Code-assisted programming method, system and related device based on double-end collaboration

Abstract

This disclosure provides a code-assisted programming method, system, and related equipment based on dual-terminal collaboration, relating to the field of artificial intelligence technology. The method includes: generating anonymized code-assisted programming information in response to a code programming trigger request; sending the anonymized code-assisted programming information to a second terminal device located within the same secure communication network as a first terminal device, wherein the second terminal device is used to generate a code programming result with general functions based on a locally or edge-deployed AI large-scale model according to the anonymized code-assisted programming information; and receiving the code programming result returned by the second terminal device. This disclosure enables code-assisted programming using AI computing resources while ensuring data security.

Description

[0001] Cross-reference to related applications

[0002] This disclosure claims priority to Chinese Patent Application No. 202610117575.4, filed on January 28, 2026, entitled "Code-Assisted Programming Method, System and Related Device Based on Dual-End Collaboration", the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to the field of artificial intelligence technology, and in particular to a code-assisted programming method, system and related equipment based on dual-end collaboration. Background Technology

[0004] With the rapid development of artificial intelligence technology, the application scenarios of large language models are constantly expanding. AI-powered knowledge-based question-answering products such as ChatGPT have been gradually implemented and widely used. As a special form of expression of human natural language, program code's compatibility with large language models is being gradually discovered, enabling large language models to be effectively applied in the field of code programming.

[0005] Applying large language models and AIGC (Artificial Intelligence Generated Content) technology to the field of code programming has revolutionized the working mode of code developers. With the help of AIGC technology, code developers can complete various code writing-related tasks such as code completion, code generation, comment generation, code explanation, and unit test generation, significantly improving programming efficiency. Therefore, AI-assisted programming technology has become an important development direction in the field of programming.

[0006] However, the large models used in AI-assisted programming often require a lot of computing power when performing inference operations. The terminal devices used by code developers to write code often do not have AI computing capabilities and cannot meet the computing power requirements of AI-assisted programming. If AI computing power resources provided by the server side are relied upon, there may be a risk of leakage of core or sensitive data on the terminal side.

[0007] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0008] This disclosure provides a code-assisted programming method, system, and related equipment based on dual-end collaboration, which at least to some extent overcomes the technical problem of data leakage risk in AI-assisted programming schemes that rely on server-side computing resources.

[0009] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0010] According to one aspect of this disclosure, a code-assisted programming method based on dual-terminal collaboration is provided, applied to a first terminal device, comprising: generating de-identified code-assisted programming information in response to a code programming trigger request; sending the de-identified code-assisted programming information to a second terminal device located in the same secure communication network as the first terminal device, wherein the second terminal device is used to generate a code programming result with general functions based on a locally or edge-deployed artificial intelligence (AI) large model according to the de-identified code-assisted programming information; and receiving the code programming result returned by the second terminal device.

[0011] In some embodiments, in response to a code programming trigger request, generating desensitized code-aided programming information includes: in response to a code programming trigger request, dividing the code programming task requested by the code programming trigger request into multiple functional subtasks; and desensitizing the code-aided programming information of each functional subtask to obtain desensitized code-aided programming information for each functional subtask.

[0012] In some embodiments, after receiving the code programming result returned by the second terminal device, the method further includes: performing a security review and validity verification on the code programming result returned by the second terminal device; and integrating the verified code programming result into the business code of the first terminal device.

[0013] In some embodiments, before sending the desensitized code-assisted programming information to a second terminal device located in the same secure network as the first terminal device, the method further includes: establishing a secure communication network between the first terminal device and the second terminal device.

[0014] In some embodiments, the anonymized code-aided programming information includes at least one of the following: non-sensitive interface description information, abstract syntax structure, and anonymized code snippets; the code programming results include at least one of the following: code programming suggestions, general function code snippets, code descriptions, and code test cases.

[0015] According to another aspect of this disclosure, a code-assisted programming system based on dual-terminal collaboration is also provided, comprising: a first terminal device and a second terminal device located within the same secure communication network; wherein, the first terminal device is configured to generate de-identified code-assisted programming information in response to a code programming trigger request and send it to the second terminal device; the second terminal device is configured to generate code programming results with general functions based on the de-identified code-assisted programming information and a locally or edge-deployed artificial intelligence (AI) large model, and return them to the first terminal device.

[0016] According to another aspect of this disclosure, a code-assisted programming device based on dual-terminal collaboration is also provided, applied to a first terminal device, comprising: a code information desensitization processing module, used to generate desensitized code-assisted programming information in response to a code programming trigger request; a code programming request module, used to send the desensitized code-assisted programming information to a second terminal device located in the same secure communication network as the first terminal device, wherein the second terminal device is used to generate a code programming result with general functions based on a locally or edge-deployed artificial intelligence (AI) large model according to the desensitized code-assisted programming information; and a code programming response module, used to receive the code programming result returned by the second terminal device.

[0017] According to another aspect of this disclosure, an electronic device is also provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the code-assisted programming method based on dual-end collaboration described above by executing the executable instructions.

[0018] According to another aspect of this disclosure, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the code-assisted programming method based on dual-end collaboration as described in any one of the preceding claims.

[0019] According to another aspect of this disclosure, a computer program product is also provided, comprising: a computer program or instructions that, when executed by a processor, implement the code-assisted programming method based on dual-end collaboration as described above.

[0020] The code-assisted programming method, system, and related devices based on dual-terminal collaboration provided in this disclosure embodiment include a first terminal device responding to a code programming trigger request by sending de-identified code-assisted programming information to a second terminal device, so that the second terminal device generates the code programming result requested by the code programming trigger request based on a large model according to the de-identified code-assisted programming information, and returns it to the first terminal device.

[0021] This disclosed embodiment is based on a dual-terminal collaborative AI-assisted programming architecture. A first terminal device provides anonymized code programming assistance information, and a second terminal device located on the same secure communication network as the first terminal device provides AI computing power resources. The first and second terminal devices collaborate to complete the AI-assisted programming of the code. This not only solves the problem of insufficient computing power resources for the first terminal device to implement AI-assisted programming, but also avoids the risk of data leakage that may occur if the first terminal device relies on external computing power resources to implement AI-assisted programming.

[0022] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0024] Figure 1 The diagram illustrates a scheme for AI-assisted programming that utilizes server-side computing resources in related technologies. Figure 2 This diagram illustrates a code-assisted programming system architecture based on dual-end collaboration in an embodiment of this disclosure. Figure 3 This diagram illustrates a code-assisted programming method based on dual-end collaboration in an embodiment of the present disclosure. Figure 4 This diagram illustrates a flowchart of generating code-aided programming information according to an embodiment of the present disclosure; Figure 5 A flowchart of an optional code-assisted programming method based on dual-end collaboration is shown in an embodiment of this disclosure; Figure 6 This diagram illustrates a system architecture for assisted programming using the AI ​​computing power resources of a user's mobile phone, as described in an embodiment of this disclosure. Figure 7 This diagram illustrates a flowchart of a user computer performing assisted programming using the AI ​​computing power resources of a user's mobile phone, according to an embodiment of this disclosure. Figure 8 This diagram illustrates a code-assisted programming device based on dual-end collaboration in an embodiment of the present disclosure. Figure 9 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0025] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0026] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0027] To facilitate understanding, before introducing the embodiments of this disclosure, the following explanations are provided for several terms involved in the embodiments of this disclosure: AI: Artificial Intelligence.

[0028] ChatGPT: Chat Gnerative Pre-trained Transformer, is one of the commercial applications of LLM.

[0029] LLM: Large Language Model refers to a large-scale machine learning model, mainly used in the field of Natural Language Processing (NLP). These models consist of millions or even billions of parameters and are capable of performing various complex tasks, such as language translation, text generation, and programming.

[0030] AIGC: Artificial Intelligence Generated Content, refers to content automatically created or generated using artificial intelligence technologies, including various forms such as text, images, audio, and video. The core of AIGC lies in using machine learning models (especially LLM) to analyze and understand large amounts of data and generate new content based on that. These models typically learn patterns and rules specific to a particular domain by being trained on large-scale datasets.

[0031] SaaS (Software as a Service) is an online software distribution method where the software application is hosted by a third-party provider and provided to users via the internet. Users typically obtain the service through a subscription, eliminating the need for local installation or maintenance; they can simply access the service via the network.

[0032] IDE: Integrated Development Environment, is a software application that provides programmers with a convenient environment for writing, modifying, testing, and debugging code. IDEs typically include a code editor, compiler or interpreter, debugging tools, and automated tools such as plugins, making the software development process more efficient and simplified.

[0033] SDK: Software Development Kit, is a collection of tools and libraries used to help develop specific software packages, software frameworks, hardware platforms, computer systems, etc. SDKs typically include programming interfaces (APIs), programming tools, documentation, and sample code, enabling developers to create applications and software for a specific platform or system.

[0034] GPU: Graphics Processing Unit, is a type of computer hardware specifically designed for rapidly processing image and video-related computations. Due to its highly parallel architecture, GPUs are also widely used to accelerate various computationally intensive tasks, including machine learning and deep learning.

[0035] The specific implementation methods of the embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0036] Currently, the industry's conventional AI-assisted programming product solutions are primarily implemented using cloud-based SaaS (Software as a Service) services. Since AIGC technology relies heavily on computing resources, service providers need to build large-scale GPU server clusters or similar accelerated chip server clusters to meet the computing power demands of large model inference operations, and then provide these services to users through API interfaces. For programmers, when writing code in an integrated development environment (IDE), they need to use the IDE's built-in AI-assisted programming plugin to call these API interfaces from the cloud via the internet, thereby enabling various AIGC-assisted programming functions.

[0037] Existing cloud-based SaaS service solutions have many shortcomings in practical applications, including: 1) High operating and usage costs: To ensure stable service operation, cloud-based AIGC service providers need to continuously maintain large-scale GPU server clusters or similar accelerated chip server clusters to meet the computing power requirements of large model inference operations, resulting in high service operating costs. This cost is ultimately passed on to users, who need to pay corresponding fees based on the amount of code written or the duration of service usage. The higher the usage frequency, the greater the cost, increasing the financial burden on users.

[0038] 2) Service availability is limited by network conditions: Access to cloud-based AIGC services relies entirely on internet transmission, and service response speed is directly related to network quality. In scenarios with poor network conditions, excessive service latency can easily occur, leading to issues such as code completion lag and interrupted interactive dialogue, severely impacting the user's programming experience. More importantly, in offline scenarios without an internet connection, users will be completely unable to use AI-assisted programming functions, failing to meet users' needs in diverse network environments.

[0039] 3) Significant core data security risks: When using cloud-based AIGC services, users need to transmit code-related data to the cloud via the internet. For enterprise users, their core business code and various sensitive information inevitably face the risk of leakage. Especially for special user groups engaged in confidential work, data leakage may cause serious business security incidents, causing irreparable losses to users and failing to meet users' high-level data security requirements.

[0040] To address the aforementioned issues, this disclosure provides a collaborative architecture in an enterprise-controlled network or local environment where a first terminal device (e.g., device A) leverages the AI ​​computing resources of a second terminal device (e.g., device B) to generate / optimize code. The first terminal device (e.g., device A) retains and isolates core business code and sensitive data, transmitting only non-sensitive interface descriptions, abstract syntax structures, or desensitized code snippets to the second terminal device (e.g., device B) through task segmentation and data anonymization. The second terminal device (e.g., device B) utilizes its local or edge AI model to generate and optimize general sub-functions and returns the results. The first terminal device (e.g., device A) performs security review, verification, and integration locally, forming a closed loop.

[0041] Figure 2 This diagram illustrates a code-assisted programming system architecture based on dual-end collaboration, as shown in an embodiment of this disclosure. Figure 2 As shown, the system architecture may include a first terminal device 10 and a second terminal device 20 located within the same secure communication network.

[0042] In this embodiment of the disclosure, the first terminal device 10 is a device that provides AI application services, and the second terminal device 20 is a device that provides AI computing power.

[0043] The first terminal device 10 is used to generate de-identified code-assisted programming information in response to a code programming trigger request and send it to the second terminal device 20; the second terminal device 20 is used to generate code programming results with general functions based on a locally or edge-deployed AI large model according to the de-identified code-assisted programming information and return them to the first terminal device 10.

[0044] In some embodiments, the first terminal device 10 is further configured to: in response to a code programming trigger request, divide the code programming task requested by the code programming trigger request into multiple functional subtasks; and perform desensitization processing on the code-assisted programming information of each functional subtask to obtain desensitized code-assisted programming information of each functional subtask.

[0045] In some embodiments, the first terminal device 10 is further configured to: perform security review and validity verification on the code programming results returned by the second terminal device; and integrate the verified code programming results into the business code of the first terminal device.

[0046] In some embodiments, the first terminal device 10 is further configured to establish a secure communication network between the first terminal device 10 and the second terminal device 20.

[0047] In some embodiments, the above-mentioned desensitized code-aided programming information may include, but is not limited to, at least one of the following: non-sensitive interface description information, abstract syntax structure, and desensitized code snippets; the above-mentioned code programming results may include, but are not limited to, at least one of the following: code programming suggestions, general function code snippets, code descriptions, and code test cases.

[0048] The communication medium between the first terminal device 10 and the second terminal device 20 can be a wired network or a wireless network.

[0049] In the embodiments of this disclosure, the first terminal device 10 and the second terminal device 20 refer to entities on the user side used to receive or transmit signals, and may also be referred to as terminal devices, mobile stations (MS), mobile terminals (MT), etc. The embodiments of this disclosure do not limit the specific technology or device form used by the terminal; in principle, any electronic device can be used.

[0050] In some embodiments of this disclosure, the first terminal device 10 refers to an electronic device for user code programming, such as a user computer, including but not limited to laptops, handheld computers, and foldable computers; the second terminal device 20 refers to an electronic device with AI computing power resources, such as a user mobile phone, including but not limited to various mobile phones with AI chips.

[0051] Those skilled in the art will know that Figure 2 The number of first and second terminal devices is merely illustrative; any number of first and second terminal devices can be used as needed. This disclosure does not limit this.

[0052] Under the above system architecture, this disclosure provides a code-assisted programming method based on dual-end collaboration, which can be executed by any electronic device with computing capabilities.

[0053] In some embodiments, the code-assisted programming method based on dual-terminal collaboration provided in this disclosure can be executed by the first terminal device of the above-described system architecture; in other embodiments, the code-assisted programming method based on dual-terminal collaboration provided in this disclosure can be implemented by the first terminal device and the second terminal device in the above-described system architecture through interaction.

[0054] Figure 3 This diagram illustrates a code-assisted programming method based on dual-end collaboration in an embodiment of the present disclosure, as shown below. Figure 3 As shown, the code-assisted programming method based on dual-end collaboration provided in this embodiment includes the following steps: S302, in response to a code programming trigger request, generates de-identified code-aided programming information.

[0055] It should be noted that the aforementioned code programming trigger request refers to a request triggered by the first terminal device that requires the AI ​​computing resources of the second terminal device for AI-assisted programming. This request can be initiated by the user or automatically triggered by the system (for example, if the user suddenly stops typing during code writing and the duration exceeds a preset threshold, a request to execute the AI ​​programming assistance task is automatically triggered). The aforementioned code assistance information refers to the information required for generating code through assistance programming. The second terminal device needs code assistance information to execute the AI-assisted programming task, but if the first terminal device directly sends the code assistance information to the second terminal device, there may be a risk of sensitive data leakage.

[0056] In this embodiment of the disclosure, the first terminal device responds to the code programming trigger request and generates desensitized code-assisted programming information, which can ensure that the second terminal device completes the AI-assisted programming task and also prevent the leakage of sensitive data (such as core business data or code) of the first terminal device.

[0057] In some embodiments, after receiving a code programming trigger request (such as a code completion request, a syntax correction request, or a functional module generation request), the first terminal device performs desensitization processing on the original code, context information, sensitive configurations, and other data carried in the code programming trigger request, removing or replacing the privacy data, core business data, and sensitive identifiers contained therein, to form code-assisted programming information that conforms to the secure transmission and use specifications.

[0058] S304, the de-identified code-assisted programming information is sent to a second terminal device located in the same secure communication network as the first terminal device. The second terminal device is used to generate code programming results with general functions based on a locally or edge-deployed AI large model according to the de-identified code-assisted programming information.

[0059] In the above steps, the first terminal device transmits the anonymized code-assisted programming information to the second terminal device through the same secure communication network (such as an enterprise intranet, dedicated VPN, or edge computing cluster intranet). This allows the second terminal device to generate code programming results with general functions using a locally or edge-deployed AI large model. Since the first terminal device does not rely on public network service-side AI computing resources in this embodiment, the risk of data leakage is significantly reduced.

[0060] S306 receives the code programming result returned by the second terminal device.

[0061] In the above steps, after the first terminal device sends the de-identified code assistance information to the second terminal device, the second terminal device returns the general function code programming result generated based on the local or edge AI large model to the first terminal device through the same secure communication network, thus realizing the closed loop of the code assistance programming request.

[0062] In some embodiments, the above-mentioned desensitized code-aided programming information may include, but is not limited to, at least one of the following: non-sensitive interface description information, abstract syntax structure, and desensitized code snippets; the above-mentioned code programming results may include, but are not limited to, at least one of the following: code programming suggestions, general function code snippets, code descriptions, and code test cases.

[0063] In this embodiment, non-sensitive interface description information refers to information after the interface description information has been anonymized (sensitive information removed). The interface description information retains only the general attributes of the interface and does not disclose any details related to specific business or privacy. It allows the AI ​​large-scale model to understand interface requirements but cannot deduce sensitive information. Abstract syntax structure refers to the abstract extraction of the syntactic logic framework of the code, stripping away specific variable names, function names, business parameter values, etc., retaining only the structured information of the code's hierarchical structure, statement types, and logical relationships (such as conditional judgments, loop execution, and function call relationships). The abstract syntax structure does not involve any specific business semantics or sensitive data and can guide the AI ​​large-scale model to generate code that conforms to syntactic specifications and logical requirements. Anonymized code snippets refer to code snippets obtained after anonymizing the original code snippets. Sensitive elements are removed through replacement, hiding, deletion, etc., while retaining the syntactic correctness and functional logic integrity of the code. Sensitive elements include, but are not limited to, privacy data, core business parameters, key information, and internal system identifiers. The first terminal device provides the second terminal device with de-identified code snippets, which retains sufficient context for the AI ​​large model to refer to while avoiding the leakage of sensitive information. It can be directly used as input reference for AI-generated code.

[0064] In this embodiment, code programming suggestions refer to guiding suggestions for code writing generated by the AI ​​model based on de-identified auxiliary information. They do not directly provide complete code, but rather focus on providing reference solutions for code architecture design, syntax optimization, performance improvement, and security specifications. Code programming suggestions help developers optimize their coding approach rather than directly producing runnable code. General-purpose functional code snippets refer to reusable code blocks generated by the AI ​​model that possess general business functions and are decoupled from specific sensitive information. These code snippets do not contain any private data, keys, or exclusive business identifiers; they only implement standardized functional logic. Developers can quickly adapt them to actual business scenarios by adding personalized parameters. Code descriptions refer to explanatory descriptions written by the AI ​​model for the generated code or programming suggestions. These descriptions include, but are not limited to, the code's functional purpose, logical flow, parameter descriptions, usage methods, and precautions, helping developers understand and correctly call the code. Code test cases refer to standardized test scripts or test scenarios written by the AI ​​model based on the generated general-purpose code functionality, used to verify the code's correctness, robustness, and boundary condition handling capabilities. Test cases typically include elements such as input data, expected output, and test steps.

[0065] This application achieves a triple technical effect of security protection, efficiency improvement, and enhanced versatility by limiting the specific content types of the anonymized code-assisted programming information and code programming results. Specifically, this can be summarized in the following three points: ① The anonymized code-assisted programming information only includes non-sensitive interface descriptions, abstract syntax structures, and anonymized code snippets, eliminating sensitive information such as privacy data, core business parameters, and keys from the original requests. This avoids the risk of sensitive data leakage during transmission and AI model processing, while also meeting relevant compliance requirements for data security and privacy protection. ② Using structured and anonymized information as input to the AI ​​large-scale model allows the model to focus on the functional logic and syntax specifications of the code, reducing interference from sensitive information. The code programming results include code programming suggestions, general functional code snippets, code descriptions, and test cases, covering the entire process of guidance-coding-interpretation-verification. This directly matches the actual needs of developers, significantly improving the efficiency of assisted programming and the reusability of the code. ③ The code programming assistance information is decoupled from sensitive information such as specific business and terminal environment. The general functional code fragments included in the code programming results do not depend on specific sensitive configurations. This makes the code-assisted programming scheme provided in this embodiment adaptable to the intranet / edge computing scenarios of different enterprises. It does not require a lot of customized development for different business scenarios, which significantly improves the universality and deployment flexibility of the technical solution.

[0066] In some embodiments, such as Figure 4 As shown in the embodiments of this disclosure, the code-assisted programming method based on dual-end collaboration can generate anonymized code-assisted programming information through the following steps: S402, in response to the code programming trigger request, divides the code programming task requested by the code programming trigger request into multiple functional subtasks; S404, desensitize the code-aided programming information of each functional subtask to obtain the desensitized code-aided programming information of each functional subtask.

[0067] In this embodiment, a functional subtask refers to the smallest executable unit obtained by breaking down a complete and complex code programming task according to dimensions such as functional modules, logical flow, and technical dimensions. Each functional subtask can complete a single or independent code function, and the code programming information of each functional subtask only contains the minimum range of code-aided programming information required to complete that functional subtask. For example, if a functional task is "generating code for a user login module", then the segmented functional subtasks may include: ① a mobile phone number format verification subtask; ② a user password hash encryption subtask; ③ a database user information query subtask; and ④ a login result return and exception handling subtask.

[0068] Through the above embodiments, upon receiving a code programming trigger request (such as generating code for a certain business module), instead of directly anonymizing the code-aided programming information for the entire task, the complete code programming task is first broken down into multiple independent functional subtasks, each corresponding to a specific sub-functional module. For each functional subtask, only the minimum range of code-aided programming information required to complete the functional subtask is extracted, and this information is anonymized separately (such as replacing sensitive parameters in the subtask or abstracting the syntax structure of the subtask), ultimately obtaining the anonymized code-aided programming information for each functional subtask.

[0069] The above embodiments can achieve, but are not limited to, the following technical effects: ① By splitting functional subtasks, the scope of desensitization can be precisely controlled, maximizing data security; ② AI-assisted programming processing at the level of small-granularity subtasks can reduce the processing burden of AI models and improve code generation efficiency; ③ Enhanced fault tolerance facilitates differentiated customization and later maintenance.

[0070] In some embodiments, after receiving the code programming result returned by the second terminal device, such as Figure 5 As shown in the embodiments of this disclosure, the code-assisted programming method based on dual-end collaboration further includes the following steps: S308, perform security review and validity verification on the code programming results returned by the second terminal device; S310 integrates the verified code programming results into the business code of the first terminal device.

[0071] In the above embodiments, after receiving the code programming results returned by the second terminal, the first terminal device performs dual verification on the general code programming results (such as code snippets, programming suggestions, etc.) generated by the second terminal device. On the one hand, it conducts a security review to detect whether there are security vulnerabilities, malicious logic, unauthorized calls, or other security risks in the code; on the other hand, it performs a validity verification to check whether the code conforms to the preset syntax specifications, whether it matches the functional requirements of the original programming task, and whether it can run normally and output as expected. Finally, only the code programming results that have passed the security review and validity verification are integrated and embedded into the existing business code of the first terminal device, completing the transformation from general code results to deployable business code.

[0072] In this embodiment, the AI ​​big model of the second terminal device generates general functional code based solely on anonymized code-aided programming information. It cannot perceive the actual business environment and security specifications of the first terminal device, and the generated code may contain hidden syntax vulnerabilities, logical defects, or even (in rare cases) the risk of malicious calls. Through security audits (such as static code scanning and vulnerability detection), insecure code can be identified and blocked in advance, avoiding the integration of risky code into the business system, which could lead to data leakage, system crashes, and other problems. At the same time, validity verification can exclude invalid code that is syntactically correct but deviates from the functional requirements, ensuring that the integrated code truly matches the original programming task objectives.

[0073] In traditional AI-assisted programming workflows, developers need to manually verify the code generated by large AI models and adapt it to the local business architecture, which is time-consuming and error-prone. The above-described embodiments of this disclosure standardize and automate security review and validity verification. By using preset verification rules (such as syntax verification templates and function matching algorithms), usable code is quickly filtered out. The verified code is then embedded into the business code of the first terminal device (such as automatically completing private parameter placeholders and connecting to local function call logic), which greatly reduces the workload of developers in manual adaptation and achieves efficient transformation from receiving code to integrating it into the business.

[0074] In some embodiments, before sending the desensitized code-assisted programming information to a second terminal device located in the same secure network as the first terminal device, the code-assisted programming method based on dual-terminal collaboration provided in this disclosure embodiment further includes the following step: establishing a secure communication network between the first terminal device and the second terminal device.

[0075] In this embodiment of the disclosure, a secure communication network refers to a dedicated communication link with multi-level security protection mechanisms built for a first terminal device (such as an office computer of a research and development personnel) and a second terminal device (such as an edge server / internal network server that deploys a large AI model). This is different from networks in open public network environments, such as enterprise-built internal VPNs, dedicated R&D networks based on VLANs, private local area networks composed of edge computing nodes, and short-range communication networks.

[0076] Before transmitting the anonymized code-assisted programming information, a dedicated secure communication network for both terminals must be established (or the validity of an existing network must be confirmed). This network must possess security features such as physical / logical isolation, encrypted transmission, and authentication. Only after this dedicated secure network is established and verified to be usable will the first terminal device transmit the anonymized code-assisted programming information to the second terminal device, rather than transmitting it through the public network or an unprotected ordinary network. The secure communication network, through physical isolation and end-to-end encryption, eliminates the possibility of data interception or tampering at the transmission link level. Even anonymized information can avoid the leakage of core programming logic, further ensuring the uniqueness and security of data transmission. For example, in scenarios with extremely high data security requirements, such as finance, government, and large enterprises, even anonymized data must not touch the public network throughout the entire process. This embodiment of the disclosure, by establishing a compliant secure communication network, ensures that the entire dual-terminal collaboration solution complies with industry regulations and internal enterprise security standards, and can be deployed at scale to all R&D terminals within the enterprise, rather than being limited to small-scale testing scenarios.

[0077] With the rapid development of mobile phone hardware technology, more and more mobile phones are beginning to incorporate independent chips with AI capabilities. Therefore, in some embodiments, the computing power resources of the AI ​​chip built into the user's mobile phone can be used to support the computing power required for large model inference operations used in AIGC-assisted programming.

[0078] In practical implementation, an application (such as an APP or mini-program) providing AIGC services can run on the mobile phone. The IDE plugin on the user's computer and the APP on the mobile phone form a peer-to-peer connection through a secure communication network, allowing the user to leverage the AI ​​chip computing power on the mobile phone while on the computer. The network connection between the user's mobile phone and the user's computer can be achieved through, but is not limited to, the following two methods: ① using the mobile phone's own hotspot to connect the user's computer, forming a peer-to-peer network based on WiFi; ② the user's mobile phone and computer are connected to the same WiFi local area network, establishing a peer-to-peer connection through the local area network.

[0079] The following describes in detail the code-assisted programming method based on dual-terminal collaboration provided in this embodiment, taking the first terminal device as a user's computer and the second terminal device as a user's mobile phone as an example.

[0080] Figure 6 This illustration shows a system architecture diagram of a user computer using the AI ​​computing power resources of a user's mobile phone for assisted programming, as shown in an embodiment of this disclosure. Figure 6 As shown, the user's computer contains an IDE for code writing and its AI-assisted programming plugin; the user's mobile phone contains a large model and related SDKs, and can run AIGC service applications (such as apps or mini-programs) based on the large model.

[0081] In practice, users develop code within their computer's IDE. The AI ​​plugin's functionality can be triggered automatically by the IDE or manually by the user. To mitigate the risk of sensitive information leakage, the AI ​​plugin performs task segmentation and sensitivity assessment before initiating collaboration, minimizing and de-identifying code snippets and requirements to only generate necessary interface descriptions, abstract syntax structures, or de-identified fragments. Subsequently, the AI ​​plugin negotiates capabilities with the user's mobile phone and establishes an end-to-end encrypted device-to-device secure connection, sending non-sensitive task descriptions and constraints to the user's mobile phone's AIGC service. Upon receiving the code programming results from the user's mobile phone, the AI ​​plugin is displayed in the IDE and performs static / dynamic review, testing, and integration locally, ensuring that core business code remains on the user's computer and is never distributed externally.

[0082] In some embodiments, the user's mobile phone provides network services to the first terminal device via API. After receiving a non-sensitive task description and constraints from the user's computer, the app performs inference on a local large model to generate corresponding code snippets, instructions, and test cases, and returns the results to the user's computer. The inference operation of the large model is supported by the computing power of the user's mobile phone's CPU and AI chip. The inference process is executed in a restricted sandbox environment, prohibiting access to unauthorized data and uncontrolled external networks, further enhancing security.

[0083] The user's computer and mobile phone establish a peer-to-peer connection through a controlled network, which may include, but is not limited to, direct connection to a mobile hotspot, the same Wi-Fi LAN, Wi-Fi Direct, Bluetooth, or Ethernet. During communication, two-way authentication and end-to-end encryption (e.g., TLS / QUIC) are employed, along with data minimization and access control policies to ensure that non-sensitive information is shared only when necessary. This connection supports low-latency collaboration in weak network / local environments and records communication and operation audit logs to meet compliance requirements.

[0084] In this embodiment of the disclosure, the user's mobile phone may also be other terminal devices that can provide AIGC services or have AI computing power, such as tablets, GPU workstations, edge servers, and other controlled terminals with AI computing power.

[0085] Through the aforementioned terminal-to-terminal collaboration architecture and security mechanism, this invention achieves efficient collaboration and local closed-loop integration of AI-assisted programming while ensuring that core enterprise code and sensitive information are not leaked, thus balancing compliance, low latency, and availability.

[0086] Figure 7This diagram illustrates a flowchart of a user computer using the AI ​​computing power resources of a user's mobile phone for assisted programming, as described in an embodiment of this disclosure. Before using the user's mobile phone for assisted programming, the user can turn on their mobile hotspot and connect their computer to the hotspot network; or the user can connect their mobile phone and computer to the same local area network (such as a WiFi network); the user opens the AIGC service application on their mobile phone and, according to the instructions on the AIGC service application, connects the IDE plugin on their computer to the application. Figure 7 As shown, code-assisted programming specifically includes the following steps: S702, users use the IDE to write code or perform other programming-related tasks; S704, AI-assisted programming function can be automatically triggered or manually triggered by the user; S706, the IDE plugin collects relevant code snippets of the user's current location and calls the AIGC service on the mobile phone side; In S708, after receiving a request, the AIGC service application on the mobile side understands the request type and content, and calls the relevant functions of the large model SDK according to different types and content. S710 calls a large model to perform inference calculations; S712, the large model returns the inference results to the large model SDK, which then encapsulates them into a complete content response; S714 After receiving the content response from the SDK, the AIGC service application returns it to the computer-side IDE plugin through a secure communication network. S716: After receiving the content results returned by the AIGC service, the IDE plugin displays them within the IDE. S718 allows users to view results in the IDE or perform corresponding operations based on the results (such as accepting code generated from a large model).

[0087] Optionally, in some embodiments, the user's computer side (IDE plugin) may display, but is not limited to, the following: 1) Device connection status indicator (not connected / connecting / connected / restricted), current B device name, and device switching dropdown menu; 2) The "Create Collaborative Task" button can be, but is not limited to, a button that performs the following functions: generate unit tests, implement functions, explain code, etc. 3) Task queue list, which may include, but is not limited to: task name, language, status (including but not limited to queued, processing, completed, failed, etc.), task duration, and token usage summary; 4) Historical request list (searchable and filterable by task type, file path, and time). 5) Model output code snippets.

[0088] Optionally, in some embodiments, the user's computer side (IDE plugin) may display, but is not limited to, the following: 1) List of currently connected devices (device name, status, last request time, etc.); 2) Request ID, source device, task type, language, duration, status, etc.; 3) Click to view request details, which include input / output information.

[0089] Taking a user writing code in the IDE editor as an example, when the user stops typing, the AI-assisted programming plugin automatically triggers the code completion function. The AI-assisted programming plugin automatically uses the code snippet at the user's current position, performs anonymization processing, and sends it to the AIGC server. The AIGC server returns the AI-generated completed code. The AI-assisted programming plugin receives the relevant completed code and displays it in the IDE for the user to view or use.

[0090] As can be seen from the above, the code-assisted programming solution based on dual-end collaboration provided in this disclosure utilizes the mobile phone's AI chip to meet the computing power requirements of the large model inference operations used in AIGC-assisted programming. This allows users to use AI-assisted programming without being limited by usage volume and duration, achieving almost zero-cost unlimited use. The direct connection between the user's computer and mobile phone significantly reduces the impact of network on the user experience, lowers latency, and avoids lag issues. Furthermore, users can use the AI-assisted programming function even without an internet connection. Since the user's computer and mobile phone are typically their own devices, user data is not collected or uploaded to the cloud, greatly reducing the security risk of data leakage.

[0091] It should be noted that the user mobile phone providing AI computing power in this embodiment can also be replaced with other electronic devices with AI computing power, such as AI computing sticks (e.g., Intel hardware products) or AI development boards (e.g., NVIDIA hardware products). Furthermore, the dual-end collaboration solution provided in this embodiment can be applied not only to code-assisted programming but also to other AIGC-related applications, such as text-to-image processing, AI question answering, and image processing applications and services. Such applications can also utilize similar technical frameworks and designs to achieve and fulfill the objectives of this invention.

[0092] It should be noted that the acquisition, storage, use, and processing of data in this disclosed technical solution comply with the relevant provisions of laws and regulations. All types of data, such as personal identity data, operational data, and behavioral data related to individuals, customers, and groups, obtained in this disclosed embodiment have been agreed upon by the users.

[0093] Based on the same inventive concept, this disclosure also provides a code-assisted programming device based on dual-end collaboration. This device may be, but is not limited to, the first terminal device described in the above embodiments. See the following embodiments. Since the principle by which this device embodiment solves the problem is similar to that of the above method embodiments, the implementation of this device embodiment can refer to the implementation of the above method embodiments, and repeated details will not be elaborated further.

[0094] Figure 8 This illustration shows a schematic diagram of a code-assisted programming device based on dual-end collaboration in an embodiment of this disclosure, such as... Figure 8 As shown, the device includes: a code information desensitization processing module 801, a code programming request module 802, and a code programming response module 803.

[0095] The code information desensitization processing module 801 is used to generate desensitized code-aided programming information in response to a code programming trigger request; the code programming request module 802 is used to send the desensitized code-aided programming information to a second terminal device located in the same secure communication network as the first terminal device, wherein the second terminal device is used to generate code programming results with general functions based on a locally or edge-deployed artificial intelligence AI model according to the desensitized code-aided programming information; and the code programming response module 803 is used to receive the code programming results returned by the second terminal device.

[0096] In some embodiments, the code information desensitization processing module 801 is further configured to: in response to a code programming trigger request, divide the code programming task requested by the code programming trigger request into multiple functional subtasks; and desensitize the code auxiliary programming information of each functional subtask to obtain desensitized code auxiliary programming information of each functional subtask.

[0097] In some embodiments, the code-assisted programming device based on dual-terminal collaboration provided in this disclosure further includes: a code verification module 804, used to perform security review and validity verification on the code programming results returned by the second terminal device; and a code integration module 805, used to integrate the verified code programming results into the business code of the first terminal device.

[0098] In some embodiments, the code-assisted programming device based on dual-end collaboration provided in this disclosure further includes: a secure communication network establishment module 806, used to establish a secure communication network between the first terminal device and the second terminal device.

[0099] In some embodiments, the above-mentioned desensitized code-aided programming information may include, but is not limited to, at least one of the following: non-sensitive interface description information, abstract syntax structure, and desensitized code snippets; the above-mentioned code programming results may include, but are not limited to, at least one of the following: code programming suggestions, general function code snippets, code descriptions, and code test cases.

[0100] It should be noted that the examples and application scenarios implemented by the modules in the above device embodiments and the corresponding steps in the method embodiments are the same, but are not limited to the content disclosed in the above method embodiments. It should also be noted that the above modules, as part of the device, can be executed in a computer system such as a set of computer-executable instructions.

[0101] Those skilled in the art will understand that various aspects of this disclosure can be implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, which can be collectively referred to herein as a "circuit", "module" or "system".

[0102] Based on the same inventive concept, this disclosure also provides an electronic device, which includes: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any of the above-described code-assisted programming methods based on dual-end collaboration by executing the executable instructions. Since the principle by which this electronic device solves the problem is similar to that of the above-described method embodiments, the implementation of this electronic device embodiment can refer to the implementation of the above-described method embodiments, and repeated details will not be elaborated further.

[0103] The following reference Figure 9 To describe an electronic device 900 according to such an embodiment of the present disclosure. Figure 9 The electronic device 900 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0104] like Figure 9 As shown, the electronic device 900 is manifested in the form of a general-purpose computing device. The components of the electronic device 900 may include, but are not limited to: at least one processing unit 910, at least one storage unit 920, and a bus 930 connecting different system components (including storage unit 920 and processing unit 910).

[0105] The storage unit stores program code that can be executed by the processing unit 910, causing the processing unit 910 to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure. For example, the processing unit 910 can perform the following steps of the above method embodiment: in response to a code programming trigger request, generating de-identified code-aided programming information; sending the de-identified code-aided programming information to a second terminal device located in the same secure communication network as the first terminal device, wherein the second terminal device is used to generate code programming results with general functions based on a locally or edge-deployed AI large model according to the de-identified code-aided programming information; and receiving the code programming results returned by the second terminal device.

[0106] Storage unit 920 may include readable media in the form of volatile storage units, such as random access memory (RAM) 9201 and / or cache memory 9202, and may further include read-only memory (ROM) 9203.

[0107] The storage unit 920 may also include a program / utility 9204 having a set (at least one) program module 9205, such program module 9205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0108] Bus 930 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0109] Electronic device 900 can also communicate with one or more external devices 940 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 900, and / or with any device that enables electronic device 900 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 950. Furthermore, electronic device 900 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 960. As shown, network adapter 960 communicates with other modules of electronic device 900 via bus 930. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 900, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0110] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0111] Based on the same inventive concept, this disclosure also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements any of the above-described code-assisted programming methods based on dual-end collaboration. Since the principle by which this computer-readable storage medium embodiment solves the problem is similar to that of the above-described method embodiments, the implementation of this computer-readable storage medium embodiment can refer to the implementation of the above-described method embodiments, and repeated details will not be elaborated further.

[0112] More specific examples of computer-readable storage media in this disclosure may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0113] In this disclosure, a computer-readable storage medium may include a data signal propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of transmitting, propagating, or transmitting a program for use by or in connection with an instruction execution system, apparatus, or device.

[0114] Optionally, the program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0115] In practical implementation, program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0116] Based on the same inventive concept, this disclosure also provides a computer program product, including a computer program or instructions, which, when executed by a processor, implements any one of the code-assisted programming methods based on dual-end collaboration in the above method embodiments. Since the principle by which this computer program product embodiment solves the problem is similar to that of the above method embodiments, the implementation of this computer program product embodiment can refer to the implementation of the above method embodiments, and repeated details will not be elaborated further.

[0117] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0118] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0119] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, mobile terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0120] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the appended claims.

Claims

1. A code-assisted programming method based on dual-end collaboration, characterized in that, Applied to the first terminal device, including: In response to a code programming trigger request, generate anonymized code-aided programming information; The anonymized code-assisted programming information is sent to a second terminal device that is in the same secure communication network as the first terminal device. The second terminal device is used to generate code programming results with general functions based on a locally or edge-deployed AI large model according to the anonymized code-assisted programming information. Receive the code programming result returned by the second terminal device.

2. The code-assisted programming method based on dual-end collaboration according to claim 1, characterized in that, In response to a code programming trigger request, anonymized code-aided programming information is generated, including: In response to a code programming trigger request, the code programming task requested by the code programming trigger request is divided into multiple functional subtasks; The code-aided programming information for each functional subtask is anonymized to obtain the anonymized code-aided programming information for each functional subtask.

3. The code-assisted programming method based on dual-end collaboration according to claim 1, characterized in that, After receiving the code programming result returned by the second terminal device, the method further includes: The code programming results returned by the second terminal device are subject to security review and validity verification. The verified code programming results are integrated into the business code of the first terminal device.

4. The code-assisted programming method based on dual-end collaboration according to claim 1, characterized in that, Before sending the de-identified code-assisted programming information to a second terminal device located in the same secure network as the first terminal device, the method further includes: Establish a secure communication network between the first terminal device and the second terminal device.

5. The code-assisted programming method based on dual-end collaboration according to any one of claims 1 to 4, characterized in that, The anonymized code-aided programming information includes at least one of the following: non-sensitive interface description information, abstract syntax structure, and anonymized code snippets; the code programming results include at least one of the following: code programming suggestions, general function code snippets, code descriptions, and code test cases.

6. A code-assisted programming system based on dual-end collaboration, characterized in that, include: The first and second terminal devices are located within the same secure communication network; The first terminal device is used to generate de-identified code-aided programming information in response to a code programming trigger request, and send it to the second terminal device; The second terminal device is used to generate code programming results with general functions based on the de-identified code-assisted programming information and a large AI model deployed locally or at the edge, and then return the results to the first terminal device.

7. A code-assisted programming device based on dual-end collaboration, characterized in that, Applied to the first terminal device, including: The code information desensitization module is used to generate desensitized code-aided programming information in response to code programming trigger requests; The code programming request module is used to send the de-identified code-assisted programming information to a second terminal device that is in the same secure communication network as the first terminal device. The second terminal device is used to generate code programming results with general functions based on the de-identified code-assisted programming information and a large AI model deployed locally or at the edge. The code programming response module is used to receive the code programming results returned by the second terminal device.

8. An electronic device, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to execute the code-assisted programming method based on dual-end collaboration as described in any one of claims 1 to 6 by executing the executable instructions.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the code-assisted programming method based on dual-end collaboration as described in any one of claims 1 to 6.

10. A computer program product, comprising: A computer program or instruction, characterized in that, when executed by a processor, the computer program or instruction implements the code-assisted programming method based on dual-end collaboration as described in any one of claims 1 to 6.

Images