Contract full-life-cycle intelligent management robot system and implementation method thereof

Abstract

The invention relates to the technical field of contract management, and discloses a contract full-life-cycle intelligent management robot system which comprises a presentation layer, an application layer, a data layer and an infrastructure layer. According to the contract full-life-cycle intelligent management robot system and the implementation method thereof, the problems of comparison, modification and recognition of contract files in complex formats are solved through the intelligent file comparison module; the OCR recognition unit supports multiple file formats, and a Tesseract OCR engine is combined with a convolutional neural network to optimize the character recognition accuracy; the image preprocessing unit executes graying, noise reduction, binaryzation and character segmentation operations to ensure high-quality image input; the comparison algorithm unit adopts a page-by-page comparison method, records modification information of texts and images, and calculates similarity by using an editing distance and structural similarity algorithm; finally, the PDF document containing the modification information and the similarity report is generated, rapid and accurate analysis of contract content is achieved, and enterprise decision-making efficiency and accuracy are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of contract management, and specifically to an intelligent management robot system for the entire life cycle of a contract and its implementation method. Background Art

[0002] At present, with the booming development of Internet operations, contract management has become a key link in enterprise operations; the number of contracts of enterprises in the Internet environment is increasing day by day, and the scope of business involved is constantly expanding. The processes of contract creation, storage, approval, and performance require an efficient and accurate management system to support, so as to ensure the smooth development of business and compliance with laws and regulations.

[0003] However, the existing Internet contract management technologies have obvious defects; most traditional methods rely too much on manual operations and are seriously insufficient in intelligent processing. Especially, there are no effective means for comparing and identifying modified complex format contract files, making it difficult to quickly and accurately analyze contract content and make intelligent decisions based on this, greatly affecting the efficiency and quality of contract management, unable to meet the growing business needs of enterprises, bringing many inconveniences and potential risks to enterprise operations. Therefore, we propose an intelligent management robot system for the entire life cycle of a contract and its implementation method. Summary of the Invention

[0004] (1) Technical Problems to be Solved

[0005] Aiming at the deficiencies of the existing technology, the present invention provides an intelligent management robot system for the entire life cycle of a contract and its implementation method, which has the advantages of intelligent contract processing and accurate risk control, and solves the problems that the existing Internet contract management technologies have obvious defects; most traditional methods rely too much on manual operations and are seriously insufficient in intelligent processing. Especially, there are no effective means for comparing and identifying modified complex format contract files, making it difficult to quickly and accurately analyze contract content and make intelligent decisions based on this, greatly affecting the efficiency and quality of contract management, unable to meet the growing business needs of enterprises, and bringing many inconveniences and potential risks to enterprise operations.

[0006] (2) Technical Solutions

[0007] To achieve the above purposes of intelligent contract processing and accurate risk control, the present invention provides the following technical solutions: An intelligent management robot system for the entire life cycle of a contract, comprising:

[0008] Presentation layer: Adopting a responsive web design framework, enabling the system interface to adapt to the screen sizes of desktop computers, tablet computers, and smart phones. The above terminals can be accessed through browsers or embedded HTML5 APPs via HTML5, CSS3, and JavaScript, supporting cross-platform access;

[0009] Application layer: It includes a contract management module, a business partner management module, an organizational structure management module, and an early warning management module. Data interaction is carried out between modules through RESTful API interfaces;

[0010] Data layer: The relational database MySQL is used to store contract information, business partner information, risk information, and log information. Redis is adopted as the data caching technology, and page caching, object caching, and query result caching are set up;

[0011] Infrastructure layer: The Docker container orchestration tool Kubernetes is used to implement microservice deployment and service - to - service calls. Requests are distributed to different microservice instances through a load balancer.

[0012] Preferably, the contract management module includes:

[0013] Contract template library: The version control system Git is used to manage the update and historical versions of contract templates, and template data is stored in XML format;

[0014] Draft storage unit: Automatically saves contract draft materials. The file storage system HDFS is used to store draft files. When storing, file naming and directory storage are carried out according to the contract unique identifier and timestamp;

[0015] The business partner management module includes:

[0016] The docking interface with the authoritative enterprise credit data platform, which conducts data interaction through the SOAP protocol and supports real - time data update and batch synchronization;

[0017] Local business partner profile database: The database table structure includes business partner name, unified social credit code, registered address, business scope, credit rating, historical transaction data, and risk indicator data;

[0018] Risk assessment algorithm model: It is calculated based on the credit status, business data, and legal litigation records of business partners. This algorithm model uses the decision tree algorithm for risk assessment, is trained with historical business partner data, and converts multi - dimensional data into feature vectors that can be processed by the algorithm through feature engineering.

[0019] Preferably, the early warning management module includes:

[0020] An architecture developed based on Spring Boot, and the Spring Cloud microservice framework is used to achieve service governance;

[0021] Quartz scheduled task scheduling framework, and the Cron expression is used to accurately set the execution period of scheduled tasks;

[0022] MybatisPlus data access framework, which optimizes database connections using the connection pool technology HikariCP;

[0023] Aviator expression engine, which supports dynamic expression configuration of warning rules;

[0024] Warning rule configuration unit, which can configure warning items, warning data sources, warning time periods, warning objects and message sending methods. The configuration information is stored in a dedicated configuration database table, and SQL statements are used to operate on the configuration information;

[0025] Message push component, which supports WeChat, DingTalk, and email message push methods, uses the message queue RabbitMQ for message storage and distribution, and sends messages asynchronously.

[0026] Preferably, the intelligent file comparison module includes:

[0027] OCR recognition unit, which supports the recognition of files in JPG, GIF, PNG, BMP, TIF, and PDF formats, uses the TesseractOCR engine for text recognition, and optimizes the recognition accuracy using the convolutional neural network algorithm;

[0028] Image preprocessing unit, which performs grayscale conversion, noise reduction, binarization, and character segmentation operations on the file, uses the Gaussian filtering algorithm for noise reduction, and the adaptive threshold algorithm for binarization processing;

[0029] Comparison algorithm unit, which compares contract files page by page, records modified, deleted, and added content, and uses a combination of the text-based longest common subsequence algorithm and the image-based perceptual hash algorithm for comparison;

[0030] Comparison report generation unit, which generates a report containing file modification information and similarity, uses HTML and CSS to generate a visual report, and stores the report in a dedicated report storage area in PDF format.

[0031] Preferably, the contract management module further includes:

[0032] Contract approval process, which automatically transfers contracts to the corresponding departments and personnel according to preset approval rules and processes, uses the workflow engine Activiti to achieve process automation, and the approval process is defined and visually displayed through the BPMN2.0 specification;

[0033] The risk assessment algorithm model includes:

[0034] Credit status data collection module, which collects the latest credit data by calling an external credit assessment API and performs data cleaning and normalization on the data;

[0035] The business data collection module collects data from the enterprise financial system and sales system, and uses the ETL tool Kettle for data extraction and transformation;

[0036] The legal litigation record collection module establishes a data link with the court announcement system and regularly obtains litigation information through web crawler technology.

[0037] Preferably, the image preprocessing unit includes:

[0038] The grayscale module converts the color image into a grayscale image using the weighted average method;

[0039] The denoising module uses the median filtering algorithm to remove image noise;

[0040] The binarization module performs image binarization processing according to the global threshold algorithm;

[0041] The character segmentation module uses the connected component analysis algorithm to split the text line into individual characters.

[0042] Preferably, the comparison algorithm unit calculates the file similarity in a page-by-page comparison manner. For the text part, the edit distance algorithm is used to calculate the similarity, and for the image part, the structural similarity algorithm is used to calculate the similarity.

[0043] Preferably, the presentation layer adopts a user behavior analysis module. By collecting the operation logs of users in the system, including click, browse, and search operations, and using the big data analysis technology Apache Flink for analysis, the user interface layout and function recommendations are optimized according to the analysis results;

[0044] The data layer uses the data encryption technology AES encryption algorithm to encrypt and store the stored contract content and customer information, and the encryption key is managed through the hardware security module HSM.

[0045] Preferably, the message push component of the warning management module adopts a message template mechanism. Users can customize the message template, use the Freemarker template engine to generate personalized message content, and send different messages according to different warning types and recipient roles.

[0046] A method for implementing a contract full-life cycle intelligent management robot includes a contract full-life cycle intelligent management robot system, and also includes the following steps:

[0047] S1. Start the responsive web design framework to ensure that the system interface can adapt to the screen sizes of desktop computers, tablets, and smartphones; achieve cross-platform access through HTML5, CSS3, and JavaScript, and support browser or embedded HTML5 APP-side access;

[0048] S2. Initialize the contract management module, customer management module, organizational structure management module, and warning management module; data interaction is carried out between modules through RESTful API interfaces to ensure the security and consistency of data transmission;

[0049] S3. Use the Docker container orchestration tool Kubernetes to deploy microservices, and distribute requests to different microservice instances through a load balancer; configure the infrastructure to ensure high availability and scalability;

[0050] S4. Configure the relational database MySQL to store contract information, customer information, risk information, and log information; use Redis as a data caching technology, set up page caching, object caching, and query result caching to improve performance; encrypt and store sensitive data using the AES encryption algorithm, and the key is managed by a hardware security module (HSM);

[0051] S5. Load the Tesseract OCR engine and its optimized convolutional neural network model to support the recognition of multiple file formats; load the edit distance and structural similarity algorithms to provide support for subsequent file comparison;

[0052] S6. Configure the message template mechanism of the message push component to allow users to customize the message format; integrate WeChat, DingTalk, and email channels, and use RabbitMQ to asynchronously distribute messages to ensure instant and reliable notifications;

[0053] S7. Configure the Apache Flink big data analysis technology to collect and analyze the operation logs of users in the system (click, browse, search operations); optimize the user interface layout and function recommendations according to the analysis results to improve the user experience;

[0054] S8. Conduct system connectivity and functionality tests to ensure normal data interaction between modules; ensure that the intelligent file comparison module can accurately identify and compare files in multiple formats; verify the function of the warning management module to send message notifications according to predetermined rules.

[0055] (III) Beneficial Effects

[0056] Compared with the prior art, the present invention provides a contract full-life-cycle intelligent management robot system and its implementation method, which have the following beneficial effects:

[0057] 1. The intelligent management robot system for the full life cycle of the contract and its implementation method solve the problems of comparison and modification recognition of contract documents in complex formats through the intelligent file comparison module; the OCR recognition unit supports multiple file formats and uses the Tesseract OCR engine combined with a convolutional neural network to optimize the accuracy of text recognition; the image preprocessing unit performs grayscale conversion, noise reduction, binarization, and character segmentation operations to ensure high-quality image input; the comparison algorithm unit adopts a page-by-page comparison method, records the modification information of the text and images, and calculates the similarity using the edit distance and structural similarity algorithms; finally, a PDF document containing the modification information and similarity report is generated, realizing the rapid and accurate analysis of the contract content, and significantly improving the efficiency and accuracy of enterprise decision-making.

[0058] 2. The intelligent management robot system for the full life cycle of the contract and its implementation method effectively improve the contract management efficiency and reduce risks through the early warning management module; developed based on Spring Boot, combined with Spring Cloud to achieve service governance, and the Quartz framework accurately sets the timing task cycle; the Aviator expression engine supports dynamic early warning rule configuration, allowing non-technical personnel to easily set personalized conditions; the message push component integrates WeChat, DingTalk, and email channels, and uses RabbitMQ to asynchronously distribute messages to ensure instant and reliable notifications; through the automated early warning mechanism and flexible message push function, this module enables enterprises to receive reminders at the budding stage of problems, take preventive measures in advance, avoid potential risks, and greatly enhance the quality and response speed of contract management. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic structural diagram of the system of the present invention;

[0060] Figure 2 It is a schematic structural diagram of the contract management subsystem of the present invention;

[0061] Figure 3 It is a sequence diagram of the merchant risk module of the present invention;

[0062] Figure 4 It is a flowchart of the early warning engine of the present invention;

[0063] Figure 5 It is a sequence diagram of the intelligent file comparison of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the embodiments and the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0065] Please refer to Figures 1-5 , a contract full-life cycle intelligent management robot system, including:

[0066] Presentation layer: Adopting a responsive web design framework, enabling the system interface to adapt to the screen sizes of desktop computers, tablets, and smartphones. The above terminals can access through browsers or embedded HTML5 APPs via HTML5, CSS3, and JavaScript, supporting cross-platform access;

[0067] Application layer: Including a contract management module, a business partner management module, an organizational structure management module, and an early warning management module. Data interaction is carried out between the modules through RESTful API interfaces;

[0068] Data layer: Using the relational database MySQL to store contract information, business partner information, risk information, and log information, adopting Redis as a data caching technology, and setting page caching, object caching, and query result caching;

[0069] Infrastructure layer: Using the Docker container orchestration tool Kubernetes to implement microservice deployment and service-to-service calls, and distributing requests to different microservice instances through a load balancer.

[0070] A method for implementing a contract full-life cycle intelligent management robot, including a contract full-life cycle intelligent management robot system, and further including the following steps:

[0071] S1. Start the responsive web design framework to ensure that the system interface can adapt to the screen sizes of desktop computers, tablets, and smartphones; implement cross-platform access through HTML5, CSS3, and JavaScript, and support access through browsers or embedded HTML5 APPs;

[0072] S2. Initialize the contract management module, the business partner management module, the organizational structure management module, and the early warning management module; data interaction is carried out between the modules through RESTful API interfaces to ensure the security and consistency of data transmission;

[0073] S3. Deploy microservices using the Docker container orchestration tool Kubernetes, and distribute requests to different microservice instances through a load balancer; configure the infrastructure to ensure high availability and scalability;

[0074] S4. Configure the relational database MySQL to store contract information, customer information, risk information, and log information; use Redis as a data caching technology, and set up page caching, object caching, and query result caching to improve performance; encrypt and store sensitive data using the AES encryption algorithm, and manage the key by a Hardware Security Module (HSM);

[0075] S5. Load the Tesseract OCR engine and its optimized convolutional neural network model to support the recognition of multiple file formats; load the edit distance and structural similarity algorithms to support subsequent file comparison;

[0076] S6. Configure the message template mechanism of the message push component to allow users to customize the message format; integrate WeChat, DingTalk, and email channels, and use RabbitMQ to asynchronously distribute messages to ensure instant and reliable notifications;

[0077] S7. Configure the Apache Flink big data analysis technology to collect and analyze the operation logs (click, browse, search operations) of users in the system; optimize the user interface layout and function recommendations based on the analysis results to enhance the user experience;

[0078] S8. Conduct system connectivity and functionality tests to ensure normal data interaction between modules; ensure that the intelligent file comparison module can accurately identify and compare files in multiple formats; verify the function of the early warning management module to send message notifications according to predefined rules.

[0079] Example 1: System Architecture and Module Design

[0080] The intelligent contract robot system of the present invention aims to solve the deficiencies in existing contract management technologies, improve the efficiency and quality of contract management through intelligent processing and precise risk control. The system structure is as Figure 1 shown, including a presentation layer, an application layer, a data layer, and an infrastructure layer.

[0081] 1. Presentation Layer

[0082] The presentation layer adopts a responsive web design framework to ensure that the system interface can adapt to the screen sizes of desktop computers, tablets, and smartphones. Users can access the browsers of the above terminals or the embedded HTML5 APP through HTML5, CSS3, and JavaScript to achieve cross-platform access. In addition, the system also integrates a user behavior analysis module, which uses Apache Flink for big data analysis to optimize the user interface layout and function recommendations based on user operation logs (clicks, browsing, searches, etc.).

[0083] 2. Application Layer

[0084] The application layer contains four main modules: the contract management module, the business partner management module, the organizational structure management module, and the early warning management module. Data interaction between modules is carried out through RESTful API interfaces to ensure the security and consistency of data transmission. Among them:

[0085] Contract management module: This module is responsible for the creation, storage, approval, and performance process management of contracts. It includes a contract template library that uses the Git version control system to manage and update contract templates and stores template data in XML format. The draft storage unit automatically saves contract draft materials, and the file name and directory are named according to the contract unique identifier and timestamp and stored in the HDFS file system. The contract approval process automatically flows to relevant departments and personnel according to preset rules, and the workflow engine Activiti realizes the automation of this process.

[0086] Business partner management module: Connects to the authoritative enterprise credit data platform through the SOAP protocol to support real-time update and batch synchronization of business partner information. The local business partner archive database records detailed information such as business partner name, unified social credit code, and registered address. The risk assessment algorithm model uses the decision tree algorithm to calculate the credit status, business data, and legal litigation records of business partners to provide accurate risk assessment.

[0087] Early warning management module: Developed based on Spring Boot and combined with the Spring Cloud microservice framework to achieve service governance. The Quartz timed task scheduling framework is used to set precise timed task execution cycles, and the Mybatis Plus framework optimizes database connections. The Aviator expression engine supports dynamic configuration of early warning rules. The early warning rule configuration unit allows setting early warning items, data sources, time periods, objects, and message sending methods, and these configuration information are stored in a dedicated database table. The message push component supports message pushes via WeChat, DingTalk, and email, and distributes messages asynchronously through RabbitMQ.

[0088] Intelligent File Comparison Module: The OCR recognition unit can process files in multiple formats (JPG, GIF, PNG, BMP, TIF, PDF), and uses the Tesseract OCR engine and convolutional neural network algorithm to improve the accuracy of text recognition; the image preprocessing unit performs operations such as grayscaling, noise reduction, binarization, and character segmentation to prepare high-quality images for the comparison algorithm unit to process; the comparison algorithm unit uses a page-by-page comparison method to calculate the similarity of the text and image parts, and generates a PDF document containing modification information and a similarity report.

[0089] 3. Data Layer

[0090] The data layer selects MySQL as the relational database to store contract information, customer information, risk information, and log information; to improve performance, Redis cache technology is adopted, and page cache, object cache, and query result cache are set; in addition, all sensitive data such as contract content and customer information are encrypted and stored using the AES encryption algorithm, and the key is managed by the Hardware Security Module (HSM).

[0091] 4. Infrastructure Layer

[0092] The infrastructure layer uses the Docker container orchestration tool Kubernetes to deploy microservices to ensure efficient invocation between services; the load balancer evenly distributes requests to each microservice instance to ensure the high availability and scalability of the system.

[0093] Example Two:

[0094] The contract management module is responsible for various affairs during the contract lifecycle, such as creation, storage, approval, and performance, etc.; in particular, this example focuses on the automated implementation of the contract approval process.

[0095] Contract Template Library: Use the Git version control system to manage and update contract templates, and store template data in XML format to ensure the consistency and traceability of templates.

[0096] Draft Storage Unit: All contract draft materials will be automatically saved to the HDFS file system, and the file naming rule combines the contract unique identifier and timestamp for easy retrieval and management.

[0097] Automated Approval Process: The workflow engine Activiti transfers the contract to the corresponding departments and personnel according to the preset approval rules, and the whole process follows the BPMN2.0 specification, visually displaying the status changes of each step.

[0098] Implementation Process:

[0099] When a new contract is submitted, the system immediately starts the approval process and automatically assigns it to the appropriate approver according to the contract type;

[0100] Each approval node has a clear time limit. If the time limit is exceeded, the system will automatically remind the relevant responsible person to speed up the progress;

[0101] After the approval is completed, the system automatically generates a formal contract text, archives and stores it, and notifies the relevant personnel at the same time.

[0102] As Figure 2 shown, the contract management subsystem includes functional modules such as payment contracts, collection contracts, contracts not involving amounts, contract stamping, contract disputes, contract cancellation, contract filing, contract performance plans, contract closure, and contract changes; these modules together constitute an information-based, digital, and intelligent management system for the entire contract life cycle.

[0103] Example 3:

[0104] The merchant management module not only maintains the archive information of enterprise customers but also provides risk assessment services; this example details how to use the decision tree algorithm for accurate risk assessment.

[0105] Real-time docking interface: Establish a connection with the authoritative enterprise credit data platform through the SOAP protocol to ensure the timeliness and accuracy of merchant information.

[0106] Local database: Records basic information such as merchant name, unified social credit code, and registered address, as well as advanced information such as credit ratings, historical transaction data, and risk indicators.

[0107] Risk assessment model: A risk assessment model built based on the decision tree algorithm, which comprehensively considers factors such as the credit status, business data, and legal litigation records of merchants and outputs the final risk score.

[0108] Implementation details:

[0109] The system regularly fetches the latest credit data from the external credit assessment API every day, updates it to the local database after cleaning and normalization processing;

[0110] For important customers, the system generates a detailed credit report every month for management reference;

[0111] If the risk score of a certain merchant reaches the warning threshold, the system will automatically send an alert to the person in charge of the relevant department to take preventive measures.

[0112] See Figure 3 the sequence diagram of the merchant risk module. The implementation effect of the merchant risk module includes the merchant risk click module and the details display page, providing an intuitive operation interface to configure and view risk information.

[0113] Example 4:

[0114] The early warning management module aims to help enterprises identify potential risks in advance and make timely response strategies; the following is the specific implementation plan for dynamically configuring early warning rules and message push.

[0115] Early warning rule configuration unit: Allows administrators to flexibly define early warning items, data sources, time periods, objects, and message sending methods. These configuration information are securely stored in a dedicated database table for easy query and modification in the future.

[0116] Message push component: Supports multiple communication channels (WeChat, DingTalk, email), and asynchronously distributes messages through RabbitMQ to ensure the immediacy and reliability of information transmission.

[0117] Aviator expression engine: Used to parse and execute complex early warning logics, enabling non-technical personnel to easily set personalized early warning conditions.

[0118] Implementation details:

[0119] Administrators can adjust early warning rules at any time according to business needs, and the changes will take effect without restarting the service;

[0120] The system will automatically generate personalized message content according to different early warning types and recipient roles, improving communication efficiency;

[0121] In case of an emergency, the system can quickly send warnings to designated personnel, reducing response time and minimizing losses.

[0122] Figure 4 Shows the flowchart of the early warning engine, where steps such as scheduler scheduling, configuration loading, rule verification, and message sending clearly present the working principle of the early warning mechanism;

[0123] Scheduling is the engine that drives the entire early warning engine and the starting point of the power for the operation of the entire early warning rule; scheduling is the frequency or timing that supports the operation of early warning items. Currently, Quartz is used to implement the scheduling of timed operations, and it supports subscribing to MQ events to trigger.

[0124] In the design of early warning matters, matters have static attributes and dynamic attributes. Static attributes refer to static data obtained through data source interfaces, and dynamic attributes refer to information generated during the scheduling operation. These information can be supplemented to matter attributes, such as the occurrence of an early warning item event, the generation time of the early warning event, etc.

[0125] The selection of the warning data source of the warning platform is to select the data source through the configuration of the data service data source. The data source is the main object served by the warning engine and the fuel for generating warning messages. The rule refers to a series of constraint conditions for the list of event sources, aiming to filter out a subset of event sources that meet the constraint conditions. The rule is a combinatorial operation on the event attributes. Not only the static attributes of the event but also the dynamic attributes of the event participate in the combinatorial operation. The warning platform not only sets rules for the time point when an event occurs but also can perform free rule filtering through the Aviator expression. The warning platform can configure the expected values through the Aviator expression engine to perform secondary screening on the data source. The time agreement is a time range agreed upon by both parties of the warning event at a certain signed time point. It is offset based on the time attribute of the event source, compares the current time, and filters out events that meet the time period conditions.

[0126] To ensure that warning messages are not sent indefinitely, the number of times the warning event of the rule under the current warning item is triggered is configured controllably.

[0127] The message body is designed to meet the different message content reminders for different users. It is set that the message template can have the characteristics of a fixed message format and the filling of message variables for message construction. The warning module constructs a message entity by agreeing on the message template of the message platform, takes the subset of events filtered by rules and time as the warning events to be sent, and calls the message platform SDK to send messages. The warning platform itself needs to record the sending situation as corresponding logs for easy tracking and searching. Read the template from the warning task configuration, support the setting of message topics, message types, and message contents, construct reminder messages according to the incoming event source and message template, and set the reminder content attributes of the message body.

[0128] Example Five:

[0129] The intelligent file comparison module solves the problem that traditional methods are difficult to quickly and accurately analyze contract files in complex formats. The following is its implementation solution:

[0130] OCR Recognition Unit: Supports multiple file formats such as JPG, GIF, PNG, BMP, TIF, and PDF. Adopts the Tesseract OCR engine combined with the convolutional neural network algorithm to improve the accuracy of text recognition.

[0131] Image Preprocessing Unit: Includes four steps: grayscale conversion, noise reduction, binarization, and character segmentation. Uses the weighted average method, Gaussian filtering algorithm, adaptive threshold algorithm, and connected region analysis algorithm respectively to ensure the quality of the input image.

[0132] Comparison algorithm unit: Compare the contract documents page by page, record the modified, deleted, and added content; for text parts, use the edit distance algorithm to calculate similarity; for image parts, use the structural similarity algorithm; finally, generate a PDF report containing file modification information and similarity.

[0133] Implementation details:

[0134] After the user uploads the file to be compared, the system first performs OCR processing on it, and then splits the recognized text information into multiple segments according to the page numbers;

[0135] Next, the system calls the comparison algorithm unit to compare the original file and the modified file page by page, and marks all the differences;

[0136] Finally, the system generates a detailed comparison report, intuitively presenting the differences between the two files to help the user quickly locate the problems.

[0137] According to Figure 5 the intelligent file comparison sequence diagram, the process of the user performing file comparison involves components such as the OCRProvider service, the FileSpotService server, and the CompareService. When the user performs file comparison, the data of the standard file and the scanned file is transmitted to the OCRProvider service. The FileSpotService server reads the file and performs file recognition, and then returns the text information obtained from the file recognition to the OCRProvider service. The OCRProvider service divides the text information by page unit and distributes the data according to the data information of each page of the standard file and all the scanned file page data information to the Compare Service for page comparison to obtain the similarity of each unit page; at this time, the similarity matrix information of each unit page of the scanned copy and the standard copy is obtained, and the double comparison algorithm is executed to generate comparison report information, and the comparison report information is stored in the Mysql database.

[0138] When the user queries the comparison report, the system queries the data from the Mysql database according to the File Code of the page file and the primary key of the contract where it is located as the query conditions for the contract comparison file, obtains the report information, and returns it to the user; if the query is successful, it jumps to the page, and if it fails, it prompts the user that the query fails.

[0139] In summary, the intelligent management robot system for the full life cycle of the contract and its implementation method solve the problems of comparison and modification recognition of contract documents in complex formats through the intelligent file comparison module; the OCR recognition unit supports multiple file formats, and uses the Tesseract OCR engine combined with a convolutional neural network to optimize the accuracy of text recognition; the image preprocessing unit performs grayscale conversion, noise reduction, binarization, and character segmentation operations to ensure high-quality image input; the comparison algorithm unit adopts a page-by-page comparison method, records the modification information of the text and images, and calculates the similarity using the edit distance and structural similarity algorithms; finally, a PDF document containing the modification information and similarity report is generated, realizing the rapid and accurate analysis of the contract content, and significantly improving the efficiency and accuracy of enterprise decision-making.

[0140] Moreover, the intelligent management robot system for the full life cycle of the contract and its implementation method effectively improve the contract management efficiency and reduce risks through the early warning management module; developed based on Spring Boot, combined with Spring Cloud to achieve service governance, and the Quartz framework accurately sets the timing task cycle; the Aviator expression engine supports dynamic early warning rule configuration, allowing non-technical personnel to easily set personalized conditions; the message push component integrates WeChat, DingTalk, and email channels, and uses RabbitMQ to asynchronously distribute messages to ensure instant and reliable notifications; through the automated early warning mechanism and flexible message push function, this module enables enterprises to receive reminders at the budding stage of problems, take preventive measures in advance, avoid potential risks, greatly enhance the quality and response speed of contract management, and solve the obvious defects existing in the existing Internet contract management technologies; most traditional methods rely too much on manual operations and are seriously insufficient in intelligent processing, especially lacking effective means for the comparison and modification recognition of contract documents in complex formats, and it is difficult to quickly and accurately analyze the contract content and make intelligent decisions based on this, greatly affecting the efficiency and quality of contract management, unable to meet the growing business needs of enterprises, and bringing many inconveniences and potential risks to enterprise operations.

[0141] The relevant modules involved in this system are all hardware system modules or functional modules that combine computer software programs or protocols in the prior art with hardware. The computer software programs or protocols themselves involved in this functional module are all well-known technologies to those skilled in the art, and they are not the improvements of this system; the improvement of this system is the interaction relationship or connection relationship between the modules, that is, the overall structure of the system is improved to solve the corresponding technical problems to be solved by this system.

[0142] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A contract life cycle intelligent management robot system, characterized in that: include: Presentation layer: Adopting a responsive web design framework, the system interface can adapt to the screen size of desktop computers, tablet computers, and smart phones. Through HTML5, CSS3, and JavaScript, the above terminals can access the system through a browser or an embedded HTML5 APP, supporting cross-platform access. Application layer: including contract management module, customer management module, organizational structure management module and early warning management module. Data exchange between modules is carried out through RESTful API interface; Data layer: Use the relational database MySQL to store contract information, customer information, risk information and log information, use Redis as data cache technology, set page cache, object cache and query result cache; Infrastructure layer: Use Docker container orchestration tool Kubernetes to implement microservice deployment and service calls, and distribute requests to different microservice instances through load balancers.

2. According to claim 1, a contract life cycle intelligent management robot system is characterized in that: The contract management module includes: Contract template library: uses the version control system Git to manage contract template updates and historical versions, and stores template data in XML format; Draft storage unit: automatically saves contract draft data, uses the file storage system HDFS to store draft files, and names and stores files in directories based on the contract unique identifier and timestamp; The merchant management module includes: The docking interface with the authoritative enterprise credit data platform exchanges data through the SOAP protocol, supporting real-time data updates and batch synchronization; Local merchant archive database: The database table structure includes merchant name, unified social credit code, registered address, business scope, credit rating, historical transaction data and risk indicator data; Risk assessment algorithm model: Calculated based on the credit status, business data and legal records of merchants, this algorithm model uses a decision tree algorithm for risk assessment, uses historical merchant data for training, and converts multi-dimensional data into feature vectors that can be processed by the algorithm through feature engineering.

3. According to claim 1, a contract life cycle intelligent management robot system is characterized in that: The early warning management module includes: Based on the architecture developed by SpringBoot, the SpringCloud microservice framework is used to implement service governance; Quartz scheduled task scheduling framework, using Cron expressions to accurately set the scheduled task execution cycle; MybatisPlus data access framework uses connection pool technology HikariCP to optimize database connections; Aviator expression engine, which supports dynamic expression configuration of warning rules; The warning rule configuration unit can configure the warning items, warning data source, warning time period, warning object and message sending method. The configuration information is stored in a special configuration database table and SQL statements are used to operate the configuration information. The message push component supports WeChat, DingTalk, and email message push methods, uses the message queue RabbitMQ to store and distribute messages, and sends messages asynchronously.

4. According to claim 1, a contract life cycle intelligent management robot system is characterized in that: The intelligent file comparison module includes: OCR recognition unit, supports the recognition of JPG, GIF, PNG, BMP, TIF and PDF format files, uses Tesseract OCR engine for text recognition, and uses convolutional neural network algorithm to optimize recognition accuracy; The image preprocessing unit performs grayscale, noise reduction, binarization and character segmentation operations on the file, uses Gaussian filtering algorithm for noise reduction, and uses adaptive threshold algorithm for binarization; The comparison algorithm unit compares the contract documents page by page, records the modifications, deletions, and additions, and uses a combination of the text-based longest common subsequence algorithm and the image-based perceptual hash algorithm for comparison; The comparison report generation unit generates a report containing file modification information and similarity, and uses HTML and CSS to generate a visual report. The report is stored in a special report storage area in the PDF format.

5. According to claim 1, a contract life cycle intelligent management robot system is characterized in that: The contract management module also includes: The contract approval process automatically transfers the contract to the corresponding department and personnel according to the preset approval rules and processes. The workflow engine Activiti is used to automate the process. The approval process is defined and visualized through the BPMN2.0 specification. The risk assessment algorithm model includes: The credit status data collection module collects the latest credit data by calling the external credit assessment API, and performs data cleaning and normalization on the data; The business data collection module collects data from the enterprise financial system and sales system, and uses the ETL tool Kettle to extract and convert data; The legal litigation record collection module establishes a data link with the court announcement system and regularly obtains litigation information through web crawler technology.

6. The intelligent management robot system for the entire life cycle of a contract according to claim 1 is characterized in that: The image preprocessing unit comprises: Grayscale module, which converts color images into grayscale images using weighted average method; Noise reduction module, which uses median filtering algorithm to remove image noise; Binarization module, which performs image binarization processing according to the global threshold algorithm; The character segmentation module uses the connected component analysis algorithm to segment text lines into individual characters.

7. A contract life cycle intelligent management robot system according to claim 4, characterized in that: The comparison algorithm unit calculates the file similarity by comparing page by page. For the text part, the similarity is calculated by using the edit distance algorithm, and for the image part, the similarity is calculated by using the structural similarity algorithm.

8. The intelligent management robot system for the entire life cycle of a contract according to claim 1 is characterized in that: The presentation layer uses a user behavior analysis module to collect user operation logs in the system, including click, browse, and search operations, and uses big data analysis technology Apache Flink for analysis, and optimizes the user interface layout and function recommendations based on the analysis results; The data layer uses the data encryption technology AES encryption algorithm to encrypt and store the stored contract content and customer information, and the encryption key is managed by the hardware security module HSM.

9. The intelligent management robot system for the entire life cycle of a contract according to claim 3 is characterized in that: The message push component of the warning management module adopts a message template mechanism. Users can customize message templates and use the Freemarker template engine to generate personalized message content, and send different messages according to different warning types and recipient roles.

10. A method for implementing a contract full life cycle intelligent management robot, characterized in that: The intelligent management robot system for the entire life cycle of a contract as described in claims 1 to 9 further includes the following steps: S1. Launch a responsive web design framework to ensure that the system interface can adapt to the screen sizes of desktop computers, tablets and smartphones; achieve cross-platform access through HTML5, CSS3 and JavaScript, and support browser or embedded HTML5 APP access; S2. Initialize the contract management module, customer management module, organizational structure management module and early warning management module; each module interacts with each other through the RESTful API interface to ensure the security and consistency of data transmission; S3. Use Docker container orchestration tool Kubernetes to deploy microservices and distribute requests to different microservice instances through load balancers; configure infrastructure to ensure high availability and scalability; S4. Configure the relational database MySQL to store contract information, customer information, risk information, and log information; use Redis as data caching technology, set up page caching, object caching, and query result caching to improve performance; use the AES encryption algorithm to encrypt and store sensitive data, and the key is managed by the hardware security module (HSM); S5. Load the Tesseract OCR engine and its optimized convolutional neural network model to support the recognition of multiple file formats; load the edit distance and structural similarity algorithms to provide support for subsequent file comparison; S6. Configure the message template mechanism of the message push component to allow users to customize the message format; integrate WeChat, DingTalk and email channels, and use RabbitMQ to asynchronously distribute messages to ensure instant and reliable notifications; S7. Configure Apache Flink big data analysis technology to collect and analyze user operation logs (click, browse, and search operations) in the system; optimize the user interface layout and function recommendations based on the analysis results to improve user experience; S8. Conduct system connectivity and functionality tests to ensure normal data interaction between modules; ensure that the intelligent file comparison module can accurately identify and compare files of various formats; and verify the function of the early warning management module to send message notifications according to predetermined rules.

Images