Project development data management method, system and equipment based on artificial intelligence and medium

Abstract

The invention relates to a project development data management method and system based on artificial intelligence, equipment and a medium. The method comprises the steps of establishing a resource index database based on a preset container mirror image database in combination with received task description and project anchor points, generating a development scheme, executing data modification operation in an independent container based on the development scheme, and generating a detailed operation log; receiving a process instruction of a user in a project development process, and generating an instruction context in combination with a current project state; a response result is generated through the AI model module, automatic operation in the independent container is executed, and an instruction execution log is generated; when project development is completed, whether project resources such as detailed operation logs and instruction execution logs in the independent containers are complete or not is verified; and if so, packaging the project resources and submitting the project resources to a project warehouse. By adopting the method, the problems of environment isomerism, data dispersion, unsmooth AI interaction and the like in traditional development can be effectively solved, and the cooperation efficiency is remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of project management technology, and in particular relates to artificial intelligence-based project development data management methods, systems, equipment and media. Background Technology

[0002] With the deep integration of artificial intelligence (AI) technology into project development, AI-assisted intelligent development models are gradually becoming a key means of improving efficiency. These technologies integrate AI capabilities such as machine learning models and natural language processing to provide intelligent support for development stages such as code generation, defect detection, and resource optimization, significantly reducing the need for manual intervention. Currently, the common practice is to introduce independent AI tool plugins into the local development environment or rely on cloud-based AI service interfaces for specific task assistance, forming a basic collaborative model of "local environment + external AI services."

[0003] In traditional technologies, project data management primarily relies on version control systems (such as Git and SVN) for file-level change tracking and collaborative modification. Development team members must configure their local development environments, manually pull code, run AI tools, process the generated results, and then commit the modified content to a central repository. During this process, the AI ​​model is typically invoked as an external service, and its generated data, intermediate results, and contextual information are scattered across different members' local working directories, lacking a unified and structured management mechanism. The integration and synchronization of project data heavily depend on manual operations and third-party communication tools, resulting in a loose coupling between the AI ​​model and the development environment.

[0004] However, current AI-assisted development methods have several prominent problems: heterogeneous local environments lead to inconsistent operating conditions for AI tools, resulting in inconsistent data formats and quality, making it difficult to guarantee the overall consistency and reproducibility of project data; intermediate assets such as temporary data, solution suggestions, and test results generated by AI models during development lack effective organization and version management, making them prone to loss or conflict in multi-person collaboration, hindering the smooth progress of task relay; in addition, the separation between AI and the development process also makes data transfer dependent on manual intermediaries, which not only increases operational complexity but also fails to achieve end-to-end automated data management from task assignment and intelligent development to result submission. Summary of the Invention

[0005] Therefore, it is necessary to provide an AI-based project development data management method that can achieve intelligent management of project data throughout its entire lifecycle, ensuring data consistency, traceability, and security, in response to the aforementioned technical problems.

[0006] Firstly, this application provides an artificial intelligence-based method for managing project development data, including: Based on a pre-defined container image library, a resource index library is built by combining the received task description and project anchor points; each container image in the pre-defined container image library has an AI agent client pre-installed. Based on the resource index and task description, a development plan is generated through the AI ​​model module, and data modification operations are performed in an independent container based on the development plan, generating detailed operation logs. Receive process instructions from users during project development, and generate instruction context based on the process instructions and the current project status; Based on process instructions and instruction context, response results are generated through AI model modules, and automated operations within independent containers are executed according to the response results, generating instruction execution logs; It receives the user's submission instruction when the project is completed, and verifies the integrity of the project resources in the independent container based on the submission instruction, and obtains the verification result; the project resources include deliverables, configuration files, test reports, detailed operation logs and instruction execution logs; If the verification result shows that the project resources in the independent container are complete, a data storage instruction is generated. The data storage instruction is used to instruct the project resources to be packaged and submitted to the project repository.

[0007] Furthermore, based on a pre-defined container image library, and combined with the received task description and project anchor points, a resource index library is established; wherein each container image in the pre-defined container image library is pre-installed with an AI agent client, including: Based on the task description and project anchor submitted by the user through the console, the corresponding container image is obtained from the preset container image library; Create an independent container based on the container image, and activate the independent container based on the security authentication key and the project repository access token to obtain the activated container; Using the activated container, obtain the project's basic resources based on the project anchor point, and build a resource index library based on the project's basic resources.

[0008] Furthermore, through the activated container, based on the project anchor point, the project's basic resources are obtained, and based on these resources, a resource index library is built, including: Based on the project repository address and base branch version in the project anchor, and combined with the project repository access token, retrieve the project files from the project repository; A full scan of the project files is performed using the activated container to obtain a file scan dataset. The file scan dataset includes the file content and metadata of each file, including file size, modification time, and permission settings. The file content is categorized and parsed to obtain the parsing results. Based on the parsing results, a dependency graph is generated for each project file. The dependency graph is used to record the reference and inclusion relationships between files. Based on the file scanning dataset, parsing results, and dependency graph, a structured resource index library is constructed.

[0009] Furthermore, based on the resource index and task description, a development plan is generated through the AI ​​model module, and data modification operations are performed within an independent container based on the development plan, generating detailed operation logs, including: An encrypted interaction link between the AI ​​agent client and the AI ​​model module is established through a preset communication protocol and security authentication key; The task description and resource index are packaged together to generate a standardized data package; the standardized data package includes task keywords, a list of resource paths, and a data summary. Standardized data packets are transmitted to the AI ​​model module via an encrypted interactive link to generate a development plan; the development plan includes code modification sections, new file templates, or configuration adjustment instructions. The operation instructions in the development plan are parsed to obtain the operation instruction parsing results. Based on the operation instruction parsing results, data modification is performed in an independent container to obtain all data modification operations corresponding to the development plan. A detailed operation log is generated based on all data modification operations; the detailed operation log includes timestamps, operation types, and a list of affected files.

[0010] Furthermore, it receives process instructions from the user during project development and generates an instruction context based on the process instructions and the current project status, including: Receive process instruction texts from users during project development, and perform intent parsing on the process instruction texts to obtain instruction types; among them, instruction types include distinguishing between resource modification instructions, query instructions, and test instructions; Retrieve the current project status information from the resource index; the current project status information includes a list of recently modified files, dependency versions, and test results; By combining the instruction type and the current project status information, an instruction context is generated; the instruction context includes the current operation file path, related dependent modules, and historical modification records.

[0011] Furthermore, based on process instructions and instruction context, a response result is generated through an AI model module, and automated operations within an independent container are executed according to the response result, generating an instruction execution log, including: Input the process instructions and instruction context into the AI ​​model module to generate response results; the response results include specific modification suggestions, code snippets or test cases; The response results are verified to obtain a response result verification report; the verification includes running unit tests and static code analysis. Based on the response result verification report, automated operations are executed within an independent container to obtain automated operation results; among them, automated operations include file replacement and configuration update; Based on the results of automated operations, an instruction execution log is generated.

[0012] Secondly, this application also provides an artificial intelligence-based project development data management system, including: The index library building module is used to build a resource index library based on a preset container image library, combined with the received task description and project anchor point; each container image in the preset container image library has an AI agent client pre-installed. The modification log generation module is used to generate a development plan based on the resource index library and task description through the AI ​​model module, and to perform data modification operations in an independent container based on the development plan, generating detailed operation logs. The context generation module is used to receive process instructions from users during project development and generate instruction context based on the process instructions and the current project status. The execution log generation module is used to generate response results based on process instructions and instruction context through the AI ​​model module, and to perform automated operations within an independent container based on the response results, thereby generating instruction execution logs. The resource verification module is used to receive the user's submission instruction when the project development is completed, and based on the submission instruction, verify whether the project resources in the independent container are complete and obtain the verification result; among them, project resources include deliverables, configuration files, test reports, detailed operation logs and instruction execution logs; The resource storage module is used to generate a data storage instruction if the verification result shows that the project resources in the independent container are complete. The data storage instruction is used to instruct the project resources to be packaged and submitted to the project repository.

[0013] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores at least one instruction, at least one program, a code set, or an instruction set, and the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by the processor to implement any of the personalized learning path generation methods based on learning behavior profiles described in the embodiments of this application.

[0014] Fourthly, this application also provides a computer-readable storage medium storing at least one piece of program code, which is loaded and executed by a processor to implement the personalized learning path generation method based on learning behavior profiles as described in any of the embodiments of this application.

[0015] The aforementioned AI-based project development data management methods, systems, devices, and media establish a resource index library through a pre-defined container image library, task descriptions, and project anchor points. Based on the resource index library and task descriptions, an AI model module generates development plans, executes data modification operations within independent containers, and generates detailed operation logs. Based on user process commands and the current project status during project development, a command context is generated. Based on the process commands and command context, the AI ​​model module generates response results and executes automated operations within independent containers, generating command execution logs. Based on the user's submission command upon project completion, the integrity of project resources within independent containers is verified. Upon successful verification, project resources are packaged and submitted to the project repository. This effectively solves problems such as heterogeneous environments, scattered data, and poor AI interaction in traditional development, significantly improving collaboration efficiency, data security, and task relay capabilities, while reducing manual intervention and achieving end-to-end automated management. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies 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.

[0017] Figure 1 This is a flowchart illustrating a data management method for project development based on artificial intelligence, as shown in one embodiment. Figure 2 This is a flowchart illustrating the steps of establishing a resource index library based on a preset container image library, combined with the received task description and project anchor point, in one embodiment. Figure 3 This is a schematic diagram of the structure of an AI-based project development data management system in one embodiment. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] In one embodiment, an AI-based project development data management method is provided. This embodiment illustrates the method by applying it to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. Figure 1 As shown, in this embodiment, the method includes the following steps: Step S101: Based on the preset container image library, and combined with the received task description and project anchor point, establish a resource index library; wherein, each container image in the preset container image library has an AI agent client pre-installed.

[0020] The task description includes key information such as task objectives, development requirements, and technical parameters; the project anchor is the identification information pointing to the project's basic resources, including the project repository address, basic branch version, and images with pre-installed AI (Artificial Intelligence) model agent clients; the preset container image repository refers to a pre-configured dedicated repository for storing and managing container images, which provides out-of-the-box image acquisition sources for the deployment of containerized applications, avoiding the trouble of users manually building or searching for third-party image repositories.

[0021] For example, the system receives a task description and project anchor point submitted by the user on the control panel. Based on the task description, it automatically matches and pulls the corresponding container image from a pre-set container image library. This image comes pre-installed with the AI ​​agent client to ensure environment consistency. An independent container is created based on the image, and a security authentication key and a project repository access token are dynamically injected into the container to activate the AI ​​agent client's functionality. After activation, basic resources are loaded from the project repository according to the project anchor point, and a resource index library is built. This index library records metadata such as file paths, types, and dependencies in a structured format. The control panel refers to the dedicated interface for users to view, edit, track, collaborate on, and manage tasks after creating them on the task panel, similar to task cards in Lark. The pre-installed AI agent client has integrated several core capabilities during the image building phase: First, the communication protocol and basic security authentication configuration with the AI ​​model module (the dynamic key will be added after the container starts) to ensure interaction security and compatibility; second, the interface permission template with the project repository (including the repository access token injection interface and branch operation permission configuration items) and data transmission protocols (such as SVN and Git), supporting direct push of resources to the project repository; third, the adoption of a tool blacklist mechanism, by default only requiring users to confirm whether to execute version management tools; the project repository is a digital space used to centrally store, manage, and track project-related files such as code, documents, and design drafts.

[0022] Step S102: Based on the resource index library and task description, a development plan is generated through the AI ​​model module, and data modification operations are performed in an independent container based on the development plan to generate detailed operation logs.

[0023] The AI ​​model module is a functional and reusable component that makes up the AI ​​system. Different modules perform their respective functions and work together to complete the core process of "receiving input → processing data → outputting results". Its design follows the principle of "modular decomposition", which reduces development complexity and facilitates subsequent optimization or replacement. For example, an encrypted interaction link between the AI ​​agent client and the AI ​​model module is established based on a preset communication protocol and a dynamically injected security authentication key to ensure secure data transmission. The task description and resource index are packaged into a standardized data packet, such as JSON (JavaScript Object Notation) format containing task keywords and a list of resource paths, and transmitted to the AI ​​model module. The AI ​​model module analyzes the received data, using natural language processing techniques to parse the task intent and, combined with the resource index, assesses the project status to generate a development plan. This plan may include code modification sections, new file templates, or configuration adjustment instructions. The operation instructions in the plan are parsed, and data modification operations, such as file content updates, are automatically executed within an independent container. Detailed operation logs are generated, recording timestamps, operation types, and a list of affected files. Among them, the preset communication protocol refers to the set of communication rules that have been adapted, tested for compatibility and integrated with functions in advance before the hardware device, software system or component leaves the factory / deploys; Natural Language Processing (NLP) technology is one of the core branches of artificial intelligence, with the goal of enabling computers to understand and use human natural languages, such as Chinese, English and other everyday communication languages, and breaking down the language barrier between humans and machines.

[0024] Step S103: Receive process instructions from the user during the project development process, and generate an instruction context based on the process instructions and the current project status.

[0025] For example, the system receives user-input process command text, such as modification requests or inquiries, in real time through a control panel. Intent parsing is then performed on the commands, and classification models (such as keyword matching or natural language processing) can be used to distinguish command types, such as resource modification, query, or test commands. Simultaneously, current project status information, including a list of recently modified files, dependency versions, and test results, is retrieved from a resource index to ensure real-time status information. Combining the command type and current status, a command context is generated, containing the current operation file path, related dependency modules, and historical modification records. Intent parsing is one of the core tasks in natural language processing, aiming to extract the user's true behavioral purpose or demand direction from their natural language input. Essentially, it maps ambiguous language text to clear, machine-understandable intent categories. Classification models, in machine learning, are used to divide input data into predefined categories, aiming to learn the "feature → category" mapping rules to solve the "which category does it belong to" problem.

[0026] Step S104: Based on the process instructions and instruction context, generate response results through the AI ​​model module, and execute automated operations within an independent container according to the response results, generating instruction execution logs.

[0027] For example, process instructions and instruction context are input into the AI ​​model module. The AI ​​model module generates response results, such as specific modification suggestions, code snippets, or test cases, using sequence-to-sequence models or similar algorithms. The response results are verified, including running unit tests to check functional correctness or using static code analysis tools to verify syntax compliance. After successful verification, automated operations, such as file replacement or configuration updates, are executed within a separate container, and an instruction execution log is generated, recording operation details and verification results. Sequence-to-Sequence (Seq2Seq) is a deep learning architecture that maps an input sequence to another output sequence, solving the problem of "input and output lengths not being fixed and no direct positional correspondence." It is one of the fundamental models in fields such as natural language processing and speech processing. Unit testing is the independent verification of the smallest testable unit in software (usually a function, method, or class). Running unit tests involves executing these pre-written test codes, with the core purpose of verifying whether the logic of the target unit meets expectations. Static code analysis tools are tools that scan, parse, and inspect source code (or bytecode / intermediate code) without running the code. Their core goal is to discover problems in the code in advance, rather than verifying the logic through execution.

[0028] Step S105: Receive the user's submission instruction when the project development is completed, and based on the submission instruction, verify whether the project resources in the independent container are complete, and obtain the verification result; wherein, the project resources include deliverables, configuration files, test reports, detailed operation logs and instruction execution logs.

[0029] For example, upon receiving a submission command sent by a user through the control panel after project development is completed, a resource integrity verification is performed on all project resources within the container. These resources include deliverables (such as final code), configuration files, test reports, and detailed operation logs and command execution logs generated earlier. The verification result is then obtained. Resource integrity verification is a technical means to ensure that digital resources such as files, data, and software installation packages have not been unauthorized tampered with, damaged, or replaced during transmission, storage, or use. Its core is to verify whether resources maintain their original state through feature comparison.

[0030] Step S106: If the verification result shows that the project resources in the independent container are complete, a data storage instruction is generated. The data storage instruction is used to instruct the project resources to be packaged and submitted to the project repository.

[0031] For example, if the verification result indicates that the resources are complete, a data storage instruction is generated to package the project resources into a version management format, such as a Git archive or ZIP compression, generating a resource list including commit information, a file list, and version identifiers (such as semantic version numbers). Using a dynamically injected project repository access token, the packaged resources are directly pushed to the specified branch of the project repository via protocols such as Git or SVN. After pushing, the repository returns commit confirmation information (such as commit hash values). The commit result is recorded and fed back to the console, completing data storage and version management. Git archive is one of the core functions provided by Git. It is used to package files of a specified version in a repository into a compressed package / archive file, without including the .git directory related to Git version control. Essentially, it is a clean snapshot of files without version information. Git is a distributed version control system used to track the modification history of files (especially code), manage different versions, and support collaborative development by multiple people. ZIP is a widely used data compression and archiving format used to reduce file size through algorithms and package multiple files / folders into a single file for easy storage and transmission. Version control format is a rule system used to standardize the version numbering of software, documents, and other files. Its core purpose is to clearly identify the version iteration order, distinguish update types, and help teams / users quickly understand the differences between versions. Packaging refers to the process of integrating various scattered resource files into one or more packaged files of a unified format through specific tools or methods to optimize resource management and usage efficiency. In this embodiment, a resource index library is established using a pre-defined container image library, task description, and project anchor points. Based on the resource index library and task description, an AI model module generates a development plan, executes data modification operations within an independent container, and generates detailed operation logs. An instruction context is generated based on the user's process instructions during project development and the current project status. Based on the process instructions and instruction context, the AI ​​model module generates response results and executes automated operations within the independent container, generating instruction execution logs. Based on the user's submission instruction upon project completion, the integrity of project resources within the independent container is verified. Upon successful verification, the project resources are packaged and submitted to the project repository. This effectively solves problems such as heterogeneous environments, scattered data, and poor AI interaction in traditional development, significantly improving collaboration efficiency, data security, and task relay capabilities, while reducing manual intervention and achieving end-to-end automated management.

[0032] In one embodiment, such as Figure 2 As shown, a resource index library is established based on a pre-defined container image library, combined with the received task description and project anchor points; each container image in the pre-defined container image library comes pre-installed with an AI agent client, including: Step S201: Based on the task description and project anchor point submitted by the user through the operation console, obtain the corresponding container image from the preset container image library.

[0033] For example, the system receives task descriptions and project anchor point information submitted by users through the console. Based on the image identifier specified in the project anchor point, it matches and pulls the corresponding container image from a pre-set container image library. This image has an AI agent client pre-installed during the construction phase and integrates basic functions such as communication protocols and warehouse docking modules. Among these, the system is a functional module used to achieve data synchronization and process collaboration between the project warehouse and external development / management systems, eliminating information silos in project collaboration and improving R&D / project management efficiency; the logistics management system is a digital tool used to coordinate, monitor, and optimize the entire logistics process, improving logistics efficiency, reducing costs, and ensuring the transparency and controllability of goods flow through information technology; the warehouse internal operation system is the core management system supporting the daily operation of the warehouse, covering three dimensions: process specifications, technical tools, and personnel collaboration, such as the warehouse management system; the logistics and distribution system is a comprehensive operational system supporting the flow of goods from the supply side (such as warehouses and factories) to the demand side (such as consumers and stores), achieving "efficient, accurate, and low-cost" goods delivery through process optimization, resource integration, and technical support.

[0034] Step S202: Create an independent container based on the container image, and activate the independent container based on the security authentication key and the project repository access token to obtain the activated container.

[0035] For example, an independent container instance is created based on the acquired container image. The container uses virtualization technology to achieve resource isolation and has an independent operating system, storage, and network space. After the container is created, a security authentication key and a project repository access token are dynamically injected into the AI ​​agent client pre-installed within the container. The security authentication key is used to encrypt the interaction data of the AI ​​model module, and the project repository access token is used to grant the client direct access to the repository. The dynamic injection mechanism is implemented through environment variables or configuration files, ensuring that the key and token are dynamically loaded at runtime, avoiding the risks of static storage. For example, the container activation status signal is received, and the container ID and resource allocation information are recorded. At this point, the AI ​​agent client function is fully activated, forming an isolated environment that can perform secure interactions. Virtualization technology is a technology that uses software or hardware to simulate multiple independent virtual resource environments on a single physical resource (such as a server, computer, or storage device). Its core is to break the one-to-one usage limitation of physical hardware and achieve efficient resource sharing and flexible management. The operating system (OS) is the core system software that manages computer hardware and software resources and is also the bridge between the user and the computer hardware.

[0036] Step S203: Using the activated container, obtain the basic resources of the project based on the project anchor point, and build a resource index library based on the basic resources of the project.

[0037] For example, through the activated AI agent client within the container, based on the repository address and branch version in the project anchor point, combined with a dynamically injected project repository access token, the client downloads basic project resources from the project repository via version control protocols such as Git or SVN. These resources include source code, dependency libraries, and configuration files. After downloading, a full scan of the resource files is performed, reading the metadata (such as file size and modification time) and content data of each file. The file content is then categorized and parsed: key symbols are extracted from code files, key-value pairs are parsed from configuration files, and hash identifiers are calculated for binary files. After parsing, the client constructs a structured resource index based on file paths, types, dependencies, and parsing results. The index is stored in JSON format and records data lineage information, providing a standardized data foundation for subsequent AI model analysis. SVN (Subversion) is a classic centralized version control system used to help teams or individuals efficiently manage the modification history of files (such as code and documents) and solve problems such as file conflicts and version tracking when collaborating with multiple people. Full scan is a common term in the fields of data processing, information retrieval, or system testing. Its core is to perform a comprehensive and thorough inspection or reading of all objects / data within the target scope without omission, rather than just targeting a part of the content (i.e., the corresponding concept of incremental scan). Classification and analysis is a structured information processing and analysis method. It first divides the research objects into clear categories based on their commonalities and differences, and then conducts targeted analysis for each category to achieve the goal of simplifying complexity and focusing precisely.

[0038] In this embodiment, based on the task description and project anchor points submitted by the user through the console, the corresponding container image is obtained from a preset container image library. An independent container is created based on the container image, and a security authentication key and project repository access token are injected to activate the AI ​​agent client within the independent container. Through the activated AI agent client, basic project resources are obtained according to the project anchor points, and a resource index library is established. This effectively solves the problems of cumbersome environment configuration and difficult data management in traditional development, significantly improving the efficiency and reliability of the data preparation phase.

[0039] In one embodiment, the project's basic resources are retrieved from the activated container based on the project anchor point, and a resource index library is built based on these basic resources, including: Step S301: Based on the project repository address and base branch version in the project anchor point, and combined with the project repository access token, retrieve the project files from the project repository.

[0040] For example, the AI ​​agent client within the activated container initiates an authentication request to the remote project repository based on the project repository address and base branch version information in the user-submitted project anchor, combined with a dynamically injected project repository access token. After passing the project anchor data to the container, the AI ​​agent client can establish a connection with the project repository via a secure protocol such as HTTPS or SSH based on the token, and download the complete set of project files, including source code, dependencies, configuration files, etc. During the download process, the transmission status is monitored, and exceptions such as network interruptions or authentication failures are handled to ensure that the files are stored intact in the container. Among them, HTTPS (Hypertext Transfer Protocol Secure) is an encrypted and secure version of the HTTP protocol, used to protect data in network communication from being stolen, tampered with, or forged, providing security for interactions between users and websites; SSH (Secure Shell) is an encrypted network protocol used for remote login and management of computers / servers, used to provide security for data transmission in insecure network environments (such as the Internet), preventing information from being eavesdropped on or tampered with; HTTP (Hypertext Transfer Protocol) is the core application layer protocol used on the Internet for transmitting data between clients and servers; version control commands are sets of instructions used to manage code / file versions (core tools such as Git and SVN).

[0041] Step S302: Perform a full scan of the project files using the activated container to obtain a file scan dataset; wherein, the file scan dataset includes the file content and metadata of each file, and the metadata includes file size, modification time and permission settings.

[0042] For example, after the project files are downloaded, all project files are traversed, and the content data (such as text code or binary streams) and metadata of each file are read. All data are then aggregated to obtain a file scan dataset, which is temporarily stored in the container's memory in a structured format (such as a dictionary or list). The metadata includes file size, last modified timestamp, and permission settings, which can be obtained in real time by calling the operating system API. The operating system API (Application Programming Interface) is a function call interface provided by the operating system for applications (such as browsers), and it serves as a bridge between the application and the operating system kernel.

[0043] Step S303: Classify and parse the file content to obtain the parsing results, and generate a dependency graph for each project file based on the parsing results; wherein, the dependency graph is used to record the reference and inclusion relationships between files.

[0044] For example, an AI agent client can invoke pre-integrated analysis tools to classify files into types such as code, configuration, and documentation based on file extensions and content characteristics (such as file header magic numbers or syntax patterns) using classification algorithms. For code files, a syntax analyzer can extract key symbols such as function definitions and class structures. For configuration files, regular expressions can be used to match key-value pairs. For binary files, a SHA-256 hash value can be calculated as a unique identifier. After parsing, a dependency graph is constructed based on the reference relationships between files, where nodes represent files, edges represent dependency directions, and reference frequency and type are recorded. Pre-integrated analysis tools refer to analysis software / modules that have completed technical adaptation, functional combination, and data link integration in advance. Users can directly use them without needing to develop complex code, debug interfaces, or connect from scratch, focusing on data interpretation rather than technical construction. Classification algorithms are one of the core technologies of supervised learning in machine learning. Their core objective is to learn a classification rule (mathematical model) based on training data with known categories, and then use this rule to determine the category of new data with unknown categories. Parsing is a core component in compiler principles or natural language processing. Based on preset grammar rules, it performs structural analysis on the input lexical unit sequence (such as words / sub-words in NLP during compilation), determines whether it conforms to grammatical rules, and generates a structured intermediate representation (such as a syntax tree). Regular expressions are a set of rule-based syntax for text matching and processing. Their core is to use concise symbol combinations to describe text patterns, enabling operations such as searching, replacing, validating, and extracting strings. SHA-256 is one of the core algorithms in SHA-2 (Secure Hash Algorithm 2 family), short for "Secure Hash Algorithm". "256-bit" is essentially a one-way cryptographic hash function that can convert input data of any length into a fixed 256-bit (32-byte) output.

[0045] Step S304: Based on the file scanning dataset, parsing results, and dependency graph, construct a structured resource index library.

[0046] For example, file paths, types, metadata, parsed symbol lists, and dependency graph data are encoded into JSON format. Each file generates an index entry, and the entries are connected by dependency edges to form a mesh structure. Database technology can be used to store this mesh structure, resulting in a structured resource index library to support fast querying and updates. Simultaneously, data lineage information, such as the file source repository and scan timestamp, is injected to ensure traceability. Database technology is a technical system used for the efficient storage, management, organization, and retrieval of data, addressing the challenges of how to store massive amounts of data in an orderly manner, access it quickly, and maintain it securely.

[0047] In this embodiment, based on the project repository address and base branch version in the project anchor point, project files are retrieved from the project repository and a full scan of the project files is performed to obtain a file scan dataset. The file content is then categorized and parsed to obtain the parsing results, and dependency analysis is performed to generate a dependency graph for each project file. Based on the file scan dataset, parsing results, and dependency graph, a structured resource index library is constructed. This effectively solves the problem of AI analysis difficulties caused by scattered and inconsistent data formats in traditional development, enhancing data consistency and traceability.

[0048] In one embodiment, a development plan is generated using an AI model module based on a resource index and task description. Data modification operations are then performed within an independent container based on the development plan, generating detailed operation logs, including: Step S401: Establish an encrypted interaction link between the AI ​​agent client and the AI ​​model module through a preset communication protocol and security authentication key.

[0049] For example, the validity of the security authentication key is verified, that is, the legality (not forged, not tampered with) and timeliness (within the validity period, not expired) of the key are confirmed. After the key is confirmed to be valid, the handshake process between the AI ​​agent client and the AI ​​model module is initiated according to the preset communication protocol (such as HTTPS based on TLS). The two parties exchange digital certificates and session keys, negotiate encryption algorithms (such as AES-256) and authentication mechanisms, thereby establishing a two-way, low-latency dedicated encrypted channel, that is, an encrypted interaction link. The handshake process refers to the initialization process in which both parties establish a communication connection, confirm capability matching, and agree on interaction rules. HTTPS based on TLS is a secure transmission protocol formed by adding TLS (Transport Layer Security) to HTTP. It achieves "authentication, data encryption, and integrity verification" through TLS, solving the security vulnerabilities of HTTP plaintext transmission, such as data being eavesdropped on, tampered with, or impersonated. The authentication mechanism is the core rule and process used to confirm the "authenticity of user / entity identity." Essentially, it solves the trust problem of "who you are," ensuring that the entity initiating the operation is indeed the identity it claims, and not a forger or illegal impersonator. Digital certificates are similar to online ID cards, issued by authoritative institutions, and contain the public key and identity information of the communicating parties. A session key is a temporarily generated encryption key that is only valid during a specific communication session, used to ensure the security of a single data interaction.

[0050] Step S402: Package the task description and resource index library to generate a standardized data package; wherein, the standardized data package includes task keywords, resource path list and data summary.

[0051] For example, key elements such as task keywords (identified through natural language processing), resource path lists (file location information in the index), and data digests (such as the SHA-256 hash values ​​of the file set) are extracted from the received task description and resource index. Serialization tools (such as JSON) can be used to package the data into standardized data packets. The packet structure includes metadata segments (such as version numbers and timestamps) and payload segments (keyword lists, path mapping tables, and digest data), and compression optimization is performed to reduce transmission load. Serialization tools are tools / libraries used to implement serialization and deserialization operations, solving data format adaptation problems in storage and transmission scenarios; essentially, they bridge the gap between in-memory data structures and persistent / transferable formats. Compression optimization is a process of reducing the storage space or transmission volume of data such as files, images, videos, and text through technical means, while preserving as much core data information as possible.

[0052] Step S403: Transmit standardized data packets to the AI ​​model module via an encrypted interactive link to generate a development plan; wherein, the development plan includes code modification sections, new file templates or configuration adjustment instructions.

[0053] For example, data packets are segmented into encrypted data blocks, encrypted using a session key, and sent to the AI ​​model module via streaming. Upon receiving the data, the AI ​​model module decrypts it, verifies its integrity, and parses the data packets: task keywords are used to understand the development intent, resource path lists are used to locate the project context, and data summaries are used to assess differences in project status. Finally, a development solution is generated, such as code patches, configuration file templates, or test case scripts. Encryption is a technology that uses specific rules (algorithms) to convert raw information (plaintext) into a garbled form (ciphertext) that is difficult to understand directly. Its core purpose is to protect information security and prevent unauthorized theft or tampering. Decryption breaks the encryption of information, restoring information that has been processed by specific rules (such as passwords, codes, hidden logic, etc.) and is only understandable to specific individuals into its original form or clear content that can be directly interpreted by the general public or target audience.

[0054] Step S404: Parse the operation instructions in the development plan to obtain the operation instruction parsing results, and based on the operation instruction parsing results, perform data modification in an independent container to obtain all data modification operations corresponding to the development plan.

[0055] For example, parsing operation instructions in the development scheme can employ a syntax analyzer (such as regular expressions) to identify instruction types (e.g., file modification, creation, or deletion) and extract parameters (e.g., target file path, modified content, or dependencies). Within a separate container, relevant tools are invoked to perform data modifications: for instance, patching tools are called when applying code differences to source files, a template engine is used when generating new files, and key-value pairs are updated when adjusting configurations. During execution, the feasibility of operations is verified (e.g., checking file permissions or dependency compatibility), and erroneous operations are rolled back in real time to ensure atomicity of modifications, and all modification results are collected. Among them, patching tools are tools specifically designed for handling difference merging (such as Git's built-in patching function). They are used to read code difference files, locate the parts of the source file that need to be modified, automatically perform add / delete / modify operations, "apply" the differences to the source file, and generate the updated file. Template engines are tools or technologies that separate template files and data for processing, and finally generate target text (such as configuration files). They are used to solve the coupling problem between data and presentation format, improve development efficiency and code maintainability. Rollback refers to restoring data from the current state to a previous stable and normal historical version to solve errors, faults or unexpected problems in the current version.

[0056] Step S405: Generate a detailed operation log based on all data modification operations; the detailed operation log includes timestamps, operation types, and a list of affected files.

[0057] For example, based on all data modification operations, metadata for each operation is collected: timestamp, operation type (e.g., "file update"), and a list of affected files (file path, hash value before / after modification). The collected data is formatted into a standard structure, such as JSON log entries, containing an operation ID, context label, and result status, and persistently stored in a log file within a container to obtain a detailed operation log. Formatting refers to organizing messy, inconsistent, or non-standard raw data into a structured form with fixed rules and general applicability.

[0058] In this embodiment, an encrypted interaction link is established between the AI ​​agent client and the AI ​​model module through a preset communication protocol and security authentication key. The task description and resource index are standardized and packaged, and the standardized data packet is transmitted to the AI ​​model module through the encrypted interaction link to generate a development plan. The operation instructions in the development plan are parsed, and based on the parsing results, data modification operations are executed within an independent container, generating detailed operation logs. This effectively solves the problems of AI interaction delays and opaque plan execution in traditional development. By combining automation and intelligence, it significantly improves development efficiency, data consistency, and the traceability of team collaboration.

[0059] In one embodiment, process instructions from the user during project development are received, and an instruction context is generated based on the process instructions and the current project state, including: Step S501: Receive the process instruction text from the user during the project development process, and perform intent parsing on the process instruction text to obtain the instruction type; wherein, the instruction type includes distinguishing between resource modification instructions, query instructions, and test instructions.

[0060] For example, the system captures user-input process command text in real time, such as natural language expressions like "query the current dependency version status," and transmits this raw text data to an AI agent client within an independent container. The command text is parsed using a pre-integrated natural language processing engine: key verbs, noun phrases, and other technical terms are extracted; and the commands are mapped to a pre-defined set of command types using a classification algorithm, including resource modification commands (involving file addition, deletion, and modification operations), query commands (requesting status information), and test commands (triggering verification processes). The identified command type labels are temporarily stored as structured data. The pre-integrated natural language processing engine, based on conventional NLP capabilities, pre-integrates technical modules and prepares for scenario adaptation, lowering the barrier to entry for users and enabling non-professionals or enterprises to quickly implement NLP applications; the pre-defined set of command types refers to the sum of pre-defined command frameworks with clear functional classifications.

[0061] Step S502: Obtain the current project status information from the resource index library; wherein, the current project status information includes a list of recently modified files, dependency versions, and test results.

[0062] Among them, the recently modified file list refers to the paths of the latest changed files arranged in reverse chronological order by timestamp; the dependency version refers to the mapping table of library names and version numbers parsed from the dependency graph; and the test results refer to quality indicators such as unit test pass rate and coverage report.

[0063] For example, the system extracts current project status information such as a list of recently modified files, dependency versions, and test results from the resource index. After receiving the status information, it displays it in a visual dashboard and encapsulates the data into a status object. The visual dashboard format is an interface that integrates scattered data, key metrics, or business information through visual elements such as charts, graphs, and indicator cards. Encapsulation refers to consolidating scattered states, state change logic, and state verification / listening capabilities into a single object, rather than distributing them throughout the code.

[0064] Step S503: Combine the instruction type and the current project status information to generate an instruction context; wherein, the instruction context includes the current operation file path, related dependent modules, and historical modification records.

[0065] Among them, the relevant dependent modules refer to the affected components traced through the dependency graph; the historical modification record is a summary of recent changes extracted from the operation log.

[0066] For example, the context framework is determined based on the instruction type: for resource modification instructions, the focus is on the current file path (e.g., the location of the file being edited in the index) and related dependent modules; for query instructions, the emphasis is on historical modification records; and for test instructions, the test coverage and environment configuration are considered. Context data is organized into structured objects, containing metadata such as timestamps, session IDs, and associated file references, and serialized using JSON format. Finally, the context object and the original instruction are bound together to form a complete instruction request package. Here, "organization" refers to an ordered whole constructed through rules, relationships, and goals.

[0067] In this embodiment, the system receives process command text from the user during project development, performs intent parsing on the command text to obtain the command type, retrieves the current project status information from the resource index, and generates a command context by combining the command type and the current project status information. This effectively solves the problem of misoperation caused by ambiguous commands and disconnected states in traditional development, and significantly improves human-computer interaction efficiency through automated context construction.

[0068] In one embodiment, based on process instructions and instruction context, a response result is generated through an AI model module, and automated operations within an independent container are executed according to the response result, generating an instruction execution log, including: Step S601: Input the process instructions and instruction context into the AI ​​model module to generate response results; wherein, the response results include specific modification suggestions, code snippets or test cases.

[0069] For example, the instruction context (including the current operation file path, dependent module information, etc.) and process instruction text are serialized and converted into a standardized input format (such as a JSON object) that the AI ​​model can recognize. The data packet is transmitted to the AI ​​model module through the established encrypted interaction link. Based on the sequence-to-sequence model, the input data is analyzed in multiple dimensions. Combined with the pre-trained project development knowledge base, a response result containing specific modification suggestions (such as code optimization snippets), executable code blocks, or targeted test cases is generated. Serialization is a core technology in the computer field used for data storage and transmission. Essentially, it is the process of converting structured data such as objects, arrays, and dictionaries in memory that are easy for programs to manipulate into a linear format (such as JSON) that can be persistently stored (e.g., stored as files) or transmitted over a network (e.g., cross-service communication). The reverse operation (restoring the linear format back to structured data) is called deserialization. Multidimensional analysis is a way of thinking and research that goes beyond a single perspective and starts from multiple related or independent dimensions to comprehensively deconstruct, explore, and evaluate things or problems. The pre-trained project development knowledge base is an intelligent knowledge system built by combining pre-training technology and knowledge in the project development domain. It first pre-learns the general rules and professional knowledge of project development through large-scale data, and then supports subsequent scenario-based applications.

[0070] Step S602: Verify the response result and obtain a response result verification report; wherein, the verification includes running unit tests and static code analysis.

[0071] Static code analysis is a technique / process that scans, inspects, and analyzes source code (or bytecode, intermediate code) without running the code. Its core objective is to discover problems in the code in advance, rather than to verify the code's execution results.

[0072] For example, the executable content in the response result is parsed: For code snippets or modification suggestions, the pre-built static code analysis tools within the container can be invoked for syntax checking, potential defect detection, and specification compliance verification; for modifications involving functional changes, the associated unit test case set is automatically run, tests are executed, and data such as test coverage and pass rate are collected. During the verification process, the execution status is monitored in real time, and exceptions or error information is captured; finally, a structured response result verification report is generated, including verification timestamps, various check results, test metrics, and problem summaries. Among them, syntax checking is the process of verifying the compliance of the grammatical rules of language expression, with the core purpose of correcting grammatical errors and improving the accuracy and standardization of expression; specification compliance verification is a systematic check and confirmation process to ensure that products, processes, or services comply with preset standards, regulations, industry specifications, or internal requirements; the unit test case set is a collection of all test cases designed for the smallest testable unit of software (usually functions, methods, classes, etc.), with the core purpose of verifying whether the logic of a single unit meets the design expectations and can correctly handle various input scenarios, and is a key carrier for left-shift testing (finding problems as early as possible) in software testing.

[0073] Step S603: Based on the response result verification report, execute the automated operations in an independent container to obtain the automated operation results; wherein, the automated operations include file replacement and configuration update.

[0074] For example, based on the results indicated in the verification report, such as passing the test or finding no serious errors in static analysis, an execution command is sent to the AI ​​agent client. The client automatically calls the file operation interface within the container: for file replacement operations, version control tools (such as Git) can be used to apply differential patches or directly overwrite the target file; for configuration updates, the corresponding key-value pairs are modified through the configuration file parser, ensuring format compatibility. During execution, an atomic operation guarantee mechanism is implemented to ensure that every modification is traceable and rollbackable; at the same time, the project state changes within the container are synchronized in real time, and relevant records in the resource index are updated. The execution progress is monitored, and automated operation results, such as a list of modified files and a summary of configuration changes, are collected after the operation is completed. The file operation interface is a set of standardized functions / methods provided by a programming language / operating system for programs to interact with files in storage media (disks, memory files, etc.); the atomic operation guarantee mechanism ensures that an operation or a set of operations is uninterrupted and indivisible during execution, either completing completely or not executing at all, without any intermediate state of execution, thus avoiding problems such as data races and consistency violations in concurrent scenarios.

[0075] Step S604: Generate an instruction execution log based on the results of the automated operation.

[0076] For example, based on the results of automated operations, key execution metadata is extracted, including the operation type (file replacement or configuration update), affected file paths, hash values ​​of the content before and after modification, and execution timeline. This data is then integrated in JSON format to generate instruction execution log entries. The log content covers the operation context, execution result status, possible error codes, and suggested solutions.

[0077] In this embodiment, process instructions and instruction context are input into the AI ​​model module to generate response results. These results are then verified to obtain a response result verification report. Based on this report, automated operations are executed within an independent container, generating instruction execution logs. This effectively solves the problems of low iteration efficiency caused by the difficulty in directly applying AI output and the lack of verification in traditional development. By combining automation and verification, the accuracy and execution efficiency of development responses are significantly improved, providing solid data support for team collaboration and project management.

[0078] To further illustrate the solutions of the embodiments of this application, a specific example is provided below.

[0079] Step S1: Create a task on the task panel, and then on the task or issue page (similar to Lark task cards, hereinafter referred to as Workbench A), the user submits the task description and project anchor points required for the current project development; wherein, the task description includes key information such as task objectives, development requirements, and technical parameters; the project anchor points are identification information pointing to the basic resources of the project, including the project repository address, basic branch version, and the image of the pre-installed AI model agent client, etc., used to locate the basic content of the project to be loaded, and at the same time provide a location basis for the AI ​​agent client pre-installed in the container to read resources and directly connect to the project repository.

[0080] Step S2: The remote server pre-stores multiple container images adapted to different development scenarios. All container images come pre-installed with the AI ​​agent client—this client integrates several core capabilities during the image building phase: First, the communication protocol and basic security authentication configuration for the AI ​​model module (the dynamic key will be added after the container starts) to ensure interaction security and compatibility; second, the interface permission template for the project repository (including the repository access token injection interface and branch operation permission configuration items) and data transmission protocols (such as SVN, Git, HTTPS) to support directly pushing resources to the project repository; third, a tool blacklist mechanism is adopted, by default only version management tools need to prompt the user to confirm whether to execute. When the user clicks the trigger button on console A, console A sends an environment creation request to the remote server. The remote server automatically matches and pulls the corresponding pre-installed AI agent client container image according to the task description submitted in step S1, and creates an independent container based on the image. The container is an isolated running environment based on virtualization technology, with independent operating system, memory, storage and network resources, which can avoid resource conflicts and data interference with other projects or server systems. After the container starts, the remote server dynamically injects a security authentication key (used to interact with the AI ​​model module) and a project repository access token (used to interface with the repository) into the AI ​​agent client to complete the client function activation.

[0081] Step S3: The remote server controls the created container and, based on the project anchor point submitted in Step S1 and the repository access token dynamically injected in Step S2, loads the project's basic resources from the corresponding project repository storage location. This includes loading the project version file, dependency libraries, configuration files, and other content pointed to by the project anchor point into the container. Simultaneously, the pre-installed AI agent client (already activated) automatically starts the resource listening function, reading the loaded project's basic resources in real time and establishing a resource index library to prepare for rapid transmission of project information during subsequent interaction with the AI ​​model module. Ultimately, this creates a basic project environment within the container that allows for direct development operations, possesses AI interaction capabilities, and supports direct repository connections.

[0082] Step S4: The remote server sends an association command to the AI ​​agent client. After receiving the command, the AI ​​agent client establishes a dedicated interaction link with the AI ​​model module through a pre-integrated communication protocol and a dynamically injected security authentication key, completing identity authentication and connection initialization. Subsequently, the AI ​​agent client packages the project resource index library established in Step S3 and the task description in Step S1 into a standardized data format (such as a JSON package containing resource paths, task keywords, and data summaries) and transmits it to the AI ​​model module. Based on the received information, the AI ​​model module analyzes the current status of the project, generates a preliminary development plan, and feeds the plan back to the AI ​​agent client. According to the plan, the AI ​​agent client automatically makes preliminary modifications to the project within the container, including basic development operations such as resource generation, file modification, and module building, while recording modification logs and synchronizing them to the operation console A.

[0083] Step S5: During project development, any content entered by the user into the module on console A is recognized as an instruction to the AI ​​model module. Console A transmits the input content to the remote server in real time, and the remote server forwards it to the pre-installed AI agent client within the container. The AI ​​agent client parses the instructions (e.g., distinguishing between resource modification instructions, problem inquiry instructions, and test requirement instructions), combines the current project status within the container (e.g., modified files, modules to be improved), supplements the generated instruction context information (e.g., current operation file path, dependent module version), and packages the "instruction + context" before transmitting it to the AI ​​model module. After receiving the complete information, the AI ​​model module generates a response result (including resource file modification schemes, suggestions, problem feedback, etc.) and feeds it back to the AI ​​agent client. The AI ​​agent client verifies the response result. After successful verification, it converts the result into a visual format and feeds it back to console A in real time via the remote server. Simultaneously, it automatically executes automated operations within the container (e.g., file replacement, configuration update) based on the response result, realizing a two-way real-time interactive closed loop between the user, the AI ​​agent client, and the AI ​​model module.

[0084] Step S6: After the user confirms the completion of the project development through console A, a project commit request is sent to the remote server. The remote server forwards the request to the pre-installed AI agent client within the container. The AI ​​agent client first performs a completeness check on the completed project resources (including deliverables, configuration files, test reports, modification logs, etc.) within the container (e.g., checking for missing deliverables, core resource modules, and complete test reports). After the check passes, the resources are packaged according to the format supported by the project repository (e.g., Git, SVN, and other version control software commit specifications), generating a resource list containing "commit information, file list, and version identifier." Subsequently, the AI ​​agent client directly calls the pre-integrated project repository integration capability and pushes the packaged resources and resource list to the preset project repository branch via protocols such as HTTPS / Git. After the push is completed, the AI ​​agent client receives the commit confirmation information returned by the project repository (e.g., commit hash value, branch update time, etc.) and organizes the complete commit results to console A for the user to view, while simultaneously completing the storage and version management of the project development results.

[0085] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0086] Based on the same inventive concept, this application also provides an AI-based project development data management system for implementing the aforementioned AI-based project development data management method. The solution provided by this system is similar to the implementation described in the above method; therefore, the specific limitations of one or more AI-based project development data management system embodiments provided below can be found in the limitations of the AI-based project development data management method described above, and will not be repeated here.

[0087] In one exemplary embodiment, such as Figure 3 As shown, an AI-based project development data management system 300 is provided, including: The index library construction module 301 is used to build a resource index library based on a preset container image library, combined with the received task description and project anchor point; wherein, each container image in the preset container image library has an AI agent client pre-installed. Modify the log generation module 302, which is used to generate a development plan through the AI ​​model module based on the resource index library and task description, and perform data modification operations in an independent container based on the development plan to generate detailed operation logs; The context generation module 303 is used to receive process instructions from the user during the project development process, and generate an instruction context based on the process instructions and the current project status; The execution log generation module 304 is used to generate response results based on process instructions and instruction context through the AI ​​model module, and to perform automated operations within an independent container based on the response results to generate instruction execution logs; The resource verification module 305 is used to receive the user's submission instruction when the project development is completed, and based on the submission instruction, verify whether the project resources in the independent container are complete and obtain the verification result; wherein, the project resources include deliverables, configuration files, test reports, detailed operation logs and instruction execution logs; The resource storage module 306 is used to generate a data storage instruction if the verification result shows that the project resources in the independent container are complete. The data storage instruction is used to instruct the project resources to be packaged and submitted to the project repository.

[0088] In one embodiment, the index building module 301 is further configured to: Based on the task description and project anchor submitted by the user through the console, the corresponding container image is obtained from the preset container image library; Create an independent container based on the container image, and activate the independent container based on the security authentication key and the project repository access token to obtain the activated container; Using the activated container, obtain the project's basic resources based on the project anchor point, and build a resource index library based on the project's basic resources.

[0089] In one embodiment, the index building module 301 is further configured to: Based on the project repository address and base branch version in the project anchor, and combined with the project repository access token, retrieve the project files from the project repository; A full scan of the project files is performed using the activated container to obtain a file scan dataset. The file scan dataset includes the file content and metadata of each file, including file size, modification time, and permission settings. The file content is categorized and parsed to obtain the parsing results. Based on the parsing results, a dependency graph is generated for each project file. The dependency graph is used to record the reference and inclusion relationships between files. Based on the file scanning dataset, parsing results, and dependency graph, a structured resource index library is constructed.

[0090] In one embodiment, the modified log generation module 302 is further configured to: An encrypted interaction link between the AI ​​agent client and the AI ​​model module is established through a preset communication protocol and security authentication key; The task description and resource index are packaged together to generate a standardized data package; the standardized data package includes task keywords, a list of resource paths, and a data summary. Standardized data packets are transmitted to the AI ​​model module via an encrypted interactive link to generate a development plan; the development plan includes code modification sections, new file templates, or configuration adjustment instructions. The operation instructions in the development plan are parsed to obtain the operation instruction parsing results. Based on the operation instruction parsing results, data modification is performed in an independent container to obtain all data modification operations corresponding to the development plan. A detailed operation log is generated based on all data modification operations; the detailed operation log includes timestamps, operation types, and a list of affected files.

[0091] In one embodiment, the context generation module 303 is further configured to: Receive process instruction texts from users during project development, and perform intent parsing on the process instruction texts to obtain instruction types; among them, instruction types include distinguishing between resource modification instructions, query instructions, and test instructions; Retrieve the current project status information from the resource index; the current project status information includes a list of recently modified files, dependency versions, and test results; By combining the instruction type and the current project status information, an instruction context is generated; the instruction context includes the current operation file path, related dependent modules, and historical modification records.

[0092] In one embodiment, the execution log generation module 304 is further configured to: Input the process instructions and instruction context into the AI ​​model module to generate response results; the response results include specific modification suggestions, code snippets or test cases; The response results are verified to obtain a response result verification report; the verification includes running unit tests and static code analysis. Based on the response result verification report, automated operations are executed within an independent container to obtain automated operation results; among them, automated operations include file replacement and configuration update; Based on the results of automated operations, an instruction execution log is generated.

[0093] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the artificial intelligence-based project development data management method described above.

[0094] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0095] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and 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 modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0096] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A project development data management method based on artificial intelligence, characterized in that, The method includes: Based on a pre-defined container image library, and combined with the received task description and project anchor points, a resource index library is established; wherein, each container image in the pre-defined container image library has an AI agent client pre-installed. Based on the resource index and the task description, a development plan is generated through the AI ​​model module, and data modification operations are performed in an independent container based on the development plan, generating detailed operation logs. Receive process instructions from the user during project development, and generate an instruction context based on the process instructions and the current project status; Based on the process instructions and the instruction context, the AI ​​model module generates a response result, and executes automated operations within the independent container according to the response result, generating an instruction execution log; The system receives a submission instruction from the user upon completion of project development, and verifies the integrity of project resources within the independent container based on the submission instruction, obtaining a verification result. The project resources include deliverables, configuration files, test reports, detailed operation logs, and instruction execution logs. If the verification result indicates that the project resources within the independent container are complete, a data storage instruction is generated. The data storage instruction is used to instruct the project resources to be packaged and submitted to the project repository.

2. The method according to claim 1, characterized in that, The system establishes a resource index library based on a pre-defined container image library, combined with the received task description and project anchor points; wherein each container image in the pre-defined container image library is pre-installed with an AI agent client, including: Based on the task description and project anchor submitted by the user through the console, the corresponding container image is obtained from the preset container image library; An independent container is created based on the container image, and the independent container is activated based on the security authentication key and the project repository access token to obtain the activated container. Using the activated container, the project's basic resources are obtained based on the project anchor point, and the resource index library is established based on the project's basic resources.

3. The method according to claim 2, characterized in that, The step of obtaining basic project resources through the activated container based on the project anchor point, and establishing the resource index library based on the basic project resources, includes: Based on the project repository address and base branch version in the project anchor point, and combined with the project repository access token, retrieve the project files from the project repository; A full scan of the project files is performed using the activated container to obtain a file scan dataset; wherein, the file scan dataset includes the file content and metadata of each file, and the metadata includes file size, modification time and permission settings; The file content is categorized and parsed to obtain parsing results, and a dependency graph for each project file is generated based on the parsing results; wherein, the dependency graph is used to record the reference and inclusion relationships between files; Based on the file scanning dataset, the parsing results, and the dependency graph, a structured resource index library is constructed.

4. The method according to claim 1, characterized in that, The process involves generating a development plan using the AI ​​model module based on the resource index and the task description, executing data modification operations within the independent container based on the development plan, and generating detailed operation logs, including: An encrypted interaction link between the AI ​​agent client and the AI ​​model module is established using a preset communication protocol and security authentication key. The task description and the resource index are packaged together to generate a standardized data package; wherein the standardized data package includes task keywords, a resource path list and a data summary; The standardized data packet is transmitted to the AI ​​model module through the encrypted interaction link to generate a development plan; wherein, the development plan includes code modification sections, new file templates, or configuration adjustment instructions; The operation instructions in the development scheme are parsed to obtain the operation instruction parsing results. Based on the operation instruction parsing results, data modification is performed in the independent container to obtain all data modification operations corresponding to the development scheme. Based on all the data modification operations, a detailed operation log is generated; wherein the detailed operation log includes timestamps, operation types, and a list of affected files.

5. The method according to claim 1, characterized in that, The step of receiving process instructions from the user during project development and generating an instruction context based on the process instructions and the current project status includes: The system receives process instruction text from the user during project development and performs intent parsing on the process instruction text to obtain instruction types; wherein, the instruction types include distinguishing between resource modification instructions, query instructions, and test instructions; Obtain the current project status information from the resource index; wherein, the current project status information includes a list of recently modified files, dependency versions, and test results; The instruction context is generated by combining the instruction type and the current project status information; wherein, the instruction context includes the current operation file path, related dependent modules, and historical modification records.

6. The method according to claim 1, characterized in that, The process involves generating a response result through the AI ​​model module based on the process instructions and the instruction context, and then executing automated operations within the independent container according to the response result, generating an instruction execution log, including: The process instructions and the instruction context are input into the AI ​​model module to generate a response result; wherein, the response result includes specific modification suggestions, code snippets or test cases; The response result is verified to obtain a response result verification report; wherein, the verification includes running unit tests and static code analysis; Based on the response result verification report, automated operations are performed within the independent container to obtain automated operation results; wherein, the automated operations include file replacement and configuration update; Based on the results of the automated operation, the instruction execution log is generated.

7. A project development data management system based on artificial intelligence, characterized in that: The system includes: The index library construction module is used to build a resource index library based on a preset container image library and combined with the received task description and project anchor point; wherein, each container image in the preset container image library has an AI agent client pre-installed. The modification log generation module is used to generate a development plan through the AI ​​model module based on the resource index library and the task description, and to perform data modification operations in an independent container based on the development plan to generate detailed operation logs; The context generation module is used to receive process instructions from the user during the project development process, and generate an instruction context based on the process instructions and the current project status; An execution log generation module is used to generate a response result through the AI ​​model module based on the process instruction and the instruction context, and to execute automated operations within the independent container according to the response result, thereby generating an instruction execution log; The resource verification module is used to receive the user's submission instruction when the project development is completed, and based on the submission instruction, verify whether the project resources in the independent container are complete, and obtain the verification result; wherein, the project resources include deliverables, configuration files, test reports, detailed operation logs and instruction execution logs; The resource storage module is used to generate a data storage instruction if the verification result indicates that the project resources in the independent container are complete. The data storage instruction is used to instruct the project resources to be packaged and submitted to the project repository.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images