Method and device for extracting Chinese text test requirements based on document standards and semantics

Abstract

The invention discloses a Chinese text test requirement extraction method based on document standards and semantics, which comprises the following steps of: for a software requirement document with a large text quantity, disassembling a document structure according to the corresponding standards, and establishing a keyword and synonym word library, a sentence pattern library and a semantic library to carry out formatting disassembly on software requirement description; and establishing a test demand feature library, and generating a corresponding test demand document by matching the conditions of software demand disassembly with the test demand features. The time for manually reading software requirements and establishing test requirements is greatly shortened, the correctly disassembled test requirement text description term is more standard, the semantic ambiguity of manually writing documents is reduced, the execution of test contents by testers is facilitated, and the test is ensured to be smoothly carried out.

Description

Technical Field

[0001] The present invention belongs to the technical field of software testing, and in particular relates to a method and device for extracting Chinese text testing requirements based on document standards and semantics. Background Art

[0002] The increasing size of software has led to an increasing amount of software requirement documents. The increasing integration and complexity of software requirements have made it increasingly important to determine software testing requirements. The industry generally models software requirements and then uses the model to generate software requirements and its internal testing requirements, which is not feasible for third-party testing organizations that use software requirement documents as input.

[0003] With the continuous improvement of software industry documentation standards, the descriptive terms of software requirement documents have tended to be more similar. Therefore, it is of practical significance to format and decompose software requirement documents based on industry standards and text semantics to generate corresponding test requirements. Summary of the invention

[0004] The purpose of the present invention is to provide a method and device for extracting Chinese text test requirements based on document standards and semantics, which can realize the decomposition of software requirements and basic test requirement analysis and basic test data design based on software development document standards and text semantic analysis.

[0005] To solve the above problems, the technical solution of the present invention is: A method for extracting Chinese text test requirements based on document standards and semantics, comprising: Based on the document standard, the structure of the software requirement text is disassembled to obtain the subject structure of the software requirement text and the text, table and figure with titles under each subject structure; Format and disassemble the software requirement text according to the preset keywords and synonyms library, sentence library, and semantic library to obtain the requirement description in a standard format; According to the preset test requirement feature library, the requirement description is structured and the corresponding test requirement document is generated according to the preset test requirement text format; Among them, the key structure of the software requirement text is located according to the preset keywords and synonyms library, the text matching of keywords and synonyms is carried out according to the title level, and the successfully matched title and its subordinate texts, tables, and figures are regarded as matching objects of the same keyword.

[0006] According to an embodiment of the present invention, the subject structure of the software requirement text is targeted for extraction based on the directory structure of the word text; Based on the table class that can be read by the word reference library, extract the table from the paragraphs under the topic structure, and extract the preceding paragraphs simultaneously, and combine them according to the text style of the paragraphs to determine the caption of the corresponding table; name the extracted table information with the prefix table format or table note format or other customized format of the document paragraph; Based on the image class that can be read by the word reference library, images are extracted from the paragraphs under the topic structure, and the subsequent paragraphs are extracted simultaneously. The captions of the corresponding images are determined based on the text style combination of the paragraphs; the extracted image information is named in the document paragraph with the prefix image format or caption format or other customized formats.

[0007] According to an embodiment of the present invention, a database is constructed to establish an equivalent relationship between keywords and synonyms, with keywords as nodes, similar synonyms are set, and similarities are attached; the keywords include software function requirement description, function, performance, interface, security requirement, adaptability requirement and computer resource requirement; The software requirement text is matched with keywords according to the title level. If the match is successful, it is determined to be the keyword corresponding to the requirement type; if the match is unsuccessful, the software requirement text is matched with synonyms.

[0008] According to an embodiment of the present invention, formatting and decomposing the software requirement text according to the preset sentence library and semantic library further includes: According to the preset disassembly standards, the software requirement text is disassembled including prepositions, nouns, verbs, auxiliary words, adjectives, and punctuation marks; According to the punctuation priority, the software requirement text is segmented using the punctuation with the highest priority to obtain multiple complete sentences; Remove the prepositions in each sentence, match the text according to the verb library and the corresponding sentence pattern; split the matched text into corresponding phrases according to the sentence pattern settings, and mark the attributes; All the phrases obtained by disassembly are disassembled according to the auxiliary words to obtain the combination of adjectives and nouns; Chinese and Arabic numerals are used to extract the conditions of adjectives and obtain the corresponding conditional formulas.

[0009] According to an embodiment of the present invention, each test requirement in the test requirement feature library includes at least a requirement description feature, a test method type, and a test value formula; Each requirement description features associated sentences and adjective features.

[0010] According to an embodiment of the present invention, the conditional formula is matched with the demand description feature; Get the corresponding test method type according to the matching requirement description features; Generate a test input value or a test input condition according to the matching test value formula and condition formula.

[0011] According to an embodiment of the present invention, a test requirement text description is generated according to the matched test requirements; According to the preset test requirement document format, the test requirement text description is output in sequence according to the test method type.

[0012] A Chinese text test requirement extraction device based on document standards and semantics, comprising: The document preprocessing module is configured to disassemble the structure of the software requirement text based on the document standard to obtain the subject structure of the software requirement text and the text, table, and figure with titles under each subject structure; The document decomposition module is configured to format and decompose the software requirement text according to preset keywords and synonyms, sentence structure and semantic libraries to obtain a standard format requirement description; The requirement structured output module is configured to perform structured processing on the requirement description according to the preset test requirement feature library, and generate the corresponding test requirement document according to the preset test requirement text format; Among them, the document disassembly module is also configured to locate the key structure of the software requirement text according to the preset keywords and synonyms library, perform text matching of keywords and synonyms according to the title level, and use the successfully matched title and its subordinate texts, tables, and figures as matching objects for the same keyword.

[0013] According to an embodiment of the present invention, the document preprocessing module includes a subject extraction unit, a table extraction unit and a picture extraction unit; The subject extraction unit is configured to perform target extraction on the subject structure of the software requirement text based on the directory structure of the word text; The table extraction unit is configured to extract the table from the paragraphs under the topic structure based on the table class that can be read by the word reference library, and simultaneously extract the preceding paragraph, and combine and determine the caption of the corresponding table according to the text style of the paragraph; and name the extracted table information in the document paragraph in the table format or table note format with the prefix or in other customized formats; The image extraction unit is configured to extract images from paragraphs under the topic structure based on the image class that can be read by the word reference library, and simultaneously extract subsequent paragraphs, and determine the caption of the corresponding image based on the text style combination of the paragraph; the extracted image information is named as a document paragraph in a prefixed image format or a caption format or other customized formats.

[0014] According to an embodiment of the present invention, the demand structured output module is further configured to: Get the corresponding test method type according to the matching requirement description features; Generate a test input value or a test input condition according to the matching test value formula and condition formula; Generate a text description of the test requirements based on the matched test requirements; According to the preset test requirement document format, the test requirement text description is output in sequence according to the test method type.

[0015] Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the prior art: In one embodiment of the present invention, a Chinese text test requirement extraction method based on document standards and semantics is used to disassemble the document structure according to the corresponding standards for software requirement documents with large text volume, establish a keyword and synonym library, a sentence library and a semantic library to format and disassemble the software requirement description; establish a test requirement feature library, and generate a corresponding test requirement document by matching the conditions of software requirement disassembly with the test requirement features. The time for manually reading software requirements and establishing test requirements is greatly reduced, and the correctly disassembled test requirement text description terms are more standardized, reducing the semantic ambiguity of manually written documents, which is conducive to the execution of test content by testers and ensures the smooth progress of the test. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flow chart of a method for extracting Chinese text test requirements based on document standards and semantics in one embodiment of the present invention; Figure 2 This is a block diagram of a device for extracting Chinese text test requirements based on document standards and semantics in one embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following is a further detailed description of a method and device for extracting Chinese text test requirements based on document standards and semantics proposed by the present invention in conjunction with the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description and claims.

[0018] Please see Figure 1 This embodiment provides a method for extracting Chinese text test requirements based on document standards and semantics, comprising the following steps: S1: Decompose the structure of the software requirement text based on the document standard to obtain the subject structure of the software requirement text and the text, table, and figure with titles under each subject structure; S2: Format and disassemble the software requirement text according to the preset keywords and synonyms library, sentence library, and semantic library to obtain the requirement description in a standard format; S3: According to the preset test requirement feature library, the requirement description is structured and the corresponding test requirement document is generated according to the preset test requirement text format; Among them, the key structure of the software requirement text is located according to the preset keywords and synonyms library, the text matching of keywords and synonyms is carried out according to the title level, and the successfully matched title and its subordinate texts, tables, and figures are regarded as matching objects of the same keyword.

[0019] Typically, a software development document contains the following basic parts: ‌Introduction‌: Overview and goals of the project, purpose and scope of the document, version history of the requirements document‌.

[0020] ‌Project Overview‌: The name and identifier of the project, an introduction to the relevant stakeholders, the context and background of the project‌.

[0021] ‌Functional Requirements‌: Detailed description of the functions that the system needs to implement, usage scenarios and use case descriptions, and the priority and importance of functional requirements‌.

[0022] ‌Non-functional requirements‌: performance requirements (such as response time, throughput), security and privacy requirements, reliability and availability requirements, compatibility and scalability requirements‌.

[0023] ‌Interface requirements‌: user interface design and user experience requirements, interface specifications with other systems or components, and format requirements for reports and logs‌.

[0024] ‌Data requirements‌: data model and database design, data input and output formats, data security and consistency requirements‌.

[0025] ‌Constraints and Limitations‌: Technical and hardware limitations, regulatory and standards compliance requirements, project time and budget constraints‌.

[0026] ‌Quality attributes‌: Requirements for software quality attributes such as maintainability, testability, and scalability, and indicators related to code quality and performance‌.

[0027] ‌Change Control‌: the process of handling requirement changes, the approval and recording mechanism for changes, and the change history‌.

[0028] Acceptance Criteria: Define the criteria for requirements to be met, and the plan and process for acceptance testing. The following example shows the structure of a software requirement text, as shown below: 1. Brief description of XX software XXXXXX 2. XX software requirement description 2.1 XX software function 2.1.1 Initialization function After the software is powered on or reset, set the receive interrupt to open; 2.1.2 Information receiving function Table 1 Information receiving function 2.2 Interface 2.2.1 XX software interface Figure 1 XX software interface 2.2.2 Host computer interface 2.3 Performance 2.3.1 Host computer communication performance 2.4 Security Requirements 2.4.1 Interface security requirements For the software requirement text in word format, this embodiment extracts its subject structure and the text, table and picture under each subject structure through step S1. The subject structure is extracted based on the directory structure of the word text, and the individual paragraphs are analyzed to see whether they are themes. For example, those that meet the format of Chapter x, Section xx, etc. are identified as subject structures.

[0029] Based on the table class that can be read by the word reference library, extract the table from the paragraphs under the topic structure, and extract the preceding paragraphs simultaneously, and combine them according to the text style of the paragraphs to determine the caption of the corresponding table; name the extracted table information with the prefix table format or table note format or other customized document paragraph format. Usually, the preceding paragraph of the table is a description of the table, such as table notes, table headers, captions, etc.

[0030] Based on the image class that can be read by the word reference library, extract the image from the paragraphs under the topic structure, and extract the subsequent paragraphs simultaneously, and determine the caption of the corresponding image based on the text style combination of the paragraph; name the extracted image information with the prefix image format or caption format or other customized document format. The subsequent paragraphs of the image, such as the caption, are also considered as part of the image.

[0031] The structure of the software requirement text obtained in step S1 is as follows: After the subject structure is extracted, step S2 locates the key structure of the software requirement text according to the preset keywords and synonyms. Specifically, the text matching of keywords and synonyms is performed according to the title level, and the successfully matched title and its subordinate texts, tables, and figures are used as matching objects of the same keyword.

[0032] Among them, the keyword and synonym database is pre-built, and the equivalent relationship between keywords and synonyms is established in the keyword and synonym database. The keyword is used as a node, and its similar synonyms are set, and the similarity between the keyword and the synonym is attached. For example, if keyword a has an equivalent synonym b (similarity 100%), and b has a synonym c (similarity 80%), then keyword a is simultaneously identified as having a synonym c (80%).

[0033] In this embodiment, the keywords include software functional requirement description, function, performance, interface, security requirement, adaptability requirement and computer resource requirement. The setting of specific keywords is related to the field involved in the software. For example, if it is applied to communication services, answering machine services, etc., it is necessary to set service-related words.

[0034] According to the title level, the software requirement text is matched with keywords. If the match is successful, it is identified as the keyword corresponding to the requirement type; if the match is unsuccessful, the software requirement text is matched with synonyms. In other words, the text title is matched with keywords and synonyms starting from the lowest title level; while the sequence is compared in numerical order, the text after the sequence number is identified as the requirement explanation, which is first compared with the keyword. If successful, it is identified as the keyword corresponding to the requirement type. If there is no match, it is compared with the synonyms with the highest similarity of all keywords. For example, keyword a has 90% synonym b, keyword c has 80% synonym d, and keyword e has 90% synonym f. If a, c, and e have no matches, the second round should match b and f.

[0035] The structure obtained after key structure positioning is as follows: In step S2, formatting and decomposing the software requirement text according to the preset sentence library and semantic library further includes: According to the preset disassembly standards, the software requirement text is disassembled including prepositions, nouns, verbs, auxiliary words, adjectives, and punctuation marks; According to the punctuation priority, the software requirement text is segmented using the punctuation with the highest priority to obtain multiple complete sentences; Remove the prepositions in each sentence, match the text according to the verb library and the corresponding sentence pattern; split the matched text into corresponding phrases according to the sentence pattern settings, and mark the attributes; All the phrases obtained by disassembly are disassembled according to the auxiliary words to obtain the combination of adjectives and nouns; Chinese and Arabic numerals are used to extract the conditions of adjectives and obtain the corresponding conditional formulas.

[0036] Specifically, the prepositions commonly used in software requirements texts include "will", "through", etc. Nouns, including device (A1), test input (A2), test output type (A3); Verbs, including topping up, sending, receiving, executing, receiving, etc.; Sentence patterns, including "through (device A1), send (test input A2), return (test output A3)", "top up (test input A2), execute (test output A3) after receiving (by device A1)", "execute (test output A3) after receiving (test input A2)", etc.; Auxiliary words are "de", "di", etc.; What comes before the auxiliary word is an adjective; Punctuation marks and their priorities are: ".", ";", ",", ":"; The format for disassembling an adjective (phrase) is "adjective" "auxiliary word" "noun".

[0037] Specifically in the above example, perform the processing of step S2 on Table 1 of the information receiving function to obtain the following results: Processing: check<host computer information>{[frame header]: "55AAH"; [frame tail] "3456H"; [checksum] ""; [function number] ">=03H, <-0AH"} =>right{<correct response frame> action: send; <correct communication frame count> action: ++} || wrong{<incorrect communication frame count> action: ++}.

[0038] In step S3, according to the preset test requirement feature library, perform structured processing on the requirement description, and generate the corresponding test requirement document according to the preset test requirement text format.

[0039] Each test requirement in the test requirement feature library in this embodiment includes at least requirement description features, test method types, and test value formulas; each requirement description feature is associated with sentence patterns and adjective features.

[0040] According to the preset test requirement feature library, perform structured processing on the requirement description. Specifically, split the text according to the punctuation mark priority using the punctuation mark with the highest priority; remove the prepositions in the text; match the text according to the verb library and the corresponding sentence patterns, and after matching, split the corresponding nouns according to the sentence patterns and mark them according to the set attributes; disassemble all the disassembled noun phrases according to the auxiliary words into combinations of adjectives and nouns; extract the conditions of the adjectives using Chinese and Arabic numerals to form a general condition formula.

[0041] Match the conditional formula with the requirement description features; obtain the corresponding test method type based on the matched requirement description features; generate the test input value or test input condition based on the matched test value formula and conditional formula. The conditional formula here is used to match the execution method, such as condition a executes in order a, and condition b executes in order b.

[0042] The information receiving function processed in step S2 above is subjected to test requirement feature matching to obtain: <Functional test> <Information receiving function test> body<host computer information>: right{<correct response frame> action:send; <correct communication frame count> action:++}|| wrong{<error communication frame count>action:++} --point<frame header>{right<55AAH>;wrong<non-55AAH>} --point<frame end>{right<3456H>;wrong<non-3456H>} --point<checksum>{right<correct>;wrong<abnormal>} --point<function number>{right<[03H,0AH]>;wrong<”<=02H,>=0BH”>} <Boundary Test> <Information receiving function number boundary test> body<host computer information>: right{<correct response frame> action:send; <correct communication frame count> action:++}|| --point<function number>{down: 03H, up: 0AH, step: 1}.

[0043] Then, based on the matched test requirements, a test requirement text description is generated; according to the preset test requirement document format, the test requirement text description is output in sequence according to the test method type. The following test requirement text can be obtained: Test item description: Test the information receiving function, including: 1) Test the information receiving function of the host computer to see whether the frame header, frame tail, checksum, and function number are correctly judged.

[0044] Test method description: Send host computer information through ground test equipment, check whether the correct communication frame count and error communication frame count are correct, and whether the correct response frame is sent through the ground test equipment.

[0045] The host computer information sent is as follows: 1) Correct host computer information (frame header: 55AAH, frame tail 3456H, checksum is correct, function number is in the range of [03H, 0AH]) 2) Host computer information of error frame header (not 55AAH); 3) Host computer information at the end of the error frame (not 3456H); 4) Error checksum host computer information; 5) Host computer information of error function number (<=02H,>=0BH).

[0046] Adoption criteria: 1) After receiving the correct information from the host computer, the ground test equipment receives the correct response frame, and the correct communication frame count is accumulated; 2) After receiving the wrong information from the host computer, the ground test equipment did not receive the correct response frame, and the error communication frame count accumulated.

[0047] The above-mentioned Chinese text test requirement extraction method based on document standards and semantics can be implemented through pure software programming, or a neural network model can be used to set corresponding feature extraction rules and classification models to classify and extract the subject structure and text, tables, and figures under the software requirement text; then a text segmentation model based on machine learning is used to format and disassemble the software requirement text from the annotated keywords and synonyms, sentence structure, and semantic library to obtain a standard format requirement description; then the text is converted into a vector through a text feature matching model, and then a dense vector is obtained through multi-layer nonlinear projection, and finally the text similarity is calculated through cosine similarity, thereby matching the requirement description with the test requirement features and outputting the test requirement document. Based on the same concept, this embodiment also provides a Chinese text test requirement extraction device based on document standards and semantics, please refer to Figure 2 , the device comprises: The document preprocessing module is configured to disassemble the structure of the software requirement text based on the document standard to obtain the subject structure of the software requirement text and the text, table, and figure with titles under each subject structure; The document decomposition module is configured to format and decompose the software requirement text according to preset keywords and synonyms, sentence structure and semantic libraries to obtain a standard format requirement description; The requirement structured output module is configured to perform structured processing on the requirement description according to the preset test requirement feature library, and generate the corresponding test requirement document according to the preset test requirement text format; Among them, the document disassembly module is also configured to locate the key structure of the software requirement text according to the preset keywords and synonyms library, perform text matching of keywords and synonyms according to the title level, and use the successfully matched title and its subordinate texts, tables, and figures as matching objects for the same keyword.

[0048] Furthermore, the document preprocessing module includes a topic extraction unit, a table extraction unit and an image extraction unit; wherein the topic extraction unit is configured to perform target extraction on the topic structure of the software requirement text based on the directory structure of the word text; The table extraction unit is configured to extract tables from paragraphs under the topic structure based on the table class that can be read by the word reference library, and simultaneously extract the preceding paragraphs, and combine and determine them as the captions of the corresponding tables according to the text styles of the paragraphs; the extracted table information is named as a document paragraph in a table format or a table note format with a prefix or in other custom formats; The image extraction unit is configured to extract images from paragraphs under the topic structure based on the image class that can be read by the word reference library, and simultaneously extract subsequent paragraphs, and determine the caption of the corresponding image based on the text style combination of the paragraph; the extracted image information is named as a document paragraph in a prefixed image format or caption format or other customized formats.

[0049] The requirements structured output module is further configured as follows: Get the corresponding test method type according to the matching requirement description features; Generate a test input value or a test input condition according to the matching test value formula and condition formula; Generate a text description of the test requirements based on the matched test requirements; According to the preset test requirement document format, the test requirement text description is output in sequence according to the test method type.

[0050] The device is used to implement the above-mentioned Chinese text test requirement extraction method based on document standards and semantics. Its specific implementation method is similar and will not be repeated here.

[0051] In summary, for software requirement documents with large amounts of text, this embodiment provides a Chinese text test requirement extraction method based on document standards and semantics, which disassembles the document structure according to the corresponding standards, establishes a keyword and synonym library, a sentence library and a semantic library to format and disassemble the software requirement description; establishes a test requirement feature library, and generates a corresponding test requirement document by matching the conditions of software requirement disassembly with the test requirement features. This greatly reduces the time for manually reading software requirements and establishing test requirements. The correctly disassembled test requirement text description is more standardized, reducing the semantic ambiguity of manually written documents, which is beneficial to the tester's execution of the test content and ensures the smooth progress of the test.

[0052] The embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments. Even if various changes are made to the present invention, if these changes fall within the scope of the claims of the present invention and their equivalents, they still fall within the protection scope of the present invention.

Claims

1. A method for extracting Chinese text test requirements based on document standards and semantics, characterized in that: include: Based on the document standard, the structure of the software requirement text is disassembled to obtain the subject structure of the software requirement text and the text, table and figure with titles under each subject structure; Format and disassemble the software requirement text according to the preset keywords and synonyms library, sentence library, and semantic library to obtain the requirement description in a standard format; According to the preset test requirement feature library, the requirement description is structured and the corresponding test requirement document is generated according to the preset test requirement text format; Among them, the key structure of the software requirement text is located according to the preset keywords and synonyms library, the text matching of keywords and synonyms is carried out according to the title level, and the successfully matched title and its subordinate texts, tables, and figures are regarded as matching objects of the same keyword.

2. The method for extracting Chinese text test requirements based on document standards and semantics as claimed in claim 1, characterized in that: The topic structure of the software requirement text is extracted based on the directory structure of the word text; Based on the table class that can be read by the word reference library, extract the table from the paragraphs under the topic structure, and extract the preceding paragraphs simultaneously, and combine them according to the text style of the paragraphs to determine the caption of the corresponding table; name the extracted table information with the prefix table format or table note format or other customized format of the document paragraph; Based on the image class that can be read by the word reference library, images are extracted from the paragraphs under the topic structure, and the subsequent paragraphs are extracted simultaneously. The captions of the corresponding images are determined based on the text style combination of the paragraphs; the extracted image information is named in the document paragraph with the prefix image format or caption format or other customized formats.

3. The method for extracting Chinese text test requirements based on document standards and semantics as claimed in claim 1, characterized in that: Construct a database to establish the equivalent relationship between keywords and synonyms, take keywords as nodes, set similar synonyms, and attach similarity; the keywords include software functional requirement description, function, performance, interface, security requirement, adaptability requirement and computer resource requirement; The software requirement text is matched with keywords according to the title level. If the match is successful, it is determined to be the keyword corresponding to the requirement type; if the match is unsuccessful, the software requirement text is matched with synonyms.

4. The method for extracting Chinese text test requirements based on document standards and semantics as claimed in claim 1, characterized in that: Formatting and decomposing the software requirement text according to the preset sentence library and semantic library further includes: According to the preset disassembly standards, the software requirement text is disassembled including prepositions, nouns, verbs, auxiliary words, adjectives, and punctuation marks; According to the punctuation priority, the software requirement text is segmented using the punctuation with the highest priority to obtain multiple complete sentences; Remove the prepositions in each sentence, match the text according to the verb library and the corresponding sentence pattern; split the matched text into corresponding phrases according to the sentence pattern settings, and mark the attributes; All the phrases obtained by disassembly are disassembled according to the auxiliary words to obtain the combination of adjectives and nouns; Chinese and Arabic numerals are used to extract the conditions of adjectives and obtain the corresponding conditional formulas.

5. The method for extracting Chinese text test requirements based on document standards and semantics as claimed in claim 4, characterized in that: Each test requirement in the test requirement feature library at least includes a requirement description feature, a test method type and a test value formula; Each requirement description features associated sentences and adjective features.

6. The method for extracting Chinese text test requirements based on document standards and semantics as claimed in claim 5, characterized in that: Match the conditional formula with the demand description features; Get the corresponding test method type according to the matching requirement description features; Generate a test input value or a test input condition according to the matching test value formula and condition formula.

7. The method for extracting Chinese text test requirements based on document standards and semantics as claimed in claim 6, characterized in that: Generate a text description of the test requirements based on the matched test requirements; According to the preset test requirement document format, the test requirement text description is output in sequence according to the test method type.

8. A Chinese text test requirement extraction device based on document standards and semantics, characterized in that: include: The document preprocessing module is configured to disassemble the structure of the software requirement text based on the document standard to obtain the subject structure of the software requirement text and the text, table, and figure with titles under each subject structure; The document decomposition module is configured to format and decompose the software requirement text according to preset keywords and synonyms, sentence structure and semantic libraries to obtain a standard format requirement description; The requirement structured output module is configured to perform structured processing on the requirement description according to the preset test requirement feature library, and generate the corresponding test requirement document according to the preset test requirement text format; Among them, the document disassembly module is also configured to locate the key structure of the software requirement text according to the preset keywords and synonyms library, perform text matching of keywords and synonyms according to the title level, and use the successfully matched title and its subordinate texts, tables, and figures as matching objects for the same keyword.

9. The Chinese text test requirement extraction device based on document standards and semantics as claimed in claim 8, characterized in that: The document preprocessing module includes a subject extraction unit, a table extraction unit and a picture extraction unit; The subject extraction unit is configured to perform target extraction on the subject structure of the software requirement text based on the directory structure of the word text; The table extraction unit is configured to extract the table from the paragraphs under the topic structure based on the table class that can be read by the word reference library, and simultaneously extract the preceding paragraph, and combine and determine the caption of the corresponding table according to the text style of the paragraph; and name the extracted table information in the document paragraph in the table format or table note format with the prefix or in other customized formats; The image extraction unit is configured to extract images from paragraphs under the topic structure based on the image class that can be read by the word reference library, and simultaneously extract subsequent paragraphs, and determine the caption of the corresponding image based on the text style combination of the paragraph; the extracted image information is named as a document paragraph in a prefixed image format or a caption format or other customized formats.

10. The Chinese text test requirement extraction device based on document standards and semantics as claimed in claim 8, characterized in that: The demand structured output module is further configured to: Get the corresponding test method type according to the matching requirement description features; Generate a test input value or a test input condition according to the matching test value formula and condition formula; Generate a text description of the test requirements based on the matched test requirements; According to the preset test requirement document format, the test requirement text description is output in sequence according to the test method type.

Images