This is a pre-production deployment of Warehouse. Changes made here affect the production instance of PyPI (
Help us improve Python packaging - Donate today!

Data scraping for a more civilized age

Project Description

Struct-o-Miner is an elegant Python library for extracting structured data from HTML or XML documents. It’s ideal for situations where you have your document in a string and just want the data out of it, something like a fancy type casting operation.


Declarative syntax. The format of data is static, so any imperative code you have to write to extract it is just boilerplate. Instead, declare the structures you’re interested in much in the same way you define models in Django or SQLAlchemy, and let Struct-o-Miner take care of the boring parts.

Rich data types. Obtain your data directly as Python types using fields like TextField, IntField or DateTimeField. You can even have lists of dictionaries using StructuredListField.

Organized. The most cumbersome part of scraping is data cleanup. All the exceptional cases and real-world considerations can rapidly degenerate into complicated and unmaintanable spaghetti. Struct-o-Miner provides tools to separate this code by field and by semantic concern.

Focused. Struct-o-Miner adheres to the Unix philosophy of doing one thing and doing it well: you give it a document and it gives you structured data. Scraping is not exclusively part of web crawling, and Struct-o-Miner is a small library that enables you to do it in any project, with no additional cruft.


For a quick example, consider the following HTML snippet:

    <span class="foo">Foo</span> <a href="">Example: Bar</a>
        <li><span>2014-03-01</span>: 1 (one)</li>
        <li><span>2014-03-05</span>: 3 (three)</li>

Here is a document that targets some of the data we might be interested in:

class Stuff(Document):
    foo = TextField('//div/span[@class="foo"]')
    bar_name = TextField('//div/a')
    bar_url = URLField('//div/a')  # Same xpath, but URLField extracts the href
    things = StructuredListField('//div/ul/li', structure=dict(
        # A StructuredField for each element selected by the xpath above
        # Sub-element xpaths are relative to their respective parent
        date = DateField('./span'),
        number = IntField('.')))

    def _extract_the_bar_name(value, **kwargs):
        # Remove 'Example: ' after the field is parsed
        return value.split(' ')[-1]

    def _uppercase_the_bar_name(value, **kwargs):
        # Handle the field after the previous processor ran
        return value.upper()

    def _clean_numbers(value, **kwargs):
        # Isolate the numeric part before the field is parsed as an int
        return value.strip(': ').split(' ')[0]

Now we just pass the HTML to this object for parsing, and data is then available using typical Python element access. In Struct-o-Miner, we call this value access.

>>> data = Stuff(html)

>>> pprint(dict(data))
{'bar_name': 'Bar',
 'bar_url': '',
 'foo': 'Foo',
 'things': [{'date':, 3, 1), 'number': 1},
            {'date':, 3, 5), 'number': 3}]}

>>> data['things'][0]['date'], 3, 1)

You can also reach the field object for each datum using parentheses (i.e. function calls). Field access may seem un-pythonic at first, but every field containing some kind of structure (ListField, DictField, StructuredField and variants) is also a callable that returns the requested child object.

>>> data('things')(0)['date'], 3, 1)

>>> data('things')(0)('date')
<structominer.fields.DateField object at 0x10efae7d0>

Finally, the third axis of access allows you to reach the objects used as structural templates in fields such as lists and dictionaries. Structure access is what enabled us to define the preprocessor on things.number. Notice how the following are distinct:

<structominer.fields.DateField object at 0x10efa1250>

>>> data('things')(0)('date')
<structominer.fields.DateField object at 0x10efae7d0>


The Python ecosystem is rich in solutions for or related to data scraping and web crawling. This is a survey of possible alternatives, highlighting the unique ways Struct-o-Miner contributes to the scene.

lxml and Beautifoul Soup are the standard building blocks of Python scrapers: they both parse markup documents and provide an interface to query and manipulate them. Using them directly can be cumbersome though, as data needs to be selected manually. Struct-o-Miner provides a declarative interface for targetting the elements, then uses lxml under the hood to select all the data.

pyquery wraps lxml.etree with a jQuery-inspired API more familiar to web developers. Apart from the convenience of selecting elements using CSS, pyquery provides little advantage in scraping over lxml. Similarly, cssselect converts CSS selectors to XPath queries which can then be used with lxml. There are plans to support it directly within Struct-o-Miner so that fields can be specified using CSS.

Scrapy is a complete web crawling framework. It can be used to build a reliable crawling operation and benefits from a large community as well as commercial support from ScrapingHub, including a PaaS for running massive Scrapy projects. Despite differences in stylistic approach, Struct-o-Miner is comparable in purpose to Scrapy Items and ItemLoaders. It was however designed to provide this functionality as a standalone library, with an arguably more pythonic flavour.


You can install Struct-o-Miner from PyPI with pip:

$ pip install structominer

or from GitHub with git:

$ git clone
$ cd structominer && python install
Release History

Release History

This version
History Node


Download Files

Download Files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
structominer-0.2.0.tar.gz (16.1 kB) Copy SHA256 Checksum SHA256 Source Apr 19, 2014

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting