Skip to content
Powerful command-line tool for slicing & dicing HTML
Branch: master
Clone or download


Powerful HTML slicing and dicing at the command line.

Build Status Coverage Status

hq is a Python-based command-line tool for querying HTML, manipulating data and producing results as HTML, JSON or any other format. It's based on a compact, flexible expression language that started out as an XPath implementation but ended up going a few different places, so I'm going ahead and calling it HQuery.

HQuery is 99% compliant with the XPath 1.0 standard, minus some features not applicable to HTML. That's nice for querying, but you need more power to take control of the shape and format of the data you produce. To that end, HQuery also includes...

  • Nuggets of XQuery — only a few of the good parts! Just enough for iteration, branching and the like.
  • XPath expansions for HTML — including a class:: axis and class() function, plus abbreviated axes to keep things terse.
  • Super-charged string interpolation — with powerful filters that you can chain together to transform data as you produce it.
  • Computed constructors for HTML and JSON — so you can programmatically assemble and output new HTML or JSON objects and arrays.
  • Out-of-left-field union decomposition — enabling amazingly terse and powerful mappings from clauses in a union to different expressions.

Installing hq

pip install hq

Running hq

cat /path/to/file.html | hq '`Hello, ${/html/head/title}!`'


hq -f /path/to/file.html '`Hello, ${/html/head/title}!`'

To print usage information:

hq --help

Running hq in a container

There's a Docker image project available that makes it super-easy to try out hq without installing any software (aside from Docker):

cat /path/to/file.html | docker run -i frew/hq '//some/hquery'

Thanks, Frew!

Learning hq

The wiki discusses the motivations guiding the HQuery language's design and provides a language reference.

Contributing to hq

hq is tested against Python 2.7 and recent generations of Python 3 (3.4 and 3.5, as of this writing). The file structure and script for the project are based on this blog post.

hq's dependencies are split into a "base" file, the subset needed to run the application, and a "dev" file providing the tools necessary to run tests and the like. To do development:

pip install -r requirements/dev.txt

The parsing logic in is based on the top-down operator precendence approach.

Running Tests


The "dev.txt" dependencies also include pytest-cov, so you can generate a nice coverage report (which you'll find in the htmlcov directory):

py.test --cov=hq --cov-report html

If you want to turn verbosity on to figure out what's going on in a test, you need the --gabby flag (since py.test owns its own -v flag). You'll probably also want to run just one test at a time, because --gabby is way gabby:

py.test --gabby -vv -k some_particular_test_function

Uploading to PyPI

This and other aspects of project setup, including running the CLI locally and using, are covered in the blog post linked above. I'm copying the PyPI upload stuff here for my own convenience, but I ask, of course, that you please submit pull requests rather than uploading to PyPI yourself:

$ python sdist
$ ls dist

$ pip install twine
$ twine upload dist/hq-0.0.4.tar.gz
Uploading distributions to
Uploading hq-0.0.4.tar.gz
You can’t perform that action at this time.