PyDocEnhancer

AI-powered Python plugin to enhance documentation with summaries, code explanations, examples, semantic search, automated example testing, and multilingual documentation.

Features

• Auto-Generated Summaries: Summarize modules, classes, and functions.
• Code Explanations: Plain-English explanations of code logic.
• Semantic Search: Query documentation with natural language (e.g., "find data processing functions").
• Auto-Generated Examples: Create working code examples from docstrings.
• Automated Example Testing: Extracts and runs code examples from docstrings, reporting results in the docs.
• Multilingual Documentation: Generate documentation in multiple languages (e.g., English, French, Spanish, Chinese) using LLM translation.
• Local LLM Support: Privacy-first processing with local models (e.g., llama 3.2, ctransformers backend).
• Integrations: Works with Sphinx, MkDocs, and Jupyter Notebooks.

LLM Requirement

PyDocEnhancer requires a real LLM provider and model.

• Specify a valid provider (local for ctransformers/Ollama, or openai for OpenAI API) and a model (e.g., ollama/llama3.2:latest).
• Mock mode is not supported.
• This is enforced in both the Python API and CLI.

Installation

pip install pydocenhancer[local]

For most users, the [local] extra is recommended. This uses the ctransformers backend, which does not require C++ build tools and works on most platforms. Only use the base install or cloud extras if you specifically need those features.

If you want to use the llama-cpp-python backend, install with:

pip install pydocenhancer[llama]

Note: This requires C++ build tools on Windows (see Troubleshooting below).

Quick Start

from pydocenhancer import DocEnhancer

# Initialize with a real LLM (Ollama or LLaMA)
enhancer = DocEnhancer(provider="local", model="ollama/llama3.2:latest", language="fr")
enhancer.generate_docs(module_path="my_project/utils.py", output_dir="docs", language="fr")

# Search documentation
results = enhancer.search_docs("file handling functions", "docs")
print(results)

Note: You must specify a real LLM provider and model. The tool will not work without a valid provider (e.g., local or openai) and model (e.g., ollama/llama3.2:latest).

CLI Usage

# Generate documentation with Ollama in English, with example testing
pydocenhancer enhance --module my_project/utils.py --output docs/ --provider local --model ollama/llama3.2:latest --language en

# Search documentation
pydocenhancer search --query "data processing functions" --docs-dir docs/

Requirements

• Python 3.8+
• Local LLM (e.g., LLaMA 3.2 via ctransformers)
• Optional: Sphinx or MkDocs for integration

Windows Users

If you want to use local LLMs, install with:

pip install pydocenhancer[local]

No C++ build tools required for ctransformers wheels.

Troubleshooting Installation (Windows)

Some features (such as local LLMs using llama-cpp-python) require compiling native code. If you see errors like:

CMake Error: CMAKE_C_COMPILER not set, after EnableLanguage
CMake Error: CMAKE_CXX_COMPILER not set, after EnableLanguage
CMake Error at CMakeLists.txt:3 (project):
  Running 'nmake' '-?' failed with: no such file or directory

This means your system is missing the required C/C++ build tools for compiling Python packages with native code.

How to Fix (Windows)

1. Install Visual Studio Build Tools
   • Download from: https://visualstudio.microsoft.com/visual-cpp-build-tools/
   • During installation, select:
     • "Desktop development with C++"
     • Ensure "C++ build tools", "Windows 10 SDK", and "CMake" are checked.
2. Restart your terminal (or use the "Developer Command Prompt for VS").
3. Retry installation:

   pip install pydocenhancer

Debugging Tips

• If you see errors about missing nmake or C/C++ compilers, the build tools are not installed or not in your PATH.
• Try installing llama-cpp-python directly to see detailed errors:

  pip install llama-cpp-python

• If you want to avoid C++ build tools, use the [local] extra to install with ctransformers backend (pre-built wheels):

  pip install pydocenhancer[local]

• For advanced debugging, check the full error log and search for the first error message.

Alternative: Use WSL

If you have trouble with Windows build tools, consider using Windows Subsystem for Linux (WSL) for easier compilation of native code.

Common Errors & Solutions

Error Message	Cause	Solution
ImportError: ctransformers is required for local LLMs.	You tried to use a local LLM without installing ctransformers.	Run pip install pydocenhancer[local]
FileNotFoundError: [Errno 2] No such file or directory	The module path you provided does not exist.	Check the path and try again.
RuntimeError: Local LLM is not initialized.	The local model failed to load or is not available.	Check your model name and installation.
requests.exceptions.ConnectionError	Ollama is not running or not reachable.	Start Ollama and ensure the model is pulled.
Error: ... in Example Test Result	The example code in the docstring is invalid or raises an exception.	Fix the example code in your docstring.
CMake Error: CMAKE_C_COMPILER not set...	Missing C++ build tools on Windows.	See Troubleshooting Installation (Windows) above.
ModuleNotFoundError: No module named 'llama_cpp_python'	You tried to use the llama backend without installing it.	Run pip install pydocenhancer[llama]

Additional Debugging Steps

• Verbose Output: Run with -v or --verbose if available, or set environment variable PYTHONVERBOSE=1 for more details.
• Check Python Version: Ensure you are using Python 3.8+.
• Check Dependencies: Run pip check to see if any dependencies are missing or incompatible.
• Update pip: Sometimes, upgrading pip helps: python -m pip install --upgrade pip.

Packaging Note for Windows Users

• If you want to use local LLMs without C++ build tools, install with:

  pip install pydocenhancer[local]

  This uses the ctransformers backend, which provides pre-built wheels and does not require compilation.
• If you want to use the llama-cpp-python backend, you must have Visual Studio Build Tools installed as described above, and install with:

  pip install pydocenhancer[llama]

Documentation

Full documentation is available in this README and at GitHub.

License

MIT © Abdoullah Ndao abdoullahaljersi@gmail.com