The software code intelligent review system of this work is committed to promoting the process of information reform in China, the system through the analysis of software requirements documents and software source code, to government agencies and institutions to provide software cost estimates and software compliance line security analysis, while providing a series of corresponding response methods. Provides software cost estimation, intellectual property compliance analysis, security vulnerability analysis, and code traceability services. It aims to help government units, enterprises and other relevant departments to ensure the economic feasibility, system availability, network security and legitimacy of software development investments, promote information technology investment management and performance evaluation, and provide technical support to regulatory agencies and third-party consulting companies. The following is a creative description of the work: 1. Open source code base built by automatic iteration: automatic maintenance of old code, daily pull of popular projects, establishment of version index tree, merge similar code as the relevant main branch; 2. Project Dependency Vulnerability Retrieval: Retrieve dependency vulnerabilities according to project Maven and Gradle configuration files, generate dependency reference trees and risk assessment reports, and try to provide solutions; 3. Multi-dimensional cost calculation model: it is suitable for various cost calculation models of different sizes, different industries (or special industries with abnormal cost calculation set up separately: such as classified industries), different customer tendencies, different regions (different development costs in different regions: average salary of developers in different regions)-to evaluate the software development costs of the whole industry, region, type and quantity; 4. Distributed service architecture: to ensure that the modules such as requirement detection, cost analysis, code compliance analysis and code traceability are independent of each other and do not affect each other; 5. Requirements detection: according to the dictionary rules to detect the compliance of the requirements book content, generate detection reports and provide solutions; 6. Multi-thread check: the target detection set slice segmentation, distributed on the super-fusion server, multi-machine multi-threaded, additional integrated concurrent lock control, in order to ensure efficient check at the same time also ensure the consistency of the transaction; 7. Privatization deployment: Provide privatization deployment services for customers with high security and stability needs such as government units and enterprises to ensure that all data are not released or leaked, and hardware systems can be flexibly designed according to users' needs. The following is the project features: 1. Software function point extraction: Based on the weighted analysis of requirements by NLP traditional engineering, the NESMA function point estimation method issued by the Netherlands Software Metrics Association is followed to accurately extract the applicable function points for workflow in the target document. 2. Code duplicate checking: This function uses a combination of multi-group and greedy string matching to efficiently check the duplicate of the target detection code uploaded by the user. The results are displayed in charts, such as large-scale data distribution charts, code multi-dimensional classification pie charts, and specific analysis charts. 3. Hosting address tracing: according to the workflow pre-check results, trace the hosting address of the open source code used. 4. Vulnerability retrieval: retrieve relevant vulnerabilities according to the version number and branch dependency of the open source code part, and generate template report according to the data tabulation comparison issued by the authority. 5. Copyright search: search the copyright requirements such as commercial license, secondary development, open source extension, etc. declared in the open source code, and give the evaluation conclusion and template report. 6. Automatically build the duplicate check code base: the system classifies and indexes the duplicate check target base according to the policy set by the user, and crawls hot items regularly to ensure the timeliness and effectiveness of the duplicate check code base. 7. Workflow and report integration: The system is highly decoupled, each workpiece does not interfere with each other, and the internal logic is strong, through the workflow design series of each module. At the same time, the report module also relies on this to ensure that the interface is highly available. The following is a summary of the characteristics of the project: 1. The system uses a zero-trust authorization model, "continuous verification, never trust" is the core concept of zero-trust authorization model. Instead of defaulting to trust authorities or anyone, devices, and systems inside and outside the enterprise network, this pattern rebuilds the trust foundation of access control based on authentication and authorization to ensure the trustworthiness of identities, devices, applications, and links. The special requirements of government agencies for network security can highly guarantee terminal security, link security and access control security through the principle of zero trust, and effectively prevent major network security incidents. 2. The system business is deployed in the high-speed memory cache. In order to ensure the efficient operation of the intelligent code analysis system, the main business part adopts the super-converged server memory disk persistence method, thus realizing ultra-high-speed read and write operations, greatly improving the business performance and ensuring the good experience of users. 3. The service cluster adopts redundant backup. In addition to the main hyperconverged cluster, the system also deploys backup on two metal bare metals, and the registration center monitors and controls. When the business entity cluster is detected to be offline, the system will automatically switch to the standby server and complete the session transfer. At the same time, the main database synchronization is completed at high speed, realizing disaster recovery switch without awareness. In addition, the system also regularly backs up important data to prevent accidental data loss or system crash, and the system can be restored quickly.