overview.txt~

This library is intended to be used with the LinkedIn API.  It is not
an application in and of itself, so you will need to obtain your own
API key to use the library properly.

This overview will go over the library in the following sections:

I.     Design Philosophy
II.    Library Structure
III.   Additional Information

I.  Design Philosophy

The library is highly object-oriented in nature.  Pretty much everything
you can do will end up being passed through a series of objects.  The
client interface is an object, all parsers are objects, and even XML nodes
from API returns are objects.

Most importantly, all returns from the LinkedIn API come back as objects,
not XML string literals.  This includes API errors caused by bad URL
formats, bad access tokens, etc.  As of right now, the original XML
used to construct the objects will not be available directly for you
to inspect, although it can easily be accessed by calling
etree.tostring() from the lxml library on the return object's xml
attribute.

In keeping with that and the Python design philosophy, most interactions
with the API are done with standard Python objects and data types.  To use
the search API, you pass a dict with your search criteria as key-value pairs.
To use field selectors, you simply pass a list of strings.  To specify timestamps,
you can either pass the UTC string/integer or a standard Python datetime object.
Lowlevel type conversions and URL formatting should be done for you.

If it should work, it probably does.  If it doesn't, let me know or contribute
code to make it work :).

II.  Library Structure

	liclient/
		__init__.py - contains the LinkedInAPI class itself and is
			      the entry point for the library
		analysis/
			__init__.py
			nlp.py  - contains utilities for NLP analysis
				  of profiles (requires NLTK)
		oauth2/
			__init__.py - 3rd party module for dealing with
				      oauth, MIT license is included in
				      the module itself
		parsers/
			__init__.py
			lixml.py   - contains the parsers for XML returns
			helpers.py - helper functions for data conversion
			mappers.py - classes for mapping XML returns

III.  Additional Information

I work at a small recruiting firm and wrote the library to help us do
LinkedIn integrations with our website and CRM system.  Some of the design
choices probably do reflect that a little bit (like the update_content
attribute that appears on each network update - it's not a part of the
update itself but is a generated string based on the type of update).

Any questions or comments can be emailed to me (my address is in the 
readme).