Contract inspection method and device, storage medium, equipment and program product

Abstract

The application discloses a contract inspection method, a contract inspection device, a computer readable storage medium, a computer device and a computer program product. The contract inspection method comprises the following steps: obtaining a contract file to be inspected; using a preset analysis model to analyze the contract file to obtain an analysis result, the analysis result comprising at least one of a text analysis result and a content analysis result; using a preset verification model to verify the analysis result to obtain a verification result; and in the case that the verification result is a verification failure, performing visual feedback at a corresponding position of the contract file according to abnormal information in the verification result.

Description

Technical Field

[0001] This application relates to the field of contract inspection technology, specifically to a contract inspection method, a contract inspection device, a computer-readable storage medium, a computer device, and a computer program product. Background Technology

[0002] In the field of contract management, inspecting contracts, such as copyright contracts, is a crucial step in ensuring compliance and consistency. Currently, common inspection methods include in-contract inspection and post-contract inspection. In-contract inspection typically uses optical character recognition (OCR) technology to extract text from scanned contracts, and then relies on basic natural language processing methods to achieve only simple character recognition and grammatical logic checks. Post-contract inspection, on the other hand, primarily focuses on verifying the correctness of key fields in the contract through form-level comparisons, and alerting and replacing any abnormal contracts.

[0003] The aforementioned technologies generally rely on content extraction and comparison at the text level, which cannot effectively correct errors or analyze content in contracts with diverse structures and complex content. For example, they suffer from low inspection efficiency, poor accuracy, and lack of content understanding in documents containing multiple tables, multi-level markings, or complex logic and high contextual relationships. Furthermore, the inspection process typically only indicates the presence of anomalies but cannot pinpoint specific clauses or wording issues, leading to high costs for manual review. Summary of the Invention

[0004] This application provides a contract inspection method, a contract inspection device, a computer-readable storage medium, a computer device, and a computer program product. By sequentially parsing and verifying the contract documents, it checks whether there are any problems with the text content or legal clauses of the contract documents, and marks the issues that fail the verification in the contract documents for subsequent processing, thereby improving the efficiency of contract inspection.

[0005] On the one hand, embodiments of this application provide a contract inspection method, the contract inspection method comprising: obtaining a contract document to be inspected; parsing the contract document using a preset parsing model to obtain a parsing result, the parsing result including at least one of text parsing result and content parsing result; verifying the parsing result using a preset verification model to obtain a verification result; and, if the verification result is a verification failure, providing visual feedback at the corresponding position of the contract document based on the abnormal information in the verification result.

[0006] On the other hand, embodiments of this application provide a contract inspection device. The inspection device includes a display unit and a control unit. The control unit is connected to the display unit. The control unit is used to: acquire the contract document to be inspected; parse the contract document using a preset parsing model to obtain a parsing result, the parsing result including at least one of text parsing result and content parsing result; verify the parsing result using a preset verification model to obtain a verification result; and, if the verification result is a verification failure, provide visual feedback at the corresponding position of the contract document based on the abnormal information in the verification result.

[0007] On the other hand, embodiments of this application provide a computer-readable storage medium storing a computer program adapted for loading by a processor to execute a contract inspection method. The contract inspection method includes: acquiring a contract document to be inspected; parsing the contract document using a preset parsing model to obtain a parsing result, the parsing result including at least one of text parsing result and content parsing result; verifying the parsing result using a preset verification model to obtain a verification result; and, if the verification result is a verification failure, providing visual feedback at a corresponding location in the contract document based on the abnormal information in the verification result.

[0008] On the other hand, embodiments of this application provide a computer device, the computer device including a processor and a memory, the memory storing a computer program, the processor executing a contract inspection method by calling the computer program stored in the memory, the contract inspection method including: obtaining a contract document to be inspected; parsing the contract document using a preset parsing model to obtain a parsing result, the parsing result including at least one of text parsing result and content parsing result; verifying the parsing result using a preset verification model to obtain a verification result; and, if the verification result is a verification failure, providing visual feedback at the corresponding position of the contract document based on the abnormal information in the verification result.

[0009] On the other hand, embodiments of this application provide a computer program product, including computer instructions. When the computer instructions are executed by a processor, they implement a contract inspection method. The contract inspection method includes: acquiring a contract document to be inspected; parsing the contract document using a preset parsing model to obtain a parsing result, the parsing result including at least one of text parsing result and content parsing result; verifying the parsing result using a preset verification model to obtain a verification result; and, if the verification result is a verification failure, providing visual feedback at the corresponding position of the contract document based on the abnormal information in the verification result.

[0010] In the contract inspection method, device, computer-readable storage medium, computer equipment, and computer program product provided in this application, the acquired contract document is parsed using a parsing model to obtain the parsing result, and then the parsing result is verified to determine whether there are textual and / or content problems in the contract document. If the verification fails (i.e., a problem is determined to exist), the location of the abnormal information in the verification result corresponding to the location in the contract document is visually displayed, allowing technicians to check or correct the contract document. Therefore, the contract inspection method of this application can not only perform textual inspection on structurally complex contract documents but also inspect specific content, achieving high inspection efficiency and accuracy. Furthermore, by accurately locating the specific location of abnormal information, it significantly reduces the cost of manual review. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of a contract inspection system provided in an embodiment of this application.

[0013] Figure 2 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0014] Figure 3 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0015] Figure 4 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0016] Figure 5 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0017] Figure 6 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0018] Figure 7 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0019] Figure 8 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0020] Figure 9 This is a flowchart illustrating the contract inspection method provided in an embodiment of this application.

[0021] Figure 10 This is a schematic diagram of the structure of the contract inspection device provided in the embodiments of this application.

[0022] Figure 11 A schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] It should be noted that the terms "first," "second," and "third," etc., used in the various parts of the embodiments and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Such terms can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in a sequence other than that shown or described herein. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes 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 these processes, methods, products, or devices.

[0025] The inspection process for contracts such as copyright agreements generally relies on content extraction and comparison at the text level. This approach is insufficient for effectively correcting errors and analyzing content in contracts with diverse structures and complex content. This results in low inspection efficiency, poor accuracy, and a lack of content understanding. Furthermore, it can only indicate the presence of anomalies but cannot pinpoint specific clauses or wording issues, leading to high costs for manual review. To address this problem, this application provides a contract inspection method, a contract inspection device, a computer-readable storage medium, a computer device, and a computer program product.

[0026] Specifically, the contract inspection method in this application embodiment can be executed by a computer device, which can be a terminal or a server. The terminal can be a terminal device such as a smartphone, tablet, laptop, smart TV, wearable smart device, or smart vehicle terminal. The terminal can also include a client, which can be a game client, browser client, instant messaging client, or mini-program. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0027] For example, when the contract inspection method runs on a terminal device, the terminal device may include a display screen and a processor. The display screen is used to present a graphical user interface (GUI) and receive user commands to the GUI. The GUI may include a contract inspection interface, etc. The processor is used to store the contract inspection program, run the contract inspection method, generate the GUI, respond to identification requests, and control the display of the GUI on the display screen. When the user operates the GUI through the display screen, the GUI can control the local content of the terminal device in response to the received operation commands. The terminal device can provide the GUI to the user in various ways, such as rendering it on the terminal device's display screen or presenting the GUI through holographic projection.

[0028] For example, when this contract inspection method runs on a server, it can be implemented and executed based on a cloud generation system. A cloud generation system refers to a generation method based on cloud computing. The cloud generation system includes a server and client devices. The main body running the contract inspection program and the main body presenting the graphical user interface are separate. The storage and operation of the contract inspection method are completed on the server. The presentation of the graphical user interface is completed on the client. The client is mainly used for receiving and sending data and for the presentation of the graphical user interface. For example, the client can be a display device with data transmission capabilities located close to the user, such as a mobile terminal, television, computer, PDA, personal digital assistant, head-mounted display device, etc. However, the terminal device for data processing is the server in the cloud. During contract inspection, the user operates the client to send instructions to the server. The server controls the execution of the contract inspection method according to the instructions, encodes and compresses the data such as the graphical user interface, returns it to the client via the network, and finally, the client decodes and outputs the graphical user interface.

[0029] It should be noted that, in this embodiment, the entity executing the contract inspection method can be a terminal device or a server. The terminal device can be a local terminal device or a client device in the aforementioned cloud generation system. This embodiment does not limit the type of the executing entity.

[0030] For example, in conjunction with the above description, Figure 1 This application illustrates a contract inspection system 1000 for implementing a contract inspection method, as provided in an embodiment of this application. The contract inspection system 1000 may include at least one terminal 1001, at least one server 1002, at least one database 1003, and a network. The user-held terminal 1001 can connect to different servers via the network. The terminal can be any device with computing hardware capable of supporting and executing software application tools corresponding to the contract inspection method.

[0031] In the aforementioned contract inspection system 1000, terminal 1001 is used to install and run the contract inspection program. In some cases, terminal 1001 may not need to have the contract inspection program or its corresponding client pre-installed; users can directly access and perform contract inspections through a browser or other client. During the contract inspection process, terminal 1001 interacts with server 1002. Terminal 1001 sends various information to server 1002. Server 1002 determines the display data for terminal 1001 based on its storage mechanism and the received information, and sends the display data back to terminal 1001 so that terminal 1001 can display the data sent by server 1002 to the user. The network can be a wireless network or a wired network, such as a wireless local area network (WLAN), local area network (LAN), cellular network, 2G network, 3G network, 4G network, 5G network, etc. Additionally, the terminal can also use its own Bluetooth network or hotspot network to connect to other terminals or to the server. In addition, the contract inspection system 1000 can include multiple databases, which are coupled to different servers.

[0032] It should be noted that, Figure 1 The schematic diagram of the contract inspection system shown is merely an example. The contract inspection system 1000 described in this application embodiment is intended to more clearly illustrate the technical solution of this application embodiment and does not constitute a limitation on the technical solution provided in this application embodiment. As those skilled in the art will know, with the evolution of contract inspection systems and the emergence of new business scenarios, the technical solution provided in this application embodiment is also applicable to similar technical problems.

[0033] It should be noted that the triggering operations mentioned in the subsequent detailed description of the contract inspection method provided in this application can all be regarded as triggering operations performed by the user through a finger or by controlling a medium such as a mouse, keyboard, or stylus. The specific medium used can be determined according to the type of computer device. For example, when the computer device is a touchscreen device such as a mobile phone, tablet, or game console, the user can operate on the touchscreen using any suitable object or accessory such as a finger or stylus. When the terminal device is a non-touchscreen terminal device such as a desktop computer or laptop, the user can operate using an external device such as a mouse or keyboard.

[0034] The technical solution of this application will be described in detail below through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0035] Please see Figure 2 , Figure 2This is a flowchart illustrating a contract inspection method provided in an embodiment of this application. It should be noted that the steps shown may be performed in a logical order different from that shown in the flowchart. The contract inspection method may include the following steps: Step 01: Obtain the contract documents to be inspected; Step 03: Use a preset parsing model to parse the contract document to obtain the parsing results, which include at least one of the text parsing results and the content parsing results; Step 05: Verify the parsing results using a preset verification model to obtain the verification results; and Step 071: If the verification result is that the verification failed, provide visual feedback in the corresponding location of the contract document based on the abnormal information in the verification result.

[0036] Specifically, a contract refers to an agreement between natural persons, legal persons, or other organizations of equal status to establish, modify, or terminate civil rights and obligations. In the field of contract management, contract inspection is a crucial process for ensuring compliance during the signing process and consistency of documents, especially for legally binding contracts such as copyright agreements, which holds significant practical importance. Contract inspection refers to a systematic and automated process of monitoring and checking contract content using systematic methods to identify potential errors, inconsistencies, or violations.

[0037] Currently, mainstream contract inspection methods can be divided into two categories: in-contract inspection and post-contract inspection. In-contract inspection refers to real-time checks implemented during the contract signing process. This process typically employs Optical Character Recognition (OCR) technology, a technique that converts text information in scanned documents or images into machine-encodeable text. For example, OCR technology can be used to process scanned contracts and extract raw text data. Subsequently, the system applies basic Natural Language Processing (NLP) techniques to perform preliminary analysis of the text content. However, these methods are often limited to shallow grammatical and logical checks, such as spell correction or simple syntactic verification.

[0038] Post-contract inspection focuses on the batch review process after contract signing. This method primarily relies on form-level comparisons, extracting key fields from the contract (such as signing date, amount, and party information) and performing consistency checks against preset templates or a database. For example, after contract documents are archived, the system can automatically compare information such as the signing entity or amount with a standard form and trigger alerts and replacement mechanisms for anomalies. However, this form-based inspection is essentially a shallow, data-driven verification that lacks an understanding of the overall contract content.

[0039] The aforementioned technologies generally rely on content extraction and comparison at the text level, and their core shortcoming lies in their inability to adapt to contract documents with diverse structures and complex content. For example, for contract documents containing multiple tables, multi-level markings, or complex logic with high contextual relevance, existing methods suffer from low inspection efficiency and poor accuracy, stemming from a lack of deep content understanding capabilities. Furthermore, the inspection process often remains at the level of anomaly alerts, merely identifying problems in the contract document without pinpointing specific problematic clauses or ambiguous expressions. This significantly increases the burden and cost of manual review, limiting the effectiveness of automated inspection. To address this issue, this application proposes a contract inspection method that can consider both text and content, and accurately locate and visualize anomaly information.

[0040] It is understood that a contract document is the carrier of a contract. In this application, it refers to an electronic contract document. A contract document can be an electronic version stored using a specific contract template, or an electronic version obtained by scanning a paper document. Contract documents can be, but are not limited to, PDF, TXT, XML, JPG, or PNG files. The contract document stores the specific content of the contract, such as the contract title and preamble, which clearly state the nature and background of the contract; party information, accurately stating the identities and contact information of each party; substantive terms, specifically stipulating the subject matter, quantity and quality, price and remuneration, performance period, place and method, liability for breach of contract, and dispute resolution mechanisms; and concluding clauses, including elements such as contract effectiveness, number of copies, attachments, and a signature section (including signature, seal, and date).

[0041] Taking a copyright contract as an example, the contract documents at this time include the following: Contract parties, referring to the copyright holder (musician or signing company), who must clearly define the legal status and qualifications of each party; Object of rights, which clearly stipulates the specific content, scope, and form of the work, such as literary works, audio-visual products, etc.; Content of rights, which details the scope of exercise of specific rights such as the right of reproduction, the right of performance, and the right of communication to the public via information networks; Conditions of use, including restrictive conditions such as the term of use, the region of use, and the method of use; Economic clauses, which involve the distribution of economic benefits such as the license agreement amount, royalty rate, payment method, and settlement cycle; Liability for breach of contract, which clearly defines the legal consequences and remedies for violating the contract terms.

[0042] It is understandable that a complete and comprehensive contract inspection needs to include both text inspection, such as whether the sentences are complete and whether there are any typos, and content inspection, such as whether the subject and object information is complete, whether the legal clauses are reasonable, and whether the rights and obligations are clear. Given the contract documents to be inspected, the contract inspection method of this application can perform a comprehensive analysis of both aspects based on the content extracted from the contract documents.

[0043] The parsing model is used to interpret and analyze contract documents. It is a pre-set, fixed model, which can be an unmodifiable fixed model stored in the memory of the Contract Inspection System 1000 before leaving the factory, or a model that can be written, adjusted, or replaced by the operator in the memory. When a contract document is input into the parsing model, the model uses pre-set algorithms, such as semantic analysis algorithms, content extraction algorithms, and image recognition algorithms, to comprehensively analyze and process the text, table, image, or seal content in the contract document, thereby outputting the parsing results of the contract document to be inspected.

[0044] Text parsing results focus on the precise extraction of structured data from contract documents, representing the most basic part of the parsing results. For example, text parsing results include, but are not limited to, textual content (contract amount, contract number, etc.), tabular information, signature and seal locations, and lists of attachments. Content parsing results, on the other hand, focus on the meaning and true intent of contract documents, representing a more advanced part of the parsing results. For example, content parsing results include, but are not limited to, the legal meaning and true intent of contract clauses, implicit rights and obligations, and logical consistency between clauses.

[0045] After obtaining comprehensive parsing results that include both superficial and / or in-depth content, the existence of errors, inconsistencies, or violations in the contract documents can be determined by verifying these results. The verification model is used to check and verify the parsing results. The parsing results are preset, fixed models, which can be unmodifiable fixed models stored in the memory of the Contract Inspection System 1000 before leaving the factory, or models written, adjusted, or replaced by operators in the memory.

[0046] When the parsed results are input into the verification model, the model uses built-in algorithms to perform multi-dimensional verification and proofreading of the parsed results, thereby outputting the verification results for the contract documents to be inspected. It can be understood that the verification process includes both superficial text verification and in-depth content verification, ensuring the logical rigor, comprehensiveness, and formal correctness of the contract inspection process. Verification results include "verification passed" and "verification failed," corresponding to the absence of problems in the contract documents and the presence of problems such as errors, inconsistencies, or violations, respectively.

[0047] For example, the verification model verifies the following based on the parsing results: the compliance of the contracting parties, confirming that the contracting parties have legal qualifications and the ability to perform; the completeness and correctness of the contract terms, focusing on whether core elements such as the subject matter, quantity, quality, price, performance period, location, and method are clearly defined and consistent with the content of the contract registration; the equivalence and legality of rights and obligations, reviewing whether the breach of contract liability and dispute resolution clauses are fair and do not violate the mandatory provisions of laws and administrative regulations; the rigor of the text, ensuring that the expression is clear and unambiguous, the attachments are complete, and the signing procedure and form comply with legal requirements; and the standardization of the form, ensuring that there are no formal problems such as typos, garbled characters, or incomplete signatures and seals. If all of the above are found to be without problems, the verification result is determined to be verification passed; otherwise, the verification result is determined to be verification failed.

[0048] If the verification fails, it can be determined that there is a problem with the contract document. The verification result also includes the specific problem that caused the failure, and this part of the verification result is defined as an anomaly. Anomaly information is used to precisely locate the problematic content in the contract document. At this point, based on the anomaly information, the specific clause or content with the problem can be uniquely and accurately identified in the contract document (e.g., a problem with the appendix or signature section).

[0049] For example, if the content regarding the price and performance period in Contract Clause 1.1.8 of the contract document is unclear or inconsistent with the content of the contract registration, resulting in the verification failing, the verification result will store this part of the content as abnormal information. If the content in Contract Clause 3.2.4 of the contract document may be a potential risk item for "our party", thus causing the verification to fail, the verification result will store this part of the content as abnormal information.

[0050] Meanwhile, to better indicate the location of abnormal information and help verification personnel quickly locate problematic clauses, thereby improving the efficiency of subsequent manual review and modification, the corresponding content of the abnormal information in the original contract document (hereinafter referred to as the feedback content) will be visually displayed in its corresponding position within the original contract document. Visual feedback refers to the process of marking content that needs to be emphasized and precisely located with a significant visual effect. For example, visual feedback methods may include, but are not limited to, highlighting the feedback content, bolding the feedback content, annotating the feedback content, and selecting the feedback content by boxing, etc.

[0051] At this point, with the contract document receiving visual feedback, verification personnel can accurately locate and quickly identify potentially problematic feedback content. This allows them to review the feedback content to more accurately determine whether it's a false alarm or a genuine anomaly, or to correct the feedback content to ensure the accuracy and reliability of the contract document. Alternatively, if there is a large amount of feedback or the content cannot be corrected, the original contract document can be scrapped and a new contract document initiated to effectively protect the legal rights and interests of both parties. It is understandable that, compared to simply indicating contract document anomalies, the contract inspection method in this application, through the step of visual feedback, greatly simplifies the workflow of subsequent verification personnel and significantly improves work efficiency.

[0052] In the contract inspection method provided in this application embodiment, the obtained contract document is parsed using a parsing model to obtain the parsing result, and then the parsing result is verified to determine whether there are textual and / or content problems in the contract document. If the verification fails (i.e., a problem is determined to exist), the location of the abnormal information in the verification result corresponding to the contract document is visualized as feedback, allowing technical personnel to check or correct the contract document. Therefore, the contract inspection method of this application can not only perform textual inspections on structurally complex contract documents, but also inspect specific content, with high inspection efficiency and accuracy. Furthermore, by accurately locating the specific location of abnormal information, it significantly reduces the cost of manual review.

[0053] Please see Figure 2 and Figure 3 In some implementations, the parsing model includes a text parsing model and a content parsing model. Step 03 can be implemented through steps 031, 033, and 035, specifically as follows: Step 031: Identify the category of the contract document to determine whether it belongs to the first or second category of contracts. The second category of contracts is more complex than the first category of contracts. Step 033: If the contract document is a Type I contract, use a text parsing model to parse the Type I contract to obtain the text parsing result, which will then be used as the parsing result; and Step 035: When the contract document is a type II contract, the content parsing model is used to parse the type II contract to obtain text parsing results and content parsing results, and the text parsing results and content parsing results are combined to form the parsing result.

[0054] Specifically, it can be understood that before generating the contract document, there is a contract initiation process, which can be considered the source of the contract document. The contract initiation process is initiated by the contract management system and is the part of contract document drafting and creation. The contract initiation process mainly includes entering the contract content, going through the approval process, and generating the contract document. During the contract document generation stage, to improve efficiency and ensure the contract's standardization and accuracy, a pre-defined formatted template is typically used to assist in the contract document generation process. It can be understood that contract documents are generally standardized documents formed by filling in the formatted template.

[0055] Based on the complexity of the contract—such as whether it contains special legal clauses, the complexity of the logical relationships between clauses, and the content of attachments—contract documents can be divided into two categories: Category I contracts, which are standardized in format and have low complexity, and Category II contracts, which are more complex in format. Understandably, to improve inspection efficiency, different inspection methods can be adopted for Category I and Category II contracts, which differ in complexity. For example, some procedures during the inspection process may differ, or the entire inspection process may differ.

[0056] Category identification is the process of classifying contract documents based on factors such as format and content. In some implementations, contract documents are marked as either Category I or Category II contracts during the contract initiation process, and the category of the contract document can be determined by obtaining the marking information during subsequent category identification. In other implementations, category identification is performed by a pre-defined category identification model. When a contract document is input for category identification, the category identification model parses the contract document, performing tasks such as semantic understanding and comparison with a formatted template, and then determines the category of the contract document based on the parsing results.

[0057] When the contract document is determined to be a Type I contract, the complexity of the contract document is low and it is consistent with the formatted template. At this time, it is not easy for content-level anomalies to occur, such as missing essential clauses, unclear rights and obligations, poor compliance with laws and regulations, or potential risks. Instead, it is more likely to have text-level anomalies such as typos or incorrect punctuation, partial formatting errors, misaligned signatures or seals, or missing attachments.

[0058] Text parsing models refer to the precise extraction of structured data. They are models that use rule engines to accurately extract tabular data, numerical information, and formatted content, focusing on shallow text-level content extraction. It should be noted that signatures and seals also contain textual content, such as the text within the signature or seal; therefore, the parsing of signatures and seals also falls under the category of text parsing. The text parsing model is a pre-defined, fixed model. It can be an unmodifiable fixed model stored in the memory of the Contract Inspection System 1000 before leaving the factory, or a model that can be written, adjusted, or replaced by operators in the memory.

[0059] In some cases, when the contract document is a PDF file, the iText text parsing model, through the iText software library, extracts the content of the contract document. By calling the iText parsing engine, it can identify and extract the text content and its hierarchical structure within the contract document, including the position and formatting information of elements such as paragraphs and tables. Its core technology lies in the decoding and rendering rule analysis of the object tree within the contract document, accurately handling documents with complex encoding or typesetting, providing a stable and reliable technical solution for automated text parsing.

[0060] In other examples, the text parsing model can also be a combination of OCR and NLP technologies as described above. OCR technology extracts text from contract documents, digitizing the content for subsequent review. Then, NLP technology uses basic natural language processing techniques to analyze the text, identifying entity information (such as contracting parties, amounts, and clauses) and performing basic text analysis to check for simple grammatical and logical errors at the text level.

[0061] The text parsing results output by the text parsing model serve as the parsing results for the first type of contract. It can be understood that, since the above text parsing process does not require analysis of deeper legal content within the first type of contract, the overall parsing process is relatively short and efficient, making it quite complete for handling the less complex first type of contract.

[0062] If the contract document is determined to be a type II contract, the contract document is highly complex and may have added a lot of content on top of the formatted template. In this case, in addition to text-level anomalies, content-level anomalies may also occur.

[0063] Content parsing models are models focused on semantic understanding for intelligent analysis of contract document content. Beyond shallow text-level content extraction, they can also deeply understand the legal meaning and true intent of contract clauses, identify implicit rights and obligations, analyze logical consistency between clauses, and assess contract integrity and compliance. The content parsing model is a pre-set, fixed model, which can be an unmodifiable fixed model stored in the memory of the Contract Inspection System 1000 before leaving the factory, or a model that can be written to, adjusted, or replaced by operators in the memory.

[0064] In some cases, the content parsing model is a large language model. Large language models use their deep neural network structure to perform multi-level semantic encoding on contract documents, thereby extracting textual content. In more complex content analysis scenarios, the model can extract structured information from clause texts, identify key legal entities such as subjects, objects, obligations, and conditions, and their relationships. Then, based on semantic knowledge gained from training with massive amounts of legal texts, it constructs a deep semantic representation of the clauses, analyzes the precise connotations and potential interpretive space of legal terms, and thereby infers the true intent and implicit rights and obligations of the clauses, thus systematically assessing the overall rigor and legal risks of the contract.

[0065] In some cases, the content parsing model is a combination of specialized natural language processing (NLP) models. In addition to text extraction, the NLP model combination uses named entity recognition (NER) and relation extraction models to accurately extract legal entities, key clauses, and explicit rights and obligations in the contract. Then, it uses pre-trained language models (such as Legal-BERT and CaseLaw BERT) fine-tuned with a large number of legal documents to perform deep encoding and reasoning on the semantics of the clauses, legal intent, and potential interpretations.

[0066] The content parsing model outputs both text parsing results and content parsing results, and uses the combination of the two as the parsing result for the second type of contract. It can be understood that the content parsing model performs both shallow text parsing and deeper content parsing on the second type of contract. The overall parsing process is time-consuming, but the parsed content is more complete, making it very suitable for handling highly complex second-type contracts.

[0067] To improve the convenience and reliability of subsequent verification processes for the parsed results, the parsed results can be in JSON (JavaScript Object Notation) format. This lightweight data exchange format is easy for humans to read and write, and also easy for machines to parse and generate. In some implementations, the JSON-formatted parsed results are also stored in a backend database for later data verification and problem troubleshooting.

[0068] Therefore, in the contract inspection method provided in this application embodiment, a text parsing model is used to obtain the text parsing result for the first type of contract and the result is used as the parsing result. For the second type of contract, a content parsing model is used to obtain the text parsing result and the content parsing result are combined to form the parsing result. The contract inspection method can flexibly adapt to different types of contract documents and provide parsing processes with different levels of complexity. This ensures sufficient parsing depth while avoiding the reduction in inspection efficiency caused by over-parsing, thus fully improving the reliability of the parsing process and the practicality of the contract inspection method.

[0069] Please see Figure 3 and Figure 4 In some implementations, step 031 can be achieved through steps 0311 and 0313, specifically as follows: Step 0311: Extract the content of the contract document and obtain the content extraction results; and Step 0313: Based on the content extraction results and the preset category recognition model, classify the contract documents into either Category I or Category II contracts. The criteria for the category recognition model include at least one of the following: standardization level, number of core clauses, number of special clauses, and complexity of attachments.

[0070] Specifically, as mentioned above, the classification of contract documents primarily relies on their complexity. To better understand the content of contract documents and accurately analyze their complexity, content extraction is necessary before performing category identification on each document. The content extraction process involves steps such as content segmentation, semantic understanding, and keyword extraction to comprehensively extract content from multiple aspects, including structure and content. The result obtained after content extraction is the content extraction result. It can be understood that the content extraction result serves as the benchmark for classifying contract documents.

[0071] For example, the content extraction process obtains the format of the contract document by analyzing its structure. The content extraction process also extracts the number of core clauses and determines whether the contract document contains special legal or commercial clauses. The content extraction process can also be used to extract attachments.

[0072] The category recognition model is used to classify contract documents based on content extraction results. The category recognition model is a preset, fixed model, which can be an unmodifiable fixed model stored in the memory of the Contract Inspection System 1000 before leaving the factory, or a model that can be written, adjusted, or replaced by operators in the memory. The category recognition model sets the criteria for classifying contract documents; these criteria are the specific dimensions upon which the category recognition model performs the category recognition process.

[0073] Standardization level refers to the degree of standardization of the contract document compared to the contract template. It characterizes the level of format standardization of the identified contract document and is typically obtained by comparing the format of the contract document extracted from the content with the contract template. Number of core clauses refers to the specific number of critical core clauses in the contract document, characterizing the importance of the identified contract document. Number of special clauses refers to whether the contract document contains special legal or commercial clauses, and the specific number of such clauses, characterizing the special nature and complexity of the identified contract document. Annex complexity refers to the complexity of the annexes in the contract document, characterizing the complexity of the content of the identified contract document.

[0074] In some implementations, the category identification model assigns different weights (the weight can be 0, i.e., the judgment criterion is not considered) to the standardization degree, the number of core clauses, the number of special clauses, and the complexity of the appendices, so as to flexibly and accurately identify the categories of contract documents, and then execute corresponding contract inspection methods for different categories of contract documents to improve inspection efficiency.

[0075] Therefore, in the contract inspection method provided in this application embodiment, the content extraction results of the contract documents are input into the category recognition model to achieve accurate and reliable classification of the contract documents. The category recognition model can evaluate the complexity of the contract documents from multiple aspects, such as the degree of standardization, the number of core clauses, the number of special clauses, and the complexity of attachments, thereby improving the accuracy and flexibility of contract classification. This is conducive to adopting targeted inspection methods for contracts with different complexities, and significantly improving inspection efficiency and reliability.

[0076] Please see Figure 2 and Figure 3 In some implementations, the content parsing model is a large language model, and the configuration process of the large language model includes at least one of configuring temperature parameters, configuring the maximum number of tokens, and configuring confidence threshold.

[0077] Specifically, when the content parsing model is a large language model, a professional configuration process is required to ensure the accuracy and reliability of processing the second type of contract. Large language model configuration refers to configuring the model's core parameters. The configuration process directly determines the model's performance, inference efficiency, and application reliability, ensuring the accuracy and reliability of the generated parsing results.

[0078] The temperature parameter controls the randomness of the model's output generation, adjusting the smoothness of the output probability distribution. A higher temperature parameter flattens the probability distribution, increasing the exploration of low-probability words during generation, resulting in more diverse and novel outputs, but potentially accompanied by incoherence. A lower temperature parameter, on the other hand, makes the probability distribution sharper, with the model highly focused on the highest-probability candidate words, resulting in more deterministic, conservative, and coherent outputs, but potentially lacking originality or becoming repetitive. Technicians can configure the temperature parameter appropriately based on actual needs such as the importance of the contract.

[0079] The maximum token count is the upper limit on the number of tokens allowed to be output in a single generation process. It limits the physical length of the generated text and primarily serves to prevent the generation of lengthy, irrelevant, or looping text, while also managing computational costs and response time. Technical personnel can configure the maximum token count based on the type of contract document and the corresponding specific task requirements to avoid truncated responses due to an insufficient maximum token count or resource waste due to an excessive maximum token count.

[0080] A confidence threshold is the minimum probability or score that a model must assign to a specific output (such as a term or a classification label) to ensure the reliability of the parsed output. A higher confidence threshold ensures high certainty in the output, improving accuracy but potentially missing some correct results; a lower confidence threshold improves recall but may introduce errors. Technical personnel can configure the confidence threshold appropriately based on actual needs such as the importance of the contract.

[0081] Therefore, in the contract inspection method provided in this application embodiment, by configuring at least one of temperature parameters, configuring the maximum number of tokens, and configuring confidence threshold, the content parsing model generation behavior can be finely and purposefully controlled to ensure that the content parsing model can meet the comprehensive requirements of the richness and reliability of the generated parsing results under different application scenarios.

[0082] Please see Figures 2 to 5 In some implementations, step 035 can be achieved through steps 0351 and 0353, specifically as follows: Step 0351: Generate content parsing prompts based on the preset prompt template. The prompt template includes at least one of the following: role positioning, task description, parsing dimensions, and format requirements. The parsing dimensions include at least one of the following: compliance, reasonableness, completeness, and consistency of contract terms; and Step 0353: Input the content parsing prompts and the contract document into the content parsing model to generate parsing results, which include text parsing results and content parsing results.

[0083] Specifically, content parsing prompts are words, phrases, or statements that define the content and form of the generated parsing results. They reflect compliance with the parsing objectives and improve the accuracy and efficiency of the content parsing model in interpreting contract documents. Prompt templates are formatted templates used to enhance the standardization and interpretability of contract parsing prompts. To standardize the content and form of contract parsing prompts, prompt templates must include at least one of the following: a role definition to enhance parsing professionalism, a description of the purpose and task, a clear parsing dimension, and formatting requirements to standardize the format.

[0084] Role positioning refers to the specific requirements for the role setting of the content parsing prompts, used to guide the contract parsing model in its parsing of contract documents. Task description refers to the prompts that indicate the steps to be taken in executing the contract parsing process, used to guide the contract parsing model in generating the parsing results. Parsing dimension refers to the dimensional requirements for the content to be considered during the parsing process of the content parsing prompts, used to standardize the specific thinking direction and parsing perspective during the parsing process. Format requirements refer to the specific requirements for the format of the generated parsing results. It should be noted that the prompt template can also take other forms, such as including reference cases, precautions, etc., and this application does not impose any restrictions on this.

[0085] Among these, the analytical dimensions are crucial, directly impacting the accuracy and comprehensiveness of the analysis results. These dimensions primarily restrict the angles from which contract clauses in a contract document can be analyzed. Compliance refers to whether the clauses comply with current laws, regulations, regulatory requirements, and industry standards, aiming to identify and mitigate potential legal risks. Reasonableness refers to whether the clauses are fair, feasible, and consistent with business practices and the purpose of the transaction, ensuring substantive justice. Completeness refers to reviewing whether the contract covers all essential elements of the transaction, with no missing key information. Consistency refers to whether there are logical contradictions or conflicting expressions between different clauses, ensuring the contract's internal unity. These four dimensions are interconnected and together form the basic framework for comprehensively assessing the rigor and enforceability of contract documents.

[0086] For example, when the contract document is a copyright contract document, compliance may include, but is not limited to, compliance with laws and regulations; reasonableness may include, but is not limited to, the reasonableness of the royalty clauses and the adequacy of the liability for breach of contract; completeness may include, but is not limited to, the completeness of the rights authorization; and consistency may include, but is not limited to, the logical consistency between the clauses.

[0087] For example, the prompt template includes role positioning, task description, analysis dimensions, and format requirements. The role positioning is: "You are a senior copyright law expert with over 15 years of experience in intellectual property law practice, specializing in copyright licensing, licensing agreements, and related legal risk assessments." The task description is: "From the perspective of a professional copyright law expert, please conduct a comprehensive legal analysis of the provided contract content, focusing on assessing the compliance and risk points of copyright-related clauses." The analysis dimensions are: "1. **Completeness Analysis of Rights Authorization** - Scope of authorization, type of rights, authorization chain, etc.; 2. **Reasonableness Assessment of Royalty Clauses** - Royalty calculation, payment methods, comparison with industry standards; 3. **Sufficiency Review of Breach of Contract Liability** - Breach of contract circumstances, remedies, dispute resolution; 4. **Compliance Check with Laws and Regulations** - Consistency with relevant laws and regulations." The format requirements are: "Analysis results must be output strictly according to the following JSON format."

[0088] Furthermore, the generated content parsing prompts and the input contract document are fed into the content parsing model. The content parsing model then performs semantic parsing and content integration to generate a parsing result that meets the requirements. It can be understood that the parsing result here is a comprehensive result combining the text parsing results and the content parsing results.

[0089] Therefore, in the contract inspection method of this application, with the assistance of content parsing prompts, the generated parsing results have the advantages of high standardization of format and accurate and detailed content, thereby providing a high-quality data source for subsequent verification steps and ensuring the reliability and accuracy of the inspection results.

[0090] Please see Figure 2 and Figure 6 In some implementations, step 05 can be achieved through steps 051 and 053, specifically as follows: Step 051: Input the parsing results into the validation model. The validation model performs at least one of the following on the parsing results: field integrity check, data format validation, logical consistency analysis, and legal clause compliance review; and Step 053: Generate a verification result containing whether the verification passed or failed, based on the verification status.

[0091] Specifically, upon obtaining the parsing results, these results are input into the verification model to further verify the correctness and compliance of the contract documents. The specific properties of the verification model are as described above and will not be repeated here. The verification model performs a multi-dimensional verification mechanism on the parsing results to fully ensure the accuracy and completeness of the verification process.

[0092] More specifically, the validation model examines the parsed results from at least one of the following dimensions. Field integrity checking is the process of checking for omissions in characters, sentences, paragraphs, etc., in the parsed results, aiming to ensure no critical information is missing. Field integrity checking verifies each predefined necessary data field (such as contract subject information, amount, date, breach of contract clauses, etc.) to confirm whether any essential clauses are missing or other important content is lacking. Field integrity checking provides the data foundation for subsequent parsing in other dimensions.

[0093] Data format verification is a step that verifies the structural standardization of the data content in the parsed results. It verifies whether the content format of specific fields is correct according to predefined rules, such as whether the date conforms to the standard format, whether the amount contains the correct currency unit, whether the ID number or unified social credit code conforms to the encoding rules, whether there are garbled characters in the text, and whether there are non-compliant expressions, so as to fully ensure that the content of the contract documents is standardized and reliable.

[0094] Logical consistency analysis is a step in analyzing the logic of specific clauses and wording in a contract document. It is used to check whether there are logical conflicts between different fields or clauses in the analysis results, such as inconsistencies between the contract amount in the main text and in the capitalized form, the effective date being later than the termination date, whether the name of the contracting party is consistent with the official registration information, and whether there are logical conflicts between the preceding and following clauses.

[0095] Legal clause compliance review is a step of analyzing the legal clauses involved in the contract documents. Based on current effective laws, regulations, regulatory policies and standard contract templates, the legal clause compliance review conducts a substantive analysis of the specific content of the legal clauses designed in the analysis results. The focus is on identifying whether the clause content violates mandatory legal provisions, whether it is missing essential statutory content, and whether its rights and obligations are obviously unfair or pose significant legal risks.

[0096] Furthermore, the verification result is generated based on the specific verification process performed by the verification model on the parsing results across multiple dimensions. It can be understood that when the verification model finds a problem in the parsing result (i.e., the contract document) during the verification process, the verification result is "verification failed," and corresponding exception information associated with the problem is generated; when the verification model finds no problem in the parsing result (i.e., the contract document), the verification result is "verification passed."

[0097] Therefore, in the contract inspection method of this application, the verification model ensures that the parsing results are not only accurate and complete, but also comply with current laws and regulations and meet actual business compliance and risk control requirements by checking at least one of the following: field integrity check, data format verification, logical consistency analysis, and legal clause compliance review. This significantly improves the integrity and reliability of the contract inspection process and effectively protects the legal benefits of the contract documents.

[0098] Please see Figure 2 and Figure 7 In some implementations, step 071 can be achieved through steps 0711 and 0713, specifically as follows: Step 0711: Generate a problem list based on the verification results, and locate the corresponding text positions of the abnormal information in the original contract document. The problem list should include at least one of the following: missing fields, inconsistent content, and formatting errors; Step 0713: Add visual markers at the text locations to generate a revised version. Visual markers include highlights, annotations, or selection boxes.

[0099] Specifically, when the verification model discovers problems in the parsed results (i.e., the contract documents) during the verification process, it generates a problem list based on the anomaly information associated with the problems. In essence, the problem list is a comprehensive and systematic summary of the specific problems existing in the contract documents. The problem list can be stored in a backend database for subsequent verification and update processes, or it can be pushed to relevant verification personnel along with the contract documents used for visual feedback, for data alignment.

[0100] Among these, "Missing Fields" refers to incomplete key fields in the contract document, such as recording "Contract Number: xxxxxxx" as "Contract Code: xxxxxxxx". "Content Discrepancies" refers to discrepancies between the content in the contract document and the content recorded in the backend system, or discrepancies between the legal provisions cited in the contract clauses and the applicable legal provisions. For example, the contract amount recorded in the backend database as 1 million might show as 100,000 in the parsing results. "Formatting Errors" refers to errors in the format of the contract document, such as missing signatures or seals, or the contract not using the correct language.

[0101] It's understandable that the text location corresponding to the abnormal information in the original contract document is the same as the text location of the feedback content mentioned above. Locating this text location assists in subsequent visual feedback steps. At this point, adding visual markers to the text location makes the feedback content stand out from other content. For example, if the contract amount recorded in the backend database is 1 million, but the data extracted from the analysis results shows 100,000, the text location corresponding to this number in the original contract document can be highlighted in a striking yellow, green, or other colors, or an abnormal annotation can be added to this number, or a rectangular or circular hollow box in a striking red or other color can be used to select it. Simultaneously, a revised version of the contract document with the visual markers is generated to avoid confusion with the original contract text, effectively helping verification personnel quickly locate the problematic clauses.

[0102] Therefore, in the contract inspection method of this application, a problem list is generated based on the verification results. At the same time, visual markers are added to the text positions corresponding to the abnormal information, and a revised version of the contract document is generated. The entire visual feedback process strictly maintains the precise correspondence with the original contract document, ensuring the accuracy of problem location and the pertinence of modification suggestions, thereby helping verification personnel to quickly locate problematic clauses and improve the efficiency of subsequent manual review and modification.

[0103] Please see Figure 2 and Figure 8 In some implementations, the contract inspection method further includes: Step 073: If the verification result is "verification passed", update the inspection status of the contract document to "inspection completed". The inspection status also includes at least one of the following: pending parsing, parsing in progress, parsing completed, pending verification, verification in progress, and verification completed.

[0104] Specifically, to improve the systematic nature of the contract document inspection process, enhance inspection efficiency, and facilitate data management, the backend database stores inspection status data that records the specific execution status of the inspection process for each contract document. In other words, the inspection status represents information indicating whether each contract document has been inspected, at what stage of the inspection, whether the inspection has been completed, and the inspection results.

[0105] According to the specific processes performed in sequence during the inspection, namely the parsing and verification steps mentioned above, the inspection status includes: "Pending Parsing" when the inspection process has not started, "Parsing in progress" when the parsing process is in progress, "Parsing Completed" when the parsing process is completed, "Pending Verification" when the verification process has not started, "Verification in progress" when the verification process is in progress, "Verification Completed" when the verification process is completed, and "Inspection Completed" when it is confirmed that the entire inspection process has been completed and there are no errors.

[0106] It should be noted that "parsing completed" and "pending verification" can be understood as two states that occur in sequence. If no other steps are set between parsing and verification, the inspection status of the contract document can be automatically updated to "pending verification" after being updated to "parsing completed".

[0107] It should be noted that the inspection status of contract documents is not limited to the types mentioned above. The inspection status can also include "Parsing Error" (indicating an anomaly in the parsing process), "Verification Error" (indicating an anomaly in the verification process), or further subdividing "Verification Completed" into "Verification Passed" and "Verification Failed," etc. The above is for illustrative purposes only and does not constitute a limitation on the inspection status.

[0108] When the verification model finds no problems in the parsed results (i.e., the contract document) during the verification process, the verification result is "verification passed." At this time, the inspection status of the contract document is updated to "inspection completed" to avoid repeated inspections of this contract document in the future, which would lead to a waste of human or material resources and improve the standardization of contract inspection.

[0109] Therefore, in the contract inspection method of this application, the inspection status data provides a basis for controlling the status of the contract inspection process. Specifically, when the verification result is "inspection passed," the inspection status of the contract document is updated to "inspection completed" to avoid duplicate inspections and significantly improve inspection efficiency. The introduction of inspection status automatically synchronizes status information, enhances the transparency of the inspection process, and provides reliable support for decision-making through accurate and timely status data, which is conducive to further optimizing the contract management process.

[0110] Please see Figure 2 and Figure 9 In some implementations, the contract inspection method further includes: Step 09: In response to the verification result being unsuccessful, the contract document and verification result are pushed to the designated personnel for review. The review includes at least one of the following operations: updating the inspection status of the contract document to inspection completed, prompting to correct the contract document, or discarding the contract document and reminding to recreate a new contract document.

[0111] Specifically, after parsing and verifying the contract documents and providing visual feedback on some of them, the contract inspection method of this application also includes an alarm handling step, namely step 09, to build a complete problem handling closed loop for contract documents that fail the inspection.

[0112] When a risk is detected in a contract document, i.e., the verification result fails, the system immediately pushes the verification result and a revised version of the text contract with added visual markers to the designated personnel, such as the relevant verification personnel. For example, at this time, a graphical user interface is provided to the verification personnel through the display unit 201 described below. The verification personnel can browse the abnormal information marked with visual markers through the graphical user interface, allowing them to determine, based on their professional judgment, whether it is a false alarm or a real compliance risk. This step is called "review".

[0113] If the verification personnel confirm through review that the abnormal information in the verification results is a false alarm, the inspection status of the contract document will be changed to "Inspection Completed" and the inspection process will end. If the verification personnel confirm through review that there is a problem, but the problem can be resolved by correction, the relevant person in charge (another designated person, who may be the same person as the verification personnel or a different person) will be prompted to correct the contract document. If the verification personnel confirm through review that there is a problem, but the problem cannot be resolved by correction, the original contract document will be discarded, and the relevant person in charge will be prompted to create a new contract document. It should be noted that whether to use "prompt to correct the contract document" or "discard the contract document and remind to create a new contract document" can be identified and classified by the system through algorithms, or it can be judged manually by relevant personnel. After the above two steps occur, the "inspection status of the contract document can be updated to inspection complete" again. This application does not restrict this.

[0114] For example, in cases where an anomaly occurs such as a missing seal or garbled characters, a prompt is sent to instruct technical personnel to correct the contract document. If the anomaly is a data mapping issue, the original contract document is invalidated, a new contract document is initiated, and the inspection status of the new contract document is updated to "inspection complete," ensuring the integrity and traceability of the entire contract document review process.

[0115] Therefore, in the contract inspection method of this application, if the verification result is unsuccessful, the contract document and the verification result are jointly sent to designated personnel for review. This further improves the handling method of contract documents with abnormal information, enhances the accuracy of contract inspection, and constructs a complete closed loop for handling abnormal information. This greatly optimizes the contract management process and helps to ensure the rights and interests of both parties.

[0116] All of the above technical solutions can be combined in any way to form optional embodiments of this application, and will not be described in detail here.

[0117] In the contract inspection method provided in this application embodiment, the obtained contract document is parsed using a parsing model to obtain the parsing result, and then the parsing result is verified to determine whether there are textual and / or content problems in the contract document. If the verification fails (i.e., a problem is determined to exist), the location of the abnormal information in the verification result corresponding to the contract document is visualized as feedback, allowing technical personnel to check or correct the contract document. Therefore, the contract inspection method of this application can not only perform textual inspections on structurally complex contract documents, but also inspect specific content, with high inspection efficiency and accuracy. Furthermore, by accurately locating the specific location of abnormal information, it significantly reduces the cost of manual review.

[0118] To facilitate better implementation of the contract inspection method of this application embodiment, this application embodiment also provides a contract inspection device 200. This contract inspection device 200 can be applied to a server in software or hardware to perform contract inspection work. Since the device embodiment is basically similar to the method embodiment, the description is relatively simple. For details of the relevant technical features, please refer to the corresponding descriptions of the method embodiments provided above. The following description of the device embodiment is merely illustrative.

[0119] Please see Figure 10 Understanding this embodiment, the figure is a unit block diagram of the contract inspection device 200 provided in this embodiment. As shown in the figure, the contract inspection device 200 provided in this embodiment includes: Display unit 201; and The control unit 203 is connected to the display unit 201. The control unit 203 is used to acquire the contract document to be inspected; to parse the contract document using a preset parsing model to obtain the parsing result, which includes at least one of text parsing result and content parsing result; to verify the parsing result using a preset verification model to obtain the verification result; and, if the verification result is that the verification fails, to provide visual feedback at the corresponding position of the contract document based on the abnormal information in the verification result.

[0120] In some embodiments, the control unit 203 is used to classify the contract document to determine whether it belongs to a first type of contract or a second type of contract with higher complexity. The control unit 203 is also used to parse the first type of contract using a text parsing model when the contract document is a first type of contract, to obtain a text parsing result and use it as the parsing result; and, when the contract document is a second type of contract, to parse the second type of contract using a content parsing model to obtain both a text parsing result and a content parsing result, and to combine the text parsing result and the content parsing result to form the parsing result.

[0121] In some embodiments, the control unit 203 is used to generate content parsing prompts according to a preset prompt template. The prompt template includes at least one of role positioning, task description, parsing dimensions and format requirements. The parsing dimensions include at least one of the following: compliance, reasonableness, completeness and consistency of contract terms. The control unit 203 is used to input the content parsing prompts and contract documents into the content parsing model to generate content parsing results.

[0122] In some embodiments, the control unit 203 is used to input the parsing result into the verification model, and the verification model performs at least one of the following on the parsing result: field integrity check, data format verification, logical consistency analysis and legal clause compliance review, and is used to generate a verification result containing whether the verification passed or failed based on the verification status.

[0123] In some embodiments, the control unit 203 is used to generate a problem list based on the verification results and locate the text position corresponding to the abnormal information in the original contract document. The problem list includes at least one of missing fields, inconsistent content, and format errors. It is also used to add visual markers at the text positions to generate a revised version. The visual markers include highlighting, annotation, or selection.

[0124] In some embodiments, the control unit 203 is used to update the inspection status of the contract document to inspection completed when the verification result is that the verification is passed. The inspection status also includes at least one of: pending parsing, parsing in progress, parsing completed, pending verification, verification in progress, and verification completed.

[0125] In some embodiments, the control unit 203 is used to push the contract document and the verification result to a designated person for review via the display unit 201 in response to the verification result being unsuccessful. The review includes at least one of the following operations: updating the inspection status of the contract document to inspection completed, prompting to correct the contract document, discarding the contract document and reminding to recreate a new contract document.

[0126] Each unit in the aforementioned inspection device 200 can be implemented entirely or partially through software, hardware, or a combination thereof. Each unit can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each unit.

[0127] The inspection device 200 for the contract can be integrated into a terminal or server that has storage and a processor and thus computing power, or the inspection device 200 for the contract can be the terminal or server.

[0128] Optionally, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0129] Figure 11 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. The computer device may be a terminal or a server. Figure 11 As shown, the computer device 300 includes a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, and a program for a contract inspection method stored in the memory 302 and executable on the processor. The processor 301 is electrically connected to the memory 302. Those skilled in the art will understand that the computer device structure shown in the figures does not constitute a limitation on the computer device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0130] The processor 301 is the control center of the computer device 300. It connects various parts of the computer device 300 through various interfaces and lines. By running or loading software programs and / or modules stored in the memory 302, and calling data stored in the memory 302, it performs various functions of the computer device 300 and processes data, thereby performing overall processing of the computer device 300.

[0131] In this embodiment, the processor 301 in the computer device 300 loads the instructions corresponding to the processes of one or more contract inspection method programs into the memory 302 according to the following steps, and the processor 301 runs the contract inspection method program stored in the memory 302 to realize various functions: Obtain the contract documents to be inspected; The contract document is parsed using a preset parsing model to obtain parsing results, which include at least one of text parsing results and content parsing results. The parsing results are verified using a pre-defined verification model to obtain the verification results; and If the verification result is that the verification fails, visual feedback will be provided in the corresponding location of the contract document based on the abnormal information in the verification result.

[0132] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0133] Optional, such as Figure 11As shown, the computer device 300 also includes: a display screen 303, a radio frequency circuit 304, an audio circuit 305, an input unit 306, and a power supply 307. The processor 301 is electrically connected to the display screen 303, the radio frequency circuit 304, the audio circuit 305, the input unit 306, and the power supply 307. Those skilled in the art will understand that... Figure 11 The computer device structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0134] The display screen 303 can be used for a graphical user interface (GUI) and to receive operation commands generated by the user interacting with the GUI. The display screen 303 may include a display panel and a touch panel. The display panel can be used to display information input by the user or information provided to the user, as well as various graphical user interfaces of the computer device. These graphical user interfaces can be composed of graphics, text, icons, video, and any combination thereof. The touch panel can be used to collect touch operations performed by the user on or near it (such as operations performed by the user using a finger, stylus, or any suitable object or accessory on or near the touch panel), and generate corresponding operation commands, which then execute the corresponding program. Optionally, the touch panel may include a touch detection device and a touch controller. The touch detection device detects the user's touch location and the signal generated by the touch operation, transmitting the signal to the touch controller. The touch controller receives touch information from the touch detection device, converts it into touch point coordinates, sends it to the processor 301, and can also receive and execute commands from the processor 301. The touch panel can cover the display panel. When the touch panel detects a touch operation on or near it, it transmits the information to the processor 301 to determine the type of touch event. Subsequently, the processor 301 provides corresponding visual output on the display panel based on the type of touch event. In this embodiment, the touch panel and the display panel can be integrated into the display screen 303 to achieve input and output functions. However, in some embodiments, the touch panel and the display panel can be implemented as two independent components to achieve input and output functions. That is, the display screen 303 can also be used as part of the input unit 306 to achieve input functions.

[0135] The radio frequency circuit 304 can be used to transmit and receive radio frequency signals to establish wireless communication with network devices or other computer devices, and to transmit and receive signals with network devices or other computer devices.

[0136] Audio circuitry 305 can be used to provide an audio interface between a user and a computer device via a speaker and a microphone. Audio circuitry 305 converts received audio data into electrical signals, transmits them to the speaker, and the speaker converts them into sound signals for output. Conversely, the microphone converts collected sound signals into electrical signals, which are then received by audio circuitry 305, converted back into audio data, and output to processor 301 for processing. The audio data is then transmitted via radio frequency circuitry 304 to, for example, another computer device, or output to memory 302 for further processing. Audio circuitry 305 may also include an earphone jack to facilitate communication between peripheral headphones and the computer device.

[0137] The input unit 306 can be used to receive input numbers, characters, or object feature information (such as fingerprints, irises, facial information, etc.), and to generate keyboard, mouse, joystick, optical, or trackball signal inputs related to user settings and function control.

[0138] Power supply 307 is used to supply power to various components of computer device 300. Optionally, power supply 307 can be logically connected to processor 301 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. Power supply 307 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0139] although Figure 11 As not shown in the diagram, computer equipment 300 may also include a camera, sensor, wireless fidelity module, Bluetooth module, etc., which will not be described in detail here.

[0140] This application also provides a computer-readable storage medium for storing a computer program. This computer-readable storage medium can be applied to a computer device, and the computer program causes the computer device to execute the corresponding processes in the display control method of the embodiments of this application; for the sake of brevity, further details are omitted here.

[0141] This application also provides a computer program product including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the corresponding processes in the contract inspection method of this application embodiment, such as executing steps 01, 03, 031, 033, 035, 0351, 0353, 05, 051, 053, 071, 0711, 0713, 073, and 09. For simplicity, these will not be elaborated further here.

[0142] It should be understood that the processor in this application may be an integrated circuit chip with signal processing capabilities. In implementation, the steps of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor described above can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0143] It is understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0144] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0145] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0146] In the embodiments of this application, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0148] 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 network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0149] In addition, the functional units in 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.

[0150] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they 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 a portion 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 or a server) 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 USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0151] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A contract inspection method, characterized in that, include: Obtain the contract documents to be inspected; The contract document is parsed using a preset parsing model to obtain parsing results, which include at least one of text parsing results and content parsing results; The parsing results are verified using a preset verification model to obtain the verification results; and If the verification result is that the verification fails, visual feedback will be provided at the corresponding location in the contract document based on the abnormal information in the verification result.

2. The contract inspection method according to claim 1, characterized in that, The parsing model includes a text parsing model and a content parsing model; the process of parsing the contract document using a preset parsing model to obtain the parsing result includes: The contract document is categorized to determine whether it belongs to a first type of contract or a second type of contract, where the second type of contract is more complex than the first type of contract. When the contract document is a first type of contract, the text parsing model is used to parse the first type of contract to obtain the text parsing result, which is then used as the parsing result; and When the contract document is a second type of contract, the content parsing model is used to parse the second type of contract to obtain the text parsing result and the content parsing result, and the text parsing result and the content parsing result are combined to form the parsing result.

3. The contract inspection method according to claim 2, characterized in that, The step of classifying the contract document to determine whether the contract document belongs to a first type of contract or a second type of contract includes: The content of the contract document is extracted, and the content extraction results are obtained; and Based on the content extraction results and the preset category recognition model, the contract documents are classified into the first type of contract or the second type of contract. The judgment criteria of the category recognition model include at least one of the following: standardization degree, number of core clauses, number of special clauses, and complexity of attachments.

4. The contract inspection method according to claim 2, characterized in that, The content parsing model is a large language model, and the configuration process of the large language model includes at least one of the following: configuring temperature parameters, configuring the maximum number of tokens, and configuring confidence threshold.

5. The contract inspection method according to claim 4, characterized in that, The content parsing model is used to parse the second type of contract, including: Content parsing prompts are generated based on a preset prompt template. The prompt template includes at least one of the following: role positioning, task description, parsing dimensions, and format requirements. The parsing dimensions include at least one of the following: compliance, reasonableness, completeness, and consistency of contract terms. The content parsing prompts and the contract document are input into the content parsing model to generate the parsing results, which include the text parsing results and the content parsing results.

6. The contract inspection method according to claim 1, characterized in that, The step of verifying the parsing results using a preset verification model to obtain verification results includes: The parsing result is input into the verification model, and the verification model performs at least one of the following on the parsing result: field integrity check, data format validation, logical consistency analysis, and legal clause compliance review; and Generate a verification result that includes whether the verification passed or failed, based on the verification status.

7. The contract inspection method according to claim 6, characterized in that, If the verification result is a failure, visual feedback will be provided at the corresponding location in the contract document based on the abnormal information in the verification result, including: A problem list is generated based on the verification results, and the text positions corresponding to the abnormal information in the original contract document are located. The problem list includes at least one of the following: missing fields, inconsistent content, and formatting errors. Add visual markers at the text locations to generate a revised version; the visual markers include highlights, annotations, or selection boxes.

8. The contract inspection method according to claim 6, characterized in that, The contract inspection method also includes: If the verification result is that the verification is passed, the inspection status of the contract document will be updated to inspection completed. The inspection status also includes at least one of the following: pending parsing, parsing in progress, parsing completed, pending verification, verification in progress, and verification completed.

9. The contract inspection method according to claim 6, characterized in that, The contract inspection method also includes: In response to the verification result being unsuccessful, the contract document and the verification result are pushed to a designated person for review. The review includes at least one of the following operations: updating the inspection status of the contract document to inspection completed, prompting to correct the contract document, discarding the contract document and reminding to recreate a new contract document.

10. A contract inspection device, characterized in that, include: Display unit; and The control unit, connected to the display unit, is used for: Obtain the contract documents to be inspected; The contract document is parsed using a preset parsing model to obtain parsing results, which include at least one of text parsing results and content parsing results; The parsing results are verified using a preset verification model to obtain the verification results; and If the verification result is that the verification fails, visual feedback will be provided at the corresponding location in the contract document based on the abnormal information in the verification result.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted for loading by a processor to execute the contract inspection method according to any one of claims 1-9.

12. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing a computer program, and the processor executes the contract inspection method according to any one of claims 1-9 by calling the computer program stored in the memory.

13. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the contract inspection method according to any one of claims 1-9.

Images