Text analysis method and device, compliance operation detection method and device, equipment and medium

Abstract

The invention relates to the technical field of text processing, in particular to a text analysis method, a compliance operation detection method and device, equipment and a medium. According to the method, the original data is subjected to standardization processing, so that the original data is converted into the standardized text data which can be recognized by a machine, and then the reference structured text and the standardized text data are input into the analysis engine; and guiding the analysis engine to convert the standardized text data into the structured target text through the reference structured text. As the original data is a reference for judging whether the operation is compliant or not, the structured target text converted from the original data can be used as the reference for judging whether the operation is compliant or not. According to the method, the analysis engine converts the original data into the structured target text, manual analysis is replaced with the analysis engine, and due to the fact that the analysis engine has certain stability, the accuracy of the structured target text obtained through analysis can be improved through the analysis engine.

Description

Technical Field

[0001] The present invention relates to the field of text processing technology, and in particular to a text parsing method, a compliance operation detection method, a device, equipment and a medium. Background Art

[0002] The data is converted into structured text that can be recognized by the machine, and then the structured text is used as the basis for whether the business operation is compliant. In other words, the structured text is used to judge whether the business operation is compliant. For example, when a bank handles business, it requires bank personnel to handle business in accordance with regulations (handling business is also business operation). Since the original data of the regulatory text is not easy to read, the existing technology manually converts the regulatory text into structured text. The structured text can intuitively tell bank personnel what to do at each step when handling business. Since the regulatory text is manually parsed into structured text, the accuracy of the parsed structured text is low.

[0003] In summary, the accuracy of structured text parsed by the existing technology is low.

[0004] Therefore, the prior art still needs to be improved and enhanced. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a text parsing method, a compliance operation detection method, an apparatus, a device and a medium, which solve the problem of low accuracy of structured text obtained by parsing in the prior art.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a text parsing method, which comprises:

[0008] Acquire original data, and perform standardization processing on the original data to obtain standardized text data;

[0009] Constructing a baseline structured text;

[0010] The reference structured text and the standardized text data are input into a parsing engine, so that the parsing engine is guided by the reference structured text to convert the standardized text data into a structured target text.

[0011] In one implementation, the original data is standardized to obtain standardized text data, including:

[0012] Extracting text information contained in the original data to form a plain text;

[0013] Obtaining a preset number of words in each paragraph and a preset number of overlapping words between paragraphs, and dividing the plain text into a plurality of paragraphs according to the preset number of words and the preset number of overlapping words;

[0014] Each of the plurality of paragraphs is encoded to obtain an encoded text, and the encoded text is used as standardized text data.

[0015] In one implementation, extracting the text information contained in the original data to form a plain text includes:

[0016] Determine the original text and original picture in the original data;

[0017] When the original text contains a table, identifying the same table on the original text according to the number of columns and / or column width and / or table header information of the table, and converting the original text into a plain text according to the information expressed in the same table;

[0018] The text on the original image is extracted to form a plain text.

[0019] In one implementation, constructing a reference structured text includes:

[0020] Acquire a prompt word, acquire an exemplary text corresponding to the prompt word, and acquire a knowledge base having the same field as the original data;

[0021] Retrieving a knowledge fragment related to the prompt word and the exemplary text from the knowledge base;

[0022] A benchmark structured text is constructed based on the prompt words, the exemplary text, and the knowledge fragments.

[0023] In one implementation, the parsing engine is a large language model after training, and the parameters of the large language model are adjusted using LoRA fine-tuning during the training of the large language model.

[0024] In a second aspect, an embodiment of the present invention further provides a compliance operation detection method, including:

[0025] Based on the text parsing method described in claim 1, generating a structured target text;

[0026] Allocating the structured target text to a business process engine corresponding to the structured target text;

[0027] Through the structured target text, whether the operation of the business process engine is compliant is detected to obtain a detection result.

[0028] In one implementation, the method further includes:

[0029] Storing the test results on the blockchain;

[0030] The detection result is associated with the monitoring and early warning module.

[0031] In a second aspect, an embodiment of the present invention further provides a text parsing device, wherein the device comprises the following components:

[0032] A standardization processing module is used to obtain original data and perform standardization processing on the original data to obtain standardized text data;

[0033] A benchmark text construction module, used to construct a benchmark structured text;

[0034] The conversion module is used to input the reference structured text and the standardized text data into a parsing engine, so as to guide the parsing engine to convert the standardized text data into a structured target text through the reference structured text.

[0035] In a third aspect, an embodiment of the present invention further provides a terminal device, wherein the terminal device comprises a memory, a processor, and a text parsing program stored in the memory and executable on the processor, and when the processor executes the text parsing program, the steps of the above-mentioned text parsing method are implemented;

[0036] Alternatively, the terminal device includes a memory, a processor, and a compliance operation detection program stored in the memory and executable on the processor, and when the processor executes the compliance operation detection program, the steps of the compliance operation detection method described above are implemented.

[0037] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, wherein a text parsing program is stored on the computer-readable storage medium, and when the text parsing program is executed by a processor, the steps of the above-mentioned text parsing method are implemented;

[0038] Alternatively, a compliance operation detection program is stored on the computer-readable storage medium, and when the compliance operation detection program is executed by the processor, the steps of the compliance operation detection method described above are implemented.

[0039] Beneficial effects: The present invention performs standardization processing on the original data to convert the original data into standardized text data that can be recognized by a machine, and then inputs both the benchmark structured text and the standardized text data into a parsing engine, so as to guide the parsing engine to convert the standardized text data into a structured target text through the benchmark structured text. Since the original data is the benchmark for judging whether the operation is compliant, the structured target text converted from the original data can be used as a benchmark for judging whether the operation is compliant. In the present invention, the original data is converted into a structured target text by the parsing engine, and the parsing engine is used instead of manual parsing. Since the parsing engine has a certain stability, the use of the parsing engine can improve the accuracy of the structured target text obtained by parsing. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is the overall flow chart of the present invention;

[0041] Figure 2 This is a flow chart of banking compliance management in an embodiment of the present invention;

[0042] Figure 3 A structural diagram of a text parsing device provided by the present invention;

[0043] Figure 4 This is a block diagram of the internal structure principle of the terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following is a clear and complete description of the technical solution of the present invention in combination with the embodiments and the accompanying drawings. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] According to research, data can be converted into structured text that can be recognized by machines, and then the structured text can be used as the basis for whether the business operation is compliant. For example, when banks handle business, they require bank personnel to handle business in accordance with regulations (handling business is also business operation). Since the original data of the regulatory text is not easy to read, the existing technology manually converts the regulatory text into structured text. The structured text can intuitively tell bank personnel what to do at each step when handling business. Since the regulatory text is manually parsed into structured text, the accuracy of the parsed structured text is low.

[0046] In order to solve the above technical problems, the present invention provides a text parsing method, a compliance operation detection method, an apparatus, a device and a medium, which solve the problem of low accuracy of structured text obtained by parsing in the prior art.

[0047] The text parsing method of this embodiment can be applied to a terminal device, which can be a terminal product with a text processing function, such as a computer. Figure 1 As shown in , the text parsing method specifically includes the following steps:

[0048] S100, obtaining original data, and performing standardization processing on the original data to obtain standardized text data;

[0049] S200, building a benchmark structured text;

[0050] S300, inputting the reference structured text and the standardized text data into a parsing engine, so as to guide the parsing engine to convert the standardized text data into a structured target text through the reference structured text.

[0051] This embodiment can be used for intelligent compliance management in banking business, and the specific process is as follows:

[0052] like Figure 2 As shown in the figure, the original data of the bank's external regulatory data, the regulatory data of the bank's internal system, and the regulatory data of each department of the bank are input into the multi-source data processing module. The regulatory data is used to constrain or inform bank personnel which business operations are compliant and which business operations are not compliant. It is necessary to design a special interface and conversion logic to process internal data such as the bank's core system and CRM system. The data security and privacy protection layer in the multi-source data processing module will perform security and privacy protection on the regulatory data of the bank's internal system to prevent the regulatory data of the bank's internal system from being mixed with the privacy data of bank personnel and introduced into the subsequent distributed ledger system, thereby protecting the security of privacy data.

[0053] The multi-source data processing module performs standardization processing on the above-mentioned multiple regulatory data to obtain standardized text data. Standardization processing is to convert the regulatory data into a language that can be recognized by a machine.

[0054] The benchmark structured text guides the parsing engine to convert the standardized text data into structured target text. The structured target text is used to detect which bank's business operations are compliant and which are not in compliance in a machine language.

[0055] For example, regulatory data: It is prohibited to transfer more than 10,000 US dollars to high-risk areas. Convert it into structured target text: whether it is an outbound transfer, whether the recipient's area is high-risk, and whether the amount exceeds 10,000.

[0056] This embodiment can also be used to check whether the computer system update program is correct. The specific process is as follows:

[0057] The original data of the computer update process manual is obtained, and the benchmark structured text corresponding to the original data is obtained, and the original data is standardized to obtain standardized text data. Standardization is to convert the original data into a language that can be recognized by the machine, and the computer update process manual and the benchmark structured text are input into the parsing engine. The benchmark structured text guides the parsing engine to convert the standardized text data into a structured target text. Before the computer system is updated, the structured target text is imported into the computer to detect whether the update process of the computer system is correct through the structured target text.

[0058] Embodiment 1: In this embodiment, step S100 includes the following specific steps S101, S102, S103, S104, and S105:

[0059] S101, determining the original text and the original picture in the original data.

[0060] Taking the banking system as an example, a unified data access interface is established to transmit the original data such as internal system data and external system data to the multi-source data processing module in real time. The original data includes two data forms: original text and original image.

[0061] S102, when the original text contains a table, identifying the same table on the original text according to the number of columns and / or column width and / or header information of the table, and converting the original text into a plain text according to the information expressed by the same table.

[0062] This embodiment uses a deep learning model to identify and extract table structures, and to identify which tables are the same table across pages and which are the same table in a nested form.

[0063] For example, if tables appear on both the second and third pages of the original text, we need to determine whether the tables on the two pages belong to the same table. If they are the same table, we need to organize the text on the table according to the information conveyed by the same table, and remove line breaks, stop words, and special symbols in the text to form a plain text.

[0064] In this embodiment, for tables that span multiple pages, the table structure continuity of adjacent pages is analyzed by comparing the number of columns, column width, and header information to determine whether the tables that span multiple pages are the same table. The logical continuity of the contents of the tables that span multiple pages (such as serial numbers, dates, etc.) can also be checked to determine whether the tables that span multiple pages are the same table. Common cross-page patterns can also be processed using rule-based methods, such as repeating headers or footers, to determine whether the tables that span multiple pages are the same table.

[0065] For nested tables, we optimized an OCR model specifically for them, and annotated more than 200 nested table data to build a model specifically for identifying nested tables.

[0066] S103, extracting the text on the original image to form a plain text;

[0067] When the original data includes original pictures, advanced OCR technology is used on the pictures, combined with image pre-processing and post-processing to optimize the pictures, and the text on the optimized pictures is extracted.

[0068] S104, obtaining a preset number of words in each paragraph and a preset number of overlapping words between paragraphs, and dividing the plain text into a plurality of paragraphs according to the preset number of words and the preset number of overlapping words.

[0069] In this embodiment, the preset number of words is 1000, the preset number of overlapping words is 100, and the plain text is divided into several paragraphs according to 1000 words in each paragraph and 100 words overlapping the previous and next paragraphs.

[0070] S105, encoding each of the plurality of paragraphs to obtain encoded text, and using the encoded text as standardized text data.

[0071] The phrases in each paragraph are encoded as encoding units to obtain the encoded text of each paragraph, and then the encoded text of the entire plain text is obtained.

[0072] The format of the standardized text data in this embodiment is JSON format. Through multiple verifications and artificial intelligence-assisted proofreading, the accuracy of data extraction and conversion is ensured to reach more than 99.9%. Through the multi-source data processing module of this embodiment, the data integration rate can be increased to 95%, realizing comprehensive internal and external information utilization. And the data preprocessing time is reduced by 90%, significantly speeding up the analysis process.

[0073] Embodiment 2, based on embodiment 1, in this embodiment, step S200 includes the following specific steps S201, S202, S203:

[0074] S201, obtaining a prompt word, obtaining an exemplary text corresponding to the prompt word, and obtaining a knowledge base in the same field as the original data.

[0075] Design complex prompt projects, including multi-level prompts and context management, to obtain prompt words. Prompt words are used to tell the subsequent parsing engine what content needs to be output. And the example text is used to tell the subsequent parsing engine what format to output the content. Among them, implement advanced few-shot learning mechanism to dynamically select the most relevant examples to improve parsing accuracy.

[0076] S202: Retrieve knowledge fragments related to the prompt word and the exemplary text from the knowledge base.

[0077] That is, knowledge fragments related to prompt words and exemplary texts are retrieved from the expert knowledge base and financial regulations, and the knowledge fragments can enhance the understanding ability of the parsing engine in the banking field.

[0078] S203: construct a benchmark structured text based on the prompt words, the exemplary text, and the knowledge fragment.

[0079] This embodiment ensures that the accuracy of regulatory parsing reaches more than 99.5% through double-blind testing with human experts and continuous optimization.

[0080] Embodiment 3, based on embodiment 1 or embodiment 2, in this embodiment, the parsing engine is a large language model after training, and the parameters of the large language model are adjusted using LoRA fine-tuning during the training of the large language model.

[0081] The process of training the large language model in this embodiment is as follows:

[0082] Obtain sample text in a specific field (such as various regulatory texts in the banking field) and the structured text corresponding to the sample text as a label, input the sample text into the large language model, obtain the structured text output by the large language model, determine the difference between the output structured text and the structured text as a label, and adjust the parameters of the large semantic model according to the difference until the model converges to obtain the trained large language model, that is, the parsing engine.

[0083] After the parsing engine is trained, the standardized text data and the reference structured text to be parsed are input into the parsing engine, and the reference structured text is used to guide the parsing engine to convert the standardized text data into structured target text. The structured target text contains fields such as conditions, behaviors, exceptions, and penalties.

[0084] The parsing engine of this embodiment has an accuracy of 99.5% in understanding regulations, achieving parsing capabilities close to those of human experts. The time to adapt to new regulations is shortened by 98%, from weeks to hours. The ability to parse complex regulations is improved by 200%, and it can handle ambiguous and cross-referenced clauses.

[0085] Embodiment 4, based on embodiment 1 or embodiment 2 or embodiment 3, as Figure 2As shown, this embodiment inputs each structured target text output by the parsing engine into the dynamic rule adapter, and distributes each structured target text to the corresponding business process engine through the dynamic rule adapter. Taking the banking system as an example, the function of the dynamic rule adapter is to convert the structured target text into more sophisticated execution rules, and then adapt the execution rules to the specific business processes and internal policies of the bank.

[0086] The dynamic rule adapter implements its functions through the following methods:

[0087] a) Build complex decision trees and rule graph structures to accurately represent rule logic and dependencies. b) Implement advanced rule conversion algorithms to convert abstract rules into executable business rules and control logic. c) Use reinforcement learning algorithms to dynamically optimize rule application strategies, taking into account historical compliance data and business impact. d) Integrate conflict detection and resolution mechanisms to automatically handle potential conflicts between rules.

[0088] The dynamic rule adapter in this embodiment achieves 99.9% accuracy in rule adaptation and execution through large-scale simulation testing and real-time monitoring feedback. The use of dynamic rule adapters increases the speed of rule updates by 99%, ensures the real-time nature of compliance measures, reduces rule conflicts by 99.5%, significantly improves the consistency and compliance of business processes, and achieves 99.9% accuracy in rule execution, minimizing compliance risks.

[0089] like Figure 2 As shown, after the rules are assigned to each business process engine, the structured target text is used as the standard to detect whether the operation of the business process engine is compliant, and the detection results are obtained, and the detection results are saved on the blockchain; the detection results are associated with the monitoring and early warning module.

[0090] The blockchain of this embodiment is located on a distributed ledger system, which outputs an unalterable, highly secure compliance operation log to ensure the unalterability, traceability and privacy protection of the data. The distributed ledger system also supports efficient query and audit at the millisecond level.

[0091] The distributed ledger system of this embodiment adopts high-performance enterprise-level distributed ledger technology and supports tens of thousands of transactions per second. It also adopts fine-grained access control and data encryption mechanisms to ensure the security of sensitive information. Smart contracts on the distributed ledger system can automatically record key compliance operations, including decision-making basis and execution process, improve audit efficiency by 90%, and achieve near-real-time compliance verification. The distributed ledger system integrates zero-knowledge proof technology to support compliance verification while protecting privacy, improve privacy protection capabilities by 200%, and support compliance audits without leaking sensitive information.

[0092] The distributed ledger system of this embodiment ensures 100% integrity and accuracy of records through cryptographic guarantees and a multi-node consensus mechanism.

[0093] The monitoring and early warning module of this embodiment can comprehensively monitor internal business operations and external environmental changes, and timely identify and warn of potential compliance risks.

[0094] The monitoring and early warning module deploys a high-throughput real-time stream processing framework to support parallel processing of massive data. The monitoring and early warning module applies advanced time series analysis and anomaly detection algorithms to identify complex risk patterns. The monitoring and early warning module integrates predictive analysis models to predict future risks based on historical data and current trends. The monitoring and early warning module has a multi-dimensional and multi-level early warning mechanism, including instant notification, trend reporting and scenario analysis.

[0095] The monitoring and early warning module of this embodiment ensures that the accuracy of risk identification and early warning reaches more than 99% through continuous model tuning and artificial intelligence-assisted review. The monitoring and early warning module extends the lead time of risk event warning to 72 hours, greatly increasing the risk prevention window. The false alarm rate is reduced to less than 1%, significantly reducing unnecessary manual intervention.

[0096] In summary, the present invention increases the speed of compliance management processing by 80-90%, far exceeding traditional methods. The overall system's compliance judgment accuracy reaches 99.9%, almost eliminating human errors. Labor costs are reduced by 70-80%, and potential violation costs are reduced by more than 95%. The response time to new regulations is shortened from weeks to hours, improving regulatory adaptability.

[0097] This embodiment also provides a text parsing device, such as Figure 3 As shown, the device comprises the following components:

[0098] The standardization processing module 01 is used to obtain original data and perform standardization processing on the original data to obtain standardized text data;

[0099] A benchmark text construction module 02, used for constructing a benchmark structured text;

[0100] The conversion module 03 is used to input the reference structured text and the standardized text data into a parsing engine, so as to guide the parsing engine to convert the standardized text data into a structured target text through the reference structured text.

[0101] Based on the above embodiments, the present invention further provides a terminal device, whose principle block diagram can be shown as follows: Figure 4As shown. The terminal device includes a processor, a memory, a network interface, and a display screen connected via a system bus. The processor of the terminal device is used to provide computing and control capabilities. The memory of the terminal device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the terminal device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a text parsing method is implemented. The display screen of the terminal device can be a liquid crystal display screen or an electronic ink display screen.

[0102] Those skilled in the art will understand that Figure 4 The principle block diagram shown in the figure is only a block diagram of a partial structure related to the scheme of the present invention, and does not constitute a limitation on the terminal device to which the scheme of the present invention is applied. The specific terminal device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0103] In one embodiment, a terminal device is provided. The terminal device includes a memory, a processor, and a text parsing program stored in the memory and executable on the processor. When the processor executes the text parsing program, the following operation instructions are implemented:

[0104] Acquire original data, and perform standardization processing on the original data to obtain standardized text data;

[0105] Constructing a baseline structured text;

[0106] The reference structured text and the standardized text data are input into a parsing engine, so that the parsing engine is guided by the reference structured text to convert the standardized text data into a structured target text.

[0107] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0108] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A text parsing method, characterized in that: include: Acquire original data, and perform standardization processing on the original data to obtain standardized text data; Constructing a baseline structured text; The reference structured text and the standardized text data are input into a parsing engine, so that the parsing engine is guided by the reference structured text to convert the standardized text data into a structured target text.

2. The text parsing method according to claim 1, characterized in that: The original data is standardized to obtain standardized text data, including: Extracting text information contained in the original data to form a plain text; Obtaining a preset number of words in each paragraph and a preset number of overlapping words between paragraphs, and dividing the plain text into a plurality of paragraphs according to the preset number of words and the preset number of overlapping words; Each of the plurality of paragraphs is encoded to obtain an encoded text, and the encoded text is used as standardized text data.

3. The text parsing method according to claim 2, characterized in that: The step of extracting text information contained in the original data to form a plain text includes: Determine the original text and original picture in the original data; When the original text contains a table, identifying the same table on the original text according to the number of columns and / or column width and / or table header information of the table, and converting the original text into a plain text according to the information expressed in the same table; The text on the original image is extracted to form a plain text.

4. The text parsing method according to claim 1, characterized in that: The construction of the benchmark structured text includes: Acquire a prompt word, acquire an exemplary text corresponding to the prompt word, and acquire a knowledge base having the same field as the original data; Retrieving a knowledge fragment related to the prompt word and the exemplary text from the knowledge base; A benchmark structured text is constructed based on the prompt words, the exemplary text, and the knowledge fragments.

5. The text parsing method according to claim 1, characterized in that: The parsing engine is a large language model after training. During the training of the large language model, the parameters of the large language model are adjusted using LoRA fine-tuning.

6. A compliance operation detection method, characterized in that: include: Based on the text parsing method described in claim 1, generating a structured target text; Allocating the structured target text to a business process engine corresponding to the structured target text; Through the structured target text, whether the operation of the business process engine is compliant is detected to obtain a detection result.

7. The compliance operation detection method according to claim 6, characterized in that: Also includes: Storing the test results on the blockchain; The detection result is associated with the monitoring and early warning module.

8. A text analysis device, characterized in that: The device comprises the following components: A standardization processing module is used to obtain original data and perform standardization processing on the original data to obtain standardized text data; A benchmark text construction module, used to construct a benchmark structured text; The conversion module is used to input the reference structured text and the standardized text data into a parsing engine, so as to guide the parsing engine to convert the standardized text data into a structured target text through the reference structured text.

9. A terminal device, characterized in that: The terminal device comprises a memory, a processor, and a text parsing program stored in the memory and executable on the processor, wherein when the processor executes the text parsing program, the steps of the text parsing method according to any one of claims 1 to 5 are implemented; Alternatively, the terminal device includes a memory, a processor, and a compliance operation detection program stored in the memory and executable on the processor, and when the processor executes the compliance operation detection program, the steps of the compliance operation detection method as described in any one of claims 6-7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a text parsing program, and when the text parsing program is executed by the processor, the steps of the text parsing method according to any one of claims 1 to 5 are implemented; Alternatively, a compliance operation detection program is stored on the computer-readable storage medium, and when the compliance operation detection program is executed by the processor, the steps of the compliance operation detection method according to any one of claims 6 to 7 are implemented.

Images