Tree-sitter XML & DTD Grammars

0.7.0 · active · verified Sun Apr 12

tree-sitter-xml provides pre-compiled Tree-sitter grammars for XML and DTD. It enables fast, robust parsing of XML and DTD documents within Python applications by integrating with the `tree-sitter` library. The current version is 0.7.0, with updates typically coinciding with upstream Tree-sitter grammar improvements or core library changes.

Warnings

gotcha `tree-sitter-xml` itself doesn't provide parsing functions directly. It exposes `language()` and `dtd_language()` to retrieve pre-compiled grammars, which must then be used with `tree_sitter.Parser` for actual parsing. You need both libraries.
Fix: Always import `tree_sitter.Parser` and set the language on its instance: `parser = tree_sitter.Parser(); parser.set_language(tree_sitter_xml.language())`.
breaking Updates to the underlying `tree-sitter` library (especially major versions or breaking changes in its Python bindings) may introduce API changes that could affect how `tree-sitter-xml`'s grammars interact with the parser.
Fix: Always test your parsing logic after `tree-sitter` upgrades. Consult the `tree-sitter` and `tree-sitter-xml` release notes.
gotcha XML and DTD grammars are distinct. Use `tree_sitter_xml.language()` for parsing XML documents and `tree_sitter_xml.dtd_language()` for parsing DTD files; they are not interchangeable.
Fix: Ensure you're using the correct grammar function (`language()` or `dtd_language()`) based on the file type you intend to parse.
gotcha Pre-compiled grammars (`.so`, `.dylib`, or `.dll` files) are platform-specific. While `tree-sitter-xml` aims to distribute compatible binaries, issues can arise in unusual environments (e.g., exotic OS, specific Python distributions) or if custom grammar compilation is attempted.
Fix: If parsing fails with 'Language not found' or similar errors despite correct imports, check platform compatibility. Ensure your `pip` environment matches the distributed binaries. For complex setups, consider building the grammar manually via `tree_sitter.Language.build_library` if provided by the grammar's source.

Install

pip install tree-sitter-xml Install `tree-sitter-xml`
pip install tree-sitter Install core `tree-sitter` library (required)

Imports

language
```
from tree_sitter_xml import language
```
The `language()` function is directly exposed in the top-level package.
dtd_language
```
from tree_sitter_xml import dtd_language
```
The `dtd_language()` function is directly exposed in the top-level package for DTD grammar.

Quickstart

This quickstart demonstrates how to load the XML grammar using `tree_sitter_xml.language()` and then parse a simple XML string using `tree_sitter.Parser`. It also shows how to print the S-expression of the parse tree and find specific elements.

import tree_sitter
from tree_sitter_xml import language

# Load the XML grammar
XML_LANGUAGE = language()

# Initialize the parser
parser = tree_sitter.Parser()
parser.set_language(XML_LANGUAGE)

# Sample XML string
xml_code = """
<root>
  <item id="1">Value 1</item>
  <item id="2">Value 2</item>
</root>
"""

# Parse the XML
tree = parser.parse(xml_code.encode('utf8'))

# Print the S-expression (a common way to inspect the parse tree)
print(f"Parsed XML Tree S-expression:\n{tree.root_node.sexp()}")

# Example of traversing a node (e.g., finding the 'item' elements)
root_node = tree.root_node
item_nodes = [child for child in root_node.children if child.type == 'element' and child.text.decode('utf8').strip().startswith('<item')]

print(f"\nFound {len(item_nodes)} 'item' elements.")
if item_nodes:
    print(f"First item's text: {item_nodes[0].text.decode('utf8')}")

view raw JSON →