Save data.json to file system
------------------------------
A copy of the data.json file at the beginning of the download process is
saved alongside the actual downloaded data. Since `export.socrata()` uses
data.json as the index to download data, this will allow users to
cross-reference the downloaded data with other metadata associated with it
available through [Project Open Data](https://project-open-data.cio.gov).

Handle non-data file
---------------------
Socrata lists non-data files, such as Socrata Stories--HTML websites that
contain text but no machine-readable data--in the data.json file. This
causes errors when trying to download those sites because they do not have
a "distribution URL". While it's arguable that these "sites" should not be
included in the first place, the script now simply skips those files.

Since a copy of the data.json file is downloaded (see above), users will
have transparency into which URLs were not downloaded.

Socrata supports external links which direct to web pages (e.g., HTML).
These would cause an error when `export.socrata()` attempted to download
them. This fix will simply skip those files and proceed to the next file.

  * Ignores HTML files (e.g., Socrata Pages)
  * Ignores on occassions there isn't any data
  * Will download (uncompressed) PDFs, Word, Excel, PowerPoint, plain text attachments.

Rebased branch with most recent `dev` branch and generated documentation.

Merge branch 'dev' into issue126

# Conflicts:
#	DESCRIPTION
#	R/RSocrata.R

* Removed user-defined option for file output (not available yet)
* Clarified documentation where `export.socrata()` files will be located.
* Fixed incorrect date in `DESCRIPTION` file.
* Iterating build number.