pygixml: the Python Giant XML framework

Explore

Everything about pygixml, in two places

API reference and guides on one side, honest performance numbers — refreshed on a schedule, not hand-edited — on the other.

Documentation

Quickstart, XML basics, XPath, streaming, dictify/objectify/jsonify, CLI reference, and full API docs.

Read the docs

Benchmarks

Speed, memory, and install size against lxml, ElementTree, xmltodict, and xmljson — interactive, raw data one click away.

See the numbers

Source code

MIT-licensed, on GitHub. Issues, PRs, and the full build (CMake + Cython) welcome.

Open on GitHub

Why pygixml

One package, every way to read XML

Pick the layer that matches what you're actually doing with the file, not one generic tree API forced onto every use case.

Raw DOM

pugixml's tree, directly — XPath 1.0, attributes, namespaces, the whole surface.

dictify

XML → plain dict, same convention as xmltodict — drop-in, not a rewrite.

objectify

Lazy attribute-style access — root.child.grandchild — nothing materialized until touched.

jsonify

Direct XML→JSON, including stream_dump — flat memory, however big the file gets.

Streaming

iterparse / iterfind over yxml — process a file too big to ever load whole, one element at a time.

CLI

pygixml cat, query, stream, convert — XML from a shell pipeline.

Beyond speed

What each library can actually do

Speed and memory aren't the whole story, and pygixml's scope isn't either — it's a Python Giant XML framework, built on pugixml plus a streaming yxml layer for the oversized case. A factual capability comparison, not a benchmark; static, so it lives here rather than in the generated report.

Feature	pygixml	lxml	ElementTree	xmltodict	xmljson
XPath queries Selecting nodes with an XPath expression.	✓ Full XPath 1.0 (via pugixml)	✓ Full XPath 1.0, plus EXSLT extensions	~ Small subset: no position(), limited axes	✗	✗
XSLT transforms	✗	✓ via libxslt	✗	✗	✗
XML Schema (XSD) / DTD validation	✗	✓ via libxml2	✗	✗	✗
Streaming, item-by-item parse Process one element at a time without holding the whole tree.	✓ iterparse/iterfind (yxml-based)	✓ iterparse	✓ iterparse	✓ item_depth + item_callback	✗ needs an already-built tree
Constant-memory XML → single JSON Converting a file too large for RAM into one valid JSON document.	✓ jsonify.stream_dump	✗	✗	✗ streams dicts to a callback, not one JSON doc	✗
Built-in XML ↔ dict/JSON conversion	✓ dictify (xmltodict-compatible) + jsonify	✗ needs xmljson	✗	✓ its entire purpose	✓ needs lxml/ElementTree to parse first
YAML / TOON conversion	✓ pygixml convert (optional deps)	✗	✗	✗	✗
Dotted / attribute-style navigation doc.root.child.grandchild instead of find()/XPath.	✓ objectify	✓ lxml.objectify	✗	✗ plain dict access only	✗
Command-line tools	✓ cat, query, jsonify, stream, convert	✗	✗	✗	✗
Backend	✓ C++ (pugixml + inlined yxml) via Cython	✓ C (libxml2 / libxslt)	~ C accelerator, pure-Python fallback	✗ pure Python, over stdlib expat	✗ pure Python
Standard library (zero install)	✗	✗	✓	✗	✗
abi3 / stable-ABI wheels One compiled wheel across Python minor versions.	✓ cp310-abi3, covers 3.10+	✗ per-Python-version wheels	✓ n/a — ships with the interpreter	✓ n/a — pure Python	✓ n/a — pure Python

Usage

Quick example

The same document, four ways.

import pygixml

doc = pygixml.parse_string("<root><item id='1'>a</item></root>")
root = doc.root()
print(root.child("item").attribute("id").value())  # 1

from pygixml import jsonify

json_str = jsonify.dumps("<root><item>a</item></root>")

# Too big to hold in memory as a tree? Stream it instead:
jsonify.stream_dump("huge.xml", "huge.json")  # flat memory, any size

import pygixml

# yxml push parser under the hood -- never holds the full document
for record in pygixml.iterfind("huge.xml", "record"):
    print(record.tag, record.get("id"))
    record.clear()  # drop this element's memory before the next one

# pretty-print + colorize
pygixml cat data.xml

# ad-hoc XPath from a shell pipeline
pygixml query data.xml "//item[@id='1']"

# convert between XML / JSON / YAML / TOON, any direction
pygixml convert data.xml data.json

The Python Giant XML framework

Everything about pygixml, in two places

One package, every way to read XML

What each library can actually do

Quick example

Ready to try it?