What is the best way to parse C++ class architecture in Python?

Question

I want to parse c++ sources in Python. Namely, information about the structure of the classes.

I need to construct some object in Python that should contain information about c++ class. Something like this:

• [namespace]
• function name [: parent]
• [constructor([parameters])]
• [destructor([parameters])]
• public methods([parameters])
• private methods([parameters])
• [enums]
• [structures]

---

any advice for this work?

any helpful modules?

similar projects?

thx!

Answer 1

There's some good information here. In particular, I would check out PLY.

oooh: edit: this little project, built on top of PLY, parses C++ headers (no templates, i guess), and is pure python. neat!

Answer 2

Using cppheaderparser can be quite a bit of hard work. When I last checked it didn't handle even reasonable C++ code with multiple files, export macros, namespaces, STL containers and all. I've a project with similar requirement like Mike for which I'm using doxygen combined with lxml. Doxygen is a sophisticated C++ parser that dumps out XML representation of the source code. Everything that you need and more. The format is straight-forward to understand. Python programs can read the generated XML using lxml Python library, which has full-fledged XPath support for quick querying. You can easily save yourselves from the headache of parsing C++!

Answer 3

GCC already has an XML output mode, and who knows C++ better than GCC ? (think of handling #include's and macros if you roll your own parser ...)

There are even bindings for it in pygccxml, I recently used it for a little unittest and it gets the job done relatively easily.