C++ Grammar for Tree-sitter

0.23.4 · active · verified Sat Apr 11

tree-sitter-cpp provides the C++ grammar for the Tree-sitter parsing library. It enables Python applications to parse C++ code, build abstract syntax trees (ASTs), and perform code analysis tasks like syntax highlighting, linting, and refactoring. The library is actively maintained with frequent minor updates (as seen in its v0.23.x releases), and it requires the core `tree-sitter` Python package to function.

Warnings

breaking The `Language.build_library` static method was removed from the `py-tree-sitter` library around version 0.21.x.
Fix: Instead of `Language.build_library(...)`, you should `pip install` the specific grammar package (e.g., `tree-sitter-cpp`) and load it using `Language(grammar_package.language())`. For `tree-sitter-cpp`, this means `import tree_sitter_cpp as ts_cpp; cpp_language = Language(ts_cpp.language())`.
gotcha Installing `tree-sitter-cpp` alone is not sufficient to parse C++ code in Python. The core `tree-sitter` Python package, which provides the `Language` and `Parser` classes, must also be installed and imported.
Fix: Ensure both packages are installed: `pip install tree-sitter tree-sitter-cpp`. Then, import `Language` and `Parser` from `tree_sitter` and the grammar from `tree_sitter_cpp`.
gotcha The `tree-sitter` parsing functions expect source code as a `bytes` object (UTF-8 or UTF-16 encoded), not a standard Python `str`.
Fix: Encode your source strings before passing them to `parser.parse()`: `parser.parse(your_string.encode('utf8'))`.

Install

pip install tree-sitter-cpp tree-sitter Install core library and C++ grammar

Imports

Language, Parser
```
from tree_sitter import Language, Parser
```
The `Language` and `Parser` classes are part of the core `tree-sitter` library, not `tree-sitter-cpp` itself. `tree_sitter_cpp` provides the grammar object.
language()
```
import tree_sitter_cpp as ts_cpp
cpp_language = Language(ts_cpp.language())
```
For pre-compiled grammars installed via pip, the `language()` function from the grammar package (e.g., `tree_sitter_cpp.language()`) should be used directly with `tree_sitter.Language`. The `Language.build_library` method was removed from `py-tree-sitter` around version 0.21.x and is no longer the recommended approach for using distributed grammars.

Quickstart

This quickstart demonstrates how to load the `tree-sitter-cpp` grammar, initialize a `tree_sitter` parser with it, and parse a simple C++ code snippet. It then shows how to access the root node of the generated syntax tree and print basic information about it and its children. The code uses `bytes` objects for source code, as required by the `tree-sitter` library.

import tree_sitter_cpp as ts_cpp
from tree_sitter import Language, Parser

# Load the C++ grammar
CPP_LANGUAGE = Language(ts_cpp.language())

# Create a parser and set its language
parser = Parser()
parser.set_language(CPP_LANGUAGE)

# C++ code to parse
cpp_code = b"""#include <iostream>

int main() {
    std::cout << "Hello, Tree-sitter!" << std::endl;
    return 0;
}
"""

# Parse the code
tree = parser.parse(cpp_code)

# Get the root node of the syntax tree
root_node = tree.root_node

# Print the root node type and its first child (for demonstration)
print(f"Root node type: {root_node.type}")
if root_node.children:
    print(f"First child type: {root_node.children[0].type}")
    print(f"First child text: {root_node.children[0].text.decode('utf8')}")

# Example of traversing the tree (optional, for deeper analysis)
def traverse(node, depth=0):
    print('  ' * depth + f'- {node.type} [ {node.start_point} - {node.end_point} ] {node.text.decode('utf8') if node.text else ''}')
    for child in node.children:
        traverse(child, depth + 1)

# print("\nFull AST:")
# traverse(root_node)

view raw JSON →