Intelligent coding-assisted robot

Abstract

This invention relates to an intelligent coding-assisted robot, comprising a graphical dialogue interface, a security protection module, a tag building module, and a processing module, and electrically connected to at least one cloud or ground server equipped with a generative pre-trained model. The security protection module allows for setting a plurality of data features, a plurality of access permissions, and a plurality of usage restrictions. When receiving a first piece of code input from an external source, the tag building module parses and obtains a plurality of segments and their corresponding plurality of identification tags. The processing module uses the data privacy filtering to form a second piece of code, which is transmitted to the cloud or ground server along with a request dialogue to obtain response data. The processing module then uses the access permissions and usage restrictions to generate edit data, and uses the identification tags to compare and extract at least one corresponding segment, thereby generating an interactive feedback with a plurality of edit action options for selection to edit the corresponding code.

Description

Technical Field

[0001] This invention relates to a program editing system, and more particularly to an intelligent coding-assisted robot. Prior Technology

[0002] In existing technologies, when using GPT (Generative Pre-trained Transformer) or other similar AI (Artificial Intelligence) models for code writing, querying, and modification, although these models can generate code based on user input, they still have several shortcomings in practical applications. First, these AI models typically lack a deep understanding and awareness of context, especially when dealing with code snippets from large projects or multi-layered software structures. The models cannot fully grasp the overall architecture or related dependencies, leading to potentially incompatible or unexecutable code. For example, when handling multi-module projects or cross-file code, existing models often cannot automatically retrieve the required functions, categories, variables, and line numbers, nor can they effectively establish references or tag relationships between these items. This forces users to spend a significant amount of time manually integrating or debugging, increasing their workload and failing to fully leverage the beneficial capabilities of intelligent tools.

[0003] Furthermore, existing AI models often lack real-time semantic verification and syntax checking when generating code, leading to potential syntax errors, logical inconsistencies, or mismatches with user needs. This forces users to invest extra time in debugging and correction, further reducing the efficiency of the tool. On the other hand, because existing AI model-generated code cannot be executed on the user's desired terminal or dynamically tested in the user's desired environment, it cannot verify the correctness of the code in real time, nor can it provide targeted correction suggestions, proving inadequate for complex programming scenarios. Moreover, these models generally lack version control and persistent memory capabilities. When users make multiple inquiries or modifications, the model cannot retain the complete context or track the code's modification history, resulting in potentially repetitive or inconsistent responses. This deficiency not only increases the burden on developers but also severely limits the possibility of multi-person collaboration, negatively impacting programming efficiency and teamwork.

[0004] In light of this, the present invention explores the concept of a mediating coding-assisted robot that provides an intelligent mediator model combining a graphical user interface and a dialogue interface. This model first parses and semantically analyzes the user's input, then filters out sensitive data from the program code logic and the dialogue requirements. The processed information is then passed to the aforementioned AI models for generation and response. This improves upon the shortcomings of existing technologies and addresses the lack of context awareness in existing AI models. Furthermore, it protects the information security of internal enterprise systems and enhances operational accuracy and stability. Summary of the Invention

[0005] The main objective of this invention is to provide an intelligent intermediary coding assistance robot that combines a graphical user interface, cybersecurity protection, and semantic integration technology to ensure information security and provide innovative solutions for team collaborative development and efficient code generation, thereby successfully overcoming the multiple shortcomings of conventional technologies.

[0006] To achieve the above objectives, the present invention discloses an intelligent coding-assisted robot, comprising a graphical dialogue interface electrically connected to at least one cloud or ground server and at least one database. The cloud or ground server is equipped with a generative pre-trained model for intelligent natural language dialogue and for providing feedback data. The intelligent coding-assisted robot is characterized by having a security protection module, a tag building module, and a processing module. The graphical dialogue interface is electrically connected to the security protection module, the tag building module, and the processing module. The processing module is electrically connected to the security protection module and the tag building module. The security protection module is used to set a plurality of data features, a plurality of access permissions, and a plurality of usage restrictions. When the graphical dialogue interface receives a first code input from an external source, the tag building module aggregates and parses the first code to obtain a plurality of segments and their corresponding plurality of... When the graphical user interface receives a request dialogue corresponding to one of the first code segments based on the identification tags, the processing module filters the first code segment using the data features to form a second code segment, and transmits the second code segment and the request dialogue dialogue to the cloud or ground server. When the cloud or ground server responds to the second code segment based on the request dialogue dialogue, the processing module uses the access permissions and usage restrictions to process the response data and generate edit data. It further uses the identification tags to compare the edit data segment and extract at least one corresponding segment. Based on this, it integrates the edit data segment and the segment to generate an interactive feedback with multiple edit action options and displays it on the graphical user interface. When a user reads the interactive feedback through the graphical user interface and selects one of the edit action options, the processing module edits the code segment corresponding to the edit action.

[0007] In a preferred embodiment, the intelligent coding-assisted robot further includes a code editing interface electrically connected to the graphical user interface (GUI), the security module, the tag creation module, and the processing module. The GUI includes an account verification element. When the robot begins operation, the GUI requests the user to enter an account and password. The account verification element verifies the usage level corresponding to the account and password. Upon determining that it is a code editing usage level, the GUI establishes a access channel for the code editing interface, allowing the user to control the robot through the code editing interface. The processing module includes a recording element that records at least the request dialogue, the response data, the editing data, and the editing log of the first code in the database.

[0008] In another preferred embodiment, when the graphical interface receives a dialogue corresponding to the request, the processing module parses the semantics of the dialogue and updates it to conform to the data logic, then transmits it to the cloud or ground server. The intelligent coding-assisted robot further includes an intelligent repair module electrically connected to the processing module and the code repair interface. The intelligent repair module checks the repaired first code, and when at least one code optimization method is known, it outputs a correction notification with one of the repair action options through the graphical interface, allowing the user to select and repair the first code again. The intelligent repair module further uses the comparison calculation of the reply information and the first code to correct a repair calculation model in the processing module.

[0009] The data features include API (Application Programming Interface) keys, server configurations, sensitive data, and encryption protocols. These access permissions set a user's access rights to the database, and the usage restrictions are operation scope restrictions and log tracking.

[0010] In summary, this invention provides an intelligent robot with a graphical user interface, a security module, and a tag creation module. This robot acts as an intermediary between an enterprise or personal local system and a cloud or local server (such as GPT or Copilot) and its database. It provides collaborative assistance functions such as automatic adjustment of user semantic logic, filtering and editing of code content, and integration of original fragments based on cloud feedback. This allows various users to easily perform code editing tasks. In other words, this invention exhibits the following characteristics: 1. High security and data feature protection – This security module can minimize the risk of data breaches when uploaded to the cloud or local server by using user-customized data features, access permissions, and usage restrictions. 2. With automated tag detection and analysis, the tag building module can automatically aggregate and analyze the first piece of code content uploaded or created by the user, determine the content and appropriately segment it into such segments, and generate corresponding identification tags for the content, so as to better process the specific content in the future; 3. Highly efficient automated response capability - Through the message interaction and integration between the processing module and the cloud or ground server, it can automatically output corresponding answers and suggestions based on the user's input of the request dialogue, that is, the interactive feedback and the modification action options, so that the user can easily perform all the required code modification work through the graphical dialogue interface. Furthermore, the present invention provides a code editing interface, allowing users to directly edit and view the code and various executable programs in the robot interface, effectively expanding the practical scope and flexibility of the intelligent coding-assisted robot to meet the usage needs of various users. Simple Explanation of the Diagram

[0011] Figure 1 is a structural diagram of a preferred embodiment of the present invention. Figure 2 is a schematic diagram of the architecture of two preferred embodiments of the present invention. Figures 3A and 3B are flowcharts of two preferred embodiments of the present invention. Figure 4 is a schematic diagram of the graphical dialogue interface of a preferred embodiment of the present invention. Figure 5 is a schematic diagram of the code editing interface of a preferred embodiment of the present invention. Implementation

[0012] To enable those skilled in the art to clearly understand the contents of this invention, please refer to the following description and accompanying drawings.

[0013] Please refer to Figure 1, which is an architectural diagram of a preferred embodiment of the present invention. As shown in the figure, the intelligent coding-assisted robot 1 includes a graphical dialogue interface 10, a security protection module 12, a tag building module 13, and a processing module 14. It is electrically connected to at least one cloud or ground server 2 and at least one database 3. The cloud or ground server 2 is equipped with a generative pre-trained model for intelligent natural language dialogue and feedback of response data 20. The graphical dialogue interface 10 is electrically connected to the security protection module 12, the tag building module 13, and the processing module 14. The processing module 14 is electrically connected to the security protection module 12 and the tag building module 13. The security protection module 12 is used to set a plurality of data features 120, a plurality of access permissions 121, and a plurality of usage restrictions 122.

[0014] When the graphical interface 10 receives a first code C1 from an external input, the tag building module 13 aggregates and parses the first code C1 to obtain a plurality of segments 130 and their corresponding plurality of identification tags 131. When the graphical interface 10 receives a request dialogue 101 corresponding to the first code C1, the processing module 14 uses the data features 120 to filter the first code C1 to form a second code, and transmits the second code and the request dialogue 101 to the cloud or ground server 2. When the cloud or ground server 2 responds to the second code based on the request dialogue 101, the processing module 14 uses the access permissions 121 and usage restrictions 122 to calculate the response data 20 and generate edit data. It then uses the identification tags 131 to compare the edit data and extract at least one corresponding segment 130. Based on this, the edit data and the segment 130 are integrated to generate an interactive feedback 141 with a plurality of edit action options 1410, which is displayed on the graphical interface 10. When a user reads the interactive feedback 141 through the graphical interface 10 and selects one of the edit action options 1410, the processing module 14 edits the corresponding code segment 130.

[0015] Please refer to Figures 2, 3A, 3B, 4, and 5, which are respectively architectural diagrams, flowcharts, and schematic diagrams of two preferred embodiments of the present invention. As shown in the figures, the intelligent coding assistance robot 1 includes a graphical dialogue interface 10, a code editing interface 11, a security protection module 12, a tag building module 13, a processing module 14, and an intelligent repair module 15. The intelligent coding assistance robot 1 is electrically connected to at least one cloud or ground server 2 and at least one database 3, such as a local database or a cloud database. The cloud or ground server 2 may be a server such as ChatGPT or Copilot and is equipped with a generative pre-trained model for intelligent natural language dialogue and feedback of response data 20. The graphical user interface 10 is equipped with an account verification element 100 and is electrically connected to the code editing interface 11. The code editing interface 11 is electrically connected to the security protection module 12, the tag building module 13, the processing module 14, and the intelligent repair module 15. The processing module 14 is equipped with a recording element 140 and is electrically connected to the security protection module 12, the tag building module 13, and the intelligent repair module 15. The operation process of the intelligent coding-assisted robot 1 may include the following steps.

[0016] When the robot starts operating, in step S1, the graphical interface 10 requests a user to enter an account and password (not shown) for verification of the correctness of the account and password and its usage level. In step S10, the account verification element 100 verifies whether the usage level corresponding to the account and password is a modification usage level. If so, in step S11, the graphical interface 10 establishes a usage channel for the code modification interface 11, allowing the user to directly control the robot through the code modification interface 11 to perform various tasks, such as setting various security protection information and editing a first code. Conversely, in step S12, the user is restricted from controlling the intelligent coding-assisted robot 1 through the graphical interface 10, thus only being able to use limited robot functions.

[0017] In step S2, the user, according to their account usage level, uses the graphical interface 10 to control the code editing interface 11, or directly through the code editing interface 11, to use the security module 12 to set multiple data features 120, multiple access permissions 121, and multiple usage restrictions 122. The data features 120 may include API (Application Programming Interface) keys, server configurations, sensitive data, and encryption protocols. The access permissions 121 set the user's access permissions to the database 3, and the usage restrictions 122 may be operation scope restrictions, such as restrictions on whether file deletion, file migration, service restart, etc., and log tracking. In step S3, when the user uploads the first code C1 to the code editing interface 11 through the graphical dialog interface 10, or directly writes the first code C1 through the code editing interface 11, the tag building module 13 aggregates and parses the first code C1 to obtain a plurality of segments 130 and their corresponding plurality of identification tags 131.

[0018] In step S4, when the graphical interface 10 receives a request dialogue 101 corresponding to one of the first code C1, the processing module 14 uses data features 120, such as containing highly sensitive words or data, names or words of confidential projects, to filter the first code C1 and form a second code. This ensures that only appropriate data can be accessed by the cloud or ground server 2, achieving effective information security protection. In step S40, the second code and the request dialogue 101 are transmitted to the cloud or ground server 2, so that the intelligent modules such as ChatGPT or Copilot can ask questions or make requests. When the graphical interface 10 receives the request dialogue 101, the processing module 14 parses the semantics of the request dialogue 101 and updates it to a request dialogue 101 that conforms to data logic before transmitting it to the cloud or ground server 2. The cloud or ground server 2 can understand the question without semantic differences and thus perform the correct calculation and response. In step S5, the cloud or ground server 2 parses the second code and provides response data 20 corresponding to the second code according to the request dialogue 101. In step S6, the processing module 14 uses the access permissions 121 and the usage restrictions 122 to process the response data 20 and generate edit data. In other words, the processing module 14 filters out or replaces data in the feedback data from the cloud or ground server 2 that is detrimental to the information security maintenance or trade secret protection of the enterprise's internal system. Then, in step S60, the editing data is further compared using the identification tags 131 to extract at least one corresponding segment 130, and the editing data and the segment 130 are integrated to combine the processed content with the original segment, thereby generating an interactive feedback 141 with a plurality of editing action options 1410 and displaying it on the graphical dialogue interface 10, so that the user can easily browse, read and select the request options on the graphical interface.

[0019] In step S7, when the user selects one of the editing action options 1410 by reading the interactive feedback 141 through the graphical dialog interface 10, the processing module 14 edits the corresponding code segment 130, allowing the user to complete the code editing work with simple clicks and inputs, thus greatly reducing the workload of manual operations. Alternatively, the user can directly select one of the editing action options 1410 by reading the interactive feedback 141 through the code editing interface 11, as shown in Figure 5, to clearly view the original and edited first code C1.

[0020] In this embodiment, in step S8, the recording element 140 records all actions and data processed by the robot, such as the request dialogue 101, the response data 20, the modification data, and the modification log of the first code C1, into the database 3. This allows the user to retrieve and view the data at any time through the code modification interface 11, and even perform operations such as restoring the original data. Furthermore, in step S9, the intelligent repair module 15 checks the modified first code C1. When at least one code optimization method is known, it outputs a correction notification with one of the modification action options 1410 through the graphical dialogue interface 10, allowing the user to select and modify the first code C1 again. In step S90, the intelligent repair module 15 further modifies a repair calculation model in the processing module 14 by comparing the response data 20 with the initial input first code C1, thereby improving the intelligence level of the processing module 14 and perfecting the accuracy and comprehensiveness of the calculation results. Incidentally, in addition to querying the cloud or ground server 2 to obtain the response data 20 corresponding to the second code, the intelligent coding assistance robot 1 can also connect to local intelligent models, such as llama3 or Mistra7b, to perform special calculations or detection operations using the local calculation model, thereby improving the functional completeness, comprehensiveness, and detail of the intelligent coding assistance robot 1.

[0021] The modules described in this invention are implemented through hardware or software supplemented by hardware. For example, the graphical interface 10, the security protection module 12, the tag building module 13, the processing module 14, and the intelligent repair module 15 are essentially defined as the integration of various hardware devices such as display panels, CPUs, microprocessors, memory, and signal transmitters, and implemented with the assistance of software programs.

[0022] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Therefore, any equivalent changes and modifications made without departing from the spirit and scope of the present invention should be covered within the patent scope of the present invention.

[0023] C1: First Code 1: Intelligent coding-assisted robot 10: Graphical Dialogue Interface 100: Account Verification Component 101: Needs Dialogue 11: Code Editing Interface 12: Security Protection Module 120: Data Characteristics 121: Access Permissions 122: Usage Restrictions 13: Tag Building Module 130: Excerpt 131: Identify Labels 14: Processing Module 140: Recording element 141: Interactive Feedback 1410: Modify Action Options 15: Intelligent Repair Module 2: Cloud or ground server 20: Response Materials 3: Database S1~S90: Steps

Claims

1. An intelligent coding-assisted robot, comprising a graphical dialogue interface electrically connected to at least one cloud or ground server and at least one database, wherein the cloud or ground server is equipped with a generative pre-trained model for intelligent natural language dialogue and feedback of response data, characterized in that: the intelligent coding-assisted robot is equipped with a security protection module, a tag building module and a processing module; the graphical dialogue interface is electrically connected to the security protection module, the tag building module and the processing module; the processing module is electrically connected to the security protection module and the tag building module; and the security protection module is configured to set a plurality of data features, a plurality of access permissions and a plurality of usage restrictions; when the graphical dialogue interface receives a first code input from an external source, the tag building module aggregates and parses the first code to obtain a plurality of segments and their corresponding plurality of identification tags; and when the graphical dialogue interface receives a request dialogue corresponding to the first code, the processing module uses the data features to filter the first code. A second code is generated to transmit the second code and the request dialogue to the cloud or ground server. When the cloud or ground server responds to the second code based on the request dialogue, the processing module uses the access permissions and usage restrictions to calculate the response data to generate edit data. It further uses the identification tags to compare the edit data and extract at least one corresponding segment. Based on this, the edit data and the segment are integrated to generate an interactive feedback with multiple edit action options and displayed on the graphical interface. When a user reads the interactive feedback through the graphical interface and selects one of the edit action options, the processing module edits the code of the corresponding segment.

2. The intelligent coding-assisted robot as described in claim 1 further includes a code editing interface electrically connected to the graphical user interface, the security protection module, the tag creation module, and the processing module. The graphical user interface includes an account verification element. When the robot starts running, the graphical user interface requests the user to enter an account password, and the account verification element verifies the usage level corresponding to the account password. Upon learning that it is a code editing usage level, the graphical user interface establishes a usage channel for the code editing interface, allowing the user to control the robot through the code editing interface.

3. The intelligent coding-assisted robot as described in claim 1, wherein, When the graphical interface receives a dialogue corresponding to the request, the processing module parses the semantics of the request dialogue and updates it to conform to the data logic, so as to transmit it to the cloud or ground server.

4. The intelligent coding-assisted robot as described in claim 1, wherein, The processing module is equipped with a recording element that records at least the request dialogue, the response information, the modification data, and the modification log of the first code into the database.

5. The intelligent coding-assisted robot as described in claim 2 is further provided with an intelligent repair module, which is electrically connected to the processing module and the code repair interface. The intelligent repair module checks the repaired first code and, upon learning at least one code optimization method, outputs a correction notification with one of the repair action options through the graphical dialog interface, which the user can click to repair the first code again.

6. The intelligent coding-assisted robot as described in claim 5, wherein, The intelligent repair module further utilizes the comparison and calculation of the response data and the first code to correct a repair calculation model in the processing module.

7. The intelligent coding-assisted robot as described in claim 1, wherein, These data characteristics include API (Application Programming Interface) keys, server configurations, sensitive data, and encryption protocols. These access permissions set a user's access rights to the database, and these usage restrictions are operation scope restrictions and log tracking.

Images