Parker 0.6.0

A web spider for collecting specific data across a set of configured sites

Parker is a Python-based web spider for collecting specific data across a set of configured sites.

Non-Python requirements:

• Redis - for task queuing and visit tracking

• libxml - for HTML parsing of pages

Installation

Install using pip:

$ pip install parker

Configuration

To configure Parker, you will need to install the configuration files in a suitable location for the user running Parker. To do this, use the parker-config script. For example:

$ parker-config ~/.parker

This will install the configuration in your homedir and will output the related environment variable for you to set in your .bashrc.

Changes

0.6.0

• Add tracking of visited URIs as well as page hashes to the crawl worker. Use that to reduce the number of URIs added to the crawl queue.

0.5.1

• Fix an issue with the order of key-value reference resolution that prevented the effective use of unique_field if using a field that was a kv_ref.

• Add some Parker specific configuration so we can specify where to download, in case the PROJECT env variable doesn’t exist.

0.5.0

• Update ConsumeModel to post process the data. This enables us to populate specific data from a reference to a key-value field.

• Reorder changes so newest first, and rename to “Changes” in the long description.

0.4.2

• Bug fix to fix RST headers which may be the problem.

• Remove the decode/encode which is not the issue.

0.4.1

• Bug fix to see if RST in ASCII fixes issues on PyPI.

0.4.0

• Added handling for a PARKER_CONFIG environment variable, allowing users to specify where configuration files are loaded from.

• Added the parker-config script to install default configuration files to a passed location. Also prints out an example PARKER_CONFIG environment variable to add to your profile files.

• Updated documentation to use proper reStructuredText files.

• Add a CHANGES file to track updates.