Text format checking method and device, equipment and storage medium

Abstract

The invention relates to the technical field of education and data processing, and provides a text format checking method, device and equipment and a storage medium, the method comprises the following steps: understanding obtained standard text data, and generating a checking rule model based on an understanding result, the checking rule model comprising a plurality of checking rules corresponding to different standard checking types; analyzing the obtained to-be-checked text data, and splitting the to-be-checked text data into to-be-checked blocks corresponding to different standard check types based on an analysis result; and performing format check on each to-be-checked block by adopting a corresponding check rule, and annotating the to-be-checked blocks which do not conform to the corresponding check rule. According to the method provided by the invention, the check rule model is automatically generated, the workload of manual input is reduced, and manual input errors are avoided; format checking is automatically carried out, the complex process of manual checking is reduced, and the checking accuracy is improved; and format errors are automatically annotated, so that the checking efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of education and data processing technology, and specifically to a text format checking method, device, equipment and storage medium. Background Art

[0002] When writing academic papers, it is necessary to strictly abide by the established format standards. Format standards not only ensure the readability and uniformity of academic papers, but also help authors explain their research results more clearly.

[0003] Traditionally, there are several forms of formatting checks for academic papers: manual online checking, which ensures the correctness of the format by carefully reviewing online academic papers; printing checking, which prints out academic papers as paper copies and flips through them to identify formatting errors; and template refreshing checking, which uses pre-designed templates (such as word templates) to format academic papers to ensure that the format meets specific requirements. Although these methods can check the format of academic papers to a certain extent, they have limitations such as low efficiency and possible omissions.

[0004] With the advancement of the education system and the expansion of disciplinary boundaries, the specific requirements for academic papers have diversified. The format specifications of academic papers are constantly evolving and updating, making the format specifications of academic papers more and more complex and changeable. Some existing academic paper format checking tools usually target the format requirements of a specific school at a specific time, and manually enter the format requirements into the format checking tool to perform format checking. Once the format requirements change, even if the changes are very minor, the format requirements need to be manually sorted out and re-entered. Therefore, the existing academic paper format checking tools cannot adapt well to the evolving paper formats and specifications, and lack the ability to self-correct based on the actual inspection results.

[0005] Therefore, the existing academic paper format checking tools have not yet achieved good versatility and self-correction. A more flexible and continuously updated solution is needed to adapt to the continuous evolution and updating of academic paper format specifications. Summary of the invention

[0006] To solve the above problems, the embodiments of the present application provide a text format checking method, apparatus, device and storage medium, which generates a checking rule model based on the understanding of standard text data, and performs format checking based on the parsing of the text to be checked by calling the checking rule model. Through precise annotations, authors are assisted in discovering format deviations of academic papers, thereby improving the accuracy and efficiency of text format checking.

[0007] The present application embodiment adopts the following technical solutions:

[0008] In a first aspect, an embodiment of the present application provides a text format checking method, the method comprising:

[0009] Understanding the obtained standard text data, and generating an inspection rule model based on the understanding result; wherein the inspection rule model includes a plurality of inspection rules corresponding to different standard inspection types;

[0010] Parse the acquired text data to be checked, and divide the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing result;

[0011] The corresponding checking rules are used to perform format checking on each block to be checked, and the blocks to be checked that do not meet the corresponding checking rules are annotated.

[0012] In a second aspect, an embodiment of the present application provides a text format checking device, the device comprising:

[0013] A rule model generating unit, used to understand the obtained standard text data and generate an inspection rule model based on the understanding result; wherein the inspection rule model includes a plurality of inspection rules corresponding to different standard inspection types;

[0014] A to-be-checked text processing unit is used to parse the acquired to-be-checked text data and split the to-be-checked text data into to-be-checked blocks corresponding to different standard inspection types based on the parsing result;

[0015] The checking and annotation unit is used to perform format checking on each block to be checked using corresponding checking rules, and to annotate the blocks to be checked that do not meet the corresponding checking rules.

[0016] In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned text format checking method when executing the computer program.

[0017] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and the computer program implements the steps of the above-mentioned text format checking method when instructed by a processor.

[0018] At least one of the above technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects:

[0019] The text format checking method provided by the present application understands the obtained standard text data and generates a checking rule model based on the understanding results; wherein the checking rule model includes multiple checking rules corresponding to different standard checking types; the obtained text data to be checked is parsed, and the text data to be checked is split into blocks to be checked corresponding to different standard checking types based on the parsing results; the corresponding checking rules are used to perform format checking on each block to be checked, and the blocks to be checked that do not meet the corresponding checking rules are annotated. The text format checking method provided by the present application can automatically generate a checking rule model based on the understanding of the standard text data, reducing the huge workload of manual input and avoiding manual input errors; it can automatically perform format checking based on the analysis of the text to be checked, reducing the complex process of manual proofreading and improving the accuracy of the inspection; it can automatically annotate format errors to assist the author in locating the format error area and assist the author in making format modifications. The method provided by the present application improves the accuracy and efficiency of text format checking. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0021] Figure 1 A schematic diagram showing an application environment of a text format checking method according to an embodiment of the present application;

[0022] Figure 2 A schematic diagram showing a flow chart of a text format checking method according to an embodiment of the present application;

[0023] Figure 3 A schematic diagram showing a flow chart of a text format checking method according to another embodiment of the present application;

[0024] Figure 4 A schematic diagram showing the structure of a text format checking device according to an embodiment of the present application is shown;

[0025] Figure 5 A schematic diagram of the structure of a computer device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0027] Traditional academic paper format checking tools are often based on fixed checking rules. These checking rules need to be entered and modified manually, which is slow to set up and may be prone to human errors. In addition, each school, discipline, etc. usually has format requirements that adapt to their own situations. These requirements are not universal and may conflict with each other, making existing tools only applicable to specific scenarios and lacking in universality.

[0028] As the requirements for the quality of academic paper content become higher and higher, there is an urgent need for a technical solution that can quickly and automatically check the format specifications of academic papers, and has stronger versatility and automatic correction capabilities. Therefore, this application generates a check rule model based on the understanding of standard text data, and performs format checking based on the analysis of the text to be checked. Through precise annotations, it assists authors in discovering academic paper format deviations, thereby improving the accuracy and efficiency of text format checking.

[0029] The text format checking method provided in the embodiment of the present application can be applied in the following aspects: Figure 1 In an application environment, a server and a client communicate through a network.

[0030] The client is used to indicate the party that needs to perform format checking on the text data to be checked. The client can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, portable devices, and other devices with display screens and input devices. The server is used to indicate the party that provides the format checking function. The server can be implemented as an independent server or a server cluster consisting of multiple servers.

[0031] The server obtains standard text data from a public network, client or other communication terminal, understands the obtained standard text data, and generates an inspection rule model based on the understanding results; wherein the inspection rule model includes multiple inspection rules corresponding to different standard inspection types. The server obtains the text data to be checked from the client, parses the obtained text data to be checked, and splits the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing results. The server uses the corresponding inspection rules to perform format checks on each block to be checked, and annotates the blocks to be checked that do not comply with the corresponding inspection rules. The server sends the annotated text data to be checked to the client.

[0032] The present application is described in detail below through specific embodiments.

[0033] Figure 2 A flowchart of a text format checking method proposed in an embodiment of the present application is shown. Figure 2 As shown, this embodiment includes steps S210 to S230:

[0034] Step S210, understanding the obtained standard text data, and generating a check rule model based on the understanding result; wherein the check rule model includes a plurality of check rules corresponding to different standard check types.

[0035] This embodiment first understands the standard text data, thereby extracting the checking rules for format checking.

[0036] The standard text data is used to indicate the standard documents for writing academic papers. The standard text data may include but is not limited to national academic paper writing standards (national standards for academic paper writing rules), provincial academic paper writing standards, school-level academic paper writing standards, and college-level academic paper writing standards.

[0037] When understanding the obtained standard text data, the standard text data can be first segmented into several sentences according to periods, semicolons, etc. The segmented sentences are preprocessed by text classification, etc., to select the inspection rule sentences that can be converted into inspection rules. When performing sentence segmentation, a neural network model in a deep learning algorithm can be used to obtain better segmentation results.

[0038] For example, the standard text data "Rules for Writing Academic Papers" records the writing requirements for titles. After sentence segmentation and preprocessing of this part, the descriptive requirements that limit the content of the title are removed, and the checking rule statements about the title format are selected and retained, including: generally should not exceed 25 words; the paper should have a foreign language (mostly English) title; subtitles are allowed in the following situations. In addition, the standard text data "Rules for Writing Academic Papers" provides writing requirements for the font size and fonts used in academic papers. After sentence segmentation and preprocessing of this part, the checking rule statements about the title format are selected, including: Chinese title small 2 bold; English title 4 bold.

[0039] When understanding the obtained standard text data, a deep learning algorithm can be used to perform semantic recognition on the inspection rule statements. The deep learning algorithm can mark the inspection rule statements after semantic recognition with semantic labels indicating the standard inspection type and semantic labels indicating the compliance conditions. When performing semantic recognition, the convolutional neural network model in the deep learning algorithm can be used for feature extraction and classification.

[0040] The standard check type is used to indicate the standard type attribute of the field. For academic papers, the standard check type can include but is not limited to: Chinese title, subtitle, English title, author, unit, instructor, communication method, date, statement, Chinese abstract title, Chinese abstract content, English abstract title, English abstract content, keyword title, keyword content, various levels of titles, introduction, table of contents, core body, conclusion, acknowledgments, references, supplementary notes, etc.

[0041] Continuing with the previous example, the checking rule statement about the title "Generally it should not exceed 25 characters" is marked with the semantic label "Chinese title" indicating the standard check type and the semantic label "<25 characters" indicating the compliance condition; the checking rule statement about the title "The paper should have a foreign language (mostly English) title" is marked with the semantic label "English title" indicating the standard check type; the checking rule statement about the title "Sub-title is allowed in the following situations" is marked with the semantic label "sub-title" indicating the standard check type. The checking rule statement about the title "Chinese title small 2 bold" is marked with the semantic label "Chinese title" indicating the standard check type and the semantic labels "small 2" and "bold" indicating the compliance condition; the checking rule statement about the title "English title size 4 bold" is marked with the semantic label "English title" indicating the standard check type and the semantic labels "size 4" and "bold" indicating the compliance condition.

[0042] When understanding the obtained standard text data, the compliance conditions of the inspection rule statements with the same standard inspection type can be clustered, the clustered compliance conditions can be parsed to form a hierarchical relationship, and logically processed into inspection rules corresponding to the standard inspection type.

[0043] Continuing with the previous example, the compliance conditions of the check rule statement with the semantic tag "Chinese title" indicating the standard check type are clustered, and the clustering result is "{<25 characters, size 2, bold}"; the clustering result is parsed to form a hierarchical relationship, and the logical processing is "has & <25 characters & size 2 & bold". This check rule is used to indicate that when performing a rule check on "Chinese title", the Chinese title must be present, and then check whether the Chinese title is <25 characters. If it is less than 25 characters, it is compliant, and if it is not less than 25 characters, it is not compliant; then check whether the Chinese title is size 2. If it is, it is compliant, and if it is not, it is not compliant; finally check whether the Chinese title is bold. If it is bold, it is compliant, and if it is not, it is not compliant.

[0044] Continuing with the previous example, the compliance conditions of the check rule statement with the semantic tag "English title" indicating the standard check type are clustered, and the clustering result is "{size 4, bold}"; the clustering result is parsed to form a hierarchical relationship, and the logical processing is "has & size 4 & bold". This check rule is used to indicate that when performing a rule check on "English title", the English title must be present, and then check whether the English title is size 4. If it is, it is compliant, otherwise it is not compliant; then check whether the English title is bold. If it is bold, it is compliant, otherwise it is not compliant.

[0045] Continuing with the previous example, since the compliance condition of the check rule statement with the semantic tag "subtitle" indicating the standard check type is empty, the check rule can be processed as "(yes or no) & null" to indicate that when performing rule checking on "subtitle", the subtitle is not a required item and is always compliant.

[0046] This embodiment understands the standard text data, extracts the standard text data into multiple inspection rules corresponding to different standard inspection types, and generates the inspection rules into an inspection rule model based on the standard inspection type. When the format of the text data to be checked needs to be checked, the corresponding inspection rule in the inspection rule model can be called.

[0047] Step S220 , parsing the acquired text data to be checked, and dividing the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing result.

[0048] This embodiment then parses the text data to be checked, thereby splitting the text data to be checked into a plurality of blocks to be checked corresponding to different standard check types in preparation for format checking.

[0049] When parsing the acquired text data to be checked, a deep learning algorithm can be first used to perform semantic recognition on the text data to be checked, and the semantically recognized text data to be checked can be marked with semantic tags indicating the text structure. When performing semantic recognition, a recursive neural network in a deep learning algorithm can be used to perform text parsing, and a convolutional neural network can be used to perform relationship extraction and classification, and finally the text structure of the text data to be checked can be determined.

[0050] The text data to be checked usually includes the front part, the main body, the appendix, etc. The front part may include the title, bibliography, abstract, keywords, etc. The title may include the Chinese title, subtitle, English title, etc. The bibliography may include the author, unit, instructor, communication method, date, statement, etc. The abstract may include the Chinese abstract title, Chinese abstract content, English abstract title, English abstract content, etc. The keywords may include the keyword title, keyword content, etc. The main body may include the introduction, main body, conclusion, acknowledgments, references, etc. The main body may include the directory, titles at all levels, core main body, etc. The appendix may include supplementary explanations, etc. The text data to be checked may be semantically identified and semantic tags may be annotated on the text data to be checked according to the text structure.

[0051] For example, for a text data to be checked, based on the semantic recognition results, the Chinese title of the text can be marked with the semantic label "Chinese title" to indicate the standard check type; or, the quotation of the text body can be marked with the semantic label "introduction" to indicate the standard check type; or, the various levels of titles in the text body can be marked with the semantic labels "first-level title", "second-level title", "third-level title", etc. to indicate the standard check type.

[0052] For example, the body of the text data to be checked includes a part of the content: a title-core body (1 paragraph)-1.1 title-core body (2 paragraphs). Based on the semantic recognition result, "a title" is marked with the semantic label "first-level title" indicating the standard check type, "1.1 title" is marked with the semantic label "second-level title" indicating the standard check type, and "core body (1 paragraph)" and "core body (2 paragraphs)" are marked with the semantic label "core body" indicating the standard check type.

[0053] When parsing the acquired text data to be checked, the text data to be checked may be segmented according to the text structure, thereby dividing the text data to be checked into a plurality of blocks to be checked, each block to be checked corresponds to a corresponding standard inspection type.

[0054] Continuing with the previous example, when the main body of the text data to be checked includes a title-core body (1 natural paragraph)-1.1 title-core body (2 natural paragraphs), the "one title" is split into a block to be checked, and the block to be checked corresponds to the "first-level title" standard inspection type; the "core body (1 natural paragraph)" is split into a block to be checked, and the block to be checked corresponds to the "core body" standard inspection type; the "1.1 title" is split into a block to be checked, and the block to be checked corresponds to the "secondary title" standard inspection type; the "core body (2 natural paragraphs)" is split into a block to be checked, and the block to be checked corresponds to the "core body" standard inspection type.

[0055] This embodiment analyzes the text data to be checked and divides the text data to be checked into a plurality of blocks to be checked corresponding to different standard check types. When it is necessary to perform format check on the text data to be checked, the corresponding check rule can be called based on the standard check type.

[0056] Step S230 , performing format check on each block to be checked using corresponding check rules, and annotating the block to be checked that does not comply with the corresponding check rules.

[0057] Finally, this embodiment uses the check rule model to perform format check on the blocks to be checked, and then marks the blocks to be checked that do not comply with the format.

[0058] When performing format check on each block to be checked, the check rule corresponding to the same standard check type in the check rule model can be called according to the standard check type corresponding to the block to be checked.

[0059] For example, if the standard check type corresponding to a block to be checked is "Chinese title", the check rule "has &<25 characters & size 2 & bold" corresponding to the standard check type "Chinese title" in the rule check model is called to perform a format check on the block to be checked.

[0060] After performing a format check on each block to be checked, mark the non-compliant blocks to be checked based on the format check results.

[0061] Continuing with the previous example, first check whether it is <25 characters. If it is less than 25 characters, it is compliant. If it is not less than 25 characters, it is not compliant. Secondly, check whether it is size 2. If it is size 2, it is compliant. If not, it is not compliant. Finally, check whether it is bold. If it is bold, it is compliant. If not, it is not compliant.

[0062] When all are compliant, there are no format errors in the block to be checked, and no annotations are made to the block to be checked. When any item is not compliant, the reason for non-compliance is annotated in the block to be checked. For example, if the block to be checked is in "Kaiti", it does not meet the "bold" condition; the corresponding inspection rules can be annotated in the block to be checked, or the inspection rule statement "Chinese title small 2 bold" can be traced back according to the inspection rule, and the inspection rule statement can be annotated in the block to be checked.

[0063] In addition, the blocks to be checked that have format errors can be highlighted to remind the author to make modifications.

[0064] from Figure 2It can be seen from the method shown that the text format checking method provided by the present application understands the obtained standard text data and generates a checking rule model based on the understanding results; wherein the checking rule model includes multiple checking rules corresponding to different standard checking types; the obtained text data to be checked is parsed, and the text data to be checked is split into blocks to be checked corresponding to different standard checking types based on the parsing results; the corresponding checking rules are used to perform format checking on each block to be checked, and the blocks to be checked that do not meet the corresponding checking rules are annotated. The text format checking method provided by the present application can automatically generate a checking rule model based on the understanding of the standard text data, reducing the huge workload of manual input and avoiding manual input errors; it can automatically perform format checking based on the analysis of the text to be checked, reducing the complex process of manual proofreading and improving the accuracy of the inspection; it can automatically annotate format errors to assist the author in locating the format error area and assist the author in making format modifications. The method provided by the present application improves the accuracy and efficiency of text format checking.

[0065] In some cases, the standardized text data may include academic paper writing standards at the national, provincial, school, and college levels, and the precision of academic paper writing standards at each level varies. For example, the national academic paper writing standards stipulate that the title of the Chinese abstract should be in bold font size 5, while a certain school-level academic paper writing standard stipulates that the title of the Chinese abstract should be written at the top of the page and in size 5 Fangsong font. Another school-level academic paper writing standard stipulates that the title of the Chinese abstract should be written in the center, with two spaces between "Abstract" and "Key".

[0066] In order to extract universal format specifications from complex standard text data and avoid contradictory situations, in some embodiments of the present application, in the above method, step S210, understands the acquired standard text data, and generates an inspection rule model based on the understanding result, including: obtaining multiple standard text data, and prioritizing each standard text data according to the level of each standard text data; understanding each standard text data, and determining the compliance conditions corresponding to different standard inspection types in each standard text data; clustering the compliance conditions of the same standard inspection type, and cleaning the clustered compliance conditions based on priority; logically processing the retained compliance conditions to obtain the inspection rules corresponding to each standard inspection type; generating an inspection rule model according to different standard inspection types and corresponding inspection rules.

[0067] In this embodiment, multiple sets of standard text data are obtained, and the multiple sets of standard text data may include academic paper writing standards at various levels. The academic paper writing standards are prioritized according to the level of each academic paper writing standard.

[0068] For example, the national-level academic paper writing standards have the highest priority, followed by the provincial-level academic paper writing standards, then the school-level academic paper writing standards, and the college-level academic paper writing standards have the lowest priority.

[0069] In some embodiments, each standard text data is understood to determine the compliance conditions corresponding to different standard inspection types in each standard text data, including: performing sentence segmentation on each standard text data to filter out inspection rule statements; performing semantic recognition on each inspection rule statement to determine the initial inspection type and compliance conditions in each inspection rule statement; and converting the initial inspection type using the standard inspection type to obtain the standard inspection type for each inspection rule statement.

[0070] Using sentence segmentation, text classification and other preprocessing methods, we can filter out the inspection rule statements that can be converted into inspection rules from each standard text data. We can also perform semantic recognition on each inspection rule statement to determine the initial inspection type and compliance conditions indicated by the inspection rule statement.

[0071] For example, by understanding each standard text data, the initial check type and compliance conditions screened out can be: in the national academic paper writing standard, the initial check type is "Chinese abstract title", and the compliance conditions are "size 5" and "bold"; in a certain school-level academic paper writing standard, the initial check type is "Chinese abstract citation", and the compliance conditions are "full alignment", "size 5", and "Song font"; in another school-level academic paper writing standard, the initial check type is "Chinese abstract title", and the compliance conditions are "centered" and "leave two spaces between 'Abstract' and 'Main'".

[0072] The initial inspection type is converted using the standard inspection type to obtain the standard inspection type of each inspection rule statement, and then the corresponding relationship between the standard inspection type and the compliance condition in each specification text data is established.

[0073] Continuing with the previous example, in the national academic paper writing standard and another school-level academic paper writing standard, the initial inspection type "Chinese abstract title" is the standard inspection type. Therefore, no conversion is required, and the correspondence between the standard inspection type and the compliance conditions can be directly established. However, in a certain school-level academic paper writing standard, the initial inspection type "Chinese abstract citation" is inconsistent with the standard inspection type. In this case, the standard inspection type is used to convert the initial inspection type to obtain the correspondence between the standard inspection type and the compliance conditions. That is, for a certain school-level academic paper writing standard, the converted standard inspection type is "Chinese abstract title", and the compliance conditions are "full grid", "small 5", and "Fangsong".

[0074] In some embodiments, compliance conditions of the same standard check type are clustered, and the clustered compliance conditions are cleaned based on priority, including: clustering the compliance conditions of the same standard check type, and determining whether the clustered compliance conditions have conflicts with each other; if there are conflicts and the conflicting compliance conditions belong to different priorities, retaining the compliance conditions with the highest priority; if there are conflicts and the conflicting compliance conditions belong to the same priority, eliminating the conflicting compliance conditions; if there are no conflicts, retaining the deduplicated compliance conditions.

[0075] Cluster compliance conditions with the same standard check type.

[0076] Continuing with the previous example, after clustering, the compliance condition for the standard check type "Chinese abstract title" is "{size small 5, bold, top grid, size small 5, Fangsong, centered, two spaces between 'Abstract' and 'Important'}".

[0077] Based on the priority, the conflicting compliance conditions in the clustered compliance conditions are cleaned to retain the non-conflicting compliance conditions after clustering. The compliance conditions of the standard text data with low priority need to comply with the compliance conditions of the standard text data with high priority, and the compliance conditions of the standard text data with the same priority can be different. In order to obtain more general inspection rules, for the conflicting compliance conditions with different priorities, the one with a higher priority is retained, and the conflicting compliance conditions with the same priority are eliminated, and deduplication is performed to obtain the cleaned compliance conditions.

[0078] Continuing with the previous example, if "bold" and "Fangsong" conflict with each other, and the priority of "bold" (national level) is higher than that of "Fangsong" (school level), then "bold" with higher priority will be retained and "Fangsong" will be deleted; "top grid" and "center" conflict with each other, and "top grid" and "center" have the same priority (both are school level), then both "top grid" and "center" will be removed; "small 5" and "small 5" are repeated, so deduplication is performed. After cleaning, the compliance condition of the standard inspection type "Chinese abstract title" is "{small 5, bold, two spaces between 'Abstract' and 'Important'}".

[0079] In addition, the removed "top spacing" and "centering" are meaningful when performing format checks on specific academic papers. That is, the removed "top spacing" needs to be treated as a special condition when performing format checks on the school's academic papers corresponding to the standard specification text data; and the removed "centering" needs to be treated as a special condition when performing format checks on the school's academic papers corresponding to the standard specification text data. Therefore, special rules can be generated for the eliminated compliance conditions, the corresponding standard inspection types, and the corresponding specification text data, and the special rules are added to the inspection rule model for format checks on specific text data to be checked. It should be noted that when there are conflicting compliance conditions with different priorities, the deleted compliance conditions are completely erased because they do not meet the superior requirements. Unlike the eliminated compliance conditions, they will not be treated as special conditions to generate special rules.

[0080] The retained compliance conditions are parsed to form a hierarchical relationship and logically processed into inspection rules corresponding to the standard inspection types. Then, an inspection rule model is generated according to different standard inspection types and corresponding inspection rules.

[0081] Continuing with the previous example, the compliance condition of the standard check type "Chinese abstract title" is parsed to form a hierarchical relationship, and the logic is processed as "there is & small 5 & bold & two spaces between 'Abstract' and 'Issue'". This check rule is used to indicate that when performing rule checks on "Chinese abstract title", the Chinese abstract title is a must, and then check whether the Chinese abstract title is in small 5 size. If it is, it is compliant, otherwise it is not compliant; then check whether the Chinese abstract title is in bold. If it is, it is compliant, otherwise it is not compliant; finally check whether there are two spaces between 'Abstract' and 'Issue' in the Chinese abstract title. If it is, it is compliant, otherwise it is not compliant.

[0082] In some cases, the text structures of different text data to be checked are not necessarily the same. In order to call the inspection rule model to perform format checking on the text data to be checked with different text structures, the text data to be checked can be split into blocks to be checked based on the text structure, so as to perform targeted format checking on each block to be checked.

[0083] In some embodiments of the present application, in the above method, step S220, parses the acquired text data to be checked, and splits the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing results, including: acquiring the text data to be checked, performing semantic recognition on the text data to be checked, and determining the text structure of the text data to be checked; splitting the text data to be checked according to the text structure to obtain blocks to be checked corresponding to different standard inspection types.

[0084] The content of the text data to be queried is parsed through deep learning algorithms, and feature extraction and classification are automatically performed to form the text structure of the text data to be queried.

[0085] For example, for a text data to be checked, its text structure can be obtained based on the semantic recognition results: Chinese title, English title, author, unit, instructor, communication method, date, statement, Chinese abstract content, English abstract lead, English abstract content, keyword lead, keyword content, introduction, table of contents, titles at all levels, core body, conclusion, acknowledgments, references, and supplementary notes.

[0086] Based on the text structure of the text data to be checked, the text data to be checked is split into inspection blocks corresponding to different standard inspection types.

[0087] Continuing with the previous example, the text data to be checked is divided into the "Chinese title" check block, the "English title" check block, the "author" check block, the "unit" check block, the "supervisor" check block, the "communication method" check block, the "date" check block, the "statement" check block, the "Chinese abstract content" check block, the "English abstract lead" check block, the "English abstract content" check block, the "keyword lead" check block, the "keyword content" check block, the "introduction" check block, the "table of contents" check block, the "titles at all levels" check block, the "core body" check block, the "conclusion" check block, the "acknowledgements" check block, the "references" check block, and the "supplementary explanation" check block. Among them, the titles at all levels represent multiple titles at multiple levels, and each title is divided into a check block, so the "titles at all levels" check block can represent check blocks of multiple different titles. Among them, the core body can be understood as the core content within the text, which is usually interrupted into multiple parts by titles, and each part may include multiple paragraphs. Therefore, the "core body" inspection block can represent the inspection block of the core content interrupted by titles and including 1 or more paragraphs.

[0088] After the inspection rule model and the block to be checked are prepared, the inspection rule matching the standard inspection type is used to perform a format check on the block to be checked. In some embodiments of the present application, in the above method, step S230, each block to be checked is format checked using the corresponding inspection rule, and the block to be checked that does not meet the corresponding inspection rule is annotated, including: for each block to be checked, the inspection rule with the same standard inspection type as the block to be checked is called to perform a format check, if the block to be checked does not meet the inspection rule, the inspection rule is annotated on the block to be checked; after the format check is performed on each block to be checked, if there is an uncalled inspection rule in the inspection rule model, it is determined whether the standard inspection type corresponding to the uncalled inspection rule is a must-have item; the inspection rule with the standard inspection type corresponding to the uncalled inspection rule as a must-have item is annotated on the text data to be checked; all annotations of the text data to be checked are summarized, all annotations are counted according to the standard inspection type, and a statistical report is generated.

[0089] According to the standard inspection type corresponding to the block to be checked, the inspection rule corresponding to the standard inspection type in the inspection rule model is called to perform format inspection; if the block to be checked meets the inspection rule, no annotation processing is performed on the block to be checked; if the block to be checked does not meet the inspection rule, the non-compliance reason is annotated on the block to be checked. The non-compliance reason can be the corresponding inspection rule; since the inspection rule is a logical language, in order to make the author understand more clearly, the corresponding inspection rule statement can be traced back based on the inspection rule, and the inspection rule statement can be annotated on the block to be checked.

[0090] Continuing with the previous example, the check rule of standard check type matching is used to check the format of the block to be checked. If the format of the "Chinese title" block to be checked is "22 characters, size 2, bold", it does not meet the "Chinese title" check rule of "with & < 25 characters & small size 2 & bold", and the block to be checked is not compliant. Based on the check rule backtracking check rule statement, "Chinese title small size 2 bold" is annotated to the "Chinese title" block to be checked.

[0091] In some cases, the text structure of the text data to be checked may lack some mandatory items or some optional items. Therefore, after formatting each block to be checked, determine whether there are uncalled check rules in the check rule model, and determine whether the standard check type corresponding to the uncalled check rule is a mandatory item. If it is a mandatory item, an annotation is required to remind the author to add this part of the content. If it is an optional item, no processing is required.

[0092] Continuing with the previous example, after the format check of the block to be checked of the text data to be checked is completed, there are still "subtitle" check rules and "Chinese abstract title" check rules in the check rule model that have not been called. If the "subtitle" check rule is "(yes or no) & null", the "subtitle" standard check type is an optional item, so the text data to be checked is compliant with this check rule and no processing is performed. If the "Chinese abstract title" check rule is "yes & small 5 & bold & two spaces between 'Abstract' and 'Issue'", the "Chinese abstract title" standard check type is a required item, so the text data to be checked is not compliant with this check rule and needs to be annotated. When annotating, the annotation can be marked at the beginning or end of the text data to be checked, or the annotation can be marked at the corresponding position in the text data to be checked where the Chinese abstract title should be recorded. When annotating, you can directly annotate the check rule, or you can backtrack the check rule statement based on the check rule and annotate the check rule statement.

[0093] In order to assist the author or the overall format error of the text data to be checked, all annotations can be summarized and statistically analyzed based on the standard inspection type to generate a statistical report. The generated statistical report can be in the form of Json data, which is helpful for assisting the author to modify the format error and for comprehensively analyzing the format error situation.

[0094] In addition, in some cases, the rule checking model also includes special rules for format checking of specific academic papers. Therefore, after all the checking rules in the checking rule model are called to check the format of the text data to be checked, it is further determined whether there are special rules for format checking of the text data to be checked. Special rules include corresponding standard text data, standard checking types, and compliance conditions.

[0095] By performing semantic recognition on the text data to be checked, the specific level corresponding to the text data to be checked can be determined, and then the specific level can be matched with the standard text data to determine whether there are targeted special rules. If there are special rules, the special rules are used to check the format of the text data to be checked, and the special rules are annotated on the block to be checked when the block to be checked does not meet the special rules.

[0096] For example, after semantic recognition of the text data to be checked, it is determined that the text data to be checked is a school-level academic paper. The school-level academic paper corresponds to a school-level academic paper writing standard in the normative text data. The special rules include the compliance condition "top grid" for the school-level academic paper writing standard, with the standard check type of "Chinese abstract title". Therefore, this special rule is used to perform a format check on the "Chinese abstract title" block to be checked in the text data to be checked. If the block to be checked does not comply with the special rule, the special rule or the check rule statement based on the special rule backtracking is annotated on the block to be checked.

[0097] In some cases, the check rule model generated entirely based on automatic understanding of standard text data may have certain machine processing deviations; or the automatically generated check rule model fails to cover the most complete check rules. Therefore, in order to improve the intelligent improvement and update capabilities of the check rule model, in some embodiments of the present application, the above method also includes: obtaining the modification and review results of the text data to be checked; using the corresponding check rules to perform format checks on the modification and review results; if the modification and review results do not comply with the corresponding check rules, then understanding the modification and review results to generate update rules; caching the update rules in the feedback system, and when the check rule model is triggered to update, using the update rules to replace the corresponding check rules.

[0098] The method proposed in this embodiment is implemented after the text data to be checked has been modified and approved. The modification and approval is used to indicate that the changes to the text data to be checked have been finally approved and there are no format errors. The text data to be checked after the modification and approval is parsed, and based on the parsing result, the text data to be checked after the modification and approval is split into inspection blocks corresponding to different standard inspection types. The inspection rules in the inspection rule model are used to perform a format check on the corresponding inspection blocks. If the inspection blocks after the modification and approval do not comply with the inspection rules, it means that the inspection rules are deviated. The possibilities of deviations in inspection rules may include, but are not limited to: deviations in inspection rules due to insufficient accuracy in understanding the standard text data; deviations in generated inspection rules due to insufficient comprehensiveness of the obtained standard text data, etc.

[0099] Perform format analysis on the inspection blocks that do not meet the inspection rules, determine the format features of the inspection blocks, and perform logical processing on the format features to obtain the format conditions of the inspection blocks. Generate update rules based on the format conditions and the inspection type of the corresponding inspection blocks.

[0100] For example, the "Chinese abstract title" part has been added to the modified and approved text data to be checked. After a format check on the "Chinese abstract title" check block, it is determined that the check block does not meet the compliance condition of the "Chinese abstract title" check type "There are &small 5& bold& two spaces between 'Abstract' and 'Issue'". At this time, the "Chinese abstract title" part is format parsed to determine the format condition of the check block is "There are &small 5& bold& three spaces between 'Abstract' and 'Issue'". This means that the "two spaces" in the check rule is a deviation caused by automatic understanding. Based on the compliance condition of "There are &small 5& bold& three spaces between 'Abstract' and 'Issue'" and the standard check type "Chinese abstract title", an update rule is generated.

[0101] The updated rules can be counted and included in the feedback system. When the inspection rule model needs to be updated using feedback, the updated rules are used to replace the corresponding inspection rules, thereby updating the inspection rule model. Through the method of this embodiment, the modified review results are used as feedback, so that the inspection rule model can be automatically improved and updated.

[0102] Figure 3 A flowchart of a text format checking method according to another embodiment of the present invention is shown. Figure 3 As shown, the method may include the following steps S301 to S316.

[0103] Step S301, obtain multiple standard text data, and prioritize each standard text data according to its level; then go to step S302.

[0104] Step S302, perform sentence segmentation on each standard text data, filter out the inspection rule statements, perform semantic recognition on each inspection rule statement, determine the initial inspection type and compliance conditions in each inspection rule statement, use the standard inspection type to convert the initial inspection type, and obtain the standard inspection type of each inspection rule statement; go to step S303.

[0105] Step S303, clustering the compliance conditions of the same standard check type, and determining whether the clustered compliance conditions conflict with each other; if there is a conflict, proceed to step S304; if there is no conflict, proceed to step S307.

[0106] Step S304, determine whether the conflicting compliance conditions are of the same priority; if so, go to step S305; if not, go to step S306.

[0107] Step S305, eliminate conflicting compliance conditions, and generate special rules based on the eliminated compliance conditions, corresponding standard inspection types and corresponding standard text data; go to step S309.

[0108] Step S306, retain the compliance condition with the highest priority; go to step S308.

[0109] Step S307, retain the compliance conditions after deduplication; go to step S308.

[0110] Step S308, perform logical processing on the retained compliance conditions to obtain the inspection rules corresponding to each standard inspection type; and proceed to step S309.

[0111] Step S309, generating an inspection rule model according to the inspection rules and special rules corresponding to different standard inspection types; and proceeding to step S312.

[0112] Step S310, obtaining the text data to be checked, performing semantic recognition on the text data to be checked, and determining the text structure of the text data to be checked; then proceed to step S311.

[0113] Step S311, split the text data to be checked according to the text structure to obtain blocks to be checked corresponding to different inspection types; then go to step S312.

[0114] Step S312, for each block to be checked, call the check rule with the same check type as the block to be checked to perform format check. If the block to be checked does not meet the check rule, annotate the check rule to the block to be checked; then go to step S313.

[0115] Step S313, if there are uncalled check rules in the check rule model, determine whether the check type corresponding to the uncalled check rule is a must-have item, and annotate the check rule whose check type corresponding to the uncalled check rule is a must-have item on the text data to be checked; go to step S314;

[0116] Step S314, determine whether there are special rules for format checking of the text data to be checked. If there are special rules, use the special rules to check the format of the text data to be checked. When the block to be checked does not comply with the special rules, annotate the special rules on the block to be checked; then go to step S315.

[0117] Step S315, summarize all the annotations of the text data to be checked, count all the annotations according to the inspection type, and generate a statistical report; then go to step S316.

[0118] Step S316, obtain the modification and approval results of the text data to be checked, use the corresponding inspection rules to perform format inspection on the modification and approval results, if the modification and approval results do not comply with the corresponding inspection rules, understand the modification and approval results to generate update rules, cache the update rules in the feedback system, and when the inspection rule model is triggered to update, use the update rules to replace the corresponding inspection rules; go to step S309.

[0119] Figure 4 A schematic diagram of the structure of a text format checking device according to an embodiment of the present application is shown. Figure 4 As shown, the device 400 includes:

[0120] A rule model generating unit 410 is used to understand the obtained standard text data and generate an inspection rule model based on the understanding result; wherein the inspection rule model includes a plurality of inspection rules corresponding to different standard inspection types;

[0121] The to-be-checked text processing unit 420 is used to parse the acquired to-be-checked text data and split the to-be-checked text data into to-be-checked blocks corresponding to different standard inspection types based on the parsing result;

[0122] The checking and annotation unit 430 is used to perform format checking on each block to be checked using a corresponding checking rule, and to annotate the block to be checked that does not comply with the corresponding checking rule.

[0123] In some embodiments of the present application, in the above-mentioned device 400, the rule model generation unit 410 is specifically used to: obtain multiple copies of standard text data, and prioritize each standard text data according to the level of each standard text data; understand each standard text data, and determine the compliance conditions corresponding to different standard inspection types in each standard text data; cluster the compliance conditions of the same standard inspection type, and clean the clustered compliance conditions based on priority; perform logical processing on the retained compliance conditions to obtain inspection rules corresponding to each standard inspection type; generate an inspection rule model based on different standard inspection types and corresponding inspection rules.

[0124] In some embodiments of the present application, in the above-mentioned device 400, the rule model generation unit 410 is also specifically used to: perform sentence segmentation on each standard text data to filter out the inspection rule statements; perform semantic recognition on each inspection rule statement to determine the initial inspection type and compliance conditions in each inspection rule statement; use the standard inspection type to convert the initial inspection type to obtain the standard inspection type of each inspection rule statement.

[0125] In some embodiments of the present application, in the above-mentioned device 400, the rule model generation unit 410 is also specifically used to: cluster the compliance conditions with the same standard check type, and determine whether there are conflicts between the clustered compliance conditions; if there are conflicts and the conflicting compliance conditions belong to different priorities, retain the compliance conditions with the highest priority; if there are conflicts and the conflicting compliance conditions belong to the same priority, eliminate the conflicting compliance conditions; if there are no conflicts, retain the deduplicated compliance conditions.

[0126] In some embodiments of the present application, in the above-mentioned device 400, the text processing unit 420 to be checked is specifically used to: obtain the text data to be checked, perform semantic recognition on the text data to be checked, and determine the text structure of the text data to be checked; split the text data to be checked according to the text structure to obtain blocks to be checked corresponding to different inspection types.

[0127] In some embodiments of the present application, in the above-mentioned device 400, the inspection annotation unit 430 is specifically used to: for each block to be checked, call the inspection rule with the same inspection type as the block to be checked to perform a format check, if the block to be checked does not comply with the inspection rule, annotate the inspection rule on the block to be checked; after format checking each block to be checked, if there are uncalled inspection rules in the inspection rule model, determine whether the inspection type corresponding to the uncalled inspection rule is a must-have item; annotate the inspection rule with the inspection type corresponding to the uncalled inspection rule as a must-have item on the text data to be checked; summarize all annotations of the text data to be checked, count all annotations by inspection type, and generate a statistical report.

[0128] In some embodiments of the present application, the above-mentioned device 400 also includes: a model updating unit, which is specifically used to: obtain the modification and approval results of the text data to be checked; use corresponding inspection rules to perform format checks on the modification and approval results; if the modification and approval results do not comply with the corresponding inspection rules, then understand the modification and approval results and generate update rules; cache the update rules in the feedback system, and when the inspection rule model is triggered to update, use the update rules to replace the corresponding inspection rules.

[0129] It should be noted that the above-mentioned text format checking device 400 can implement the above-mentioned text format checking methods one by one, which will not be described in detail.

[0130] Figure 5 A schematic diagram showing the structure of a computer device according to an embodiment of the present application is shown. Figure 5 As shown, the internal structure of the computer device may include a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, the functions or steps of the text format checking method server are implemented.

[0131] In one embodiment, the computer device provided by the present application includes a memory and a processor, the memory stores a database and a computer program that can be run on the processor, and the processor implements the following steps when executing the computer program:

[0132] Understanding the obtained standard text data, and generating an inspection rule model based on the understanding result; wherein the inspection rule model includes a plurality of inspection rules corresponding to different standard inspection types;

[0133] Parse the acquired text data to be checked, and divide the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing result;

[0134] The corresponding checking rules are used to perform format checking on each block to be checked, and the blocks to be checked that do not meet the corresponding checking rules are annotated.

[0135] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0136] Understanding the obtained standard text data, and generating an inspection rule model based on the understanding result; wherein the inspection rule model includes a plurality of inspection rules corresponding to different standard inspection types;

[0137] Parse the acquired text data to be checked, and divide the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing result;

[0138] The corresponding checking rules are used to perform format checking on each block to be checked, and the blocks to be checked that do not meet the corresponding checking rules are annotated.

[0139] It should be noted that the above functions or steps that can be implemented by the computer device or computer-readable storage medium can refer to the relevant description in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0140] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0141] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0142] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A text format checking method, characterized in that: The method comprises: Understanding the obtained standard text data, and generating an inspection rule model based on the understanding result; wherein the inspection rule model includes a plurality of inspection rules corresponding to different standard inspection types; Parsing the acquired text data to be checked, and dividing the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing result; The format of each block to be checked is checked using the corresponding checking rule, and the block to be checked that does not comply with the corresponding checking rule is annotated.

2. The text format checking method according to claim 1, characterized in that: The step of understanding the obtained standard text data and generating a check rule model based on the understanding result includes: Acquire multiple copies of the standard text data, and prioritize each of the standard text data according to the level of each of the standard text data; Understanding each of the standard text data, and determining compliance conditions corresponding to different standard inspection types in each of the standard text data; Clustering the compliance conditions of the same standard inspection type, and cleaning the clustered compliance conditions based on the priority; Performing logical processing on the retained compliance conditions to obtain the inspection rules corresponding to each standard inspection type; The inspection rule model is generated according to the different standard inspection types and the corresponding inspection rules.

3. The text format checking method according to claim 2, characterized in that: The understanding of each of the standard text data and determining the compliance conditions corresponding to different standard inspection types in each of the standard text data include: Segmenting each of the standard text data to filter out inspection rule sentences; Performing semantic recognition on each of the inspection rule statements to determine the initial inspection type and the compliance condition in each of the inspection rule statements; The initial check type is converted using the standard check type to obtain the standard check type of each check rule statement.

4. The text format checking method according to claim 2, characterized in that: The clustering of the compliance conditions of the same standard inspection type and cleaning the clustered compliance conditions based on the priority includes: Clustering the compliance conditions of the same standard check type, and determining whether the clustered compliance conditions conflict with each other; If there are conflicting compliance conditions and the conflicting compliance conditions have different priorities, the compliance condition with the highest priority is retained; If there are conflicts between the compliance conditions and the conflicting compliance conditions have the same priority, the conflicting compliance conditions are eliminated; If there is no conflict, the compliance conditions after deduplication are retained.

5. The text format checking method according to claim 1, characterized in that: The step of parsing the acquired text data to be checked and dividing the text data to be checked into blocks to be checked corresponding to different standard inspection types based on the parsing result includes: Acquire the text data to be checked, perform semantic recognition on the text data to be checked, and determine the text structure of the text data to be checked; The to-be-checked text data is split according to the text structure to obtain the to-be-checked blocks corresponding to the different inspection types.

6. The text format checking method according to claim 1, characterized in that: The step of performing format check on each of the blocks to be checked using the corresponding check rule, and annotating the blocks to be checked that do not conform to the corresponding check rule, includes: For each of the blocks to be checked, calling the check rule with the same check type as the block to be checked to perform format check, if the block to be checked does not comply with the check rule, annotating the check rule to the block to be checked; After format checking is performed on each of the blocks to be checked, if there are uncalled checking rules in the checking rule model, it is determined whether the checking type corresponding to the uncalled checking rule is a required item; Annotate the inspection rule whose inspection type is a mandatory item corresponding to the inspection rule that has not been called on the text data to be inspected; Summarize all the annotations of the text data to be checked, count all the annotations according to the inspection type, and generate a statistical report.

7. The text format checking method according to claim 1, characterized in that: The method further comprises: Obtaining a modification and verification result of the text data to be checked; Using the corresponding checking rules to perform format checking on the modification and approval results; If the modified verification result does not conform to the corresponding inspection rule, the modified verification result is understood to generate an update rule; The update rules are cached in a feedback system, and when the check rule model is triggered to update, the update rules are used to replace the corresponding check rules.

8. A text format checking device, characterized in that: The device comprises: A rule model generating unit, configured to understand the obtained standard text data and generate an inspection rule model based on the understanding result; wherein the inspection rule model includes a plurality of inspection rules corresponding to different standard inspection types; A to-be-checked text processing unit, used for parsing the acquired to-be-checked text data, and splitting the to-be-checked text data into to-be-checked blocks corresponding to different standard inspection types based on the parsing result; The checking and annotation unit is used to perform format checking on each of the blocks to be checked using the corresponding checking rules, and to annotate the blocks to be checked that do not comply with the corresponding checking rules.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the text format checking method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is instructed by a processor, the computer program implements the steps of the text format checking method according to any one of claims 1 to 7.

Images