adblocker-rules/docs/python-abp.rst

python-abp
==========

This repository contains a library for working with Adblock Plus filter lists,
a script for rendering diffs between filter lists, and the script that is used
for building Adblock Plus filter lists from the form in which they are authored
into the format suitable for consumption by the adblocking software (aka
rendering).

.. contents::


Installation
------------

Prerequisites:

* Linux, Mac OS X or Windows (any modern Unix should work too),
* Python (2.7 or 3.5+),
* pip.

To install::

    $ pip install --upgrade python-abp


Rendering of filter lists
-------------------------

The filter lists are originally authored in relatively smaller parts focused
on particular types of filters, related to a specific topic or relevant for a
particular geographical area.
We call these parts *filter list fragments* (or just *fragments*) to
distinguish them from full filter lists that are consumed by the adblocking
software such as Adblock Plus.

Rendering is a process that combines filter list fragments into a filter list.
It starts with one fragment that can include other ones and so forth.
The produced filter list is marked with a `version and a timestamp <https://adblockplus.org/filters#special-comments>`_.

Python-abp contains a script that can do this called ``flrender``::

    $ flrender fragment.txt filterlist.txt


This will take the top level fragment in ``fragment.txt``, render it and save it
into ``filterlist.txt``.

The ``flrender`` script can also be used by only specifying ``fragment.txt``::

    $ flrender fragment.txt


in which case the rendering result will be sent to ``stdout``. Moreover, when
it's run with no positional arguments::

    $ flrender


it will read from ``stdin`` and send the results to ``stdout``.

Fragments might reference other fragments that should be included into them.
The references come in two forms: http(s) includes and local includes::

    %include http://www.server.org/dir/list.txt%
    %include easylist:easylist/easylist_general_block.txt%


The http include contains a URL that will be fetched and inserted at the point
of reference.
The local include contains a path inside the easylist repository.
``flrender`` needs to be able to find a copy of the repository on the local
filesystem. We use ``-i`` option to point it to to the right directory::

    $ flrender -i easylist=/home/abc/easylist input.txt output.txt


Now the local include referenced above will be resolved to:
``/home/abc/easylist/easylist/easylist_general_block.txt``
and the fragment will be loaded from this file.

Directories that contain filter list fragments that are used during rendering
are called sources.
They are normally working copies of the repositories that contain filter list
fragments.
Each source is identified by a name: that's the part that comes before ":" in
the include instruction and it should be the same as what comes before "=" in
the ``-i`` option.

Commonly used sources have generally accepted names. For example the main
EasyList repository is referred to as ``easylist``.
If you don't know all the source names that are needed to render some list,
just run ``flrender`` and it will report what it's missing::

    $ flrender easylist.txt output/easylist.txt
    Unknown source: 'easylist' when including 'easylist:easylist/easylist_gener
    al_block.txt' from 'easylist.txt'


You can clone the necessary repositories to a local directory and add ``-i``
options accordingly.


Generating diffs
----------------

A diff allows a client running ad blocking software such as Adblock Plus to
update the filter lists incrementally, instead of downloading a new copy of a
full list during each update. This is meant to lessen the amount of resources
used when updating filter lists (e.g. network data, memory usage, battery
consumption, etc.), allowing clients to update their lists more frequently
using less resources.

python-abp contains a script called ``fldiff`` that will find the diff between
the latest filter list, and any number of previous filter lists::

    $ fldiff -o diffs/easylist/ easylist.txt archive/*


where ``-o diffs/easylist/`` is the (optional) output directory where the diffs
should be written, ``easylist.txt`` is the most recent version of the filter
list, and ``archive/*`` is the directory where all the archived filter lists are.
When called like this, the shell should automatically expand the ``archive/*``
directory, giving the script each of the filenames separately.

In the above example, the output of each archived ``list[version].txt`` will be
written to ``diffs/diff[version].txt``. If the output argument is omitted, the
diffs will be written to the current directory.

The script produces three types of lines, as specified in the `technical
specification <https://gitlab.com/eyeo/devops/python-abp/wikis/iflu-0.1>`_:


* Special comments of the form ``! <name>:[ <value>]``
* Added filters of the form ``+ <filter-text>``
* Removed filters of the form ``- <filter-text>``


Library API
-----------

python-abp can also be used as a library for parsing filter lists. For example
to read a filter list (we use Python 3 syntax here but the API is the same):

.. code-block:: python

    from abp.filters import parse_filterlist

    with open('filterlist.txt') as filterlist:
        for line in parse_filterlist(filterlist):
            print(line)


If ``filterlist.txt`` contains this filter list::

    [Adblock Plus 2.0]
    ! Title: Example list

    abc.com,cdf.com##div#ad1
    abc.com/ad$image
    @@/abc\.com/


the output will look something like:

.. code-block:: python

    Header(version='Adblock Plus 2.0')
    Metadata(key='Title', value='Example list')
    EmptyLine()
    Filter(text='abc.com,cdf.com##div#ad1', selector={'type': 'css', 'value': 'div#ad1'}, action='hide', options=[('domain', [('abc .com', True), ('cdf.com', True)])])
    Filter(text='abc.com/ad$image', selector={'type': 'url-pattern', 'value': 'abc.com/ad'}, action='block', options=[('image', True)])
    Filter(text='@@/abc\\.com/', selector={'type': 'url-regexp', 'value': 'abc\\.com'}, action='allow', options=[])


The ``abp.filters`` module also exports a lower-level function for parsing
individual lines of a filter list: ``parse_line``. It returns a parsed line
object just like the items in the iterator returned by ``parse_filterlist``.

For further information on the library API use ``help()`` on ``abp.filters`` and
its contents in an interactive Python session, read the docstrings, or look at
the tests for some usage examples.

Blocks of filters
~~~~~~~~~~~~~~~~~

Further processing of blocks of filters separated by comments can be performed
using ``to_blocks`` function from ``abp.filters.blocks``:

.. code-block:: python

    from abp.filters import parse_filterlist
    from abp.filters.blocks import to_blocks

    with open(fl_path) as f:
        for block in to_blocks(parse_filterlist(f)):
            print(json.dumps(block.to_dict(), indent=2))

Use ``help()`` on ``abp.filters.blocks`` for more information.

Testing
-------

Unit tests for ``python-abp`` are located in the ``/tests`` directory. `Pytest <http://pytest.org/>`_
is used for quickly running the tests during development. `Tox <https://tox.readthedocs.org/>`_ is used for
testing in different environments (Python 2.7, Python 3.5+ and PyPy) and code
quality reporting.

Use tox for a comprehensive report of unit tests and test coverage::

    $ tox

Development
-----------

When adding new functionality, add tests for it (preferably first). If some
code will never be reached on a certain version of Python, it may be exempted
from coverage tests by adding a comment, e.g. ``# pragma: no py2 cover``.

All public functions, classes and methods should have docstrings compliant with
`NumPy/SciPy documentation guide <https://github.com/numpy/numpy/blob/master/doc/HOWTO_DOCUMENT.rst.txt>`_.
One exception is the constructors of classes that the user is not expected to
instantiate (such as exceptions).


Using the library with R
------------------------
Installation
~~~~~~~~~~~~
``python-abp`` can be installed from PyPI or from the source code, either
directly onto a system or in a virtual environment.

To install from PyPI::

    $ pip install -U python-abp

To install from a local source, clone the repo and then::

    $ pip install -U /path/to/python-abp

To use the virtual environment, it must first be created. Python 2 and 3 use
different scripts to create a virtualenv.

In Python 2::

    $ virtualenv env

In Python 3::

    $ python3 -m venv env

Then, use the virtualenv's version of pip to install python-abp, either from
PyPI or from source (as shown above)::

    $ env/bin/pip install -U python-abp

For more information about virtualenv, please see the `User Guide`_ and the
docs_.

Usage
~~~~~
In R, ``python-abp`` can be imported with ``reticulate``:

.. code-block:: R

    > library(reticulate)
    > use_virtualenv("~/path/to/env", required=TRUE)  # If using virtualenv
    > abp <- import("abp.filters.rpy")

Now you can use the functions with ``abp$functionname``, e.g.
``abp$line2dict("@@||g.doubleclick.net/pagead/$subdocument,domain=hon30.org")``

For more information about the reticulate package, see their guide_.

.. _User Guide: https://virtualenv.pypa.io/en/latest/userguide/#usage
.. _docs: https://docs.python.org/3/library/venv.html
.. _guide: https://rstudio.github.io/reticulate/