Test point intelligent extraction method and device for technical specification document

Abstract

This application discloses a method and apparatus for intelligent extraction of test points from technical specification documents, relating to the field of software testing. The method includes: acquiring the original content of the technical specification document; performing format noise cleanup on the original content to obtain target document content, wherein format noise cleanup includes deleting page number information that is not part of the main text and reassembling paragraphs separated by page numbers; detecting the structural type of the target document content, wherein the structural type includes a first type and a second type, the first type being a structure where technical requirements and test methods appear in pairs, and the second type being a structure including functional descriptions; generating test points according to the structural type using an extraction strategy corresponding to the structural type, obtaining a test point set; and outputting the test point set according to a preset standardized format. This application solves the technical problem that existing technologies cannot reliably generate complete and structurally accurate test point sets from technical specification documents.

Description

Technical Field

[0001] This application relates to the field of software testing, and more specifically, to a method and apparatus for intelligent extraction of test points from technical specification documents. Background Technology

[0002] In the field of software testing, test point extraction is a core step in test case design. Technical specification documents (such as technical standards, security specifications, and requirements specifications) are the primary source of test points, typically containing numerous technical requirements, functional descriptions, and verification methods. Currently, extracting test points from technical specification documents mainly relies on manual reading and writing, which suffers from low efficiency, high subjectivity, and the potential for overlooking key test points.

[0003] In recent years, automated test point generation technologies based on large language models have emerged, generating test points by inputting document content into a large language model. However, due to the complex structure and diverse presentation of technical specification documents, existing methods often struggle to accurately identify the document's internal structure when directly parsing the content, resulting in incomplete or missing content in the generated test point set. Summary of the Invention

[0004] This application provides a method and apparatus for intelligent extraction of test points from technical specification documents, which at least solves the technical problem that existing technologies cannot reliably generate a complete and structurally accurate set of test points from technical specification documents.

[0005] According to one aspect of the embodiments of this application, a method for intelligent extraction of test points from technical specification documents is provided, comprising: obtaining the original content of the technical specification document; performing format noise cleanup on the original content to obtain target document content, wherein the format noise cleanup includes deleting page number information of non-text content and reassembling paragraph content separated by page numbers; detecting the structural type of the target document content, wherein the structural type includes a first type and a second type, the first type being a structure in which technical requirements and test methods appear in pairs, and the second type being a structure including functional descriptions; generating test points according to the structural type using an extraction strategy corresponding to the structural type to obtain a test point set; and outputting the test point set according to a preset standardized format.

[0006] Optionally, test points are generated using an extraction strategy corresponding to the structure type, including: when the structure type is detected as the first type, establishing a correspondence between the technical requirements section and the test method section; extracting test content from the technical requirements section and test steps from the test method section according to the correspondence, and combining them to obtain test points.

[0007] Optionally, test points are generated using an extraction strategy corresponding to the structure type, including: when the structure type is detected as the second type, test points are intelligently generated based on a multi-dimensional testing framework, wherein the multi-dimensional testing framework includes forward testing dimension, reverse testing dimension, boundary testing dimension, permission testing dimension, and data verification testing dimension.

[0008] Optionally, the original content is cleaned of formatting noise to obtain the target document content, including: detecting multiple page number formats in the original content based on regular expressions and heuristic rules to obtain a set of suspected page numbers, including at least two of the following page number formats: pure number format, format with numbering identifier, and format enclosed by delimiters; selecting suspected page numbers located in the top or bottom area of ​​the page from the set of suspected page numbers as candidate page numbers based on the position characteristics of the page numbers at the top or bottom of the page; for each candidate page number, detecting whether the adjacent text of the candidate page number can be concatenated into a complete sentence after deleting the candidate page number; determining the candidate page numbers that cannot be concatenated into a complete sentence after deletion as page number information; deleting the page number information from the original content and re-concatenating the paragraph content separated by the page number information to obtain the target document content.

[0009] Optionally, the target document content is subjected to structure type detection to obtain the document's structure type, including: detecting whether the chapter titles of the target document content contain preset keywords related to the document's structure type, and obtaining keyword matching results, wherein the preset keywords are used to characterize the title features of the technical requirements chapter or the test method chapter; analyzing the hierarchical relationship between the technical requirements chapter and the test method chapter, and obtaining hierarchical analysis results, wherein the hierarchical analysis results are used to determine whether the technical requirements chapter and the test method chapter are at the same level and in parallel; and determining the structure type of the target document content according to preset judgment rules based on the keyword matching results and the hierarchical analysis results.

[0010] Optionally, the preset judgment rules include: when the keyword matching results indicate that both the technical requirements chapter and the testing method chapter exist, the hierarchical analysis results indicate that the technical requirements chapter and the testing method chapter are at the same level and are parallel, and the chapter numbers of the technical requirements chapter and the testing method chapter are the same or consecutive, the target document content is judged to belong to the first type; when the keyword matching results indicate that a function description chapter exists but there is no corresponding testing method chapter, the target document content is judged to belong to the second type.

[0011] Optionally, a correspondence is established between the technical requirements chapter and the testing method chapter. Based on the correspondence, test content is extracted from the technical requirements chapter and test steps are extracted from the testing method chapter, and these are combined to obtain test points. This includes: establishing a correspondence between the technical requirements chapter and the testing method chapter in the target document content based on the characteristic that the chapter numbers are the same or consecutive, resulting in at least one chapter pair; for each chapter pair, a functional description is extracted from the technical requirements chapter in the chapter pair as test content, and test steps are extracted from the testing method chapter in the chapter pair; the test content and test steps extracted from each chapter pair are combined to obtain test points.

[0012] Optionally, before extracting the functional description from the technical requirements section of the chapter pair as test content, it is checked whether the test method section of the chapter pair contains test step description features. Test step description features are used to characterize the verb usage patterns related to test operations. When the test method section is found to contain test step description features, the test method section is determined to be valid, and the operation of extracting test content continues. When the test method section is found not to contain test step description features, the test method section is determined to be invalid, and a prompt message is generated.

[0013] Optionally, extracting functional descriptions from the technical requirements section of the chapter pair as test content includes: extracting the original test content text from the technical requirements section as text to be processed; identifying redundant prefixes in the text to be processed, the redundant prefixes including preset guiding phrases; deleting the redundant prefixes from the text to be processed to obtain the text after removing redundancy; extracting key constraints from the text after removing redundancy as functional descriptions; and using the functional descriptions as test content.

[0014] Optionally, test points are intelligently generated based on a multi-dimensional testing framework, including: determining at least one functional point based on the functional point description information in the target document content; for each functional point, generating corresponding test points from the forward testing dimension, reverse testing dimension, boundary testing dimension, permission testing dimension, and data verification testing dimension, respectively, to obtain the test point set corresponding to the functional point.

[0015] Optionally, the forward testing dimension is used to verify scenarios where the function works normally; the reverse testing dimension is used to verify scenarios where abnormal input or illegal operation is rejected; the boundary testing dimension is used to verify scenarios where boundary conditions are handled; the permission testing dimension is used to verify the access control mechanism; and the data verification testing dimension is used to verify the integrity of the data.

[0016] Optionally, the test point set is output according to a preset standardized format, including: encapsulating each test point in the test point set into a JSON object in JSON format, wherein the JSON object includes the following fields: test module name, function point name, test case identifier, risk level, function description, and test point list; and combining all JSON objects into a JSON array as the standardized output result.

[0017] According to another aspect of the embodiments of this application, a test point intelligent extraction device for technical specification documents is also provided, comprising: an acquisition unit for acquiring the original content of the technical specification document; a first processing unit for cleaning up format noise in the original content to obtain target document content, wherein the format noise cleaning includes deleting page number information of non-text content and reassembling paragraph content separated by page numbers; a detection unit for detecting the structural type of the target document content, wherein the structural type includes a first type and a second type, the first type being a structure in which technical requirements and test methods appear in pairs, and the second type being a structure including functional descriptions; a second processing unit for generating test points according to the structural type using an extraction strategy corresponding to the structural type to obtain a test point set; and an output unit for outputting the test point set according to a preset standardized format.

[0018] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, which stores a computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located executes the above-described intelligent extraction method for test points of technical specification documents.

[0019] According to another aspect of the embodiments of this application, an electronic device is also provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the above-described intelligent extraction method for test points of technical specification documents.

[0020] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program or instructions, which, when executed by a processor, implement the above-described intelligent extraction method for test points of technical specification documents.

[0021] As described above, when processing the original content, this application first performs a format noise cleanup step, deleting page number information that does not belong to the main text and reassembling the paragraphs separated by page numbers to obtain the target document content. This step helps eliminate interfering information in the document, improves the continuity of paragraph content, and provides semantically complete input for subsequent processing. Secondly, after obtaining the target document content, this application further detects the structural type of the target document content, distinguishing between the first type (structures where technical requirements and test methods appear in pairs) and the second type (structures including functional descriptions). This step helps identify the inherent structural characteristics of the document, thus avoiding the use of a single processing method for all documents. Finally, based on the detected structural type, test points are generated using an extraction strategy corresponding to the structural type. Using different extraction strategies for different types of documents helps ensure that the generated test points match the actual structure of the document, thus avoiding omissions or formatting chaos caused by structural mismatches, and thereby solving the technical problem that existing technologies cannot stably generate a complete and structurally accurate set of test points from technical specification documents. Attached Figure Description

[0022] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0023] Figure 1 This is a flowchart of an optional intelligent method for extracting test points from technical specification documents according to an embodiment of this application;

[0024] Figure 2 This is a flowchart of the format noise cleanup process provided in the embodiments of this application;

[0025] Figure 3 This is a flowchart of the dual-mode structure detection process provided in the embodiments of this application;

[0026] Figure 4 This is a flowchart of the first type of document direct extraction mode provided in the embodiments of this application;

[0027] Figure 5 This is a flowchart of the second type of document intelligent generation mode provided in the embodiments of this application;

[0028] Figure 6 This is a schematic diagram of a test point intelligent extraction device for technical specification documents according to an embodiment of this application. Detailed Implementation

[0029] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0031] It should also be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, and displayed data) collected in this application are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with the relevant laws, regulations, and standards of the relevant regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation entry points are provided for users to choose to authorize or refuse. For example, this system has interfaces with relevant users or organizations. Before obtaining relevant information, a request to obtain the information needs to be sent to the aforementioned user or organization through the interface, and the relevant information is obtained only after receiving consent from the aforementioned user or organization.

[0032] According to an embodiment of this application, an embodiment of a method for intelligent extraction of test points for technical specification documents is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0033] According to the embodiments of this application, a test point intelligent extraction system for technical specification documents (hereinafter referred to as the test point intelligent extraction system) can be used as the execution subject of the test point intelligent extraction method for technical specification documents in the embodiments of this application. The system can be a software system or an embedded system combining software and hardware. Of course, the execution subject of the method in the embodiments of this application can also be other forms of execution subject, such as devices, equipment, etc. Those skilled in the art should know that this application does not particularly limit the specific form of the execution subject.

[0034] Figure 1 This is a flowchart of an optional intelligent method for extracting test points from technical specification documents according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0035] Step S101: Obtain the original content of the technical specification document.

[0036] For example, the original content of a technical specification document can refer to the initial data of the document before any processing, including all characters, numbers, symbols, and formatting marks in the document. A technical specification document can be a portable document format file, an office document format file, or other types of text file, which includes, but is not limited to, core content such as technical requirements, functional descriptions, and test methods, while also carrying non-text information such as page numbers, headers, footers, and tables of contents.

[0037] In some embodiments, the intelligent test point extraction system can read portable document format files uploaded by users through a document parsing interface. For example, the system calls a document parsing library to extract all text content from the file, including body paragraphs, table cells, header areas, footer areas, and page number markers. The system can also organize the extracted text content into an original content sequence according to page order, with each page including all text information for that page and retaining the position coordinates of characters on the page. This method helps to completely obtain the original information of the document, facilitating subsequent noise identification based on positional features.

[0038] In other embodiments, the intelligent test point extraction system can read office document format files from local or network storage. For example, the system uses a document processing component to open the file, traverses all paragraphs, text boxes, headers, footers, tables, and other elements in the document, and extracts the text from each element in the order it is read to form the original content string. The system also records the page number and position information of each text element in the document for subsequent page number recognition and paragraph merging operations. This method is suitable for obtaining original content from technical specification documents in office document format, maintaining the integrity and traceability of the original content.

[0039] Step S102: Perform format noise cleanup on the original content to obtain the target document content. The format noise cleanup includes deleting page number information that is not part of the main text and reassembling the paragraph content that was separated by page numbers.

[0040] For example, the intelligent test point extraction system can clean up formatting noise in the original content to remove page number information that is not part of the main text and reassemble paragraphs that have been split by page numbers. Formatting noise refers to auxiliary information in the document that is not part of the main text, including page numbers indicating page order, document numbers, and pagination marks. While this information is helpful in document layout, it can interfere with the correct identification of paragraph boundaries during test point extraction, causing the same paragraph to be incorrectly segmented into multiple fragments. By removing this noise and restoring paragraph continuity, the intelligent test point extraction system helps provide semantically complete document content for subsequent structural detection.

[0041] Step S103: Detect the structure type of the target document content. The structure type includes a first type and a second type. The first type is a structure in which technical requirements and testing methods appear in pairs. The second type is a structure that includes functional descriptions.

[0042] For example, the intelligent test point extraction system can detect the structural type of the target document content to distinguish whether the document belongs to the first type or the second type. The first type refers to a structure in a technical specification document where "technical requirements" and "test methods" appear in pairs. That is, each functional point or technical indicator in the document contains both a technical requirement description and a corresponding test method description, usually located at the same chapter level and numbered together. The second type refers to a structure in the document that includes functional descriptions. That is, the document mainly describes functional characteristics, business rules, or technical indicator requirements, but does not provide corresponding test methods or verification steps. By identifying the document's structural type, the intelligent test point extraction system provides a basis for subsequently selecting differentiated extraction strategies.

[0043] Step S104: Based on the structure type, generate test points using an extraction strategy corresponding to the structure type to obtain a test point set.

[0044] For example, the intelligent test point extraction system generates test points based on the detected structural type using an extraction strategy corresponding to that structural type, thus obtaining a test point set. The extraction strategy refers to the different test point acquisition methods used for documents of different structural types, including direct extraction mode and intelligent generation mode. When the document is of the first type, the system uses direct extraction mode, directly extracting test content from paired "Technical Requirements" and "Testing Methods" sections and combining them into test points. When the document is of the second type, the system uses intelligent generation mode, proactively generating test points based on the functional descriptions in the document. Through differentiated extraction strategies, the intelligent test point extraction system ensures that the test point generation method matches the actual structure of the document.

[0045] Step S105: Output the test point set according to the preset standardized format.

[0046] For example, the intelligent test point extraction system outputs the generated test point set according to a preset standardized format. This preset standardized format refers to a predefined structured data format used to unify the representation of test points, ensuring that the output test points have a consistent field structure and data type. This standardized format facilitates the import, export, and sharing of test points across different test management systems, avoiding integration difficulties caused by format differences.

[0047] In some embodiments, test points are generated using an extraction strategy corresponding to the structure type, including: when the structure type is detected to be a first type, establishing a correspondence between the technical requirements section and the test method section; extracting test content from the technical requirements section and test steps from the test method section according to the correspondence, and combining them to obtain test points.

[0048] For example, when the intelligent test point extraction system detects that a document belongs to the first type, it establishes a correspondence between the technical requirements section and the test method section. Based on this correspondence, it extracts test content from the technical requirements section and test steps from the test method section, combining them to obtain test points. The technical requirements section refers to the content module in the document that describes functional indicators or constraints, while the test method section refers to the content module of verification methods that correspond to the technical requirements. By establishing a chapter-to-chapter mapping, the intelligent test point extraction system helps ensure that each technical requirement has a corresponding test method, avoiding the omission of test points.

[0049] In some embodiments, the intelligent test point extraction system establishes a correspondence based on the same or consecutive features of chapter numbers. For example, the system traverses the document's chapter structure, and when it detects two chapters with the same number belonging to the technical requirements type and the test method type respectively, it establishes a mapping relationship between them. For example, chapters "6.1.1 Technical Requirements" and "6.1.1 Test Methods" have the same number, so the system establishes them as a chapter pair. When it detects chapters with consecutive numbers belonging to two different types, the system also establishes a mapping relationship between them, such as "6.1.1 Technical Requirements" and "6.1.2 Test Methods". After establishing the chapter pairs, the system extracts functional descriptions from the technical requirements chapters as test content and extracts operational steps from the test method chapters as test steps, combining the two to form complete test points.

[0050] In other embodiments, the intelligent test point extraction system establishes correspondences through fuzzy matching of title keywords. When the chapter numbers in a document are not standardized or are missing, the system can establish correspondences by matching title content. The system extracts the core keywords of the technical requirements chapter titles (such as "identity authentication" or "user login"), searches the document for chapters whose titles contain the same core keywords and whose type is "test method," and establishes a mapping relationship between the two. For example, if the technical requirements chapter title is "Identity Authentication Functional Requirements," the system searches for chapters whose titles contain "identity authentication" and whose type is "test method," such as "Identity Authentication Test Verification." After establishing the correspondence, the system extracts the test content and test steps respectively and combines them into test points. This method is suitable for documents with non-standard chapter numbers and helps improve the robustness of establishing correspondences.

[0051] In some embodiments, test points are generated using an extraction strategy corresponding to the structure type, including: when the structure type is detected to be the second type, test points are intelligently generated based on a multi-dimensional testing framework, wherein the multi-dimensional testing framework includes forward testing dimension, reverse testing dimension, boundary testing dimension, permission testing dimension and data verification testing dimension.

[0052] In some embodiments, the intelligent test point extraction system intelligently generates multi-dimensional test points based on the description of core business functions. The system can extract core functions (such as "user login") and related constraints (such as "account locked after three consecutive failures") from the functional description. For positive testing dimensions, the system generates test points that verify a valid username and password for successful login; for negative testing dimensions, the system generates test points that verify an incorrect username, incorrect password, or empty credentials for login failure; for boundary testing dimensions, the system generates test points that verify the account is locked after the first, second, and third failures, and that login is rejected after the fourth failure. The system organizes the generated test points by dimension to form a test point set corresponding to that functional point.

[0053] In other embodiments, the intelligent test point extraction system intelligently generates multi-dimensional test points based on the description of data management functions. The system extracts data operation types (such as "batch delete user data") and related business rules from the function description. For positive testing dimensions, the system generates test points to verify that successfully deleting selected valid data; for negative testing dimensions, the system generates test points to verify that the deletion operation is rejected when no data is selected; for permission testing dimensions, the system generates test points to verify that users without deletion permissions cannot perform deletion operations; for data verification testing dimensions, the system generates test points to verify that data is physically deleted from the database or logically marked as deleted after the deletion operation. The system ensures that each test point contains clear operation steps and expected results, making it easy for testers to use directly. This approach is suitable for documents that only have function descriptions without readily available test methods, and can generate a comprehensive set of test points.

[0054] In some embodiments, format noise cleanup is performed on the original content to obtain the target document content, including: detecting multiple page number formats in the original content based on regular expressions and heuristic rules to obtain a set of suspected page numbers, wherein the multiple page number formats include at least two of the following: pure number format, format with number identifier, and format enclosed by delimiters; filtering suspected page numbers located in the top or bottom area of ​​the page from the set of suspected page numbers as candidate page numbers based on the position characteristics of the page numbers at the top or bottom of the page; for each candidate page number, detecting whether the adjacent text of the candidate page number can be concatenated into a complete sentence after deleting the candidate page number; determining the candidate page numbers that cannot be concatenated into a complete sentence after deletion as page number information; deleting the page number information from the original content and re-concatenating the paragraph content separated by the page number information to obtain the target document content.

[0055] For example, the test point intelligent extraction system cleans up format noise in the original content through steps such as page number format recognition, position filtering, and semantic judgment. The suspected page number set refers to the set of text lines that may be page numbers initially identified through format rules, including pure numeric formats (such as the numbers "6" and "12" appearing alone), formats with number identifiers (such as the document number "Q / GDW10597—2022" containing forward slashes, hyphens, or spaces), and formats enclosed by delimiters (such as "-12-" and "[6]" with hyphens or square brackets on both sides). Candidate page numbers can be text lines further filtered from the suspected page number set based on position features. The test point intelligent extraction system helps improve the accuracy of page number recognition through a multi-level filtering mechanism.

[0056] Figure 2 This is a flowchart of the format noise cleanup process provided in an embodiment of this application. Figure 2As shown, the process includes: format recognition, which identifies the original format of the document; content extraction, which extracts the text content from the document; page number recognition, which identifies the page number information in the document; page number cleanup, which removes the identified page numbers from the document; paragraph merging, which reassembles the paragraphs that were separated by page numbers; and format standardization, which converts the cleaned content into a standardized format.

[0057] In some embodiments, the test point intelligent extraction system processes document content that can be converted from portable document formats. For example, the system first uses regular expressions to match lines of pure numbers (such as "6" and "12") and lines of numbers enclosed in separators (such as "-12-"), adding these text lines to a set of suspected page numbers. Using coordinate information obtained during document parsing, the system filters text lines located within 1 cm of the top or bottom of the page from the set of suspected page numbers as candidate page numbers. For each candidate page number, the system deletes the line and checks whether its adjacent text can be connected naturally: if the end of the previous line and the beginning of the next line can be joined to form a complete sentence, the line is confirmed as page number information and the preceding and following paragraphs are merged; if they cannot be connected naturally, the line is retained as body text. The system continues processing the next candidate page number until all candidate page numbers have been processed.

[0058] In other embodiments, the intelligent test point extraction system processes page number formats containing document numbers. The system uses regular expressions to match number formats containing numbers, forward slashes, hyphens, and spaces, such as "Q / GDW10597—2022" and "GB / T25000.51-2016," adding the matched text lines to a set of suspected page numbers. After filtering candidate page numbers based on their position at the top or bottom of the page, for each candidate page number, the system deletes the line and attempts to concatenate the preceding and following paragraphs. The system also checks whether deletion results in obvious grammatical anomalies in the paragraph, such as lowercase initial letters or incomplete subject-verb structures. If such anomalies exist, the system re-evaluates the judgment result for that line and, if necessary, restores it to the main text. This method effectively handles page numbers with complex number formats and avoids accidentally deleting standard number references in the document.

[0059] In some embodiments, structural type detection is performed on the target document content to obtain the document's structural type, including: detecting whether the chapter titles of the target document content contain preset keywords related to the document's structural type, and obtaining keyword matching results, wherein the preset keywords are used to characterize the title features of the technical requirements chapter or the test method chapter; analyzing the hierarchical relationship between the technical requirements chapter and the test method chapter, and obtaining hierarchical analysis results, wherein the hierarchical analysis results are used to determine whether the technical requirements chapter and the test method chapter are at the same level and in parallel; and determining the structural type of the target document content according to preset judgment rules based on the keyword matching results and the hierarchical analysis results.

[0060] For example, the intelligent test point extraction system performs structural type detection on the target document content through keyword matching and hierarchical relationship analysis. Preset keywords refer to characteristic words used to identify chapter types, including terms like "technical requirements," "technical indicators," and "functional requirements" for chapters representing technical requirements, and terms like "test methods," "verification methods," and "detection methods" for chapters representing testing methods. Keyword matching results refer to the chapter type labeling results obtained by detecting whether the chapter titles contain the above keywords. Hierarchical analysis results refer to the relationship judgment results between chapters obtained by analyzing the hierarchical structure of chapter numbers, used to determine whether two chapters are parallel relationships at the same parent chapter. The intelligent test point extraction system combines the above two analysis results and determines the document's structural type according to preset rules.

[0061] In some embodiments, the intelligent test point extraction system processes documents with standardized chapter numbers. The system traverses the chapter titles at all levels within the document. For each chapter title, it checks whether it contains keywords such as "technical requirements" or "technical indicators." If so, the chapter is labeled as a "technical requirements" chapter. It also checks whether it contains keywords such as "test methods" or "verification methods." If so, the chapter is labeled as a "test method" chapter. The system obtains the chapter number hierarchy; for example, "6.1.1" represents a third-level chapter, and "6.1.2" represents a parallel chapter at the same level as "6.1.1." When the system detects that a "technical requirements" chapter and a "test method" chapter are parallel chapters at the same level under the same parent chapter (such as "6.1.1" and "6.1.2"), and their numbers are the same or consecutive, the system combines the keyword matching results and the hierarchy analysis results to determine that the document belongs to the first type.

[0062] In other embodiments, the intelligent test point extraction system handles documents with non-standard or missing chapter numbers. The system first identifies the technical requirements and testing method chapters through keyword matching, recording the order in which each chapter appears in the document. When chapter numbers are non-standard, the system determines whether they are siblings based on their position in the document tree, for example, by inferring the parent-child relationship of chapters through the indentation level of the chapter titles or the document outline level. The system also checks whether the core keywords of the titles of the technical requirements and testing method chapters are consistent; for example, "identity authentication technical requirements" and "identity authentication testing methods" share the same core keyword "identity authentication." When the system detects that technical requirements and testing method chapters have the same core keyword and are at the same level in the document structure, the system determines the document belongs to the first type; when only functional description chapters are detected and no corresponding testing method chapters are detected, the system determines the document belongs to the second type. This method is suitable for technical specification documents with inconsistent formats and helps improve the adaptability of structure type detection.

[0063] In some embodiments, the preset judgment rules include: when the keyword matching result indicates that there are both a technical requirements chapter and a test method chapter, the hierarchical analysis result indicates that the technical requirements chapter and the test method chapter are at the same level and are parallel, and the chapter numbers of the technical requirements chapter and the test method chapter are the same or consecutive, the target document content is determined to belong to the first type; when the keyword matching result indicates that there is a function description chapter but there is no test method chapter corresponding to the function description chapter, the target document content is determined to belong to the second type.

[0064] For example, the intelligent test point extraction system determines the structural type of the target document content according to preset judgment rules. The first type of judgment requires three conditions to be met simultaneously: keyword matching results indicate that the document contains both a technical requirements section and a test method section; hierarchical analysis results indicate that these two types of sections are at the same level and parallel, and their section numbers are the same or consecutive. The second type of judgment condition is: keyword matching results indicate that the document contains a functional description section, but there is no corresponding test method section. Through these rules, the intelligent test point extraction system accurately distinguishes between the two structural types of the document.

[0065] Figure 3 This is a flowchart of the dual-mode structure detection process provided in an embodiment of this application. Figure 3 As shown, the process includes: keyword matching, which detects whether the chapter titles contain preset keywords; hierarchical analysis, which analyzes the hierarchical relationship between chapters (such as parallel relationships at the same level); semantic extraction, which detects the semantic features in the chapter content; and determining whether the document belongs to the first type or the second type based on the above analysis results.

[0066] In some embodiments, the test point intelligent extraction system processes documents with a standardized chapter structure. The system detects chapters "6.1.1 Identity Authentication Technical Requirements" and "6.1.2 Identity Authentication Test Methods," and keyword matching results indicate the simultaneous existence of both a technical requirements chapter and a test method chapter. Analysis of the chapter numbers reveals that both are third-level chapters, belonging to the same parent chapter "6.1" and are parallel to each other at the same level, with consecutive numbers (6.1.1 and 6.1.2), satisfying all the judgment conditions for the first type. Based on this, the system determines that the document belongs to the first type. The system continues to check for other pairings between technical requirements chapters and test method chapters; if all conditions are met, the first type determination is maintained.

[0067] In some other embodiments, the test point intelligent extraction system processes documents that only contain functional descriptions. The system detects chapters such as "User Login Function" and "Password Modification Function," and keyword matching results indicate the existence of functional description chapters. The system traverses all chapters in the document and does not detect any test method chapters whose titles contain keywords such as "Test Method" or "Verification Method," nor does it detect any test method chapters corresponding to the functional description chapters. Based on the second type's judgment rules, the system determines that the document belongs to the second type. The system records the judgment result and triggers the corresponding intelligent generation strategy.

[0068] In some embodiments, a correspondence is established between technical requirement chapters and test method chapters. Test content is extracted from the technical requirement chapters and test steps are extracted from the test method chapters based on the correspondence, and these are combined to obtain test points. This includes: establishing a correspondence between technical requirement chapters and test method chapters in the target document content based on the characteristic that the chapter numbers are the same or consecutive, to obtain at least one chapter pair; for each chapter pair, a functional description is extracted from the technical requirement chapters in the chapter pair as test content, and test steps are extracted from the test method chapters in the chapter pair; the test content and test steps extracted from each chapter pair are combined to obtain test points.

[0069] Figure 4 This is a flowchart of the first type of document direct extraction mode provided in the embodiments of this application. For example... Figure 4 As shown, the process includes: chapter pair identification, that is, identifying the technical requirements chapter and the test method chapter and establishing a pairing relationship; function point extraction, that is, extracting function points from the technical requirements chapter; test content extraction, that is, extracting test content; redundant prefix removal, that is, deleting redundant prefixes from the test content; test point conversion, that is, converting the test method description into test points in a standard format; and risk assessment, that is, automatically determining the risk level (high, medium, low) based on the importance of the function.

[0070] In some embodiments, the intelligent test point extraction system processes chapter pairs with consecutive numbers. The system traverses the document's chapter structure, and when it detects that the chapter title of chapter number "6.1.1" contains the keyword "technical requirements," it identifies it as a technical requirements chapter. The system continues to search for the next consecutively numbered chapter under the same parent chapter, i.e., chapter number "6.1.2," and checks whether its title contains the keyword "test method." If so, the chapter is identified as a test method chapter. The system establishes these two chapters as a chapter pair. For this chapter pair, the system extracts the functional description from the technical requirements chapter as test content and the operational steps from the test method chapter as test steps, then combines the two to generate test points.

[0071] In some embodiments, before extracting the functional description from the technical requirements section of the chapter pair as test content, it is detected whether the test method section of the chapter pair contains test step description features, which are used to characterize the verb usage patterns related to test operations; when the test method section is detected to contain test step description features, the test method section is determined to be valid, and the operation of extracting test content continues; when the test method section is detected not to contain test step description features, the test method section is determined to be invalid, and a prompt message is generated.

[0072] For example, before extracting test content, the intelligent test point extraction system first checks whether the test method section contains test step description features. Test step description features refer to the verb usage patterns in the text that characterize test operation behaviors, including keywords indicating specific test actions such as "check," "verify," "test," "confirm," and "execute." By detecting these features, the intelligent test point extraction system determines whether the test method section truly contains executable test steps, rather than merely containing result descriptions or vague statements. This verification step helps ensure that only complete test method sections are used for test point generation, avoiding the generation of invalid or incomplete test points.

[0073] In some embodiments, the intelligent test point extraction system performs verb feature detection on the test method section. The system scans the text content of the test method section, extracts verb words, and counts the occurrence frequency of test-related verbs such as "check," "verify," "test," "confirm," "execute," "input," "click," and "select." When at least one test-related verb is detected, the system determines that the test method section contains test step description features, classifies the section as valid, and allows the subsequent extraction of test content to continue. For example, for a test method section containing "check whether the system identifies and authenticates the user's identity," the system detects the verb "check" and classifies the section as valid.

[0074] In other embodiments, the intelligent test point extraction system handles chapters lacking test steps. When the system scans the text content of a test method chapter and fails to detect any test-related verbs, or if it only contains result-oriented descriptions such as "should meet requirements" or "must comply with specifications" without specific operational instructions, the system determines that the test method chapter does not contain test step description features. In this case, the system determines the chapter is invalid, stops the extraction operation, and generates a prompt message informing the user that the test method description for that chapter is incomplete or formatted incorrectly, suggesting manual review or supplementation of test steps. The system records this prompt message in the processing log and skips the current chapter pair, continuing to process the next chapter pair. This method effectively identifies incomplete test method chapters, avoids generating invalid test points, and provides an entry point for manual intervention, ensuring the quality of the final test point set.

[0075] In some embodiments, extracting functional descriptions from the technical requirements section of a chapter pair as test content includes: extracting original test content text from the technical requirements section as text to be processed; identifying redundant prefixes in the text to be processed, the redundant prefixes including preset guiding phrases; deleting redundant prefixes from the text to be processed to obtain text after redundancy removal; extracting key constraints from the text after redundancy removal as functional descriptions; and using the functional descriptions as test content.

[0076] For example, the intelligent test point extraction system extracts the original test content text from the technical requirements section as the text to be processed. It then obtains the functional description through steps such as identifying and removing redundant prefixes and extracting key constraints. Redundant prefixes refer to common introductory phrases in the technical requirements text, such as the opening parts of normative statements like "the system should," "the application software should," and "the product should." Key constraints refer to the core content of the technical requirements, including substantive descriptions such as functional behavior, performance indicators, and limitations. By removing redundant prefixes and retaining core constraints, the intelligent test point extraction system makes the extracted functional description more concise and focused on the key points of test verification.

[0077] In some embodiments, the intelligent test point extraction system processes technical requirement text containing standard introductory statements. The system extracts the original text from the technical requirements section, such as "Application software systems should identify and authenticate the identity of logged-in users," and treats this as text to be processed. The system identifies the redundant prefix "Application software systems should" in the text to be processed and marks it as content to be deleted. The system removes this redundant prefix from the text to obtain the redundant-free text "Identify and authenticate the identity of logged-in users." The system further extracts key constraints from this text, identifying the core functional point as "identity identification and authentication," and uses this as a functional description. The system uses this functional description as test content, which is then combined with subsequent test steps to generate test points.

[0078] In other embodiments, the intelligent test point extraction system processes technical requirement text containing multiple constraints. The system extracts the original text from the technical requirements section, such as "The system should ensure that the user password is no less than 8 characters long and should contain numbers and letters," and treats this as the text to be processed. The system identifies the redundant prefix "the system should," removes it, and obtains "ensure that the user password is no less than 8 characters long and should contain numbers and letters." The system further breaks down the multiple constraints in the text, identifying two key constraints: the password length must be no less than 8 characters, and the password must contain numbers and letters. The system merges these two constraints or uses them separately as functional descriptions to generate corresponding test content. For descriptions involving high-risk functions such as fund payment and authentication, the system can also mark the risk level in subsequent steps, providing a priority reference for test execution. This approach can handle complex multi-constraint technical requirements, ensuring the completeness of the functional description.

[0079] In some embodiments, intelligently generating test points based on a multi-dimensional testing framework includes: determining at least one functional point based on the functional point description information in the target document content; and for each functional point, generating corresponding test points from the forward testing dimension, reverse testing dimension, boundary testing dimension, permission testing dimension, and data verification testing dimension to obtain a set of test points corresponding to that functional point.

[0080] For example, the intelligent test point extraction system determines at least one functional point based on the functional point description information in the target document content, and generates corresponding test points from five testing dimensions for each functional point. A functional point refers to an independently testable functional unit extracted from the document description, such as "user login," "data export," and "permission management."

[0081] Figure 5 This is a flowchart of the second type of intelligent document generation mode provided in the embodiments of this application. For example... Figure 5 As shown, the process includes: function identification, which is to identify the core function from the function point description; module determination, which is to determine the test module to which the function belongs; content summary, which is to summarize the function description; five-dimensional generation, which is to generate corresponding test points from the forward testing dimension, reverse testing dimension, boundary testing dimension, permission testing dimension and data verification testing dimension respectively; and expected result addition, which is to add expected results to each test point.

[0082] In some embodiments, the intelligent test point extraction system generates multi-dimensional test points for the user authentication function. The system extracts the function point "User Login" from the function description. For the positive testing dimension, the system generates the test point "Login using a valid username and correct password, expecting successful login and redirection to the main page." For the negative testing dimension, the system generates the test point "Login using a valid username and incorrect password, expecting login failure and displaying an error message." For the boundary testing dimension, the system generates the test point "Entering the wrong password three times consecutively, expecting the account to be locked after the third failure, and unable to log in even with the correct password on the fourth attempt," based on the description "Account locked for 30 minutes after three consecutive failures." For the permission testing dimension, the system generates the test point "Unlogged users directly access pages requiring login, expecting redirection to the login page." For the data verification testing dimension, the system generates the test point "After successful login, check if the user information in the session matches the logged-in account." The system combines these five generated test points into a test point set for this function.

[0083] In other embodiments, the intelligent test point extraction system generates multi-dimensional test points for the batch data operation function. The system extracts the function point "batch delete user data" from the function description. For the positive testing dimension, the system generates the test point "select multiple valid user data records and perform batch deletion, expecting the selected data to be successfully deleted." For the negative testing dimension, the system generates the test point "click the batch delete button without selecting any data, expecting the deletion operation to be blocked and prompting 'Please select data first.'" For the boundary testing dimension, the system generates the test points "select a single data record and perform batch deletion, expecting that data to be successfully deleted"; "select the maximum number of data records and perform batch deletion, expecting successful deletion and normal system performance." For the permission testing dimension, the system generates the test point "users without deletion permissions log into the system, expecting the batch delete button to be invisible or grayed out." For the data verification testing dimension, the system generates the test point "after performing batch deletion and refreshing the data list, expecting the deleted data to no longer be displayed." This method ensures that even if the documentation does not provide readily available testing methods, a comprehensive set of test points can be generated.

[0084] In some embodiments, the forward testing dimension is used to verify scenarios where the function works normally; the reverse testing dimension is used to verify scenarios where abnormal input or illegal operation is rejected; the boundary testing dimension is used to verify scenarios where boundary conditions are handled; the permission testing dimension is used to verify the access control mechanism; and the data verification testing dimension is used to verify the integrity of the data.

[0085] For example, the intelligent test point extraction system employs a five-dimensional testing framework to guide the intelligent generation of test points, with each dimension verifying different aspects of the function. The forward testing dimension verifies the expected behavior of the function under normal input and standard operation, helping to ensure the function executes correctly according to the requirements specification. The reverse testing dimension verifies the function's fault-tolerance capabilities under abnormal input, illegal operation, or unexpected usage scenarios, helping the system correctly reject invalid requests and provide reasonable prompts. The boundary testing dimension verifies the function's behavior under critical value conditions, including maximum, minimum, and critical quantity boundary cases. The permission testing dimension verifies the access control mechanism under different roles or permission levels, helping to prevent unauthorized users from performing restricted operations. The data verification testing dimension verifies the integrity, accuracy, and consistency of data before and after the operation.

[0086] In some embodiments, the intelligent test point extraction system applies a five-dimensional testing framework to the file upload function. For the forward testing dimension, the system generates test points to verify that files conforming to the format and size requirements can be successfully uploaded. For the reverse testing dimension, the system generates test points to verify that uploads of empty files, virus-infected files, or files with incorrect formats are rejected and error messages are displayed. For the boundary testing dimension, the system generates test points to verify the system's behavior when uploading 0-byte files, files reaching the size limit, or files exceeding the size limit. For the permission testing dimension, the system generates test points to verify that users without upload permissions do not see the upload button or that their upload operation is rejected. For the data verification testing dimension, the system generates test points to verify the integrity of the file stored on the server after a successful upload, and the correctness of the filename and path.

[0087] In other embodiments, the intelligent test point extraction system applies a five-dimensional testing framework to the data query function. For the forward testing dimension, the system generates test points to verify that valid input query conditions return the correct result set. For the reverse testing dimension, the system generates test points to verify that the system can correctly filter or escape malicious input such as illegal characters or SQL injection code. For the boundary testing dimension, the system generates test points to verify scenarios such as empty query results, the query reaching the maximum number of returned results, and pagination to the last page. For the permission testing dimension, the system generates test points to verify that users can only query data within their authorized scope, and that data they do not have permission to access is invisible. For the data verification testing dimension, the system generates test points to verify that the data fields returned by the query are complete and that the data values ​​are consistent with the database records.

[0088] In some embodiments, the test point set is output according to a preset standardized format, including: encapsulating each test point in the test point set into a JSON object in JSON format, wherein the JSON object includes the following fields: test module name, function point name, test case identifier, risk level, function description, and test point list; and combining all JSON objects into a JSON array as the standardized output result.

[0089] For example, the intelligent test point extraction system encapsulates each test point in the test point set into a JSON object according to the JSON format, and combines all JSON objects into a JSON array as the standardized output. JSON objects are a lightweight data exchange format that organizes data using key-value pairs, offering good readability and cross-platform compatibility. A JSON object contains the following fields: Test Module Name (to identify the functional module to which the test point belongs), Function Point Name (to identify the specific function being tested), Test Case Identifier (to uniquely identify each test point), Risk Level (to identify the importance of the test point), Function Description (to describe a summary of the function being tested), and Test Point List (to store the specific test steps and expected results). Through the unified JSON format, the intelligent test point extraction system enables the output test points to be read and parsed by different test management systems.

[0090] In some embodiments, the intelligent test point extraction system generates JSON-formatted test point output for the identity authentication module. The system names the test module "Identity Authentication," the function point "User Identity Authentication," and the test case identifier "AUTH_001" according to rules. The risk level is set to "Medium" based on functional importance. The function description field stores test content extracted from the technical requirements section, such as "The system should perform identity authentication for logged-in users." The test point list field stores test steps extracted from the test method section, such as "Check whether the system has an identity authentication function, expecting users to input identity information upon login." After encapsulating this test point into a JSON object, the system combines it with other test point JSON objects from the same module into a JSON array, outputting it as a separate JSON file. The test management system can parse this JSON file, import test points in batches, and automatically create corresponding test cases.

[0091] In other embodiments, the intelligent test point extraction system generates multi-dimensional JSON-formatted test point outputs for the user login function module. The system generates multiple test points for the same function, corresponding to different dimensions such as forward testing, reverse testing, and boundary testing. Each test point is an independent entry in the test point list. The test case identifiers generated by the system increment sequentially according to the test point order, such as "LOGIN_001" corresponding to a forward test point and "LOGIN_002" corresponding to a reverse test point. The function description field remains unchanged, while the test point list field contains multiple test step entries. The system organizes all test points for this function in the same JSON object, enabling unified management and association of test points from different dimensions of the same function. The output JSON array can be imported into test management systems such as ZenTao, Jira, and TestRail, achieving automated data entry of test points and subsequent test execution tracking. This output method realizes end-to-end automated data flow from technical specifications to the test management system.

[0092] According to another aspect of the embodiments of this application, a test point intelligent extraction device for technical specification documents is also provided. Figure 6 This is a schematic diagram of a test point intelligent extraction device for technical specification documents according to an embodiment of this application, such as... Figure 6 As shown, the device includes: an acquisition unit 601 for acquiring the original content of a technical specification document; a first processing unit 602 for cleaning up formatting noise in the original content to obtain the target document content, wherein the formatting noise cleaning includes deleting page number information that is not part of the main text and reassembling the paragraph content that is separated by page numbers; a detection unit 603 for detecting the structural type of the target document content, wherein the structural type includes a first type and a second type, the first type being a structure in which technical requirements and test methods appear in pairs, and the second type being a structure that includes functional descriptions; a second processing unit 604 for generating test points according to the structural type using an extraction strategy corresponding to the structural type, thereby obtaining a test point set; and an output unit 605 for outputting the test point set according to a preset standardized format.

[0093] For example, the second processing unit 604 includes: a first processing subunit, configured to establish a correspondence between the technical requirements section and the test method section when the structure type is detected to be the first type; and a second processing subunit, configured to extract test content from the technical requirements section and test steps from the test method section according to the correspondence, and combine them to obtain test points.

[0094] For example, the second processing unit 604 includes: a third processing subunit, used to intelligently generate test points based on a multi-dimensional testing framework when the structure type is detected to be the second type, wherein the multi-dimensional testing framework includes a forward testing dimension, a reverse testing dimension, a boundary testing dimension, a permission testing dimension, and a data verification testing dimension.

[0095] For example, the first processing unit 602 includes: a first detection subunit, used to detect multiple page number formats in the original content based on regular expressions and heuristic rules to obtain a set of suspected page numbers, wherein the multiple page number formats include at least two of the following: pure number format, format with number identifier, and format enclosed by delimiters; a filtering subunit, used to filter out suspected page numbers located in the top or bottom area of ​​the page from the set of suspected page numbers based on the position characteristics of the page numbers located at the top or bottom of the page, as candidate page numbers; a second detection subunit, used to detect, for each candidate page number, whether the adjacent text of the candidate page number can be concatenated into a complete sentence after deleting the candidate page number; and to determine the candidate page numbers that cannot be concatenated into a complete sentence after deletion as page number information; and a fourth processing subunit, used to delete the page number information from the original content and re-concatenate the paragraph content divided by the page number information to obtain the target document content.

[0096] For example, the detection unit 603 includes: a third detection subunit, used to detect whether the chapter titles of the target document content contain preset keywords related to the document structure type, and obtain keyword matching results, wherein the preset keywords are used to characterize the title features of the technical requirements chapter or the test method chapter; an analysis subunit, used to analyze the hierarchical relationship between the technical requirements chapter and the test method chapter, and obtain hierarchical analysis results, wherein the hierarchical analysis results are used to determine whether the technical requirements chapter and the test method chapter are at the same level and in parallel; and a first determination subunit, used to determine the structure type of the target document content according to preset judgment rules based on the keyword matching results and the hierarchical analysis results.

[0097] For example, the preset judgment rules include: when the keyword matching result indicates that there are both a technical requirements chapter and a test method chapter, the hierarchical analysis result indicates that the technical requirements chapter and the test method chapter are at the same level and are parallel, and the chapter numbers of the technical requirements chapter and the test method chapter are the same or consecutive, the target document content is judged to belong to the first type; when the keyword matching result indicates that there is a function description chapter but there is no test method chapter corresponding to the function description chapter, the target document content is judged to belong to the second type.

[0098] For example, the second processing subunit includes: a first processing module, configured to establish a correspondence between the technical requirements section and the test method section in the target document content based on the characteristic that the section numbers are the same or consecutive, to obtain at least one section pair; a second processing module, configured to extract functional descriptions from the technical requirements section of each section pair as test content, and extract test steps from the test method section of each section pair; and a third processing module, configured to combine the test content and test steps extracted from each section pair to obtain test points.

[0099] For example, the intelligent test point extraction device for technical specification documents further includes: a first detection unit, used to detect whether the test method section in the chapter pair contains test step description features, the test step description features being used to characterize the verb usage patterns related to test operations; a first determination unit, used to determine that the test method section is valid when the test method section is detected to contain test step description features, and continue to perform the operation of extracting test content; and a second determination unit, used to determine that the test method section is invalid when the test method section is detected not to contain test step description features, and generate a prompt message.

[0100] For example, the second processing module includes: an extraction submodule for extracting the original test content text from the technical requirements section as the text to be processed; an identification submodule for identifying redundant prefixes in the text to be processed, the redundant prefixes including preset guiding phrases; a deletion submodule for deleting redundant prefixes from the text to be processed to obtain the text after redundancy removal; a first processing submodule for extracting key constraints from the text after redundancy removal as functional descriptions; and a second processing submodule for using the functional descriptions as test content.

[0101] For example, the third processing subunit includes: a function point determination module, used to determine at least one function point based on the function point description information in the target document content; and a test point generation module, used to generate corresponding test points for each function point from the forward testing dimension, the reverse testing dimension, the boundary testing dimension, the permission testing dimension, and the data verification testing dimension, respectively, to obtain a test point set corresponding to the function point.

[0102] For example, the forward testing dimension is used to verify scenarios where the function works normally; the reverse testing dimension is used to verify scenarios where abnormal input or illegal operation is rejected; the boundary testing dimension is used to verify scenarios where boundary conditions are handled; the permission testing dimension is used to verify access control mechanisms; and the data verification testing dimension is used to verify data integrity.

[0103] For example, output unit 605 includes: an encapsulation subunit for encapsulating each test point in the test point set into a JSON object in JSON format, wherein the JSON object includes the following fields: test module name, function point name, test case identifier, risk level, function description, and test point list; and an object combination subunit for combining all JSON objects into a JSON array as a standardized output result.

[0104] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, which stores a computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located executes the above-described intelligent extraction method for test points of technical specification documents.

[0105] According to another aspect of the embodiments of this application, an electronic device is also provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the above-described intelligent extraction method for test points of technical specification documents.

[0106] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program or instructions, which, when executed by a processor, implement the above-described intelligent extraction method for test points of technical specification documents.

[0107] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0108] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0109] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0110] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0111] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0112] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0113] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A test point intelligent extraction method for technical specification documents, characterized in that, include: Obtain the original content of the technical specification document; The original content is cleaned of formatting noise to obtain the target document content. The formatting noise cleanup includes deleting page number information that is not part of the main text and reassembling the paragraph content that is separated by page numbers. The structure type of the target document content is detected, wherein the structure type includes a first type and a second type, the first type is a structure in which technical requirements and testing methods appear in pairs, and the second type is a structure that includes a functional description; Based on the structure type, test points are generated using an extraction strategy corresponding to the structure type, resulting in a test point set. The test point set is output according to a preset standardized format.

2. The method of claim 1, wherein, Test points are generated using an extraction strategy corresponding to the structure type, including: If the structure type is detected to be the first type, establish a correspondence between the technical requirements section and the test method section; Based on the correspondence, test content is extracted from the technical requirements section and test steps are extracted from the test methods section, and then combined to obtain test points.

3. The method of claim 1, wherein, Test points are generated using an extraction strategy corresponding to the structure type, including: When the structure type is detected to be the second type, test points are intelligently generated based on a multi-dimensional testing framework, wherein the multi-dimensional testing framework includes forward testing dimension, reverse testing dimension, boundary testing dimension, permission testing dimension and data verification testing dimension.

4. The method of claim 1, wherein, The original content is cleaned of formatting noise to obtain the target document content, including: Based on regular expressions and heuristic rules, multiple page number formats in the original content are detected to obtain a set of suspected page numbers. The multiple page number formats include at least two of the following: pure number format, format with number identifier, and format wrapped by delimiters. Based on the position of the page number at the top or bottom of the page, suspected page numbers located in the top or bottom area of ​​the page are selected from the suspected page number set as candidate page numbers. For each candidate page number, check whether the adjacent text of the candidate page number can be pieced together to form a complete sentence after deleting the candidate page number; determine the candidate page number information if it cannot be pieced together to form a complete sentence after deletion. The page number information is deleted from the original content, and the paragraph content that was separated by the page number information is reassembled to obtain the target document content.

5. The method of claim 1, wherein, Detecting the structural type of the target document content includes: Detect whether the chapter titles of the target document content contain preset keywords related to the document structure type, and obtain keyword matching results, wherein the preset keywords are used to characterize the title features of the technical requirements chapter or the test method chapter; Analyze the hierarchical relationship between the technical requirements section and the test method section to obtain hierarchical analysis results, wherein the hierarchical analysis results are used to determine whether the technical requirements section and the test method section are at the same level and parallel. Based on the keyword matching results and the hierarchical analysis results, the structural type of the target document content is determined according to preset judgment rules.

6. The method of claim 5, wherein, The preset judgment rules include: When the keyword matching result indicates that the technical requirements section and the test method section exist simultaneously, the hierarchical analysis result indicates that the technical requirements section and the test method section are at the same level and are parallel, and the chapter numbers of the technical requirements section and the test method section are the same or consecutive, the content of the target document is determined to belong to the first type. When the keyword matching result indicates that there is a functional description section but no corresponding test method section, the content of the target document is determined to belong to the second type.

7. The method according to claim 2, characterized in that, Establish a correspondence between the technical requirements section and the testing method section. Based on this correspondence, extract test content from the technical requirements section and test steps from the testing method section, and combine them to obtain test points, including: For the technical requirements section and the test method section in the target document content, establish a correspondence between the technical requirements section and the test method section based on the feature that the section numbers are the same or consecutive, and obtain at least one section pair; For each of the chapter pairs, extract the functional description from the technical requirements chapter of the chapter pair as test content, and extract the test steps from the test method chapter of the chapter pair; The test content extracted from each chapter pair is combined with the test steps to obtain test points.

8. The method of claim 7, wherein, Before extracting functional descriptions from the technical requirements section of the aforementioned section as test content, the following is also included: Detect whether the test method chapter in the chapter pair contains test step description features, which are used to characterize the verb usage patterns related to test operations; When the test method section is detected to contain the test step description feature, the test method section is determined to be valid, and the operation of extracting test content continues. When it is detected that the test method section does not contain the test step description feature, the test method section is determined to be invalid, and a prompt message is generated.

9. The method of claim 7, wherein, Extract functional descriptions from the technical requirements section of the aforementioned chapter pair as test content, including: Extract the original test content text from the technical requirements section as the text to be processed; Identify redundant prefixes in the text to be processed, the redundant prefixes including preset guiding phrases; The redundant prefix is ​​removed from the text to be processed to obtain the text after redundancy removal; Extract key constraints from the text after removing redundancy as the functional description; The functional description is used as the test content.

10. The method of claim 3, wherein, Test points are intelligently generated based on a multi-dimensional testing framework, including: Based on the functional point description information in the target document content, at least one functional point is determined; For each functional point, corresponding test points are generated from the forward testing dimension, reverse testing dimension, boundary testing dimension, permission testing dimension, and data verification testing dimension, respectively, to obtain the test point set corresponding to that functional point.

11. The method of claim 10, wherein, The forward testing dimension is used to verify scenarios where the function works normally; the reverse testing dimension is used to verify scenarios where abnormal input or illegal operation is rejected; the boundary testing dimension is used to verify scenarios where boundary conditions are handled; the permission testing dimension is used to verify the access control mechanism; and the data verification testing dimension is used to verify the integrity of the data.

12. The method of claim 1, wherein, The test point set is output according to a preset standardized format, including: Each test point in the test point set is encapsulated into a JSON object in JSON format, wherein the JSON object includes the following fields: test module name, function point name, test case identifier, risk level, function description, and test point list; Combine all JSON objects into a JSON array as the standardized output.

13. A device for intelligent extraction of test points from technical specification documents, characterized in that, include: The acquisition unit is used to acquire the original content of the technical specification document; The first processing unit is used to perform format noise cleanup on the original content to obtain the target document content, wherein the format noise cleanup includes deleting page number information of non-text content and re-sewing together the paragraph content separated by page numbers. The detection unit is used to detect the structural type of the target document content, wherein the structural type includes a first type and a second type, the first type is a structure in which technical requirements and testing methods appear in pairs, and the second type is a structure that includes a functional description; The second processing unit is used to generate test points according to the structure type using an extraction strategy corresponding to the structure type, thereby obtaining a test point set. The output unit is used to output the set of test points according to a preset standardized format.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the computer program is executed, the device containing the computer-readable storage medium performs the intelligent extraction method for test points of technical specification documents as described in any one of claims 1 to 12.

15. An electronic device, comprising: It includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the intelligent extraction method for test points of technical specification documents as described in any one of claims 1 to 12.

16. A computer program product, characterised in that, It includes a computer program or instructions that, when executed by a processor, implement the intelligent extraction method for test points of technical specification documents as described in any one of claims 1 to 12.

Images