the technical document, or directly reading the technical document from a storage medium. In this embodiment, the channel for obtaining the technical document is not limited.

In practical applications, the technical document may specifically be a technical manual of a device or system or other technical documents. The specific contents recorded in the technical document and the specific circumstances of the division of chapters therein are not limited in this embodiment.

After obtaining the technical document, the technical document may be divided into chapters according to information such as the title format in the technical document, thereby obtaining several chapter texts of the technical document.

In one embodiment of the present invention, the technical document is decomposed into chapters to obtain chapter texts, including:

Obtain technical documents, parse them, and obtain article titles and title levels;

The technical document is decomposed into chapters based on the article title and heading level to obtain the chapter text.

For ease of description, the above steps are combined for explanation below.

After obtaining the technical document, the technical document can be parsed to obtain the article title and title level. Then, based on the article title and title level, the technical document can be divided into chapters to obtain the text of each chapter of the technical document.

In one embodiment of the present invention, obtaining a technical document includes:

Receive technical documentation from a specified interface;

Use the specified interface to obtain the title format information of the technical document; the title format information includes the title level.

For ease of description, the above two steps are combined for explanation below.

Specifically, the specified interface can be an interface such as Apache POI that can realize document uploading and obtaining document title format information. Apache POI (Poor Obfuscation Implementation) is a free, open source, cross-platform Java API written in Java. Apache POI provides an API for Java programs to read and write Microsoft Office format files.

In other words, Apache POI can be used to extract information from Word documents, upload technical data in a unified format, and store all title format information in the Word document, distinguishing title levels when storing.

In one embodiment of the present invention, obtaining title format information of a technical document includes:

In the process of receiving the technical document using the specified interface, a chapter object corresponding to the technical document is created; wherein the chapter object stores the chapter title content, the chapter text content and the sub-chapter list;

Loop through each text in the technical document and use the title format information to determine whether the current text is a title;

If not, it is determined that the current text is text content, and the current text is written into the chapter text content of the last generated title object;

If so, if the current text is a title of the same level as the title object generated last time or is a higher level title of the title object generated last time, it is determined that the text content of the title object generated last time has been completed, a new title object is generated, and the current text is written into the chapter title content; if the current text is a lower level title of the title object generated last time, a new title object is generated and the current text is written into the chapter title content, and the new title object is added to the sub-chapter list of the title object generated last time;

Using the chapter object, the technical document is converted into a JSON format object array divided according to the document structure; each value in the array corresponds to each chapter object of the technical document. If the chapter object has sub-chapter, the chapter object has a list of sub-chapter objects as an attribute.

For ease of description, the above steps are combined for explanation below.

Specifically, a chapter object may be defined to store chapter title content, chapter text content, and a sub-chapter list.

You can loop through each paragraph of text in the document and determine whether it is a title based on the format of the text. If it is not a title, it will be regarded as text content and used as the text content of the previous title object; if it is a title, and it is a title of the same level as the title object generated last time or the previous title object is a higher-level title, then the title object generated last time is regarded as completing the text content, and a new title object is generated; if the title object generated last time is a lower-level title of this title, then a new title object is generated and used as a sub-title of the previous title object.

In actual applications, the entire document can also be converted into a JSON (Java Script Object Notation, JS object notation, a lightweight data exchange format) format object array divided according to the document structure based on the document's chapter objects. Each value in the array corresponds to each chapter object of the article. If each chapter object has sub-chapter, the chapter object has a sub-chapter object list as an attribute and stores all sub-chapter objects at the same time.

In order to improve the quality of the text, after decomposing the technical documents and obtaining the chapter texts, you can also perform operations such as text cleaning, text segmentation, and stop word filtering on the text data based on the chapter texts.

S102: extracting text feature vectors of chapter texts, and classifying the chapter texts by article content based on the text feature vectors to obtain article classifications.

After obtaining the chapter text, the chapter text can be classified based on the text feature vector by obtaining the text feature vector of the chapter text in advance, thereby obtaining the article classification.

In practical applications, each chapter text can be input into a text content classification model for classification and recognition, thereby obtaining the article classification of each chapter text.

In one embodiment of the present invention, the chapter text is classified into article content based on the text feature vector to obtain article classification, including:

Using the naive Bayes algorithm and based on the Bayesian theorem, the text classification is obtained by calculating the probability of the features appearing in the text feature vector under the condition of a given category.

Among them, the naive Bayes algorithm is based on Bayes' theorem. It performs classification by calculating the probability of each feature appearing under the condition of a given category. It can be used for text classification. For a given text sample, the naive Bayes algorithm calculates the posterior probability of each category and selects the category with the highest posterior probability as the classification result.

Specifically, all chapter texts of the technical document can be extracted using a machine learning text feature extraction algorithm to obtain text feature vectors, and the article content can be classified using a naive Bayes algorithm.

S103, extracting topics from different types of chapter texts to obtain text topics.

Considering that in actual applications, although chapter texts belong to different categories, they may involve a theme. In order to realize interactive recommendation based not only on categories but also on themes in the interactive process. Therefore, in this embodiment, theme extraction is performed on chapter texts of different categories to obtain a common text theme. Specifically, algorithms such as clustering algorithms can be used to extract and obtain text themes.

In one embodiment of the present invention, topic extraction is performed on different types of chapter texts to obtain text topics, including:

Treat different types of chapter texts as text groups;

Using the topic model algorithm, the topics of the text group are trained;

After convergence, the machine learning text similarity algorithm is used to determine the topic similarity of the articles;

Based on the similarity of article topics, the text topics are determined from the trained topics.

In other words, based on the article classification results, different types of articles can be grouped together for topic extraction, and the algorithm can be trained and converged using the LDA algorithm. Then, the machine learning text similarity algorithm can be used to determine the similarity of article topics, and topics with higher similarity can be extracted and stored as topics of the same category.

Among them, the LDA topic model belongs to the topic model algorithm, which can be used to infer the topic distribution of documents. It can give the topic of each document in the document set in the form of probability distribution, so that after analyzing some documents to extract their topic distribution, topic clustering or text classification can be performed based on the topic distribution.

S104, extracting keywords from the chapter text to obtain article keywords.

Interaction based on keywords is also one of the channels to improve the effectiveness of interaction. Therefore, in this embodiment, keyword extraction can be performed on the chapter text to obtain article keywords.

In practical applications, the chapter text can be segmented, and then each word can be counted to select the article keywords based on the repetition rate.

In one embodiment of the present invention, keyword extraction is performed on the chapter text to obtain article keywords, including:

Perform word segmentation on the chapter text to obtain several words that constitute the chapter text;

Calculate the word frequency and inverse document frequency of the word, and take the product of the word frequency and the inverse document frequency as the word score;

Use word scores to select article keywords from a number of words.

For ease of description, the above steps are combined for explanation below.

TF (Term Frequency) is the term frequency, and IDF (Inverse Document Frequency) is the inverse document frequency index.

TF-IDF (term frequency–inverse document frequency) is a commonly used weighting technique for information retrieval and data mining to evaluate the importance of a word to a document set or a document in a corpus. The importance of a word increases in direct proportion to the number of times it appears in a document, but decreases in inverse proportion to the frequency of its appearance in the corpus.

That is to say, in practical applications, the TF-IDF keyword extraction algorithm can be performed on the chapters in the article to obtain the key information contained in each chapter, extract the keywords in the chapters and store them.

S105: storing the article classification, text theme and article keywords as text association information of the corresponding chapter text.

In practical applications, a file management server may be provided to store all text data and text-related information of all data files.

Text-related information includes article classification, text subject, and article keyword information.

S106 : In response to the operation behavior of the interactive electronic manual client, output the target chapter text associated with the operation behavior based on the text association information.

The interactive electronic manual client can provide an operation interface, and the user can perform related operations based on the operation interface. The interactive electronic manual client captures the user's operation behavior and feeds it back to the server, and the server can respond to the operation behavior. Specifically, the target chapter text associated with the operation behavior can be output based on the text management information.

Of course, the operation interface can also be presented based on technologies such as VR (Virtual Reality, also known as virtual technology, virtual environment), and in this embodiment, the specific presentation form of the interaction is not limited.

For example: the user can input query information in the interactive electronic manual client, perform approximate calculations based on the key content in the query information and the text association information, and select the target chapter texts in the query list based on the calculated distance information; or, when the interactive electronic manual client displays the chapter A text, obtain the text association information of the chapter A text, and use the text association information to find the chapter B text that is closely associated with the text association information of the chapter A text from the file management server.

For example: This electronic manual processing method can be applied to a VR-based vehicle equipment maintenance system, combined with interactive electronic manual technology. Specifically, the user identifies the vehicle maintenance fault point through virtual-reality fusion technology, and obtains the target chapter texts of the maintenance parts and related maintenance technologies based on the virtual identification parts operated by the user's virtual-reality interaction. Alternatively, through the relevant article keywords entered by the user, the relevant articles are directly searched, and the text association information associated with the article text theme is obtained, and the associated chapter text is obtained through the text association information, and the relevant articles and associated chapter texts are displayed on the user terminal at the same time.

The method provided by the embodiment of the present invention is applied, and the method includes: obtaining technical documents, and decomposing the technical documents into chapters to obtain chapter texts; extracting text feature vectors of the chapter texts, and classifying the chapter texts into article content based on the text feature vectors to obtain article classifications; extracting topics from chapter texts of different categories to obtain text topics; extracting keywords from chapter texts to obtain article keywords; storing article classifications, text topics, and article keywords as text association information of corresponding chapter texts; and outputting target chapter texts associated with the operation behavior based on the text association information in response to the operation behavior of the interactive electronic manual client.

In the present invention, in order to achieve effective processing of each part in the technical document, the complete technical document is first decomposed into chapters to obtain chapter texts. By classifying the chapter texts, article classifications can be obtained; by extracting keywords from the chapter texts, keywords of the chapter texts can be obtained; by extracting topics from chapter texts of different categories, text topics can be obtained. The article classifications, text topics and article keywords corresponding to the chapter texts are stored as text association information of the chapter texts. When the user operates on the interactive electronic manual client, in response to the operation behavior of the interactive electronic manual client, the target chapter text associated with the operation behavior can be output based on the text association information. Since the text association information is stored based on the chapter text dimension, and the text association information includes article classifications, text topics and article keywords, compared with realizing interactive responses based only on keywords in the complete document dimension, the present invention can provide more accurate output content during the interactive process.

Technical effect: The technical document is broken down chapter by chapter, and key information of each chapter is extracted and classified by combining different algorithms, so as to obtain text association information including article classification, text theme and article keywords. Based on this text association information, the query and interaction efficiency of the electronic manual can be improved.

Corresponding to the above method embodiment, the embodiment of the present invention further provides an interactive system. The interactive system described below and the electronic manual processing method described above can be referred to in correspondence with each other.

As shown in FIG2 , the system includes:

Interactive electronic manual client 100, file management server 200, file parsing server 300;

The file parsing server is used to obtain technical documents, decompose the technical documents into chapters, and obtain chapter texts; extract text feature vectors of the chapter texts, classify the chapter texts into article content based on the text feature vectors, and obtain article classifications; extract topics from different types of chapter texts to obtain text topics;

Extract keywords from the chapter text to obtain article keywords; use article classification, text theme and article keywords as text association information of the corresponding chapter text; send the text association information and technical documents to the file management server;

The file management server is used to receive and store text association information and technical documents; in response to the operation behavior of the interactive electronic manual client, based on the text association information, output the target chapter text associated with the operation behavior;

The interactive electronic manual client is used to provide an operation interface, interact with the file management server, and output the target chapter text fed back by the file management server.

The system provided by the embodiment of the present invention is applied to obtain technical documents, decompose the technical documents into chapters, and obtain chapter texts; extract text feature vectors of the chapter texts, classify the chapter texts by article content based on the text feature vectors, and obtain article classifications; extract topics from different types of chapter texts to obtain text topics; extract keywords from the chapter texts to obtain article keywords; store article classifications, text topics, and article keywords as text association information of corresponding chapter texts; and output target chapter texts associated with the operation behavior based on the text association information in response to the operation behavior of the interactive electronic manual client.

In the present invention, in order to achieve effective processing of each part in the technical document, the complete technical document is first decomposed into chapters to obtain chapter texts. By classifying the chapter texts, article classifications can be obtained; by extracting keywords from the chapter texts, keywords of the chapter texts can be obtained; by extracting topics from chapter texts of different categories, text topics can be obtained. The article classifications, text topics and article keywords corresponding to the chapter texts are stored as text association information of the chapter texts. When the user operates on the interactive electronic manual client, in response to the operation behavior of the interactive electronic manual client, the target chapter text associated with the operation behavior can be output based on the text association information. Since the text association information is stored based on the chapter text dimension, and the text association information includes article classifications, text topics and article keywords, compared with realizing interactive responses based only on keywords in the complete document dimension, the present invention can provide more accurate output content during the interactive process.

Technical effect: The technical document is broken down chapter by chapter, and key information of each chapter is extracted and classified by combining different algorithms, so as to obtain text association information including article classification, text theme and article keywords. Based on this text association information, the query and interaction efficiency of the electronic manual can be improved.

In a specific implementation of the present invention, the file parsing server is specifically used to obtain technical documents and parse the technical documents to obtain article titles and title levels;

The technical document is decomposed into chapters based on the article title and heading level to obtain the chapter text.

In a specific embodiment of the present invention, the file parsing server is specifically used to perform word segmentation processing on the chapter text to obtain a number of words constituting the chapter text;

Calculate the word frequency and inverse document frequency of the word, and take the product of the word frequency and the inverse document frequency as the word score;

Use word scores to select article keywords from a number of words.

In a specific embodiment of the present invention, the file parsing server is specifically used to treat different types of chapter texts as text groups;

Using the topic model algorithm, the topics of the text group are trained;

After convergence, the machine learning text similarity algorithm is used to determine the topic similarity of the articles;

Based on the similarity of article topics, the text topics are determined from the trained topics.

In a specific embodiment of the present invention, the file parsing server is specifically used to utilize the naive Bayes algorithm and, based on the Bayes theorem, classify by calculating the probability of occurrence of features in a text feature vector under the condition of a given category to obtain a text classification.

In a specific implementation of the present invention, the file parsing server is specifically used to receive the technical document from the specified interface;

Use the specified interface to obtain the title format information of the technical document; the title format information includes the title level.

In a specific embodiment of the present invention, the file parsing server is specifically used to create a chapter object corresponding to the technical document during the process of receiving the technical document using the specified interface; wherein the chapter object stores the chapter title content, the chapter text content and the sub-chapter list;

Loop through each text in the technical document and use the title format information to determine whether the current text is a title;

If not, it is determined that the current text is text content, and the current text is written into the chapter text content of the last generated title object;

If so, if the current text is a title of the same level as the title object generated last time or is a higher level title of the title object generated last time, it is determined that the text content of the title object generated last time has been completed, a new title object is generated, and the current text is written into the chapter title content; if the current text is a lower level title of the title object generated last time, a new title object is generated and the current text is written into the chapter title content, and the new title object is added to the sub-chapter list of the title object generated last time;

Using the chapter object, the technical document is converted into a JSON format object array divided according to the document structure; each value in the array corresponds to each chapter object of the technical document. If the chapter object has sub-chapter, the chapter object has a list of sub-chapter objects as an attribute.

The electronic manual processing method can be applied to the above-mentioned system. To facilitate those skilled in the art to better understand and implement the electronic manual processing method, the electronic manual processing method is described in detail below with reference to a specific application scenario as an example.

Specifically, the electronic manual processing method can be implemented in the system shown in FIG2 , and the details are as follows:

Interactive electronic manual client (hereinafter referred to as client): browse the information of the interactive electronic manual or perform interactive operations on the equipment through a visual operation interface. For example, the client uses VR to perform maintenance operations on the simulated equipment, and further displays relevant technical documents based on the interactive maintenance operation information.

File management server: used to store all text data and text-related data of all data files.

File parsing server: used to parse the searched data files and obtain relevant text-related data and article composition.

Text-related data (i.e., text-related information) includes article classification, text topics, and article keyword information.

Referring to FIG. 3 , the method includes the following steps:

S1. Upload the data file to the file parsing server. The file parsing server decomposes the file into chapters according to the parsed article titles and title levels, and obtains different chapter titles and the article contents they contain.

That is, the file parsing server obtains the positions of all titles in the document and determines the chapter level of the document based on the text style or title number of the title;

Specifically, Apache POI can be used to extract Word document information, upload Word documents in a unified format, and store all title format information in the Word document, and distinguish title levels when storing;

Furthermore, a chapter object is defined to store chapter title content, chapter text content and a sub-chapter list. Each text in the document is looped through, and it is determined whether it is a title based on the format of the text. If it is not a title, it is regarded as text content and used as the text content of the previous title object. If it is a title, and it is a title of the same level as the previously generated title object or the previous title object is a higher-level title, the text content of the previously generated title object is regarded as completed, and a new title object is generated. If the title object generated previously is a lower-level title of this title, a new title object is generated and used as a sub-title of the previous title object.

Furthermore, according to the document's chapter objects, the entire document is converted into a JSON format object array divided according to the document structure. Each value in the array corresponds to each chapter object of the article. If each chapter object has sub-chapter, the chapter object has a sub-chapter object list as an attribute and stores all sub-chapter objects at the same time.

S2. The file parsing server uses AI algorithms to analyze the chapter text, and sends the results of the analysis to the file management server.

The file parsing server is used for text analysis. First, all uploaded data files are preprocessed as articles, including text cleaning, text segmentation, stop word filtering and other operations.

When uploading text data for the first time through the file parsing server, select the text data that has been classified for uploading, and do not set the algorithm parameters (weight and smoothing factor). After completing the model training and prediction, evaluate the results through the evaluation algorithm, and adjust the parameter settings again to continue training until the evaluation value reaches the maximum. Then add new data files one by one to improve the classification model.

According to the article classification results, different types of articles are grouped together for topic extraction. The LDA algorithm is used to train the algorithm and make it converge. Then, the machine learning text similarity algorithm is used to determine the similarity of article topics, and topics with higher similarity are extracted and stored as topics of the same category.

The TF-IDF keyword extraction algorithm is performed on the chapters in the article to obtain the key information contained in each chapter, extract the keywords in the chapters and store them.

S3. The file management server stores the relevant document analysis information according to the analysis results of the file analysis server.

S4. When the client browses and interacts with information, related information is displayed.

When the client performs VR electronic document interaction, it extracts the relevant article chapters from the file management server for information display based on the text-related data associated with the maintenance interaction, and can directly browse the relevant article content.

It can be seen that the present invention solves the problem of low efficiency of manual search when storing a large number of data files in an interactive electronic manual. The article text information extraction is completed automatically through a machine learning algorithm. The shortcomings of a single algorithm are compensated by a variety of different algorithms.

Corresponding to the above method embodiment, an embodiment of the present invention further provides an electronic device. The electronic device described below and the electronic manual processing method described above can refer to each other.

As shown in FIG4 , the electronic device includes:

A memory 332, for storing computer programs;

The processor 322 is used to implement the steps of the electronic manual processing method of the above method embodiment when executing the computer program.

Specifically, please refer to FIG. 5, which is a schematic diagram of the specific structure of an electronic device provided in this embodiment. The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) (for example, one or more processors) and a memory 332, and the memory 332 stores one or more computer programs 342 or data 344. Among them, the memory 332 can be a temporary storage or a permanent storage. The program stored in the memory 332 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the data processing device. Furthermore, the processor 322 can be configured to communicate with the memory 332 to execute a series of instruction operations in the memory 332 on the electronic device 301.

The electronic device 301 may further include one or more power supplies 326 , one or more wired or wireless network interfaces 350 , one or more input and output interfaces 358 , and/or one or more operating systems 341 .

The steps in the electronic manual processing method described above can be implemented by the structure of an electronic device.

Corresponding to the above method embodiment, the embodiment of the present invention further provides a readable storage medium. The readable storage medium described below and the electronic manual processing method described above can be referred to each other.

A readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the electronic manual processing method of the above method embodiment are implemented.

The readable storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other readable storage medium that can store program codes.

In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

Claims (10)

1. A method for processing an electronic manual, comprising: Acquire a technical document, and decompose the technical document into chapters to obtain chapter texts; Extracting a text feature vector of the chapter text, and classifying the chapter text by article content based on the text feature vector to obtain an article classification; Performing topic extraction on the chapter texts of different categories to obtain text topics; Extract keywords from the chapter text to obtain article keywords; The article classification, the text theme and the article keywords are stored as text association information of the corresponding chapter text; In response to an operation behavior of the interactive electronic manual client, a target chapter text associated with the operation behavior is output based on the text association information. 2. The method according to claim 1, characterized in that the technical document is decomposed into chapters to obtain chapter texts, comprising: Obtaining technical documents, and parsing the technical documents to obtain article titles and title levels; The technical document is decomposed into chapters based on the article titles and the title levels to obtain chapter texts. 3. The method according to claim 1, characterized in that the keyword extraction of the chapter text to obtain article keywords comprises: Performing word segmentation processing on the chapter text to obtain a number of words constituting the chapter text; Calculating the word frequency and the reverse document frequency of the word, and taking the product of the word frequency and the reverse document frequency as the word score; The article keywords are selected from a plurality of the words using the word scores. 4. The method according to claim 1 is characterized in that the topic extraction is performed on the chapter texts of different categories to obtain the text topics, including: The chapter texts of different categories are regarded as text groups; Using a topic model algorithm, training the topic of the text group; After convergence, the machine learning text similarity algorithm is used to determine the topic similarity of the articles; Based on the article topic similarity, the text topic is determined from the trained topics. 5. The method according to claim 1, characterized in that the article content classification of the chapter text is performed based on the text feature vector to obtain the article classification, comprising: The naive Bayes algorithm is used and based on the Bayesian theorem, the probability of the features in the text feature vector appearing under the condition of a given category is calculated to obtain the text classification. 6. The method according to any one of claims 1 to 5, characterized in that obtaining technical documents comprises: Receiving the technical document from a specified interface; The designated interface is used to obtain title format information of the technical document; the title format information includes a title level. 7. The method according to claim 6, characterized in that, using the designated interface to obtain the title format information of the technical document comprises: In the process of receiving the technical document by using the designated interface, creating a chapter object corresponding to the technical document; wherein the chapter object stores chapter title content, chapter text content and a sub-chapter list; Loop through each text in the technical document, and use the title format information to determine whether the current text is a title; If not, it is determined that the current text is text content, and the current text is written into the chapter text content of the last generated title object; If so, if the current text is a title of the same level as the title object generated last time or is a higher level title of the title object generated last time, it is determined that the text content of the title object generated last time has been completed, a new title object is generated, and the current text is written into the chapter title content; if the current text is a lower level title of the title object generated last time, a new title object is generated and the current text is written into the chapter title content, and the new title object is added to the sub-chapter list of the title object generated last time; Using the chapter object, the technical document is converted into a JSON format object array divided according to the document structure; wherein each value in the array corresponds to each chapter object of the technical document, and if the chapter object has sub-chapter, the chapter object has a sub-chapter object list as an attribute. 8. An interactive system, comprising: Interactive electronic manual client, file management server, file parsing server; The file parsing server is used to obtain technical documents, decompose the technical documents into chapters, and obtain chapter texts; extract text feature vectors of the chapter texts, classify the chapter texts into article content based on the text feature vectors, and obtain article classifications; extract topics from the chapter texts of different categories to obtain text topics; Extract keywords from the chapter text to obtain article keywords; use the article classification, the text theme and the article keywords as text association information of the corresponding chapter text; and send the text association information and the technical document to the file management server; The file management server is used to receive and store the text association information and the technical document; in response to the operation behavior of the interactive electronic manual client, based on the text association information, output the target chapter text associated with the operation behavior; The interactive electronic manual client is used to provide an operation interface, interact with the file management server, and output the target chapter text fed back by the file management server. 9. An electronic device, comprising: Memory for storing computer programs; A processor is used to implement the steps of the electronic manual processing method as claimed in any one of claims 1 to 7 when executing the computer program. 10. A readable storage medium, characterized in that a computer program is stored on the readable storage medium, and when the computer program is executed by a processor, the steps of the electronic manual processing method according to any one of claims 1 to 7 are implemented.