DocAIGenPlus
The DocAIGenPlus project aims to investigate and develop a solution for the automatic generation of technical documentation, C4 diagrams, and sequence diagrams for software development projects.
Support the automatic creation of business content, such as technical documentation, C4 diagrams and sequence diagrams.
Reduce product delivery and maintenance times.
Improve knowledge sharing between the work teams associated with a product.
Improve the learning process of Generative AI mechanisms through human inputs.
Project Context
The problem:
The creation of business content, such as technical documentation, C4 diagrams, and sequence diagrams, is currently a process that requires significant effort, characterized by a lack of consistency and efficiency, with a propensity for human error. The maintenance of technical documentation is often neglected due to the pressure to deliver products faster, which can result in outdated or inaccurate information. There are still challenges regarding collaboration and knowledge sharing among work teams, where information is often fragmented and dispersed.
Generative AI represents one of the most remarkable evolutions in the field of AI, enabling the creation of new data and content that resemble real-world data, such as texts, images, videos, three-dimensional models, and business contents, like technical documentation, C4 diagrams, and sequence diagrams. However, there is still a way to go regarding the application of Generative AI, since many users still do not fully understand how to use it effectively, and there is a need to improve the learning process. There are solutions on the market that attempt to address some of these challenges, but many still have gaps, not offering the full range of necessary features, nor the ability to automate tasks without human intervention.
The solution:
The DocAIGenPlus project proposes the research and development of an innovative ecosystem for the automatic generation of technical documentation, C4 diagrams, and sequence diagrams, using generative AI mechanisms and human-in-the-loop techniques to improve the quality of the generated content. This ecosystem will include two interfaces for automatic content generation: a CI/CD plugin, which will be integrated into code change processes in repositories to automatically generate the desired content; and a web interface, where users can submit code updates so that the platform automatically generates the documentation and diagrams. The web interface, which can be used to generate content in legacy systems that do not have a CI/CD pipeline, will also offer other functionalities, such as the choice of repositories for storing the generated content, manual validation of the generated content, etc.
Goals
Automatic content generation for businesses
Support the automatic creation of business content, namely C4 diagrams, sequence diagrams, and technical documentation.
Product creation process optimization
Reducing product and content delivery and maintenance time, without neglecting the creation and maintenance of technical documentation.
More collaborative business environments
Improving collaboration and knowledge sharing among work teams associated with a product/project.
Contents with greater quality
Improve the learning process of generative AI mechanisms through human input.
User Interfaces
- Generative AI
- Storage
- Security
- Integration with Repositories
- Code Processing
- Human in the Loop
- System Versioning
- User Management
CI/CD plugin
The CI/CD plugin will be integrated into DevOps pipeline flows to find changes in the software code in the repositories, in order to generate new and useful content for end users.
Features
CI/CD Plugin
The CI/CD plugin will be an integrated mechanism in the frontend and backend CI/CD pipeline flows of any desired system, continuously monitoring code repositories. With each change in the software code, it will interpret the system changes, generating a new set of diagrams and technical documentation that reflects the updated solution. The new contents, generated automatically, can be accessed and evaluated by end users on the web interface.
Web interface
Web-enabled interface that will allow the automatic generation of technical documentation, sequence diagrams, and C4 diagrams, using the software’s source code, loaded into the system, as a basis. This interface also allows users to rate the quality of the content that was automatically generated by the system’s Gen AI algorithms, suggesting corrections and improvements to them.
Features
Authentication and user management in the web interface
Functionality that allows users to authenticate themselves on the web interface. Different user roles may exist on the web platform
Code upload
Web interface users can upload software code to the system in order to feed the Generative AI algorithms that will generate the desired business content, namely technical documentation, C4 diagrams, or sequence diagrams.
Choice of repository(ies) for storing generated content
Feature that allows users to choose the repositories where the generated content will be stored, whether they are local file systems, collaborative platforms, agile management systems, or cloud storage services
Manual validation of generated content (Human-in-the-Loop)
Functionality which will allow users with the appropriate permissions to access automatically generated content and validate the information present in it. Thus, it will be possible to make annotations or corrections to the technical documentation, flowcharts, or C4 diagrams that the platform automatically generated.
Code upload history
Feature in the web interface that will allow users to access a list of software code uploads that have been performed over time, making it possible to view and access the uploaded code files.
Deliverables
