Checklists-scrapers

-------------

* Fixed docs structure so it can be built by readthedocs.org.

-------------

* Added sphinx for document generation.
* Updated documentation, improving the README and adding longer descriptions
in the main documentation, also hosted on readthedocs.org.

-------------

* Corrected errors in reStructuredText formatting.
* Updates install instructions to make them more readable.

-------------

* Replaced separate settings for the directory to download checklist to by
a single setting, CHECKLISTING_DOWNLOADS_DIR that is used by all spiders.
* Added a meta dictionary attribute with the keys 'version' and 'language'
to the JSON format files containing the downloaded checklist.
* Checklist files now contain ints and floats for numbers and coordinates.
* Protocols are defined using a name rather than a pre-defined code. Same for
the codes used in the 'details' section of the checklist entries . This
avoids making assumptions on where and how the data will be loaded.
* Species records now contain only two fields: name and scientific_name.
* Removed the language suffix from comment fields. The language for comments
is given in the 'meta' attribute.
* Added tests to download and validate checklists from real sites.
* Added an activity field to the checklist.
* Source and observers are now nested dictionaries.

-------------

* Login failures on the WorldBirds spider now raises a LoginException.
* Added an extension to record spider errors. Updated the status report to
include errors.
* Added a status report which is generated and emailed when a spider closes.
* Added support for local_settings.py to override scrapy settings.
* Added a spider for downloading checklists for WorldBirds.

-------------

* Fixed the problem of checklists getting assigned to the wrong response by
limiting the number of simultaneous requests.
* Simplified the arguments used to initialize the eBird spider by moving
semi-static ones into the settings file.