A tool for generating and verifying Doxygen comments for C, C++, and C# code.
DoxyPatch is a command-line tool designed to simplify the creation and verification of Doxygen comments for C, C++, and C# code. It utilizes a combination of regular expressions and AI-powered models to generate high-quality comments that adhere to the Doxygen syntax.
- Generates empty Doxygen comment templates for functions and methods without Ollama
- Generates filled Doxygen comments with Ollama enabled, using AI-powered models to create accurate and informative comments
- Verifies consistency between function names and Doxygen comments
- Checks for empty comments (to compile manually)
- Supports C, C++, and C# languages
DoxyPatch was born out of the need for a tool to simplify the creation and verification of Doxygen comments. Initially developed as an external tool due to the lack of support in embedded compilers/IDEs, it has evolved over time to support multiple languages and AI-powered models.
- .NET 9 or later
- Ollama AI backend (download instructions below)
- 4GB of RAM on the GPU (for running the AI model)
- Install .NET 9 SDK
- Clone this repository:
git clone https://github.com/undici77/DoxyPatch.git - Run
build_all.shorbuild_all.ps1: Files available inartifactsfolder
- Install the Ollama AI backend: follow the instructions on the Ollama website
- Unzip file based on your architecture
- Folder should be like:
.
|-- DoxyPatch
|-- DoxyPatch.ini
`-- Models
|-- doxypatch
`-- doxypatch-with-contextNote: When running DoxyPatch with the Ollama feature for the first time, if the required models by doxypatch and doxypatch-with-context are not present in Ollama, they will be automatically downloaded. This process can take a long time depending on your internet connection speed (the 3b model is approximately 2GB). Alternatively, you can manually import models using the scripts import.sh or import.bat located in the Models directory.
Usage: DoxyPatch [file or directory] <options>
Arguments:
[file or directory]
Specify the file or directory path to process.
Options:
h, --help
Show this help message and exit.
r, --recursive
Process directories recursively, including all subdirectories and files.
o, --ollama
Enable **Ollama** mode to automatically generate **Doxygen** fields for the specified files or directory.
b, --rebuild
Rebuild existing **Doxygen** fields, overwriting any previously generated documentation.
c, --with-context
Pass the entire source code as context for more accurate documentation generation (experimental feature).
m, --dry-mode
Run **DoxyPatch** in dry mode, simulating the documentation generation process without making any actual changes.
d, --delay
Specify a delay in seconds between processing files to avoid overheating GPU and CPU.The DoxyPatch.ini is automatically generated at first start and contains configuration settings for the tool. You can modify this file to customize the behavior of DoxyPatch.
[Ollama]
Address="http://localhost:11434"
ModelName="DoxyPatch:latest"
ModelNameWithContext="DoxyPatch-with-context:latest"
PrePrompt="Please provide your best effort, in **English**, adhering to the rules for this method written in '{LANG}':"
PrePromptWithClass="Please provide your best effort, in **English**, adhering to the rules for this '{CLASS}' class method written in '{LANG}':"Note: {CLASS} and {LANG} are placeholders that will be replaced with the actual class name and language being processed.
DoxyPatch is designed as an external tool, which means it can be easily integrated into any Integrated Development Environment (IDE) that supports external tools. This allows you to use DoxyPatch's features directly from within your favorite IDE.
The primary purpose of having an external model is to allow users to experiment with different parameters, models, and prompts to improve the quality of generated comments.
These parameters control various aspects of the model's behavior, such as:
FROM: Specifies the base modeltemperature: Controls randomness; lower values make responses more deterministic.top_p/top_k: Manage token selection for balanced creativity and coherence.num_ctx: Defines the max input size.num_predict: Limits response length.SYSTEM: Prompt to defines the model's behavior.
For more information on customizing these parameters, adding new ones, or learning about other advanced features, please refer to the Ollama Model File documentation.
At every startup, DoxyPatch checks the hashes of model files in Models folder and automatically updates and recreates the model if any changes are detected. This ensures that the model is always up-to-date with the latest prompts and parameters.
The --with-context mode is currently an experimental feature.
As a result, demands substantially more GPU RAM capability. If your system lacks sufficient GPU resources, you may encounter performance issues or errors when attempting to utilize this feature.
A local model ensures your privacy, keeping all data processing on your machine (so private at 100%).
DoxyPatch is licensed under the GPL (General Public License). The full text of the license can be found in the LICENSE file.
The GNU General Public License is a free, copyleft license for software and other kinds of works. It ensures that the software remains free and open, and that any modifications or derivative works are also made available under the same license.
Please note that by using DoxyPatch, you agree to comply with the terms and conditions of the GPL.
This project serves as an excellent starting point for studying and learning how to modernize old tools by integrating AI. DoxyPatch demonstrates the process of taking a traditional code documentation tool and enhancing it with AI-powered capabilities. By working with small, local models, you can gain valuable experience in AI integration and optimization, ensuring that even limited resources can produce high-quality results. This project is particularly interesting for those looking to explore AI in practical, real-world applications, providing a solid foundation for more advanced AI projects in the future.
Contributions are welcome! If you have any ideas or bug fixes, please submit a pull request or issue on this repository.