HTML5 Parser for Python

Warnings

• breaking Support for Python 2.6, 3.3, and 3.4 has been dropped in recent versions. Specifically, Python 2.6 support was removed in 1.0.1, and Python 3.3/3.4 support was removed in 1.1.
  Fix: Upgrade to Python 3.5+ for html5lib 1.1+.
• deprecated The `html5lib` sanitizer (via `html5lib.serialize(sanitize=True)` and `html5lib.filters.sanitizer`) has been deprecated since version 1.1. Users are recommended to migrate to the `Bleach` library for HTML sanitization.
  Fix: Migrate to `Bleach` for sanitization. Note that `Bleach` is not a drop-in replacement and may require tuning due to different default allow lists and escaping behaviors.
• breaking The default DOM treebuilder was removed, meaning `html5lib.treebuilders.dom` is no longer directly supported as a module. Instead, `html5lib.treebuilders.getTreeBuilder("dom")` should be used, which returns a builder using `xml.dom.minidom`.
  Fix: Replace direct imports or references to `html5lib.treebuilders.dom` with `html5lib.getTreeBuilder("dom")`.
• gotcha When using `html5lib` as a backend for `pandas.read_html()`, you might encounter `ImportError: missing optional dependency html5lib` even if `html5lib` is installed. This often happens if the `html5lib` installation is not correctly recognized by the `pandas` environment or if there are conflicts with other parsers.
  Fix: Ensure `html5lib` is installed in the correct environment (e.g., using `pip install html5lib` or `conda install html5lib`). If the issue persists, try installing `BeautifulSoup4` and `lxml` alongside `html5lib` (`pip install "pandas[html]"`) and restarting your kernel/environment.
• gotcha `html5lib` is a pure-Python library and can be significantly slower than alternatives like `lxml` (which is written in C). While `html5lib` provides more specification-compliant parsing, performance-critical applications might prefer `lxml` where strict HTML5 parsing isn't the absolute highest priority.
  Fix: For performance-sensitive applications, consider using `lxml` as the parser directly or as a treebuilder with `html5lib` if some HTML5-specific error handling is still desired. `lxml` can be used as a tree format with `html5lib` by specifying `treebuilder='lxml'`.

Install

• pip install html5lib Install stable version

Imports

• html5lib

  import html5lib

• HTMLParser

  from html5lib.html5parser import HTMLParser

  While 'html5lib.HTMLParser' might work due to re-exports, direct import from 'html5lib.html5parser' is clearer and avoids potential ambiguity.
• getTreeBuilder

  from html5lib import getTreeBuilder

  The `getTreeBuilder` function is re-exported at the top-level `html5lib` module for convenience.

Quickstart

This quickstart demonstrates basic HTML parsing using `html5lib.parse` for the default `xml.etree` output, and how to use `html5lib.HTMLParser` with a custom treebuilder like `xml.dom.minidom`. It also shows how to enable strict parsing to catch HTML errors. By default, `html5lib` provides an `xml.etree` element instance, but `xml.dom.minidom` and `lxml.etree` are also supported via treebuilders.

import html5lib

# Parse a simple HTML string
document = html5lib.parse("<p>Hello <b>World</b>!</p>")
print(f"Parsed document tag: {document.tag}")
print(f"First child's tag: {document[0].tag}")
print(f"First child's text: {document[0].text}")

# Parse with a specific treebuilder (e.g., xml.dom.minidom)
parser = html5lib.HTMLParser(tree=html5lib.getTreeBuilder("dom"))
minidom_document = parser.parse("<p>Another <i>example</i>.</p>")
# Accessing an element using minidom structure
p_tag = minidom_document.getElementsByTagName('p')[0]
print(f"Minidom document P tag: {p_tag.tagName}")

# Example with strict parsing (raises exceptions on errors)
try:
    strict_parser = html5lib.HTMLParser(strict=True)
    strict_parser.parse("<div><p>Missing close tag")
except html5lib.html5parser.ParseError as e:
    print(f"Strict parsing error: {e}")